The Department of Health has asked the National Institute for Health and Care Excellence (NICE) to produce guidance on using lenalidomide in the NHS in England. The Appraisal Committee has considered the evidence submitted by the manufacturer and the views of non-manufacturer consultees and commentators, and clinical specialists and patient experts.

This document has been prepared for consultation with the consultees. It summarises the evidence and views that have been considered, and sets out the draft recommendations made by the Committee. NICE invites comments from the consultees and commentators for this appraisal (see section 10) and the public. This document should be read along with the evidence base (the evaluation report).

The Appraisal Committee is interested in receiving comments on the following:

• Has all of the relevant evidence been taken into account?
• Are the summaries of clinical and cost effectiveness reasonable interpretations of the evidence?
• Are the provisional recommendations sound and a suitable basis for guidance to the NHS?
• Are there any aspects of the recommendations that need particular consideration to ensure we avoid unlawful discrimination against any group of people on the grounds of race, gender, disability, religion or belief, sexual orientation, age, gender reassignment, pregnancy and maternity?

Note that this document is not NICE's final guidance on this technology. The recommendations in section 1 may change after consultation.

After consultation:

The Appraisal Committee will meet again to consider the evidence, this appraisal consultation document and comments from the consultees.

At that meeting, the Committee will also consider comments made by people who are not consultees.

After considering these comments, the Committee will prepare the final appraisal determination (FAD).

Subject to any appeal by consultees, the FAD may be used as the basis for NICE’s guidance on using lenalidomide in the NHS in England.

For further details, see the Guides to the technology appraisal process.

The key dates for this appraisal are:

Closing date for comments: 21 August 2014

Third Appraisal Committee meeting: 17 September 2014

Details of membership of the Appraisal Committee are given in section 9, and a list of the sources of evidence used in the preparation of this document is given in section 10.

1   Appraisal Committee’s preliminary recommendations

1.1   Lenalidomide in combination with dexamethasone is not recommended for treating multiple myeloma in people:

• whose condition has relapsed for the first time and
• who have had 1 prior treatment with bortezomib and
• for whom thalidomide is contraindicated or cannot be tolerated and
• for whom stem cell transplantation is not appropriate.

1.2   People currently receiving treatment initiated within the NHS with lenalidomide that is not recommended for them by NICE in this guidance should be able to continue treatment until they and their NHS clinician consider it appropriate to stop.

2   The technology

2.1   Lenalidomide (Revlimid) is a derivative of thalidomide and has immunomodulatory, anti-neoplastic, anti-angiogenic and pro-erythropoietic activity. It is administered orally. Lenalidomide in combination with dexamethasone has a marketing authorisation for treating multiple myeloma in adults who have received at least 1 prior therapy.

2.2   The summary of product characteristics includes the following adverse effects for lenalidomide: neutropenia, anaemia and thrombocytopenia. Because lenalidomide is structurally related to thalidomide, a known human teratogen that causes severe birth defects, a risk minimisation plan has been developed and agreed with the Medicines and Healthcare products Regulatory Agency to avoid fetal exposure to lenalidomide. For full details of adverse reactions and contraindications, see the summary of product characteristics.

2.3   Lenalidomide is available as a 21-capsule pack. The cost per pack varies according to capsule size: £3570 (5 mg), £3780 (10 mg), £3969 (15 mg) and £4368 (25 mg; excluding VAT; British National Formulary edition 65). The recommended starting dose is 25 mg orally once daily on days 1–21 of repeated 28‑day cycles. Costs may vary in different settings because of negotiated procurement discounts.

3   The manufacturer’s submission

The Appraisal Committee (section 9) considered evidence submitted by the manufacturer of lenalidomide and a review of this submission by the Evidence Review Group (ERG; section 10). The purpose of the manufacturer’s submission was to demonstrate the effectiveness and cost effectiveness of lenalidomide compared with bortezomib, bortezomib plus dexamethasone, bendamustine or chemotherapy in adults who have been treated with bortezomib first line, and so for adults in whom treatment with thalidomide or stem cell transplant was not appropriate (see NICE technology appraisal guidance 228, Bortezomib and thalidomide for first-line treatment of multiple myeloma) and who would consider second-line treatment.

3.1   The manufacturer identified, using a systematic literature review, 2 identically designed randomised controlled trials that compared lenalidomide plus dexamethasone with placebo plus dexamethasone (MM-009 and MM-010). The manufacturer did not identify any randomised controlled trials that directly compared lenalidomide with the comparators defined in the scope (bortezomib, bendamustine or chemotherapy, which included regimens based on melphalan, vincristine, cyclophosphamide and doxorubicin). The manufacturer presented data from the following studies, identified through a systematic literature review, to estimate the efficacy of bortezomib and bendamustine:

• Bortezomib: The manufacturer identified 6 observational studies. Taverna et al. (2012), a small retrospective survey of people who had received bortezomib re-treatment, was selected as the most relevant (see sections 3.13 and 3.14).
• Bendamustine: The manufacturer identified 1 retrospective observational study that included people with relapsed or refractory multiple myeloma after prior treatment (Damaj et al. 2012; see sections 3.15 and 3.16).

Lenalidomide plus dexamethasone compared with placebo plus dexamethasone (MM-009 and MM-010)

Clinical effectiveness

3.2   MM-009 and MM-010 compared treatment with lenalidomide plus dexamethasone with placebo plus dexamethasone in people with multiple myeloma who had received at least 1 prior therapy. The prior therapy did not necessarily include bortezomib (and was therefore different to the population set out in the decision problem). The dose schedule on the trial was a once-daily starting dose of 25 mg oral lenalidomide or placebo on days 1–21 of each 28‑day cycle, and a daily dose of 40 mg oral dexamethasone on days 1–4, 9–12 and 17–20 for the first 4 cycles. After the fourth cycle, 40 mg of dexamethasone was administered on days 1–4 only. Treatment was continued until the disease progressed, unless treatment was stopped because of adverse reactions. Those who had placebo plus dexamethasone were offered lenalidomide when they stopped treatment or when the study was unblinded (that is, they could cross over). The study designs of MM-009 (n=353) and MM-010 (n=351) were identical other than study location (MM-009 was carried out in the USA and Canada, whereas MM-010 was carried out in Australia, Europe [including the UK] and Israel). The primary outcome was time to disease progression. Secondary outcomes included overall survival, progression-free survival (defined as time from randomisation to disease progression or death), response rates, adverse reactions and time to decrease in performance status. Randomisation was stratified according to the serum concentration of beta-2 microglobulin, whether or not the patient had previously had stem cell transplantation, and to the number of previous anti-myeloma therapies (1 compared with 2 or more). About 35% had received only 1 prior therapy (first relapse) and about 65% had received at least 2 prior therapies. Only 2 people in the combined MM-009 and MM-010 trials reflected the population in this appraisal, that is, 1 prior treatment only, with bortezomib. Exclusion criteria included people previously treated with lenalidomide or whose disease was refractory to dexamethasone. The investigators assessed response using the European Group for Blood and Marrow Transplantation criteria. In its submission, the manufacturer identified several ‘unspecified’ (post hoc) subgroup analyses using the pooled populations of both trials, including (but not limited to) whether prior treatment with thalidomide or bortezomib had been received. The median age of the population in both studies was between 62 years and 64 years.

3.3   The manufacturer presented primary and secondary outcomes of the MM-009 and MM-010 trials analysed individually and as a pooled dataset using the intention-to-treat population, and intention-to-treat analyses. The absolute values observed in the lenalidomide plus dexamethasone group informed the modelling (see section 3.20), rather than the relative effectiveness of lenalidomide plus dexamethasone compared with placebo plus dexamethasone. At the end of the study (‘unblinding’), randomisation to lenalidomide plus dexamethasone was associated with a statistically significantly reduced median time to progression compared with placebo plus dexamethasone (48.1 weeks compared with 20.1 weeks, p<0.001 in MM-009; and 48.7 weeks compared with 20.1 weeks, p<0.001 in MM-010). In addition, lenalidomide plus dexamethasone was associated with a statistically significantly increased median progression-free survival compared with placebo plus dexamethasone (41.1 weeks compared with 20.1 weeks, p<0.001 in MM-009; and ‘not yet reached’ compared with 20.1 weeks, p<0.001 in MM-010). In an extended follow-up (median 48 months), lenalidomide plus dexamethasone was associated with statistically significantly increased median overall survival compared with placebo plus dexamethasone (29.6 months compared with 20.2 months, p<0.001 in MM-009; and ‘not estimable’ – because too few people in the treatment arm had died for it to be possible to estimate the median –compared with 20.6 months, p<0.05 in MM-010).

3.4   The manufacturer presented pooled analyses of MM-009 and MM-010. Lenalidomide plus dexamethasone was associated with a statistically significant longer median time to progression (13.4 months compared with 4.6 months, p<0.001) and median progression-free survival (11.1 months compared with 4.6 months, p<0.001) compared with placebo plus dexamethasone at unblinding (median 17.5 months; n=704). The manufacturer presented pooled data for overall survival after a median follow-up of 48 months. Overall survival was statistically significantly longer with lenalidomide plus dexamethasone than placebo plus dexamethasone (median of 38.0 months compared with 31.6 months respectively; p=0.045).

Health-related quality of life

3.5   The MM-009 and MM-010 trials did not collect information on health-related quality of life. Instead, the manufacturer presented Eastern Cooperative Oncology Group-performance status (ECOG-PS) data from the trials and it used this to assess progression of disease and impact on daily living abilities. The median time to ‘first worsening’ was statistically significantly longer with lenalidomide plus dexamethasone (36.3 weeks) than with placebo plus dexamethasone (12.1 weeks in MM-009; p=0.012). The median time to first worsening with lenalidomide plus dexamethasone was shorter than with placebo plus dexamethasone in MM-010 (lenalidomide plus dexamethasone: 10.1 weeks; placebo plus dexamethasone: 12.3 weeks). However, this difference was not statistically significant (p=0.271).

Adverse events

3.6   The manufacturer presented data on adverse events from MM-009 and MM-010 separately and pooled. The most common adverse effects associated with lenalidomide plus dexamethasone were haematological. The manufacturer suggested that clinicians could manage these by reducing the dose. Anaemia, neutropenia, thrombocytopenia, constipation, pneumonia, decreased weight, hypokalaemia, hypocalcaemia, tremor, rash and deep vein thrombosis were reported considerably more frequently in the lenalidomide plus dexamethasone group than in the placebo plus dexamethasone group. There was an increased risk of developing thromboembolic adverse events (deep vein thrombosis, pulmonary embolism) with lenalidomide plus dexamethasone compared with placebo plus dexamethasone (9.1% compared with 4.3%, and 4.0% compared with 0.9% respectively).

3.7   The manufacturer stated that bortezomib and thalidomide can be associated with peripheral neuropathy, whereas pooled analyses of MM-009 and MM-010 showed that lenalidomide plus dexamethasone neither increased nor decreased the risk of peripheral neuropathy compared with placebo plus dexamethasone.

3.8   Lenalidomide is structurally related to thalidomide, a known human teratogen that causes severe birth defects. As such, a plan was developed and agreed with the Medicines and Healthcare products Regulatory Agency to avoid fetal exposure to lenalidomide.

Subgroups

3.9   The manufacturer presented 2 subgroup analyses from MM-009 and MM-010. The first analysis was a post hoc comparison of lenalidomide plus dexamethasone with placebo plus dexamethasone after 1 prior therapy. The second analysis was a pre-specified comparison of lenalidomide plus dexamethasone after 1 prior therapy with lenalidomide plus dexamethasone after 2 or more prior therapies (Stadtmauer et al. 2009).

3.10   The first subgroup analysis showed that, in the MM-009 trial, after 1 prior therapy only, lenalidomide plus dexamethasone was associated with a statistically significantly longer median progression-free survival (16.6 months, 95% confidence interval [CI] 11.0 to 36.8 months) compared with placebo plus dexamethasone (4.6 months, 95% CI 4.0 to 5.7 months; hazard ratio [HR] 0.3, 95% CI 0.19 to 0.47; p<0.0001). Similarly, in MM-010, after 1 prior therapy, lenalidomide plus dexamethasone was associated with a statistically significantly longer median progression-free survival (13.3 months, 95% CI 5.1 to 26.9 months) compared with placebo plus dexamethasone (4.5 months, 95% CI 2.8 to 5.6 months, HR 0.39, 95% CI 0.24 to 0.62; p<0.0001). There were no statistically significant differences in median overall survival in MM-009 or MM-010 for lenalidomide plus dexamethasone compared with placebo plus dexamethasone (p>0.05).

