The Appraisal Committee considered evidence from a number of sources.
The Assessment Group considered the clinical effectiveness of increased doses of imatinib (600 or 800 mg/day) compared with sunitinib or best supportive care for the treatment of people with unresectable and/or metastatic GISTs whose disease had progressed on 400 mg/day imatinib. Studies reporting the clinical effectiveness of comparator treatments (sunitinib plus best supportive care or best supportive care alone) were also identified.
The Assessment Group did not identify any randomised controlled trials or non-randomised studies comparing the effectiveness of increased doses of imatinib (600 or 800 mg/day) with sunitinib or best supportive care. However, it identified 6 papers and ten abstracts reporting 4 separate clinical trials and 1 non-randomised retrospective cohort study that included a treatment arm of 400 mg/day imatinib and reported data separately for people who initially received 400 mg/day imatinib, then received an increased dose of imatinib when disease progressed ('crossover groups'). The outcomes for people who took increased doses of imatinib informed this appraisal.
The EORTC-ISG-AGITG randomised controlled trial (Zalcberg et al. 2005; Debiec-Rychter et al. 2006) compared 800 mg/day imatinib with 400 mg/day imatinib in people with advanced GISTs (473 people from total trial population of 946 participants). When disease progressed in people randomised to 400 mg/day imatinib the dose was increased to 800 mg/day imatinib (n=133 out of 473 people). Interim response data were reported for 97 people from this trial (Zalcberg et al. 2004). The S0033 randomised controlled trial (Blanke et al. 2008a) also compared 800 mg/day imatinib with 400 mg/day imatinib in people with advanced GISTs (345 people from total trial population of 746 participants). One hundred and seventeen people originally randomised to 400 mg/day who received an increased dose of 800 mg/day imatinib after disease progression were assessed for response. Interim response data (Rankin et al. 2004) were reported for 68 people. The B2222 randomised controlled trial (Blanke et al. 2008) compared 400 mg/day imatinib (n=73) with 600 mg/day imatinib (n=74) in people with advanced GISTs. People with disease progression on 400 mg/day imatinib, but who could tolerate a higher imatinib dose, received an increased dose of 600 mg/day. Forty-three people originally randomised to 400 mg/day imatinib and 13 people originally randomised to 600 mg/day imatinib received increased doses of imatinib. A retrospective cohort study (n=24; Park et al. 2009) reported data for people with metastatic or unresectable GISTs. People in this study received an initial dose of 400 mg/day imatinib that was increased to 600 or 800 mg/day imatinib on disease progression.
The Assessment Group identified 7 abstracts that provided interim results for an ongoing, non-comparative open-label trial on the effectiveness of sunitinib in people whose condition had failed to respond to treatment with different doses of imatinib (including doses up to 400 mg/day). An abstract by Seddon et al. (2008; n=1,117) was considered by the Assessment Group to be the primary source for this trial. The Assessment Group also used data from Reichardt et al. (2008) to obtain overall survival rates for sunitinib and data from Prior et al. (2009) and Demetri et al. (2006) to obtain response rates for sunitinib.
Overall, 28.1% (EORTC-ISG-AGITG), 34.2% (S0033) and 58.9% (B2222) of people initially randomised to receive 400 mg/day imatinib had disease progression and received either 600 or 800 mg/day imatinib. All the people in the retrospective cohort study received an increased dose of imatinib. In the sunitinib study (Seddon et al. 2008), 31.4% of people received sunitinib after disease progression on 400 mg/day imatinib.
The results for people treated with an increased dose of 600 mg/day imatinib after disease progression on 400 mg/day imatinib were reported in the B2222 trial and the retrospective cohort study. In the B2222 trial, the median length of follow-up was 63 months (maximum 71 months) from randomisation. After disease progression, 11 out of 43 people (25.6%) who increased to 600 mg/day imatinib either had a partial response or had stable disease. The Assessment Group noted that some of the people who received increased doses may have had an initial response to 400 mg/day imatinib before disease progression. In the retrospective cohort study the median length of follow-up was 8 months (range 1.4 to 22.3 months). Of the 12 people who received 600 mg/day imatinib, 5 (41.7%) either had a partial response or had stable disease after treatment.
The results for people treated with an increased dose of 800 mg/day imatinib after disease progression on 400 mg/day imatinib were reported in the S0033 trial, the EORTC-ISG-AGITG trial, and the retrospective cohort study. Of the subgroups who received increased doses in the S0033 and EORTC-ISG-AGITG trials (117 out of 345, and 133 out of 473 people respectively), 3 people in each trial had a partial response to treatment, while 33 and 36 people respectively achieved stable disease as a best response. Out of a total of 250 people in the 2 studies combined, 75 (30%) met the criteria for a response to treatment after an increase in the dose of imatinib from 400 to 800 mg/day. The retrospective cohort study reported that 4 of the 12 people (33.3%) who received an increased dose of imatinib (800 mg/day) after disease progression either had a partial response or had stable disease.
The manufacturer of imatinib reported confidential data from a meta-analysis of the S0033 and EORTC-ISG-AGITG trials in their submission, which reported the response to treatment in people who received increased doses of imatinib. However, the Assessment Group did not review these data because of differences in the numbers of people who achieved stable disease or had a partial response after increased doses of imatinib compared with the results from the same studies available as published articles.
Interim data for the EORTC-ISG-AGITG trial reported that, of 97 people whose disease had progressed on 400 mg/day imatinib, 2 (2.1%) showed a partial response, 30 (30.9%) had stable disease, and 65 (67.0%) still had progressive disease after an increase in the dose of imatinib from 400 to 800 mg/day. Interim data from the S0033 trial showed that, of 68 people whose disease had progressed on 400 mg/day imatinib, 5 (7.4%) had a partial response and 20 (29.4%) had stable disease after an increase in the dose of imatinib from 400 to 800 mg/day.
In addition, a secondary analysis of the EORTC-ISG-AGITG trial (Debiec-Rychter et al. 2006) suggested (without stating the number of people involved) that response to treatment after increasing the dose of imatinib from 400 to 800 mg/day was significantly more likely to occur in people with wild-type GISTs compared with the KIT exon 11 mutation (p=0.0012). Response after increasing the dose of imatinib was also significantly more likely to occur in people with the KIT exon 9 mutation compared with the exon 11 mutation (p=0.0017). The Assessment Group noted that it was outside the remit of this appraisal to consider outcomes for patients receiving escalated doses other than after disease progression on an initial dose of 400 mg/day imatinib. In the Debiec-Rychter et al. study, the dose of imatinib was increased to 800 mg/day shortly after starting on 400 mg/day imatinib, not necessarily after disease progression. Because of the lack of data available from the study, it was not possible for the Assessment Group to analyse the impact of increased doses of imatinib after disease progression on 400 mg/day imatinib for people with specific KIT exon mutations.
Response rates to treatment with 400 mg/day imatinib followed by 50 mg/day sunitinib on disease progression were not reported in the studies identified by the Assessment Group. The Assessment Group instead calculated a weighted average response rate from 2 studies (Demetri et al. 2002; Prior et al. 2009), but noted that there were differences between the patient groups in these 2 studies. The Prior et al. study did not report the dose of imatinib that people received before starting on sunitinib. In Demetri et al., people were randomised to receive sunitinib or placebo after treatment failure with a median dose of 800 mg/day imatinib (300 to 1,600 mg/day). Across these 2 studies 266 of a total of 382 people had a response to treatment, and a simple weighted mean was used to derive the pooled response rate (69.6%). This response rate was assumed to be unaffected by prior treatment received. The Assessment Group determined that there was no statistically significant difference in the response rates between these 2 studies.
The S0033 study reported the overall survival for people who were randomised to an initial dose of 400 mg/day imatinib which was increased to 800 mg/day after disease progression. The median follow-up for the study was 4.5 years (follow-up calculated from start of increased dose following disease progression at 400 mg/day imatinib), during which time 76 of 118 people (64.4%) had died. The median overall survival was 19 months (95% confidence interval [CI] 13 to 23 months) from the point at which the dose of imatinib was increased. Interim data for the S0033 trial were also provided by Rankin et al. (2004), who reported that median overall survival after the increase in dose of imatinib was 19 months.
Two abstracts with different follow-up periods for the same study reported overall survival data for people receiving 50 mg/day sunitinib after disease progression or for people who could not tolerate imatinib at a dose of up to 400 mg/day. Reichardt et al. (2008) analysed data after a median of 24 weeks, at which point 231 of 339 people (68.1%) were still alive. Seddon et al. (2008) analysed data after a median of 51 weeks (range 0.1 to 159 weeks) and 193 of 351 people (55.0%) were still alive. The Assessment Group used data on the proportion of people still alive (Reichardt et al.) and the median follow-up (Seddon et al.) in its analysis.