3.11   Stadtmauer et al. (2009) looked at the effectiveness of lenalidomide at different lines of therapy. Compared with lenalidomide plus dexamethasone after 2 or more prior therapies (n=220), lenalidomide plus dexamethasone after 1 prior therapy (n=133) had longer:

• median time to progression: 17.1 months compared with 10.6 months (HR 0.68, 95% CI 0.48 to 0.97; p=0.026)
• median progression-free survival: 14.1 months compared with 9.5 months (HR 0.71, 95% CI 0.2 to 0.99; p=0.047)
• median overall survival from study enrolment: 42.0 months compared with 35.8 months; p=0.041.

3.12   The manufacturer identified the following factors that increased the risk of death among people in MM-009 and MM-010 (p<0.05):

• treatment with placebo versus lenalidomide
• high versus low percent of plasma cells in the bone marrow
• high versus low serum concentrations of beta-microglobulin
• shorter versus longer duration of myeloma
• more versus fewer previous therapies for myeloma
• earlier versus later dexamethasone therapy
• higher versus lower international staging system.

The manufacturer also identified the following predictors of progression and treatment failure: beta-2 microglobulin; time since diagnosis of multiple myeloma; number of prior therapies; baseline presence or absence of bone lesions; and ECOG performance score.

Bortezomib

Clinical effectiveness and adverse events

3.13   The manufacturer presented data on the efficacy of bortezomib from 6 single-arm observational studies. The manufacturer presented a retrospective review of patient medical records by Taverna et al. (2012), which included 42 patients across 26 centres in Switzerland. People whose records were surveyed for the study had multiple myeloma that had responded to initial bortezomib therapy, but whose disease had subsequently progressed or relapsed, and were then re-treated with bortezomib, although not necessarily in second line. The study inclusion criteria specified that the initial treatment with bortezomib therapy had achieved complete response, near complete response or partial response, and that people had completed a re-treatment regimen with bortezomib after their disease had relapsed or progressed. Of those in the study, 31% had previously received a stem cell transplant, 12 people had received a different therapy between bortezomib treatments, and a third of people did not receive dexamethasone when re-treated with bortezomib. People had received a median of 2 prior therapies (range 1 to 11). The study end points were listed as complete response, near complete response and partial response, but were not clearly defined. How frequently patients had been followed was not defined, and results were not presented as a Kaplan–Meier plot.

3.14   The Taverna et al. (2012) study showed:

• Median time to progression after bortezomib re-treatment was 10.5 months (range 0.4 to 39.5 or more months).
• Median overall survival from first diagnosis was 9.3 years, after prior bortezomib 3.5 years and after bortezomib re-treatment 1.7 years.

Bendamustine

Clinical effectiveness

3.15   The manufacturer identified a retrospective study to show the clinical effectiveness of bendamustine in relapsed or refractory multiple myeloma (n=110, Damaj et al. 2012). The study consisted of a review of medical records of people who had been treated with bendamustine in a compassionate-use programme in France, and therefore had no control group. People whose records were reviewed had been previously treated with all of the following: alkylators, corticosteroids, immunomodulatory drugs and bortezomib. The study evaluated the response rate to bendamustine, the duration of response, progression-free survival and overall survival. Overall survival was calculated from the first dose of bendamustine and progression-free survival included death from any cause or progression as events.

3.16   The Damaj et al. (2012) study showed that after treatment with bendamustine, at a median follow-up of 10 months, 49 people who had received bendamustine had died after progression (or from other causes related to myeloma). Median progression-free survival was 9.3 months and median overall survival was 12.4 months.

Evidence Review Group critique of clinical effectiveness

3.17   The ERG reviewed the manufacturer’s literature review and concluded that it was broadly suitable. The ERG reviewed the designs of MM-009 and MM-010 and concluded that they were of high quality. In addition, the ERG noted several issues:

• Both trials started in 2004. Since then, management of multiple myeloma has changed (specifically, the introduction of bortezomib) and therefore the trials do not reflect current clinical practice.
• The mean age of people was lower (63 years) than that of people in clinical practice with multiple myeloma in the UK (around 70 years).
• The proportion of people who received 2 or 3 prior therapies was higher than the proportion who had 1 prior therapy only.

3.18   The ERG commented that the manufacturer did not present time to treatment failure, an end point of MM-009 and MM-010, in the clinical section of its submission, although the manufacturer included it in its model.

Cost-effectiveness evidence

3.19   The cost-effectiveness evidence presented by the manufacturer consisted of a systematic literature review and a Markov model developed by the manufacturer. The manufacturer’s systematic review did not identify relevant cost-effectiveness studies. The manufacturer’s model compared lenalidomide with bortezomib re-treatment for people with multiple myeloma for whom stem cell transplant was not suitable, for whom thalidomide was contraindicated, and who had received 1 prior therapy only with bortezomib. The model had 3 health states: pre-progression on treatment, pre-progression off treatment, and post-progression, plus death. The manufacturer chose a lifetime time horizon (25 years), a cycle length of 28 days, and discount rate of 3.5% for costs and quality-adjusted life years (QALYs). The model took an NHS and personal social services perspective. Although a patient access scheme is available for bortezomib after 1 prior therapy, the manufacturer did not include this in the model’s base case.

3.20   The transition probabilities between states in the model were based on overall survival, progression-free survival and time to treatment failure curves. To extrapolate beyond the period observed in studies, the manufacturer presented 6 parametric distributions (exponential, Weibull, log-logistic, log-normal, Gompertz and gamma) for overall survival, progression-free survival, and time to treatment failure for lenalidomide plus dexamethasone. The manufacturer used a log-logistic distribution in the model base case for progression-free survival and time to treatment failure, and used an exponential piecewise distribution for overall survival. The manufacturer considered other distributions in scenario analyses. The manufacturer calculated transition probabilities between health states as follows:

• All people started the model in the ‘progression free on treatment’ state.
• Transition into the ‘progressive disease’ state from either the ‘pre-progression on treatment’ or ‘pre-progression off treatment’ states was determined by progression-free survival.
• Transition from ‘pre-progression on treatment’ to ‘pre-progression off treatment’ was determined by the difference between progression-free survival and time to treatment failure; that is, those people whose disease had not progressed, but for whom treatment had failed because of adverse events.
• Transition from any health state to death was determined by overall survival.
• Lenalidomide: The manufacturer estimated progression-free survival, time to treatment failure and overall survival from the lenalidomide arm of the MM-010 dataset (see section 3.22 for further details).
• Comparators: The manufacturer derived a hazard ratio reflecting the effectiveness of lenalidomide compared with bortezomib. The effectiveness of lenalidomide was derived from the lenalidomide plus dexamethasone arm of the MM-010 data. The manufacturer took estimates of absolute progression-free survival and overall survival of bortezomib from the literature, as described. The manufacturer then approximated a crude hazard ratio between lenalidomide and bortezomib by comparing the median progression-free survival or overall survival estimates between the studies, using Taverna et al. (2012) in the base case. To derive overall survival, progression-free survival and time to treatment failure curves for bortezomib, the manufacturer adjusted the lenalidomide overall survival, progression-free survival and time to treatment failure curves using these crude hazard ratios. The manufacturer used the hazard ratio calculated for progression-free survival for time to treatment failure. It employed the same approach to estimate the relative effectiveness of lenalidomide compared with bendamustine. The manufacturer assumed that chemotherapy and bendamustine were  equally effective. In the ‘progressive disease’ health state, in the comparator arm only, people received lenalidomide as a third-line treatment in the manufacturer’s initial base case. This affected overall survival in that, when the third-line treatment was lenalidomide, transition to death was based on the MM-010 trial overall survival data.

Table 1 Efficacy estimates for lenalidomide compared with bortezomib re-treatment, or bendamustine, included in the initial base-case model

Variable	Evidence source	Hazard ratio (>1 favours lenalidomide)
Re-treatment with bortezomib
Overall survival	Taverna et al. 2012a (n=42; Switzerland) Median (range) prior therapies: 2 (1–11) Prior treatment with bortezomib: 100%	1.70
	White et al. 2013 (n=53; USA and Canada) Number of prior therapies: 1 prior therapy 49%; 2+ prior therapies 51% Prior treatment with bortezomib: 19%	1.42
Progression-free survival	Taverna et al. 2012a (as above)	1.15
	White et al. 2013 (as above)	1.76
	Hrusovsky et al. 2010 (n=60; Germany and Switzerland) Median (range) prior therapies: 3.7 (1–14) Prior treatment with bortezomib: 100%	1.09
	Dispenzieri et al. 2010 (n=7; USA) Median (range) prior therapies: not reported Prior treatment with bortezomib: 100%	1.28
	Petrucci et al. 2013 (n=50; Europe; time to progression used as a proxy for progression-free survival) Number of prior therapies: 1 prior therapy 12%; 2+ prior therapies 88% Prior treatment with bortezomib: 100%	1.26
	Min et al. 2007 (n=57; South Korea) Median (range) prior therapies: not reported: 2 (1–3) Prior treatment with bortezomib: 100%	0.84
Bendamustine (and chemotherapy agents)
Overall survival	Damaj et al. 2012 (n=110; France) Median (range) prior therapies: 4 (1–9) Prior treatment with bortezomib: 100%	3.00
Progression-free survival	Damaj et al. 2012 (as above)	1.09
a Base-case values n=number of people in study

3.21   ‘Modelled’ patients entered the model having received bortezomib as a first-line treatment, and then received second-line treatment with lenalidomide plus dexamethasone, or with the comparator, until their disease progressed (modelled as the duration of the pre-progression on-treatment health state), or until stopping treatment for other reasons (modelled as the duration of the pre-progression off-treatment health state). The manufacturer used bortezomib as the comparator in its base case. The manufacturer explored in scenario analyses alternative second-line treatments, including bendamustine and other chemotherapy agents. The manufacturer’s model also included the possibility that patients go on to receive third- and fourth-line treatments. For third- and fourth-line treatments, the manufacturer developed a best supportive care mix (that is, a miscellaneous ‘basket’ of standard chemotherapy), which depended on previous treatments and included:

• bortezomib
• dexamethasone
• melphalan
• cyclophosphamide
• cisplatin
• doxorubicin
• etoposide
• prednisolone
• prednisone
• lenalidomide.

3.22   The manufacturer estimated overall survival and progression-free survival from MM-010 rather than from combining the 2 studies, because MM-010 included a European population. The manufacturer stated that pooling the trials was not feasible because it would break randomisation. After a request for clarification by the ERG, the manufacturer provided scenario analyses that used MM-009 only data or combined MM-009 and MM-010 results using mean outcomes weighted by the number of people in the trials. The manufacturer obtained absolute values for overall survival and progression-free survival associated with re-treating with bortezomib from the Taverna et al. (2012) study (see sections 3.13 and 3.14). The manufacturer adjusted the MM-010 patient data for factors that could influence outcomes, so that the ‘modelled patients’ from MM-010 better resembled the patients in the Taverna et al. study. Overall survival was adjusted for patient characteristics (concentration of beta-2 microglobulin, ECOG, and presence or absence of bone lesions) in the manufacturer’s base case. However, the manufacturer did not adjust the rates of progression-free survival in MM-010 to reflect the patients in Taverna et al. re-treated with bortezomib. A scenario that included progression-free survival adjusted for duration of myeloma was provided by the manufacturer in response to a request for clarification from NICE. The manufacturer used further sources of overall survival and progression-free survival for bortezomib in scenario analyses (White et al. 2013, Petrucci et al. 2013, Hrusovsky et al. 2010, Dispenzieri et al. 2010 and Min et al. 2007; see table 1 for details). The manufacturer used the hazard ratios to adjust the extrapolated lenalidomide overall survival, progression-free survival and time to treatment failure curves (see section 3.20), to derive the comparator curves.

3.23   The manufacturer accounted for adverse events during treatment in the model, taking adverse events in the pre-progression health states from Lenalidomide for the treatment of multiple myeloma in people who have received at least one prior therapy (NICE technology appraisal guidance 171). The manufacturer estimated the event rates associated with lenalidomide plus dexamethasone from MM-010, and for bortezomib from the VISTA trial, which compared bortezomib plus melphalan and prednisone with melphalan, and prednisone. The manufacturer used VISTA’s melphalan and prednisolone arm as a proxy to estimate the adverse events rates for all other types of chemotherapy. The manufacturer estimated an average rate and applied it to each model cycle. The manufacturer assumed that people do not experience any adverse events after their disease progresses (when in the post-progression health state), but did assume that people in the comparator arm who receive lenalidomide plus dexamethasone as a third-line treatment experience adverse events. Adverse events in the model included anaemia, constipation, diarrhoea, deep vein thrombosis, hypercalcaemia, neutropenia, peripheral neuropathy, pneumonia and thrombocytopenia.

3.24   To estimate utility, the manufacturer used utility values from the literature because health-related quality-of-life data were not collected in the MM-009 and MM-010 clinical trials or in the observational studies used to describe the comparator treatments. Of the 8 studies identified by the manufacturer, only 2 (Khanna et al. 2006 [SF-36, USA], van Agthoven et al. 2004 [EQ-5D, Netherlands]) undertook primary data collection. The manufacturer modelled utility values from the study by van Agthoven et al. The value for pre-progression utility was 0.81 decreasing after 2 years to 0.77. The value for post-progression utility was 0.64. The manufacturer adjusted the utility values for age, based on published UK EQ-5D population norms. The manufacturer included decrements in utility associated with adverse events to each model cycle. The manufacturer applied these in the pre-progression health state for all treatments, as well as in the post-progression health state for patients receiving lenalidomide plus dexamethasone third line (in the comparator arm). The utility decrements associated with each treatment per model cycle were 0.013 for lenalidomide plus dexamethasone, 0.033 for bortezomib, and 0.025 for bendamustine or chemotherapy agents.

3.25   The manufacturer’s model included costs associated with treatment, resource use and adverse events (see table 2). The manufacturer obtained acquisition costs for each treatment from the Department of Health’s electronic Market Information Tool (eMIT) and the British National Formulary. The cost of lenalidomide included dose reductions and treatment interruptions based on the experience of patients in MM-010. Bortezomib has a patient access scheme, whereby the manufacturer of bortezomib reimburses the NHS for people whose condition does not respond to treatment. The manufacturer did not account for this in its base case; however, it did provide a scenario analysis that assumed a discount of 15%. The manufacturer based the cost of each adverse event on where a patient received treatment (inpatient, hospital day case, outpatient and general practice). The model included a weighted average of adverse events costs. Other costs were obtained from NHS reference costs.

Table 2: List of health states and associated costs in the economic model

Health state and item	Lenalidomide arm (cost per cycle)	Base-case comparator arm, bortezomib (cost per cycle)
Pre-progression
Technology	Lenalidomide: £3773 Dex (cycles 1–4): £7.76 Dex (cycles 5+): £2.59	£4067.30
Concomitant G-CSF and administration	£473.62	n/a
Monitoring and tests	£153.34
Administration	£161.85 (first cycle only)	£1065.76
Transport	£6.39 (first cycle only)	£17.04
Adverse events	£17.11	£29.26
Additional monitoring for lenalidomide (annual rate)	£824.26	n/a
Post-progression
3rd line treatment	£70.20 IV administration: £69.63 Transport: £3.06 Duration: maximum 4 cycles (17.2 weeks)	£3772.88 IV administration: £0.00 Transport: £0.00 Duration: Lenalidomide PFS from MM-010 Adverse events with lenalidomide: £17.11
4th line treatment	Therapy: £2277.28 IV administration: £0.00 Transport: £0.00 Duration: maximum 4 cycles (16.8 weeks)
Monitoring and tests	£175.86
Terminal care	£1235 on death
Alternative 3rd line scenarios
Changing treatment composition after bortezomib 2nd line	n/a	Drug cost £1716.99 IV administration: £49.45 Transport: £2.20
Changing treatment composition after other 2nd line comparators	n/a	£2592.00 IV administration: £203.00 Transport: £3.25
Abbreviations: Dex, dexamethasone; G-CSF, granulocyte-colony stimulating factor; IV, intravenous; PFS, progression-free survival.

Results of the economic analyses

3.26   The manufacturer presented initial results in its submission. However, the ERG noted several problems with the manufacturer’s initial base-case modelling, and requested, and received, more analyses from the manufacturer. Later, following the concerns noted by the Committee documented in the appraisal consultation document and the consultation period, the manufacturer submitted further analyses. The analyses received pre-consultation are termed as follows: ‘base case’ (reflecting initial modelling); ‘base-case A’ (modelling after a first round of clarification); and ‘base-case B’ (modelling after a second round of clarification). The analyses received post-consultation are referred to as ‘base-case C’. Each set of the analyses is presented below.

Manufacturer’s initial analyses and results before clarification and before consultation

3.27   The manufacturer presented a deterministic base-case incremental cost-effectiveness ratio (ICER), comparing lenalidomide plus dexamethasone with bortezomib re-treatment, of £14,535 per QALY gained (incremental costs £7682, incremental QALYs 0.53), and a mean probabilistic ICER of £13,930 per QALY gained. The probabilities of lenalidomide being cost effective at £20,000 and £30,000 per QALY gained were 58.6% and 73.1% respectively. The manufacturer’s deterministic sensitivity analysis showed that the ICER was most sensitive to the hazard ratio for overall survival (reflecting the relative effectiveness of lenalidomide compared with bortezomib). Other key drivers of the ICERs included which parametric function for overall survival and time to treatment failure the manufacturer chose to apply.

3.28   The manufacturer presented several scenario analyses, changing the time horizon, comparator, efficacy of lenalidomide and efficacy of the comparator. The manufacturer also used different discount levels for the price of bortezomib to approximate the complex bortezomib patient access scheme (for which the manufacturer rebates the full cost of bortezomib for people who, after a maximum of 4 cycles of treatment, have less than a partial response). The ICERs ranged from dominant (time horizon reduced from 25 years to either 5 years or 10 years; curve fit for overall survival changed from piecewise exponential to Weibull; source of progression-free survival efficacy changed from Taverna et al. (2012) to Min et al. (2007) to £56,274 per QALY gained (log–normal curve used for progression-free survival and time to treatment failure) for lenalidomide compared with bortezomib re-treatment. Other scenarios submitted by the manufacturer included:

• Reducing the proportion of people receiving lenalidomide as a third-line treatment in the comparator arm. This increased the ICER to £38,330 per QALY gained for lenalidomide compared with bortezomib re-treatment.
• Including a patient access scheme for bortezomib. This increased the ICER to between £21,885 and £27,898 per QALY gained for lenalidomide compared with bortezomib re-treatment.
• Comparing lenalidomide plus dexamethasone with:
  • bendamustine plus prednisolone. This resulted in an ICER of £80,108 per QALY gained.
  • melphalan plus prednisone. This resulted in an ICER of £60,246 per QALY gained.
  • high-dose cyclophosphamide plus dexamethasone. This resulted in an ICER of £67,660 per QALY gained.
  • a blended comparator (that is, a mixture of bortezomib, lenalidomide, bendamustine and standard chemotherapy; the manufacturer assigned proportions based on market share). This resulted in an ICER of £32,462 per QALY gained.

Manufacturer’s revised analyses after clarification round 1 (base-case A)

3.29   In its original submission, when estimating the progression-free survival hazard ratio to determine the relative effectiveness of lenalidomide and bortezomib in the model, the manufacturer had not adjusted the data from MM-010 to reflect the characteristics of the patients from Taverna et al. (2012; see section 3.20), as it had done for overall survival. Therefore, in response to the first round of clarification between the manufacturer and the ERG, the manufacturer adjusted the hazard ratio for patient characteristics, by adjusting for the duration of multiple myeloma observed in the Taverna et al. study. This reduced the hazard ratio from 1.15, which suggested that patients receiving bortezomib progress more rapidly than those receiving lenalidomide, to 0.9, which suggested that patients receiving lenalidomide progress more rapidly than those receiving bortezomib. However, these base-case A analyses showed that lenalidomide dominated bortezomib, that is lenalidomide was more effective and less costly than – bortezomib. The manufacturer explained this counterintuitive result by suggesting that people who receive bortezomib experience a longer duration of progression-free survival having previously responded to bortezomib, but die sooner than those who receive lenalidomide plus dexamethasone second line. The manufacturer stated that a person whose multiple myeloma has previously responded to bortezomib is likely to respond again (delaying progression) but that, because the patient has been exposed to treatment with bortezomib before, the benefit in progression-free survival may not translate to a benefit in overall survival.

3.30   In a scenario analysis of base-case A, the manufacturer combined the results of MM-009 and MM-010 and adjusted the progression-free survival hazard ratio for the characteristics of patients in Taverna et al. (2012). This generated an overall survival hazard ratio of 1.7 and a progression-free survival hazard ratio of 1.35, and ICERs of £12,567 per QALY gained (MM-009 only) and £3122 per QALY gained (trials combined) for lenalidomide compared with bortezomib.

3.31   In other scenario analyses of base-case A, the manufacturer included the scenarios it presented in its original analyses, as well as fitting different curves, using trial data from patients who had 1 prior therapy only, and using MM-009 rather than MM-010 (or the trials combined) to estimate lenalidomide’s clinical effectiveness. The cost-effectiveness estimates for lenalidomide plus dexamethasone compared with bortezomib re-treatment ranged from lenalidomide dominating bortezomib re-treatment to an ICER of £43,331 per QALY gained. The probabilistic results showed that lenalidomide plus dexamethasone dominated re-treating with bortezomib with a probability of cost effectiveness at £20,000 and £30,000 per QALY gained of 74.5% for lenalidomide plus dexamethasone and 85% for bortezomib. However, the manufacturer had excluded the cost of dexamethasone that is given with bortezomib.

Manufacturers revised analyses after clarification round 2 (base-case B)

3.32   During the second round of clarification between the ERG and the manufacturer, the ERG highlighted that the progression-free survival curve crossed the overall survival curve in the manufacturer’s model. This resulted in the model predicting that there were more people who remained progression free than there were people alive. The manufacturer corrected this using the minimum value between overall and progression-free survival to reflect progression-free survival. So, when the model predicted that overall survival was shorter than progression-free survival, the overall survival value was used (that is, progression-free survival was the same as overall survival; all the people alive would be progression free). In addition, to prevent the curves crossing, the manufacturer changed the function it had used to fit curves, choosing a log-logistic function for overall survival instead of a piecewise exponential curve, as in the base-case model. The manufacturer did not change the log-logistic curves it had previously fitted to progression-free survival and time to treatment failure.

3.33   In response to the second clarification in base-case B the manufacturer:

• used the log-logistic curve for overall survival, progression-free survival and time to treatment failure for the lenalidomide arm (and therefore also the comparator arm), to prevent the curves from crossing
• continued as in base-case A to adjust for patient characteristics (duration of multiple myeloma) when deriving the progression-free survival hazard ratio for bortezomib compared with lenalidomide (hazard ratio 0.9 in favour of bortezomib; see sections 3.22 and 3.29).

3.34   The revised probabilistic and deterministic base-case results from base-case B showed that lenalidomide dominated bortezomib. The probability of cost effectiveness was 100% at a threshold of £20,000 per QALY gained. The manufacturer did one-way sensitivity analyses using net monetary benefit (assigning a monetary value for costs and benefits, assuming a maximum acceptable ICER of £30,000 per QALY gained). The results were most sensitive to the hazard ratios reflecting the relative effectiveness of bortezomib and lenalidomide for progression-free survival.

3.35   For base-case B, the manufacturer conducted several scenario analyses. Lenalidomide plus dexamethasone dominated bortezomib re-treatment in the following scenarios when:

• reducing the time horizon
• including concomitant dexamethasone with bortezomib in the base case and in third- and fourth-line treatments
• using different studies to estimate the efficacy of bortezomib
• approximating the complex bortezomib patient access scheme by assuming different levels of discount.

When comparing lenalidomide plus dexamethasone with the other comparators, the manufacturer estimated ICERs that were considerably higher as follows:

• bendamustine plus dexamethasone: £23,435 per QALY gained
• melphalan plus prednisone: £28,516 per QALY gained
• high-dose cyclophosphamide plus low-dose dexamethasone: £36,718 per QALY gained.

Evidence Review Group critique of cost effectiveness

3.36   The ERG stated that the manufacturer had used appropriate search terms and databases in its systematic literature review.

3.37   The ERG reviewed the manufacturer’s approach to the Markov model and highlighted several important errors. Several of these remained unresolved after the 2 rounds of clarification, and are described below.

Data extrapolation

3.38   The manufacturer adjusted progression-free survival, time to progression and overall survival curves from MM-010, using the mean of covariates method, to the means published in Taverna et al. (2012), so that the MM-010 data reflected the patient characteristics from Taverna et al. The ERG noted that the mean of covariates method may skew results, and stated that the manufacturer could have used alternative approaches. Further, the ERG stated that it was not clear why the manufacturer chose the covariates it had chosen because some covariates, such as the ECOG score, were not statistically significant predictors of progression-free survival, time to progression or overall survival. The ERG stated that the manufacturer did not include all predictors in the model noting, for example, that the manufacturer did not include the number of prior therapies.