Progression-free survival data were not published for the B2222 trial for people receiving an initial dose of 400 mg/day imatinib which was increased to 600 mg/day after disease progression.
The S0033 and the EORTC-ISG-AGITG trials reported data on progression-free survival for people randomised to an initial dose of 400 mg/day imatinib, which was increased to 800 mg/day after disease progression. For the S0033 trial, at a median follow-up of 54 months from randomisation, the disease had progressed in 99 of 118 people (83.9%) who received 800 mg/day imatinib. Median progression-free survival was estimated to be 5 months (95% CI 2 to 10 months). Using interim data from this trial for 68 people (Rankin et al. 2004), the authors estimated a median progression-free survival after crossover of 4 months. For the EORTC-ISG-AGITG trial, at a median follow-up of 25 months (maximum follow-up 35 months), the disease had progressed in 108 of 133 people (81.2%) who received 800 mg/day imatinib. Median progression-free survival was 2.7 months.
The sunitinib trial (Seddon et al. 2008) reported no data on disease progression in people randomised to an initial dose of imatinib of up to 400 mg/day, followed by 50 mg/day sunitinib after disease progression.
The retrospective cohort study reported data on the duration of response and time to treatment failure. Of the 12 people who received an increased dose of 600 mg/day imatinib after disease progression on 400 mg/day, 1 person died of a cause unrelated to both disease and treatment, while disease progressed in the remaining 11 people after a median of 1.7 months (range 0.7 to 24.9 months).
Data from the EORTC-ISG-AGITG trial showed that of the people who had a partial response or had stable disease after receiving an increased dose of imatinib (from 400 to 800 mg/day) after disease progression, the median duration of stable disease was 153 days (range 37 to 574 days).
The sunitinib trial (Seddon et al. 2008) did not provide the specific median duration of treatment with sunitinib for people who initially received up to 400 mg/day imatinib, followed by 50 mg/day sunitinib after disease progression. However, the median duration of treatment for the whole cohort was reported as 126 days (range 1 to 618 days) and did not differ significantly between people based on the dose of imatinib they received before sunitinib treatment.
Adverse events were not reported for people receiving an increased dose of 600 mg/day imatinib after disease progression on 400 mg/day imatinib.
No information on adverse events in people receiving an increased dose of 800 mg/day imatinib after disease progression on 400 mg/day was given by Zalcberg et al. (EORTC-ISG-AGITG trial). The authors reported that the majority of people who stopped taking imatinib (88.4%) did so because of disease progression. The Assessment Group suggested that this indicates 11.6% (11 of 97) stopped treatment because of adverse events. Interim data from this study showed that 31% of people (absolute number not given) needed to reduce the dose from 800 mg/day imatinib to a dose that was not reported.
Interim data for the S0033 trial reported by Dileo et al. (2005) showed that of the 77 people who had increased their dose of imatinib from 400 to 800 mg/day, 18 (23.4%) had at least 1 delay in dose, and 14 (18%) had at least 1 dose reduction because of oedema or rash. No information was given on the reduced dose given.
No adverse event data were available for people in the sunitinib trial (Seddon et al. 2008).
The Assessment Group carried out a systematic review of the literature. Seven studies that assessed both the costs and cost effectiveness of imatinib or alternative treatments for GISTs were identified.
Three studies compared imatinib (at 400 mg/day or at increased doses) with best supportive care (Wilson et al. 2005; Mabasa et al. 2008; Huse et al. 2007). Two studies included sunitinib, increased doses of imatinib, and best supportive care or palliative care as comparators (Contreras-Hernandez et al. 2008; Teich et al. 2009). Because the results of the Teich et al. study were available in abstract form only, its definition of best supportive care was not available. Studies by Chabot et al. (2008) and Paz-Ares et al. (2008) compared treatment with sunitinib and best supportive care for people with GISTs that were resistant to or intolerant of imatinib.
The study by Wilson et al. (2005) used a modified version of the model submitted by the manufacturer in NICE's technology appraisal guidance on imatinib for the treatment of unresectable and/or metastatic gastrointestinal stromal tumours (TA86). This appraisal evaluated the cost effectiveness of increased doses of imatinib (to 600 mg/day) after disease progression on 400 mg/day in people with unresectable and/or metastatic GISTs from a UK NHS perspective. The estimates for the incremental cost per quality-adjusted life year (QALY) gained compared with best supportive care ranged from £51,515 to £98,889 at 2 years, and from £27,331 to £44,236 at 5 years.