3.39   The ERG commented that the duration of overall survival associated with lenalidomide was likely to have been overestimated in the manufacturer’s model. The ERG noted that, when the manufacturer chose a log-logistic extrapolation, 11% of people in the model were still alive after 25 years. The ERG commented that because people enteredthe model aged 63 years, the model predicted that 11% of people with multiple myeloma at first relapse live beyond the age of 88 years, which it considered to be an unrealistic assumption.

Comparative effectiveness

3.40   The ERG commented on the manufacturer’s estimate for the hazard ratio that it used in base-cases A and B to compare progression-free survival between bortezomib and lenalidomide. The ERG noted that it had changed from 1.15 (favouring lenalidomide, in the original analyses) to 0.9 (favouring bortezomib in base-case A; see section 3.29). The ERG noted that, despite this favouring bortezomib, after this change, lenalidomide dominated bortezomib (from the original base-case ICER of about £14,500 per QALY gained). The ERG’s clinical specialists advised that this hazard ratio for progression-free survival in favour of bortezomib was not plausible because, in clinical practice, multiple myeloma re-treated with bortezomib would be expected to progress more quickly than when treated with lenalidomide, despite previous response to bortezomib.

Curves crossing

3.41   The ERG considered the manufacturer’s approach to resolving the problem of the progression-free survival and overall survival curves crossing in the model. It noted that the manufacturer had used the minimum value for progression-free and overall survival in its initial base-case analyses to ensure the curves did not cross. However the ERG commented that this does not address why the underpinning survival curves crossed. The ERG noted that, despite the manufacturer changing the curve fitting for overall survival from piecewise exponential to log-logistic in base-case B, the curves continued to cross in the comparator arm of the model. The ERG commented that obtaining the best fit for a curve and the natural history of disease, rather than preventing curves from crossing, should form the basis for selecting a curve.

Model structure

3.42   The ERG noted that there were differences between how the manufacturer had modelled third- and fourth-line treatments in the lenalidomide and the comparator arms of the model, as follows:

• In the lenalidomide arm, when modelling third- and fourth-line treatments, the manufacturer modelled only the costs of the treatments. It did not model disutility or costs of adverse events. However, in the comparator arm, when including lenalidomide as a third-line treatment, the manufacturer included the cost and disutility of adverse events as well as the treatment costs. In addition, overall survival, progression-free survival and time to treatment failure in the comparator arm were then determined from the lenalidomide arm of the MM-010 trial.
• The duration of third-line treatment differed between the intervention and the comparator arms. In the intervention arm the third-line treatment was a mix of chemotherapies that did not include lenalidomide, and the duration of treatment for these was fixed. In the comparator arm, however, the model included treatment with lenalidomide, the duration of which was until treatment failure. This was derived from the progression-free survival data in the lenalidomide plus dexamethasone arm in MM-010.

3.43   The ERG noted that, for the health state defined by ‘progressive disease’, the manufacturer’s model held patients’ utility values constant. However, in this health state, patients could receive third- and fourth-line treatments that, in clinical practice, could result in a remission and an increase in utility. The ERG noted that this benefit was not captured in the manufacturer’s model.

3.44   The ERG commented that the manufacturer’s defined clinical treatment pathway differed from the third- and fourth-line treatments included in the model, and that several modelled treatments were no longer routinely used in clinical practice in the UK. The ERG questioned whether clinicians in the UK would offer third- or fourth-line treatments in clinical practice. Taking this and the other issues into account (see sections 3.42 and 3.43), the ERG therefore questioned whether these treatment lines should be included in the model.

Drug costs

3.45   The ERG noted several problems related to the costs used in the model:

• The manufacturer did not include treatment with dexamethasone with the comparators in the initial base case, base-case A or base-case B.
• The ERG could not determine how the manufacturer estimated the costs of bortezomib per cycle in the model.
• The manufacturer’s model assumed that patients remain on bortezomib second line until disease progression or they develop adverse effects of treatment. However, the ERG’s clinical specialist suggested that bortezomib in the UK is given for a fixed number of cycles (usually a maximum of 8).
• The manufacturer assumed transportation costs for half of patients to attend clinic to receive intravenous bortezomib; however, the ERG’s clinical specialists suggested that the true proportion of patients would be considerably lower.
• The ERG noted that the manufacturer’s calculations for the costs of disease monitoring were not clear, particularly which costs the manufacturer attributed to each state (that is, progression or progression-free).

Health-related quality of life

3.46   The ERG noted that, in the model, the progression-free survival state had a utility value of 0.81. This value is higher than would be expected for an average member for the UK population at the same age (expected to be 0.80).

Sensitivity analysis

3.47   The ERG presented exploratory analyses using the manufacturer’s model for lenalidomide plus dexamethasone compared with bortezomib excluding costs and effectiveness of third- and fourth-line treatments. The ERG assumed that 64% of people who receive bortezomib also receive dexamethasone. Given the ERG’s concerns about the structure and methodology of the manufacturer’s model, the ERG stated that the Committee should interpret the ERG’s results with caution. In the ERG’s analysis excluding third- and fourth-line treatments, lenalidomide plus dexamethasone dominated bortezomib re-treatment. The ERG presented further scenarios (based on the manufacturer’s base-case B model) for lenalidomide plus dexamethasone compared with bortezomib re-treatment:

• correcting an error in the manufacturer’s model related to how patients are allocated to the different health states (lenalidomide dominated)
• using a maximum duration of bortezomib re-treatment in the comparator arm of 8 cycles and correcting a mistake in the values for disutility for adverse events
• generating an ICER that incorporated all these changes of £54,369 per QALY gained.

Manufacturer response to the appraisal consultation document

3.48   The manufacturer submitted additional analyses to address concerns noted by the Committee in the appraisal consultation document. The additional analyses included:

• more evidence relating to the efficacy of lenalidomide compared with bortezomib
• more evidence relating to the efficacy of lenalidomide compared with chemotherapy
• scenarios varying health-related quality of life
• scenarios varying costs (including limiting the number of cycles of bortezomib re-treatment)
• updated cost-effectiveness analyses.

Efficacy of lenalidomide compared with bortezomib

3.49   The manufacturer presented a mixed treatment comparison comparing lenalidomide with first-time (but not first-line) bortezomib treatment for people who had at least 1 prior therapy. The manufacturer stated that the common comparator dexamethasone was used in a connected network of pairwise treatment comparisons. The mixed treatment comparison included the trials MM-009 and MM-010 (see section 3.2), as well as:

• APEX, a phase III multicentre randomised controlled trial of 669 people with multiple myeloma who had received between 1 and 3 prior treatments, which compared bortezomib plus dexamethasone with dexamethasone monotherapy. No one in the trial had received prior bortezomib treatment.
• DOXIL, a phase III multicentre randomised controlled trial of 646 people with multiple myeloma who had received at least 1 prior therapy, which compared bortezomib plus pegylated liposomal doxorubicin with bortezomib monotherapy.

The manufacturer used the mixed treatment comparison to derive hazard ratios for time to progression and overall survival (table 3) using Bucher and Bayesian methods. However, the manufacturer did not use these hazard ratios in its base-case analyses; instead, it provided them as scenario analyses (see section 3.28).

Table 3 Mixed treatment comparison results

	Bucher	Bayesian
Time to progression HR [95% CI]	0.63 [0.42 to 0.92]	0.64 [0.41 to 0.95]
Overall survival HR [95% CI]	0.68 [0.41 to 1.10]	0.72 [0.49 to 1.01]
Abbreviations: CI, confidence interval; HR: hazard ratio.

3.50   The manufacturer presented further evidence that compared lenalidomide with bortezomib after 1 prior treatment with bortezomib using an analysis of subsequent treatments given after the VISTA trial (see section 3.23). The manufacturer did not present a cost-effectiveness analysis that used these data. The proportions responding to lenalidomide, bortezomib and thalidomide estimated by the manufacturer were:

• response to lenalidomide: 73% (16 out of 22 people)
• response to bortezomib re-treatment: 41% (9 out of 22 people)
• response to thalidomide: 37% (23 out of 63 people).

3.51   The manufacturer provided updated survival analyses using pooled data from MM-009 and MM-010 to produce survival models for overall survival, time to progression, progression-free survival and time to treatment failure. The manufacturer identified the Gompertz (overall survival) and gamma (progression-free survival and time to treatment failure) curves as having the best visual fit and clinical plausibility. The manufacturer chose these curves for its updated base-case analysis. The manufacturer noted that the survival curves still crossed (see section 3.32). The models were adjusted using the mean of covariates method in the base-case analysis, and the corrected group prognosis method in a scenario analysis.

3.52   The manufacturer clarified how it used the lenalidomide curves to extrapolate the survival estimates. The manufacturer adjusted the lenalidomide curves to match the study-level patient characteristics of the comparators (for bortezomib, the manufacturer used Taverna et al. (2012) in the base case, in which only the duration of myeloma was available to adjust the pooled data). The adjusted data generated predicted median survival outcomes for lenalidomide, which the manufacturer then compared with the observed median survival outcomes for bortezomib from Taverna et al. to calculate ‘crude’ hazard ratios. The manufacturer also derived what the manufacturer referred to as an ‘unadjusted’ hazard ratio (but was, as the manufacturer stated, adjusted for prognostic factors, but not patient characteristics) by comparing pooled MM-009 and MM-010 data with the observed survival outcomes for bortezomib. The manufacturer used a hazard ratio for overall survival for the base case of its updated model of 1.89 (range 1.57–2.14); and 1.11 (range 0.90–2.06) for progression-free survival (where a value greater than 1 favours lenalidomide, that is the figure shows the comparison of bortezomib compared with lenalidomide).

Efficacy of lenalidomide compared with chemotherapy

3.53   The manufacturer widened its original systematic review and found 2 further studies that estimated the effectiveness of chemotherapy. In its original submission, the manufacturer assumed that the efficacy of chemotherapy was the same as bendamustine (see section 3.20):

• Petrucci et al. (1989), a single-arm trial of melphalan plus prednisolone that included 34 patients with multiple myeloma whose disease had relapsed or was refractory to standard chemotherapy.
• Celesti et al. (1997), a non-randomised trial including 28 people with advanced multiple myeloma that compared high-dose cyclophosphamide plus low-dose dexamethasone with low-dose cyclophosphamide plus dexamethasone.

The manufacturer used these studies to derive hazard ratios for overall survival in the same manner as for bortezomib (see section 3.51). The manufacturer used the same hazard ratio for both overall survival and progression-free survival because the studies did not report on progression-free survival. The overall survival hazard ratio for lenalidomide compared with melphalan plus prednisolone was 4.66 (range 3.85–5.25) and this was used for the base-case analysis. The manufacturer used high-dose cyclophosphamide plus low-dose dexamethasone and low-dose cyclophosphamide plus dexamethasone as comparators in scenario analyses. The manufacturer did not include any other chemotherapeutic agents, such as vincristine and doxorubicin, because of a lack of evidence.

3.54   The manufacturer stated that chemotherapy was not a relevant comparator because:

• People do not receive standard chemotherapy in clinical practice because of advances in treatment, including proteasome inhibitors, immunomodulatory agents and broader use of existing treatments.
• In clinical practice, people for whom bortezomib re-treatment is not appropriate would receive lenalidomide rather than chemotherapy through the Cancer Drugs Fund.
• A survey of 7 haematology consultants performed by the manufacturer estimated that 5–10% of patients receive chemotherapy for second-line treatment.

Health-related quality of life

3.55   The manufacturer submitted 2 additional scenarios for health-related quality of life. The first scenario assumed that health-related quality of life depends on the time since the start of treatment, rather than progression status. The scenario included a range of values from 0.52, for up to 1 month since treatment, to 0.71 for 36 months or longer since treatment. The utility values were sourced from Bortezomib and thalidomide for the first-line treatment of multiple myeloma (NICE technology appraisal 228), which used mapped values (from the EORTC-C30 to the EQ-5D). The second scenario used the utility values in the manufacturer’s original model (that is, were based on progression status rather than time since treatment) but incorporated a lower utility value (0.59), sourced from MM-003 (a randomised open-label trial of pomalidomide plus dexamethasone compared with dexamethasone alone for the treatment of multiple myeloma), for fourth-line treatment onwards.

Updated costs

3.56   The manufacturer submitted updated cost information after consultation. The manufacturer:

• Added the cost of dexamethasone given with bortezomib to the base case for 64.3% of patients (based on estimates from Taverna et al. 2012).
• Updated the method it used to calculate the number of people who survive long enough to become eligible for the patient access scheme for lenalidomide when given third line in the comparator arm (now 15.52% of patients; original estimate not stated).
• Included additional scenario analyses to explore the impact of fewer bortezomib treatment cycles.

3.57   The manufacturer stated that bortezomib is given until treatment failure, noting that up to 19 cycles of bortezomib were given in both Taverna et al. (2012) and Sood et al. (2009) (a study of 32 people with multiple myeloma who had more than 1 treatment bortezomib). It also noted that people whose multiple myeloma responded to treatment received a median of 10 cycles bortezomib in the APEX trial.

Updated cost-effectiveness results – base-case C

3.58   After consultation, the manufacturer submitted 2 new base-case analyses (hereafter both referred to as base-case C), depending on whether re-treating with bortezomib was appropriate or not; both contained additional scenario analyses. The manufacturer made the following changes to the updated base case; it:

• pooled MM-009 and MM-010 data
• corrected the patient flow sheets
• added the cost of concomitant dexamethasone for 63% of people in the bortezomib arm
• changed the method used to estimate the proportion of people surviving long enough to receive the third-line lenalidomide patient access scheme
• added the complex patient access scheme (see section 3.25) available for bortezomib to the comparator arm in the base case.

The manufacturer assumed that not all clinicians prescribing bortezomib for patients eligible for the patient access scheme would successfully claim it. Therefore the manufacturer assumed in the base case a discount of 15% on the cost of bortezomib for 55% of modelled patients.

Manufacturer’s base-case C results after consultation

3.59   Lenalidomide dominated bortezomib (deterministic analysis: incremental costs −£13,634, incremental QALYs 0.81; probabilistic analysis: incremental costs −£14,067, incremental QALYs 0.82). The manufacturer’s one-way sensitivity analysis using net monetary benefit showed a 100% chance of cost effectiveness at £20,000 per QALY gained, with the ICER most sensitive to the hazard ratio for progression-free survival for lenalidomide compared with bortezomib. In a scenario analysis in which bortezomib re-treatment was given for a specific number of cycles, the ICERs were £15,409 per QALY gained (19 cycles) and £31,999 per QALY gained (8 cycles). In another scenario analysis in which the hazard ratio from the manufacturer’s mixed treatment comparison was used for overall survival, lenalidomide dominated bortezomib. When the hazard ratio from the manufacturer’s mixed treatment comparison was used for progression-free survival, the ICER for lenalidomide was £8936 per QALY gained compared with bortezomib.

3.60   When re-treating with bortezomib was not appropriate, and therefore chemotherapy was the comparator, the base-case ICER was £54,898 per QALY gained (incremental costs £54,414, incremental QALYs 0.99). The probabilistic ICER was £53,686 per QALY gained (incremental costs £53,629, incremental QALYs 1.00), with a 0.1% chance of cost effectiveness at £30,000 per QALY gained. The sensitivity analysis showed that the ICER was most sensitive to time to treatment failure.

Evidence Review Group critique of additional analyses presented by manufacturer after consultation

3.61   The ERG critiqued the additional base-case C analyses submitted by the manufacturer after consultation. The critique focused on:

• clinical effectiveness of bortezomib
• clinical effectiveness of chemotherapy
• model structure
• costs
• utility values.

Clinical effectiveness of bortezomib

3.62   Survival – unadjusted curves: The ERG stated that it was not clear why the manufacturer had selected the curves it chose to extrapolate the unadjusted (see section 3.51) pooled MM-009 and MM-010 data for overall survival, progression-free survival and time to treatment failure. The ERG noted that the manufacturer used the Gompertz curve in its base case to extrapolate overall survival, but commented that the gamma curve appeared to fit the trial data equally well based on visual inspection. For progression-free survival and time to treatment failure, the ERG noted that the manufacturer had changed its choice of curve from log-logistic to gamma, but that it was not clear why the manufacturer chose this particular curve because other curves may have been appropriate based on visual inspection and AIC and BIC statistics.

3.63   The ERG noted that the manufacturer had provided a piecewise exponential curve for overall survival in a scenario analysis. However, the ERG noted that the methodology used by the manufacturer to extrapolate the piecewise exponential curve resembled a ‘broken curve’ approach, wherein the extrapolated curve is fitted to a small amount of data at the right tail of the dataset. Consequently, the extrapolated portion is sensitive to the point in the curve where extrapolation begins. The ERG noted that the manufacturer had chosen to extrapolate from 100 weeks, without clearly justifying this time point. The ERG also noted that the manufacturer’s approach to modelling the piecewise exponential curve differed from that used in the manufacturer’s original base case, when the manufacturer used a traditional piecewise approach (that is, fitted separate parametric models to different time periods, allowing the hazard to change over time).

3.64   Prognostic factors chosen by the manufacturer to adjust curve: The ERG raised several concerns with the manufacturer’s process of selecting which variables it would include to adjust curves for the treatment arm of the combined MM-009 and MM-010 trial to reflect the characteristics of the patients in the Taverna et al. (2012) study. These included:

• Concerns with the transparency of the process related to which variables to select:
  • The manufacturer had noted that prognostic factors and outcomes were taken from observational analyses of MM-009 and MM-010, but the ERG did not find any related analyses in the clinical study reports.
  • Some of the variables (for example, the presence or absence of lytic bone lesions) included by the manufacturer from the literature (for example, Dimopoulos [2009]) were not statistically significant predictors of clinical outcomes when analysed in MM-009 and MM-010.
  • The ERG noted discrepancies between p values presented by the manufacturer in its submission and in its model.
  • The ERG noted that the manufacturer did not present some measures of statistical significance (for example, p values) either in its submission or in its model.
• The ERG had concerns with the manufacturer’s rule used to select prognostic factors, that is, only those variables statistically significant (p<0.05) associated with both overall survival and either progression-free survival or time to treatment failure. The ERG noted that:
  •       this introduced bias by eliminating statistically significant predictors of progression-free survival and time to treatment failure
  •       different prognostic factors should be used to adjust the different overall survival, progression-free survival and time to treatment failure curves.

3.65   Survival models – adjusted intervention curves: the ERG commented that the overall survival estimated by the Gompertz curve seemed reasonable because the model estimated that fewer than 1% of patients who take lenalidomide second line would be alive after 12 years. However, the ERG noted that the manufacturer’s model remained flawed because the curve for overall survival still crossed the curves for progression-free survival and time to treatment failure. The ERG noted that survival curves using the Gompertz rather than the piecewise exponential curve now crossed even earlier in the model. For example, in the original base case, in the intervention arm the curves for overall survival and progression-free survival crossed at about 20 years. However, in the manufacturer’s base-case C, the curves crossed at after about 10 years

3.66   Mean of covariates method: the ERG noted that not all the relevant prognostic factors had median and mean values reported in the comparator studies and that the method chosen by the manufacturer needed the mean. For example, Taverna et al. (2012) reported only the median duration of myeloma. The ERG commented that this implied that the manufacturer therefore had to make assumptions, but that the manufacturer did not explicitly define these assumptions.

3.67   Corrected group prognosis method: the ERG noted that the manufacturer had used the corrected group prognosis method as an alternative to the mean of covariate methods to adjust the lenalidomide trial population in scenario analyses. The ERG commented that the manufacturer’s calculations were unclear, particularly about the way it derived weights for each patient.

3.68   Hazard ratio calculation: to estimate the effectiveness of lenalidomide and its comparators indirectly, the manufacturer calculated ‘crude’ hazard ratios. The ERG noted that, to derive the hazard ratio, the manufacturer had assumed that progression and mortality occurred at a constant rate. However, the ERG considered the manufacturer’s approach flawed because the parametric models (Gompertz, gamma) fitted to the MM-010 and MM-009 pooled data do not assume constant hazards. The ERG noted that an appropriate approach is to assume constant hazards only in an exponential model. The ERG also noted a mistake in the manufacturer’s calculation of the crude hazard ratio which, once corrected, changed the progression-free survival hazard ratio from 1.11 to 1.19 (using the mean of covariate method). The ERG corrected the error and changed the reference such that a hazard ratio of less than 0 favoured lenalidomide. This resulted in the manufacturer’s hazard ratios for lenalidomide compared with bortezomib as follows:

• overall survival: 0.53 (mean of covariates method); 0.61 (corrected group prognosis method) (range 0.47–0.64)
• progression-free survival: 0.84 (mean of covariates); 0.95 (corrected group prognosis method) (range 0.42–1.08).

Clinical effectiveness of chemotherapy

3.69   The ERG sought the opinion of a clinical specialist who stated that chemotherapy was likely to be less effective than lenalidomide, and that only about 5% of people with multiple myeloma receive chemotherapy as a second-line treatment.

3.70   Hazard ratios: the ERG noted that the manufacturer used the same methodology to derive the efficacy of lenalidomide compared with chemotherapy as it used to compare lenalidomide with bortezomib re-treatment. Therefore, the ERG’s same concerns apply (see section 3.67). In addition, the ERG noted it was likely that the manufacturer’s estimates of efficacy used for chemotherapy underestimated the true effect. In the papers used to derive the efficacy of chemotherapy (Celesti et al. 1997 and Petrucci et al. 1989), all patients had received prior chemotherapy, and the manufacturer had assumed that prior treatment did not affect the efficacy of chemotherapy. However, a clinical specialist advising the ERG questioned this assumption. The ERG also questioned the validity of assuming that lenalidomide was equally effective for delaying progression, time to treatment failure and death (applying the same hazard ratio to all), given that MM-009 and MM-010 did not support this.

Model structure

3.71   Third- and fourth-line treatments: the ERG understood that the manufacturer had included in its model the possibility that patients receive therapy third and fourth line, reflecting the complex multiple myeloma care pathway. The ERG questioned the value of including these treatments in this model because:

• The data available did not allow evaluation of the effectiveness and quality of life resulting from further treatment options in the lenalidomide arm of the model.
• The effectiveness of lenalidomide third line was assumed to be the same as second line.
• The manufacturer had not dealt with several issues in its model (see sections 3.38–3.46), including:
  • utility of progressive disease
  • inconsistency in treatment arms
  • outdated drugs in the ‘treatment basket’.

Costs

3.72   Bortezomib patient access scheme: a clinical specialist advising the ERG believed that the bortezomib patient access scheme was rarely implemented in clinical practice in England.

3.73    Duration of bortezomib re-treatment: the ERG noted that the manufacturer had cited both the Taverna et al. (2012) and Sood et al. (2009) studies to support that people with multiple myeloma may receive up to 19 cycles of treatment with bortezomib. However, the ERG noted that most studies and clinical experience suggest treatment lasts between 1 and 6 cycles:

• In the Taverna et al. study, although a maximum number of 19 cycles was given, 90.4% of people in the study received fewer than 6 cycles, and only 2% of patients received more than 10 cycles.
• In Sood et al., 50% of patients received 5 cycles, with a range of 1–12 cycles.
• In APEX, 85% of people in the trial received 8 cycles or fewer.
• In Hrusovsky et al. (2010), 43% of patients received 1–3 cycles, 41% of patients received 4–6 cycles and only 6% received more than 10 cycles (maximum 14 cycles).
• The ERG noted that the bortezomib summary of product characteristics states that a maximum number of 8 cycles should be given.

3.74   Transportation costs: the ERG noted that, in the model, half of people receiving intravenous bortezomib treatment needed transport to a hospital or clinic. For those who needed more than 1 treatment per week, the manufacturer assumed that patients were hospitalised for up to 1 week. A clinical specialist advising the ERG stated that this was not realistic, and the ERG determined that the manufacturer was likely to have overestimated the costs of bortezomib.

Utility values

3.75   Utility values dependent on time: the ERG reviewed the utility scenario in which utility values depend on time since treatment. The ERG commented that the manufacturer’s approach changed from Bortezomib and thalidomide for the first-line treatment of multiple myeloma (NICE technology appraisal 228), in which utility was a function of disease state, to this appraisal, in which utility values are a function of time. The ERG also noted that the alternative utility values were from a younger population than that being considered in this appraisal.

3.76   Utility decrement for fourth-line treatment: the ERG noted that the manufacturer had applied a utility decrement to fourth-line treatments, but stated that this did not solve the problem that the model did not include a utility value for ‘disease not yet progressed’ before a utility for ‘progressed disease’ (see section 3.43). The ERG also noted that the manufacturer did not derive its utility estimates for fourth-line treatment from the clinical trial population and there was no information about the age of patients.

Results of Evidence Review Group exploratory analyses

3.77   The ERG presented several results and scenarios for the 2 different subpopulations, that is, people who can receive bortezomib again and people who cannot. The scenarios presented included:

• correcting for errors in hazard ratio calculations
• different duration of bortezomib re-treatment
• whether third- and fourth-line treatment were included
• whether the mean of covariates or corrected group prognosis method of adjustment was used.

The ERG stated that the Committee should interpret all results with caution because the individual and cumulative impacts of the problems noted with the model were unclear.