The study by Contreras-Hernandez et al. (2008) suggested that 800 mg/day imatinib would deliver cost savings compared with best supportive care when best supportive care includes treatment with imatinib at a dose lower than 800 mg/day. Over a 5-year treatment horizon, Contreras-Hernandez et al. (2008) found that the mean life years gained was 1.40 for people receiving sunitinib, 1.31 for people receiving imatinib 800 mg/day and 1.08 for people receiving best supportive care. The study also suggested that receiving 800 mg/day imatinib incurred the highest mean treatment costs, indicating that 800 mg/day imatinib is dominated by sunitinib. Teich et al. (2009) found that 800 mg/day imatinib was less effective than sunitinib (0.02 life years gained and 0.47 progression-free life years gained) and less costly over 6 years, also indicating that 800 mg/day imatinib is dominated by sunitinib.
The manufacturer did not submit a cost-effectiveness analysis of imatinib for this appraisal. The manufacturer stated that major limitations with the available clinical data prevented it from developing a sufficiently robust health economics model.
The Assessment Group developed a Markov model to compare alternative treatments for people with KIT (CD117)-positive unresectable and/or metastatic GISTs whose disease had progressed on imatinib 400 mg/day or whose treatment with imatinib had failed because of resistance or intolerance. The Assessment Group specifically addressed the cost effectiveness of imatinib at doses of 600 or 800 mg/day compared with best supportive care or sunitinib.
The model looked at the costs and outcomes for GIST treatments. A time horizon of 10 years and a cycle length of 1 month were used to reflect the natural history of the disease. The costs and outcomes were discounted at 3.5% in accordance with the NICE reference case.
The Assessment Group considered a range of treatment pathways for people whose disease had progressed on 400 mg/day imatinib. Based on advice from its clinical advisers, the Assessment Group decided on 7 clinically plausible pathways on which to base the model:
Pathway 1: people receive best supportive care plus 400 mg/day imatinib
Pathway 2: people receive 600 mg/day imatinib, and on disease progression, they receive 800 mg/day imatinib. On further disease progression people then receive 50 mg/day sunitinib, followed by, on further disease progression, best supportive care plus 400 mg/day imatinib.
Pathway 3: people receive 600 mg/day imatinib, and on disease progression, they receive 50 mg/day sunitinib. After further disease progression on sunitinib, people receive best supportive care plus 400 mg/day imatinib.
Pathway 4: people receive 600 mg/day imatinib, and on disease progression, they receive best supportive care plus 400 mg/day imatinib.
Pathway 5: people receive 800 mg/day imatinib, and on disease progression, they receive 50 mg/day sunitinib. After disease progression on sunitinib, people receive best supportive care plus 400 mg/day imatinib.
Pathway 6: people receive 800 mg/day imatinib, and on disease progression, they receive best supportive care plus 400 mg/day imatinib.
Pathway 7: people receive 50 mg/day sunitinib, and on disease progression, they receive best supportive care plus 400 mg/day imatinib.
The Assessment Group determined clinical effectiveness using results from the systematic review and other evidence. No data were available to estimate the effectiveness of any of the treatment pathways compared with each other.
In the model the Assessment Group combined the estimates of effectiveness with data on health state utility to provide estimates of QALYs for the different treatment pathways. The Assessment Group obtained data on survival for best supportive care (which includes treatment with 400 mg/day imatinib for all patients) from 2 studies (McGrath et al. 1987; Pierie et al. 2001). Pooled weighted estimates from these studies suggested that 87.9% (51 of 58) of people died during the observation period of 60 months. Data on survival for people receiving 600 mg/day imatinib were obtained from the B2222 trial. Forty-five per cent (5 of 11) of people whose dose of imatinib was increased from 400 mg/day to 600 mg/day died during the trial period of 60 months. Data on survival for 800 mg/day imatinib came from the S0033 trial, and indicated that 64.4% (76 of 118) of people died in this group during a median follow-up period of 4.5 years. The data on survival for people receiving sunitinib came from Seddon et al. (2008) and showed that 55.0% (193 of 351) of people receiving sunitinib were still alive after a median survival period of 11.8 months. The Assessment Group derived a monthly mortality rate from these survival rates and assumed exponential rates. The Assessment Group also assumed that the monthly mortality rate for people receiving sunitinib did not depend on whether they had received previous treatment.