3.78   The ERG presented updated results for lenalidomide compared with bortezomib in people who could receive bortezomib re-treatment. The ERG corrected for an error in the manufacturer’s base case relating to the hazard ratio calculation. The updated analyses showed that lenalidomide dominated bortezomib. Further scenarios for lenalidomide compared with bortezomib re-treatment, which also accounted for this error, included:

• When bortezomib re-treatment was a maximum of 19 cycles, the ICER was £38,871 per QALY gained (incremental costs £36,673, incremental QALYs 0.94).
• When bortezomib re-treatment was a maximum of 8 cycles, the ICER was £59,856 per QALY gained (incremental costs £56,472, incremental QALYs 0.94).
• When using alternative utility values, lenalidomide dominated bortezomib re-treatment.
• When third- and fourth-line treatments were excluded, the ICER was £11,325 per QALY gained (incremental costs £10,684, incremental QALYs 0.94).
  • When the bortezomib patient access scheme and third- and fourth-line treatments were excluded, the ICER was £5428 per QALY gained (incremental costs £5121, incremental QALYs 0.94).
• When assuming a hazard ratio of 1 and a maximum of 19 cycles of bortezomib re-treatment, the ICER was greater than £1,000,000 per QALY gained.

3.79   The ERG presented the following results for lenalidomide compared with chemotherapy in people who could not receive bortezomib:

• When correcting for an error in the hazard ratio calculation, the ICER was £54,898 per QALY gained (incremental costs £54,414, incremental QALYs 0.99).
• When third- and fourth-line treatments were excluded, and the error was accounted for, the ICER was £43,708 per QALY gained.

3.80   Full details of all the evidence are in the manufacturer’s submission and the ERG report.

4   Consideration of the evidence

4.1   The Appraisal Committee reviewed the data available on the clinical and cost effectiveness of lenalidomide, having considered evidence on the nature of multiple myeloma and the value placed on the benefits of lenalidomide by people with the condition, those who represent them, and clinical specialists. It also took into account the effective use of NHS resources.

4.2     The Committee heard about the experience of patients with multiple myeloma. It heard from a clinical specialist that survival rates improved after the introduction of thalidomide, bortezomib and lenalidomide, but multiple myeloma remains an incurable disease and can be associated with renal failure and anaemia. The Committee heard how important it was for people who are unable to take thalidomide first line to have options for treatment after first relapse with bortezomib; this is true especially for people whose disease has not responded well to bortezomib, or for whom bortezomib re-treatment is not appropriate, such as people who experience adverse reactions with bortezomib. The Committee understood that, for this patient group, in the absence of lenalidomide, the treatment options are limited to standard chemotherapy and bendamustine. The Committee also heard that, in the opinion of the patient expert, using a novel agent earlier in the pathway may provide more benefit than using it later in the pathway. Finally, the Committee recognised that patients value oral treatments, such as lenalidomide.

4.3     The Committee considered the treatment pathway for the population in this appraisal, that is, people with multiple myeloma who are not eligible for stem cell transplantation and cannot receive thalidomide first line because of contraindications, and who therefore have had 1 prior treatment with bortezomib. The Committee heard from a clinical specialist that although lenalidomide and thalidomide are structurally similar, there are people for whom thalidomide treatment is not appropriate but who could take lenalidomide. The Committee heard from the clinical specialist that this population may receive bortezomib re-treatment at first relapse after initial treatment with bortezomib. However, the clinical specialist stated that this may not be appropriate for more than half of this population either because their condition does not respond to bortezomib or because of adverse reactions. The clinical specialist explained that for these people, in the absence of lenalidomide, current treatment options would be limited to standard chemotherapy and bendamustine, so clinicians would value having alternative treatments to offer. The clinical specialist and patient experts stated that bendamustine was rarely offered to patients because it is not licensed for second-line treatment. They noted that it is licensed for first-line treatment but is not used in this way, that it is available through the Cancer Drugs Fund for relapsed multiple myeloma where other treatments are not appropriate, and that it is used in clinical practice as fourth- and fifth-line treatment. The clinical specialist added that, even when bortezomib re-treatment is appropriate, clinicians may prefer to use a drug with a different mode of action after relapse, such as lenalidomide. The Committee understood that lenalidomide as second-line treatment is available through the Cancer Drugs Fund. The Committee concluded that there are 2 subpopulations depending on whether bortezomib re-treatment can or cannot be used, and that these 2 populations have different treatment options. For people who could be treated again with bortezomib, the treatment options include bortezomib re-treatment, lenalidomide and standard chemotherapy. For people for whom bortezomib re-treatment is not an option, treatment options include lenalidomide and standard chemotherapy. Based on the opinion of the clinical specialist, the Committee agreed that it was appropriate not to consider bendamustine as a second-line treatment in this appraisal.

Clinical effectiveness

4.4   The Committee discussed the evidence for lenalidomide compared with placebo and considered it robust because the data were sourced from 2 randomised controlled trials. The Committee agreed that MM-009 and MM-010 had demonstrated the effectiveness of lenalidomide plus dexamethasone compared with placebo plus dexamethasone for progression-free survival, overall survival and time to progression (see sections 3.3 and 3.4), but recognised that not offering treatment (placebo) was not a relevant comparator for the purpose of this appraisal. The Committee also recognised that the population in the trial did not match the population set out in the decision problem for this appraisal because:

• only 2 out of 353 patients had received 1 prior treatment with bortezomib
• the trials’ inclusion criteria did not specify that thalidomide treatment was inappropriate, contraindicated or could not be tolerated
• the trial patients were younger than the multiple myeloma population addressed in this appraisal
• the trials included a high proportion of people who had had 2 or more prior therapies.

The Committee heard from the clinical specialist that, based on her experience, despite these differences the study was generalisable to the population of interest, and that the younger patient age was unlikely to affect the generalisability of the study results. The Committee concluded that the MM-009 and MM-010 trial data show that lenalidomide is more effective than no therapy for the treatment of multiple myeloma after 1 prior therapy.

4.5   The Committee considered the evidence for clinical effectiveness of lenalidomide in people for whom bortezomib re-treatment is appropriate. The Committee noted that the manufacturer estimated the effectiveness of bortezomib re-treatment from small observational studies that did not include control arms. The Committee noted that, after consultation, the manufacturer provided indirect evidence from a mixed treatment comparison, and direct evidence using follow-up data from the bortezomib VISTA trial (see section 3.23). The Committee acknowledged that the mixed treatment comparison appeared to favour lenalidomide over bortezomib for progression-free survival, but not for overall survival. The Committee noted that the VISTA trial investigated bortezomib used first line, and that the follow-up data reflected the post-progression period when bortezomib re-treatment, lenalidomide or other treatments could be used second line. The Committee noted that the follow-up data from the VISTA trial showed higher response rates for lenalidomide than for bortezomib retreatment. The Committee recognised that the patient populations in the mixed treatment comparison and the VISTA trial were not directly relevant to the decision problem. The Committee agreed that the limited evidence base means there is significant uncertainty about the relative effectiveness of lenalidomide compared with bortezomib re-treatment for the treatment of multiple myeloma after 1 prior therapy. When questioned during the Committee meeting, the manufacturer could not provide any information about whether there were ongoing studies directly addressing the decision problem.

4.6   The Committee considered the clinical-effectiveness evidence for lenalidomide in people for whom bortezomib re-treatment was not appropriate (that is, compared with standard chemotherapy). The Committee recognised that there was no direct evidence (see section 3.52), and the manufacturer instead had estimated a crude hazard ratio for overall survival comparing standard chemotherapy with lenalidomide that ranged from 3.85 to 5.19 in favour of lenalidomide. It heard from the clinical specialist that, despite the lack of published efficacy data, in her experience it was more effective to use lenalidomide rather than standard chemotherapy, a view the ERG’s clinical specialist agreed with. The Committee concluded that, although there was significant uncertainty in the evidence, given the significant size of effect in favour of lenalidomide indicated by the hazard ratios, and the opinion of the 2 clinical specialists, lenalidomide is likely to be more effective than standard chemotherapy for the treatment of multiple myeloma in people who have had 1 prior therapy.

Cost effectiveness

4.7   The Committee considered the manufacturer’s economic models (initial base case, base-case A, base-case B and base-case C [the latter provided after the consultation]), and the ERG’s critique.

4.8   The Committee considered the manufacturer’s choice of parametric curves for extrapolating overall survival, noting that it chose a piecewise exponential curve for its initial base case and base-case A, a log-logistic curve for base-case B and a Gompertz curve for base-case C. The Committee recognised that all the curves provided before and after consultation presented fundamental problems because the overall survival and progression-free survival curves crossed, fitted poorly to the Kaplan–Meier data, or generated clinically implausible survival estimates. Therefore, after considering the additional evidence provided during consultation, the Committee remained concerned about the validity of the extrapolation curves used.

4.9   The Committee discussed the data used in the model to derive the hazard ratio for comparing lenalidomide with bortezomib. It understood that the evidence of effectiveness in all 4 base-case analyses for bortezomib did not come from a clinical trial with clearly defined end points, but from a review of patients’ medical records. The Committee agreed that it was not clear whether the definitions of response were the same in the lenalidomide trials as the bortezomib review of medical records, and therefore whether they were comparable. The Committee also understood that the manufacturer had not included a study by Hrusovsky et al. (2010) (which, like Taverna et al. 2012, was a retrospective review of patients’ medical records) in the base case because 26% of patients in the study received concomitant treatment. However, the Committee understood that the inclusion criteria for the 2 studies were similar. The manufacturer explained that the results in the Hrusovsky et al. study were similar to those in the Taverna et al. study, so including them was unlikely to affect the results. The Committee noted that the Hrusovsky et al. study included more patients than the Taverna et al. study but was concerned that the studies may have included the same Swiss patients, so pooling them may have been inappropriate. Further, the Committee was aware of the mixed treatment comparison and data from the VISTA trial provided by the manufacturer during consultation (see section 4.4). It noted that the hazard ratios derived from the mixed treatment comparison showed that lenalidomide was less effective compared with bortezomib re-treatment than when the relative effectiveness was estimated from the Taverna et al. study, and that these hazard ratios were used in scenario analyses of base-case C by the manufacturer. The Committee noted that the mixed treatment comparison was for a different patient population. The Committee recognised that there were limited data available to the manufacturer for this population (see section 4.5). Therefore significant uncertainty remained in the relative effectiveness of lenalidomide compared with bortezomib retreatment in all 4 base cases. The Committee agreed that the additional data provided by the manufacturer after consultation, including the mixed treatment comparison, had not reduced this uncertainty. It concluded that this uncertainty needed to be taken into account in the decision making.

4.10   The Committee discussed how the manufacturer calculated the hazard ratios comparing the effectiveness of lenalidomide with bortezomib. It noted that the crude hazard ratios had been calculated by comparing the medians from each distribution in all base cases. The Committee understood that the assumptions necessary to calculate hazard ratios in this way only hold when using exponential distributions, but that the manufacturer extrapolated progression-free survival and overall survival using a range of curves that were not exponential, including log-logistic, Gompertz and gamma. The Committee therefore concluded that the manufacturer’s approach was not appropriate. The manufacturer acknowledged these limitations, explaining that it chose this approach because the exponential curve was not a good fit. The Committee concluded that this further added to its concerns about the validity of the model.

4.11   The Committee discussed how the manufacturer had adjusted the lenalidomide data to derive the crude hazard ratios. The Committee understood that, when deriving the hazard ratios that compared overall survival, progression-free survival and time to progression of lenalidomide with bortezomib, the manufacturer had adjusted the lenalidomide data to include factors that predicted outcomes, and also to allow for differences in patient characteristics between the studies. The Committee noted the large effect that adjusting for patient characteristics could have, for example, the hazard ratio in base-case A for progression-free survival changed from favouring lenalidomide (without adjustment) to favouring bortezomib (when adjusted). However, for all 4 base cases, it was unclear which variables the manufacturer had chosen, and how it had adjusted the data. The Committee heard from the manufacturer at the second Committee meeting that its original submission stated that variables associated with overall survival, and 1 or the other or both of progression-free survival and time to progression were chosen. However, this had been incorrectly described because only prognostic factors associated with overall survival were chosen. The Committee concluded that this meant that the manufacturer was likely to have adjusted progression-free survival and time to progression for factors that were not associated with these outcomes, and omitted risk factors that were in all 4 base-case analyses. Further, in all 4 base cases, the manufacturer had identified prognostic factors significant for overall survival using data from MM-010 and MM-009, for which no information was reported in Taverna et al. (2012), and had therefore assumed the patient characteristics were the same in Taverna at el. as MM-010 and MM-009. The Committee noted that, following consultation, the manufacturer had provided hazard ratios that were described as ‘unadjusted’, but that although these hazard ratios were not adjusted for patient characteristics, they were adjusted for prognostic factors. The Committee agreed that it was not clear whether there was any value in adjusting the data, given the absence of necessary information, and questioned whether the manufacturer had adjusted the data appropriately. It concluded that these adjustments increase the noise in the manufacturer’s analyses, deepening the uncertainty of the results.