The Assessment Group took response rates for 600 mg/day imatinib from the B2222 trial. The trial reported that 25.5% (11 of 43) of people receiving 600 mg/day imatinib had a response to treatment and remained stable during a median follow-up of 63 months. Data from the S0033 and EORTC-ISG-AGITG trials provided the response rate to 800 mg/day imatinib. The trials reported that 30% (75 of 250) of people receiving 800 mg/day imatinib showed a partial response or remained stable after a median follow-up of 54 months. For response rates to sunitinib, the Assessment Group took data from 2 studies (Demetri et al. 2002; Prior et al. 2009). A simple weighted mean was used to derive a pooled response rate of 70% (266 of 382) at a median follow-up period of 3.6 months, which did not take into account previous treatment. The Assessment Group converted data reflecting no response for each treatment into monthly transition probabilities by assuming an exponential rate.
The costs of 400 mg/day, 600 mg/day and 800 mg/day imatinib and 50 mg/day sunitinib in the Assessment Group's model were calculated from costs listed in BNF 58. Because sunitinib treatment is given for 4 weeks followed by no treatment for 2 weeks, the Assessment Group calculated the costs per year, and a proportional rate per month. The Assessment Group assumed that the cost for people receiving best supportive care was equivalent to the cost of 400 mg/day imatinib (which all people on best supportive care also received).
The Assessment Group based resource costs for 600 mg/day and 800 mg/day imatinib on those reported by Wilson et al. (2005). The resource costs included GP visits and outpatient visits, including tests, computed tomography scans and the costs of managing adverse events. For sunitinib and best supportive care, the Assessment Group based resource costs on the manufacturer's submission for NICE's technology appraisal guidance on sunitinib for the treatment of gastrointestinal stromal tumours. The Assessment Group adjusted the costs from 2008 to 2009 prices using the Hospital and Community Health Services Index.
The Assessment Group derived health state utility values from the EQ-5D and Chabot et al. (2008) in the absence of data from the available clinical studies. The utility value for progression-free survival for people whose disease responded to imatinib, regardless of dose, was assumed to be 0.935. The utility value for people receiving best supportive care was assumed to be 0.577 (Chabot et al. 2008). In the absence of alternative data, the utility value for people whose disease responded to sunitinib was assumed to be the same as that for imatinib, that is, 0.935.
The base-case results show that pathway 4 (600 mg/day imatinib followed by, on disease progression, best supportive care plus 400 mg/day imatinib) has an incremental cost-effectiveness ratio (ICER) of £27,304 per QALY gained compared with pathway 7 (50 mg/day sunitinib then best supportive care plus 400 mg/day imatinib). For pathway 2 (600 mg/day imatinib, increasing to 800 mg/day imatinib, then 50 mg/day sunitinib followed by best supportive care plus 400 mg/day imatinib) the ICER was £45,850 per QALY gained (compared with the next least costly option, pathway 4). For pathway 3 (600 mg/day imatinib, then 50 mg/day sunitinib followed by best supportive care plus 400 mg/day imatinib) the ICER was £71,723 per QALY gained (compared with the next more costly option, pathway 4). Both pathway 5 (800 mg/day imatinib, then 50 mg/day sunitinib, followed by best supportive care plus 400 mg/day imatinib) and pathway 6 (800 mg/day imatinib, then best supportive care plus 400 mg/day imatinib) were dominated by pathway 4 (that is, they were more costly and less effective). The estimated survival benefit of 800 mg/day imatinib compared with best supportive care was 4.2 months.
The Assessment Group also performed sensitivity analyses to account for uncertainties in the model. The parameters varied included structure and methodological assumptions around distribution, transition probabilities of survival and response to treatment with 600 mg/day imatinib, utility values, and the costs of imatinib and sunitinib.
The Appraisal Committee reviewed the data available on the clinical and cost effectiveness of increased doses of imatinib after disease progression on 400 mg/day imatinib, having considered evidence on the nature of unresectable and/or metastatic GISTs and the value placed on the benefits of imatinib by people with the condition, those who represent them, and clinical specialists. It also took into account the effective use of NHS resources.