4.12   The Committee considered the manufacturer’s approach to modelling third- and fourth-line treatment. The Committee agreed it was important to consider treatments that would follow second-line treatment, and to include both their costs and effectiveness. It noted that the manufacturer had modelled the costs and effects differently in the lenalidomide and the comparator arms in all 4 base cases presented, and agreed that this was not appropriate and instead costs and effectiveness should be modelled consistently in the lenalidomide and comparator arms. The Committee agreed that it was not clinically plausible that, as someone’s disease progresses, his or her utility value would remain constant despite receiving third- and fourth-line treatments which is likely to increase quality of life (see section 3.43). In addition, the Committee noted that, by including third- and fourth-line treatments, lenalidomide treatment was included in the comparator arm in all 4 base cases, meaning that the model compared lenalidomide with bortezomib followed by lenalidomide. The Committee agreed that this contributed substantially to the costs incurred in the comparator arm, and has an impact on the incremental cost-effectiveness ratio (ICER) in favour of lenalidomide. The Committee also recognised that the included third- and fourth-line treatments were not representative of current clinical practice in the UK. The Committee noted that the manufacturer had explored the impact of third- and fourth-line treatment by providing scenarios that excluded these treatments in the model; this changed the ICER in base-case C to £3900 per quality-adjusted life year (QALY) gained (from lenalidomide dominating bortezomib re-treatment). The ERG presented the same scenario, after making corrections to the model, which resulted in an ICER of £11,300 per QALY gained for lenalidomide compared with bortezomib re-treatment. The Committee agreed that it is appropriate to include third- and fourth-line treatments in the economic analysis, but agreed that here this had introduced further noise into the model. The Committee concluded that the ICERs were sensitive to modelling of third- and fourth-line treatments, and it would need to take this into account when making its decisions.

4.13   The Committee discussed the number of cycles of bortezomib that people receive when re-treated with bortezomib. The Committee noted that, in all 4 manufacturer’s base cases, modelled patients receive bortezomib until disease progression which, the Committee heard from the manufacturer, was for a median of 12 cycles. The Committee was aware that the marketing authorisation for bortezomib limits the number of cycles to 8. It also heard from the clinical specialist that in clinical practice in the UK, the number of bortezomib cycles offered to patients varies from 6 cycles to until treatment progression. However, the Committee noted that, in Taverna et al. (2012), the median number of bortezomib cycles was 3, that 90% of people received 6 or fewer cycles, and, as stated by the manufacturer in the second Committee meeting, only 1 person received 19 cycles. The Committee agreed that the manufacturer’s base-case model markedly overestimated the number of cycles of bortezomib used in clinical practice in the UK, and that the number of cycles had a substantial impact on the cost effectiveness of lenalidomide. The Committee noted that, when the manufacturer limited the number of cycles of bortezomib to 19 in a scenario analysis provided after consultation (base-case C), the ICER changed from lenalidomide dominating bortezomib re-treatment to £15,400 per QALY gained. The Committee noted that reducing the maximum number of cycles of bortezomib to 8 increased the ICER to about £32,000 per QALY gained for lenalidomide compared with bortezomib in the manufacturer scenario analyses (base-case C) (see section 3.58). The Committee concluded that, although there may be people in clinical practice who receive more than 8 cycles of bortezomib, there are also people who receive fewer, and that, of all the scenarios presented, the one that had a maximum of 8 cycles of bortezomib re-treatment most plausibly reflected clinical practice.

4.14   The Committee noted 1 comment received during consultation that the patient access scheme for bortezomib is rarely implemented in clinical practice. It understood that patient access schemes are pricing agreements between the manufacturer and the Department of Health. The Committee acknowledged that there may be some uncertainty about the actual costs of bortezomib paid in the NHS, but that it had no authority to assume a different cost than that agreed between the manufacturer and the Department of Health. The Committee noted comments from consultation that bortezomib can be administered as a subcutaneous injection, but that the manufacturer’s model included only intravenous administration. The Committee agreed that, if some patients received bortezomib subcutaneously rather than intravenously, the costs of administration and transportation, and the disutility associated with intravenous therapy would fall, and the cost effectiveness of lenalidomide relative to bortezomib would likely worsen.

4.15   The Committee discussed the sources of the utility values and utility decrements used by the manufacturer in its model. The Committee noted that the source of the utility values used by the manufacturer in all its base cases was van Agthoven et al. (2004), derived from a 2002 PhD thesis which, to the Committee’s knowledge, has never been published in a peer-reviewed journal. The Committee also noted that the utility values were derived from a younger population, and were higher than the average population of the same age. In addition, the manufacturer took the utility decrements for adverse events from several different sources, which used different methods, were from other countries, and included people with different types of cancers. The Committee agreed that the utility scenarios provided by the manufacturer after consultation were of limited value because of the way in which they were applied in the model. The Committee concluded that, although the utility values from all base cases had been used in previous appraisals for multiple myeloma, significant uncertainty remained in how utility and disutility values affected the model outcomes.

4.16   The Committee discussed whether it could determine a preferred model version and most plausible ICERs. It acknowledged that there were limited data available to the manufacturer, and that the manufacturer had adjusted the model in an attempt to make it more realistic in all 4 base cases. However, the Committee agreed that this did not reduce the high levels of uncertainty associated with survival extrapolation, adjusting median survival outcomes, and modelling third- and fourth-line treatments. The Committee also agreed this uncertainty had not been resolved by the additional analyses provided by the manufacturer at consultation. In considering the cost effectiveness of:

• Lenalidomide compared with bortezomib re-treatment: the Committee considered that the most useful base case for decision-making was one in which the manufacturer had corrected an error in the hazard ratio calculation, and for which the ERG provided scenario analyses assessing the effect of different curves (base-case C). It noted that the ICER presented by the manufacturer (which included a maximum of 8 cycles of bortezomib) was £32,000 per QALY gained for lenalidomide compared with bortezomib re-treatment, and that when incorporating the ERG corrections this ICER increased to £60,000 per QALY gained. The Committee noted that the ICER was sensitive to applying different extrapolation models that could increase or decrease the ICER. The Committee also noted that taking into account the uncertainties relating to third- and fourth-line treatments, hazard ratios and bortezomib costs would increase the ICER,noting that the ICER was higher than £1 million per QALY gained when the effectiveness of lenalidomide and bortezomib were assumed to be the same.
• Lenalidomide compared with standard chemotherapy: The Committee noted that the most recently presented ICER (base-case C) was about £55,000 per QALY gained for lenalidomide compared with standard chemotherapy (see section 3.59), but agreed that this value was very uncertain.

The Committee concluded that the ICERs for lenalidomide compared with bortezomib re-treatment or with standard chemotherapy were more than £30,000 per QALY gained. It therefore concluded that it could not recommend lenalidomide as a cost-effective use of NHS resources for people with multiple myeloma for whom thalidomide treatment is not appropriate, and who have received 1 prior treatment with bortezomib.

4.17   The Committee discussed whether lenalidomide is innovative in making a significant and substantial impact on health-related benefits. It agreed that, as an oral treatment, lenalidomide would be convenient, and could save time and resources for people with multiple myeloma who had had 1 prior treatment with bortezomib, and that this may not be included in the ICERs presented. However, the Committee also appreciated that bortezomib can be administered subcutaneously, lessening the difference in convenience between treatments. The Committee concluded that it was unclear how these aspects could be modelled, and that it was unlikely to make a substantial difference to its conclusions considering the high level of uncertainty.

4.18   The Committee considered whether lenalidomide treatment meets the end-of-life criteria for people with multiple myeloma who have had 1 prior treatment with bortezomib, and for whom thalidomide treatment and stem cell transplant are not appropriate. It noted that the manufacturer had not presented data to support lenalidomide considered as an end-of-life therapy. It noted that the number of life years estimated in the comparator arm of the model was greater than 24 months, and therefore concluded that, based on the data available, the end-of-life criteria had not been met. The Committee agreed that it did not need to discuss the remaining criteria.

Summary of Appraisal Committee’s key conclusions

	Appraisal title:		Section
Key conclusion
Lenalidomide in combination with dexamethasone is not recommended within its marketing authorisation for treating multiple myeloma in people: whose condition has relapsed for the first time, and who have had 1 prior treatment with bortezomib, and for whom thalidomide is contraindicated or cannot be tolerated and for whom stem cell transplantation is not appropriate. The Committee concluded that there was uncertainty in the relative effectiveness (progression-free survival and overall survival) of lenalidomide compared with bortezomib re-treatment because direct trial data were not available, the effectiveness of bortezomib re-treatment was estimated from single-arm trials, data were conflicting in terms of whether it favoured lenalidomide or bortezomib, and indirect data were not from an appropriate population. The Committee concluded that lenalidomide was likely to be more effective than standard chemotherapy because, although the data were taken from single-arm studies, the data consistently favoured lenalidomide with a large benefit, and this was supported by clinical specialist opinion. The Committee concluded that the incremental cost-effectiveness ratios (ICERs) for lenalidomide compared with bortezomib re-treatment or with standard chemotherapy were more than £30,000 per quality-adjusted life year (QALY) gained.			1.1 4.3 4.15
Current practice
Clinical need of patients, including the availability of alternative treatments		The Committee heard how important it was for people who are unable to take thalidomide first line to have options for treatment after first relapse with bortezomib, especially for people whose disease has not responded well to bortezomib, or for whom bortezomib re-treatment is not appropriate, such as people who experienced adverse reactions with bortezomib.	4.2
The technology
Proposed benefits of the technology How innovative is the technology in its potential to make a significant and substantial impact on health-related benefits?		The Committee agreed that, as an oral treatment, lenalidomide would be convenient and could save time and resources. However, the Committee also noted that bortezomib can be administered subcutaneously, lessening the difference between treatments in terms of convenience.	4.16
What is the position of the treatment in the pathway of care for the condition?		Lenalidomide in combination with dexamethasone has a marketing authorisation for treating multiple myeloma in adult patients who have received at least 1 prior therapy.	2.1
Adverse reactions		The summary of product characteristics includes the following adverse effects for lenalidomide: neutropenia, anaemia and thrombocytopenia. Because lenalidomide is structurally related to thalidomide, a known human teratogen that causes severe birth defects, a risk minimisation plan has been developed and agreed with the Medicines and Healthcare products Regulatory Agency to avoid fetal exposure to lenalidomide.	2.2
Evidence for clinical effectiveness
Availability, nature and quality of evidence		The manufacturer presented evidence from randomised clinical trials, MM-009 and MM-010, to show the effectiveness of lenalidomide. However, the trials were placebo controlled, and as no treatment was not a comparator in the decision problem, only data from the lenalidomide arm of the trials were used. The Committee noted that the manufacturer estimated the effectiveness of bortezomib re-treatment from small observational studies, and that these studies did not include a control arm, preventing any direct or indirect comparisons. After consultation, the manufacturer provided indirect evidence using a mixed treatment comparison and direct evidence using follow-up data from the bortezomib VISTA trial. However, the populations of these were not directly relevant to the decision problem. The Committee agreed that limited data exist for lenalidomide compared with bortezomib in this setting. The Committee recognised that the manufacturer had presented no direct evidence comparing lenalidomide with standard chemotherapy. The Committee concluded that there was significant uncertainty in the evidence.	3.23 3.55 4.4 4.6 4.5 3.3 3.4
Relevance to general clinical practice in the NHS		The Committee recognised that the population in the MM-009 and MM-010 trials did not match the population set out in the decision problem for this appraisal. However, the Committee heard from the clinical specialist that, based on her experience, despite these differences, the trials were generalisable to the population of interest.	4.6
Uncertainties generated by the evidence		There was uncertainty in the relative effectiveness (progression-free survival and overall survival) of lenalidomide compared with bortezomib re-treatment or standard chemotherapy. This was determined by estimating crude hazard ratios that used data from different trials for each treatment because no direct or indirect evidence was available. In addition, these hazard ratios were calculated using median values. The assumptions needed to use median values in this way only hold when using an exponential distribution, however, the exponential distribution was not used. These hazard ratios were therefore highly uncertain.	4.4 4.5 4.6 4.10
Are there any clinically relevant subgroups for which there is evidence of differential effectiveness?		The Committee concluded that there were 2 subpopulations depending on whether bortezomib re-treatment was or was not appropriate, and that these 2 populations had different treatment options.	4.3
Estimate of the size of the clinical effectiveness including strength of supporting evidence		Based on the evidence provided by the manufacturer, the Committee could not conclude that lenalidomide second line was more effective than bortezomib re-treatment for the treatment of multiple myeloma after 1 therapy. The Committee concluded that, although there was significant uncertainty in the evidence, lenalidomide was likely to be more effective than standard chemotherapy for the treatment of multiple myeloma in those who have had 1 prior therapy.	4.4 4.5 4.6
How has the new clinical evidence that has emerged since the original appraisal (TA171) influenced the current (preliminary) recommendations?		The same clinical trials as in TA171 (MM-009 and MM-010) were presented to demonstrate the effectiveness of lenalidomide plus dexamethasone and placebo plus dexamethasone. The evidence presented in this appraisal included an extended follow-up of overall survival, which was not included in NICE technology appraisal guidance 171. This added further support to the evidence that lenalidomide was clinically effective compared with placebo. Additional non-comparative observational studies have been published for the comparators.	4.4 4.5 4.6
Evidence for cost effectiveness
Availability and nature of evidence		The manufacturer’s Markov model compared lenalidomide with bortezomib re-treatment and with chemotherapy. The manufacturer presented its original base case in the manufacturer’s submission. Following the clarification process with the ERG, the manufacturer submitted base-case A and then base-case B, each with different extrapolation methods. Following additional concerns noted by the Committee, the manufacturer provided further analyses, base case-C, after consultation.	3.19 3.26
Uncertainties around and plausibility of assumptions and inputs in the economic model		The Committee acknowledged that there were limited data available to the manufacturer, and that the manufacturer had adjusted the model in an attempt to make it more realistic. However, the Committee agreed that this did not reduce the high levels of uncertainty, noting concerns around: survival extrapolation bortezomib effectiveness estimates adjustments modelling of third- and fourth-line treatments the number of bortezomib re-treatment cycles cost associated with administration of bortezomib utility values.	4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15
Incorporation of health-related quality-of-life benefits and utility values Have any potential significant and substantial health-related benefits been identified that were not included in the economic model, and how have they been considered?		The Committee noted that the source of the utility values used in all the base cases was van Agthoven et al. (2004), derived from a 2002 PhD thesis that, to the Committee’s knowledge, had not been published in a peer-reviewed journal. The Committee also noted that the utility values were derived from a younger population andwere higher than the average population of the same age. In addition, the manufacturer took the utility decrements for adverse events from several different sources, which used different methods, were from other countries and included people with different types of cancers. The Committee concluded that, although the values had been used in previous appraisals for multiple myeloma, there was significant uncertainty in the values of utility and disutility used in the modelling. The Committee agreed that the benefit of oral treatment was not included in the analysis. However, the Committee concluded that it was unclear how this could be modelled, and that it was unlikely to make a substantive difference to its conclusions considering the high level of uncertainty.	4.11 4.14 4.16
Are there specific groups of people for whom the technology is particularly cost effective?		See above for subgroups.	4.3
What are the key drivers of cost effectiveness?		The Committee noted that the key drivers of cost effectiveness were the hazard ratio for the relative effectiveness of lenalidomide compared with bortezomib or standard chemotherapy, the extrapolation and the adjustment of the curves applied for progression-free survival and overall survival, the way in which third- and fourth-line treatments had been modelled, and the number of cycles of bortezomib re-treatment.	4.15
Most likely cost-effectiveness estimate (given as an ICER)		The Committee concluded that, for both groups, the ICERs for lenalidomide compared with bortezomib re-treatment or treating with standard chemotherapy were more than £30,000 per QALY gained. It therefore concluded that lenalidomide could not be recommended as a cost-effective use of NHS resources for people with multiple myeloma for whom thalidomide treatment and stem cell transplant were not appropriate, and who had received 1 prior treatment with bortezomib.	4.15
How has the new cost-effectiveness evidence that has emerged since the original appraisal (NICE technology appraisal guidance 171) influenced the current (preliminary) recommendations?		The manufacturer identified no new health economic studies from the literature, but presented 4 iterations of a new Markov model, and data on the effectiveness of the comparators from small observational studies. The preliminary recommendations for lenalidomide for treating multiple myeloma after its first relapse have not changed from TA171.	3.19 4.15
Additional factors taken into account
Patient access schemes (PPRS)		Not applicable.
End-of-life considerations		The Committee noted that the manufacturer had not presented data to support lenalidomide qualifying as an end-of-life therapy. The Committee recognised that the population size was likely to be small. The Committee noted that the number of life years estimated in the comparator arm of the model were greater than 24 months, and therefore concluded that, based on the data available, the end-of-life criteria had not been met.	4.17
Equalities considerations and social value judgements		Not applicable.

5   Implementation

Section 7(6) of the National Institute for Health and Care Excellence (Constitution and Functions) and the Health and Social Care Information Centre (Functions) Regulations 2013 requires clinical commissioning groups, NHS England and, with respect to their public health functions, local authorities to comply with the recommendations in this appraisal within [insert number] months of its date of publication. The normal period of compliance, of 3 months, has been extended for this technology because [insert reason]. This extension is made under Section 7(5) of the Regulations.

5.1   NICE has developed tools [link to www.nice.org.uk/guidance/TAXXX] to help organisations put this guidance into practice (listed below). [NICE to amend list as needed at time of publication]

• Slides highlighting key messages for local discussion.
• Costing template and report to estimate the national and local savings and costs associated with implementation.
• Implementation advice on how to put the guidance into practice and national initiatives that support this locally.
• A costing statement explaining the resource impact of this guidance.
• Audit support for monitoring local practice.

6   Recommendations for further research

6.1   The Committee heard from the manufacturer that registry data were available for multiple myeloma in the USA, and that a study had been recently completed comparing lenalidomide plus low-dose dexamethasone with high-dose dexamethasone in people who had multiple myeloma.

6.2   The Committee noted that, to reduce uncertainty, it would be valuable to have a trial comparing lenalidomide with bortezomib and standard chemotherapy for the treatment of multiple myeloma in people whose multiple myeloma has relapsed for the first time, who have received 1 prior treatment with bortezomib, for whom thalidomide is contraindicated or cannot be tolerated and for whom bone marrow transplantation is not appropriate.

7   Related NICE guidance

Details are correct at the time of consultation and will be removed when the final guidance is published. Further information is available on the NICE website.

Published

• Bortezomib for induction therapy in multiple myeloma before high-dose chemotherapy and autologous stem cell transplantation. NICE technology appraisal guidance 311 (2014).
• Bortezomib and thalidomide for the first-line treatment of multiple myeloma. NICE technology appraisal guidance 228 (2011).
• Lenalidomide for the treatment of multiple myeloma in people who have received at least one prior therapy. NICE technology appraisal guidance 171 (2009).
• Bortezomib monotherapy for relapsed multiple myeloma. NICE technology appraisal guidance 129 (2007).
• Improving outcomes in haematological cancer. NICE cancer service guidance (2003).

Under development

• Myeloma: diagnosis and management of myeloma. NICE clinical guideline, publication expected January 2016.

8   Proposed date for review

8.1   NICE proposes that the guidance on this technology is incorporated, verbatim, into the upcoming NICE clinical guideline on myeloma, at which point an appropriate review date will also be considered. NICE welcomes views on this proposed date. The Guidance Executive will decide whether the technology should be reviewed based on information gathered by NICE, and in consultation with consultees and commentators.

Amanda Adler
Chair, Appraisal Committee
July 2014

9Appraisal Committee members, guideline representatives and NICE project team

9.1                 Appraisal Committee members

The Appraisal Committees are standing advisory committees of NICE. Members are appointed for a 3‑year term. A list of the Committee members who took part in the discussions for this appraisal appears below. There are 4 Appraisal Committees, each with a chair and vice chair. Each Appraisal Committee meets once a month, except in December when there are no meetings. Each Committee considers its own list of technologies, and ongoing topics are not moved between Committees.

Committee members are asked to declare any interests in the technology to be appraised. If it is considered there is a conflict of interest, the member is excluded from participating further in that appraisal.

The minutes of each Appraisal Committee meeting, which include the names of the members who attended and their declarations of interests, are posted on the NICE website.

Dr Amanda Adler (Chair)
Consultant Physician, Addenbrooke's Hospital

Professor Ken Stein (Vice Chair)
Professor of Public Health, University of Exeter Medical School

Dr Ray Armstrong
Consultant Rheumatologist, Southampton General Hospital

Dr Jeff Aronson
Reader in Clinical Pharmacology, Nuffield University Department of Primary Health Care Sciences, University of Oxford

Professor John Cairns
Professor of Health Economics Public Health and Policy, London School of Hygiene and Tropical Medicine

Matthew Campbell-Hill
Lay member

Mark Chapman
Health Economics and Market Access Manager, Medtronic UK

Professor Imran Chaudhry
Lead Consultant Psychiatrist and Deputy Associate Medical Director, Lancashire Care NHS Foundation Trust

Dr Lisa Cooper
Echocardiographer, Stockport NHS Foundation Trust

Dr Maria Dyban
GP

Robert Hinchliffe
HEFCE Clinical Senior Lecturer in Vascular Surgery and Honorary Consultant Vascular Surgeon, St George's Vascular Institute

Dr Neil Losson
GP

Anne Joshua
Associate Director of Pharmacy, NHS Direct

Dr Miriam McCarthy
Consultant, Public Health, Public Health Agency, Northern Ireland

Professor Ruairidh Milne
Director of Strategy and Development and Director for Public Health Research at the National Institute for Health Research (NIHR) Evaluation, Trials and Studies Coordinating Centre at the University of Southampton

Dr Peter Norrie
Principal Lecturer in Nursing, DeMontfort University

Christopher O’Regan
Head of Health Technology Assessment & Outcomes Research, Merck Sharp & Dohme

Dr John Pounsford
Consultant Physician, Frenchay Hospital, Bristol

Dr Danielle Preedy
Lay member

Cliff Snelling
Lay member

Professor Andrew Stevens
Professor of Public Health, Department of Public Health and Epidemiology, University of Birmingham

David Thomson
Lay member

Dr Nicky Welton
Senior Lecturer in Biostatistics/Health Technology Assessment, University of Bristol

Dr Nerys Woolacott
Senior Research Fellow, Centre for Health Economics, University of York

9.2                 NICE project team

Each technology appraisal is assigned to a team consisting of 1 or more health technology analysts (who act as technical leads for the appraisal), a technical adviser and a project manager.

Carl Prescott
Technical Lead

Melinda Goodall
Technical Adviser

Jeremy Powell
Project Manager

10                  Sources of evidence considered by the Committee

A. The Evidence Review Group (ERG) report for this appraisal was prepared by Peninsula Technology Assessment Group (PenTAG):

·      Bacelar M, Durand A, Cooper C et al. (2014) The clinical and cost effectiveness of lenalidomide for people who have received at least one prior therapy with bortezomib (partial review of TA171).

B. The following organisations accepted the invitation to participate in this appraisal as consultees and commentators. They were invited to comment on the draft scope, the ERG report and the appraisal consultation document (ACD). Organisations listed in I were also invited to make written submissions. Organisations listed in II and III had the opportunity to give their expert views. Organisations listed in I, II and III also have the opportunity to appeal against the final appraisal determination.

I. Manufacturer/sponsor:

·      Celgene

II. Professional/specialist and patient/carer groups:

·      Myeloma UK

III. Other consultees:

·      Cancer Research UK

·      British Society of Haematology

·      Royal College of Nursing

·      Royal College of Physicians

·      UK Myeloma Forum

·      Department of Health

·      NHS England

·      Welsh Government

IV. Commentator organisations (did not provide written evidence and without the right of appeal):

·      Commissioning Support Appraisals Service

·      Department of Health, Social Services and Public Safety for Northern Ireland

·      Napp

·      National Collaborating Centre for Cancer

C. The following individuals were selected from clinical specialist and patient expert nominations from the consultees and commentators. They gave their expert personal view on Celgene by attending the initial Committee discussion and providing written evidence to the Committee. They are invited to comment on the ACD.

·      Dr Faith Davies, Faculty and Haematology Consultant, Institute of Cancer Research and Royal Marsden Hospital Section of Haemato-oncology, nominated by the Royal College of Physicians on behalf of NCRI/RCP/RCR/ACP/JCCO – clinical specialist

·      Judy Dewinter, Chairman, Myeloma UK, nominated by Myeloma UK – patient expert

·      Eric Low, Chief Executive, Myeloma UK, nominated by Myeloma UK – patient expert

D. Representatives from the following manufacturer/sponsor attended Committee meetings. They contributed only when asked by the Committee chair to clarify specific issues and comment on factual accuracy. They were also invited to comment on the ACD.

·      Celgene