The Committee discussed current clinical practice for the treatment of people with unresectable and/or metastatic GISTs. The Committee heard from the clinical specialists that, despite the recommendations in NICE's technology appraisal guidance TA86 on imatinib (recommendation 1.4 states that 'an increase in the dose of imatinib is not recommended for people receiving imatinib who develop progressive disease after initially responding'), people frequently receive 800 mg/day imatinib when their disease progresses on 400 mg/day imatinib if they have tolerated previous imatinib treatment. The clinical specialists reported that 600 mg/day imatinib is rarely prescribed after disease progression on 400 mg/day. The Committee was aware that NICE's technology appraisal guidance on sunitinib recommends that patients have the option to receive treatment with sunitinib after disease progression on 400 mg/day imatinib. However, the clinical specialists explained to the Committee that clinicians often consider increasing the dose of imatinib before offering treatment with sunitinib because imatinib is considered to have a more favourable adverse event profile, even at higher doses, than sunitinib. They also noted that if a person's disease progresses on 400 mg/day imatinib, it is common practice (in approximately 50% of people) to continue treatment with 400 mg/day imatinib in addition to best supportive care if the person tolerates imatinib, however this is inconsistent with NICE's technology appraisal guidance TA86 on imatinib, which does not recommend continued treatment with imatinib after disease progression (recommendation 1.3). The Committee heard from patient experts that the limited data available suggest that imatinib at higher doses may be effective in prolonging survival and improving quality of life. Clinical specialists and patient experts also highlighted the importance of providing hope to people with metastatic GISTs by offering them additional treatment options after disease progression on 400 mg/day imatinib. Despite the lack of clinical trial evidence to demonstrate the effectiveness of increased doses of imatinib treatment after disease progression on 400 mg/day imatinib, the Committee acknowledged that there is a perception among both patient experts and clinical specialists that treatment with 800 mg/day imatinib after disease progression on 400 mg/day imatinib may offer some benefit.
The Committee considered the evidence provided by the Assessment Group and the manufacturer on the clinical effectiveness of 600 and 800 mg/day imatinib after disease progression on 400 mg/day imatinib. The Committee heard from the Assessment Group that no randomised controlled trials were identified on the effectiveness of an increased dose of imatinib after disease progression on 400 mg/day imatinib compared with sunitinib or best supportive care – the 2 comparator treatments identified in the scope for this appraisal. Uncontrolled observational data were available evaluating the effectiveness of increased doses of imatinib in people with unresectable and/or metastatic GISTs who had not received previous imatinib treatment. However, the Committee was concerned that the populations in these studies differed from the population covered by this appraisal (that is, people whose disease has progressed on 400 mg/day imatinib), and that the studies did not explore the comparisons defined in the scope. The Committee also noted that the populations in the studies varied substantially and that combining results may introduce more uncertainty. The Committee noted that this appraisal is a part review of NICE's technology appraisal guidance TA86 on imatinib, published in 2004, which was intended to only address whether increased doses of 600 or 800 mg/day imatinib could be recommended after disease progression on 400 mg/day imatinib. The Committee noted that consultees had requested the review based on the belief that a large amount of clinical evidence about imatinib had been published since 2004. The Committee further noted that during the scoping process for this review consultees and commentators for this appraisal were given another opportunity to comment on the appropriateness of this review. The Committee was reminded that, at the time of the review proposal, the manufacturer of imatinib was seeking to extend the marketing authorisation for 800 mg/day imatinib for unresectable and/or metastatic GISTs and that the manufacturer supported the review going ahead. However, in their submission and during the Committee meeting, the manufacturer stated that no new evidence had emerged since 2004 on the effectiveness of increased doses of imatinib after disease progression on 400 mg/day imatinib. The Committee heard from the patient experts that they were disappointed that the available studies did not provide clear evidence about the effectiveness of higher doses of imatinib. The Committee agreed with the manufacturer and the Assessment Group that there is a lack of robust data available to demonstrate the effectiveness of increased doses of imatinib after disease progression on 400 mg/day imatinib, and that the available evidence is associated with uncertainty and potential bias.
The Committee noted that the majority of available data related to people who received an increased dose of 800 mg/day imatinib after disease progression on 400 mg/day imatinib. Only one phase 2 study (B2222) described the experience of 43 people who received an increased dose of 600 mg/day imatinib. The Committee heard that data from the Gastrointestinal Stromal Tumor Meta-Analysis Group (metaGIST) was published in March 2010. However, these data were not included in the Assessment Group's economic analyses because the population was randomised to higher doses of imatinib at baseline and had not received treatment with 400 mg/day imatinib first. Therefore, the population in the metaGIST study was different from the population for this appraisal. The Committee discussed the possible biases in the evidence for increased doses of imatinib from the uncontrolled observational data. The Committee recognised that if people with a better prognosis preferentially received increased doses of imatinib, then any improvement in outcome might not be because of the use of higher doses. For example, the Committee was aware that in the B2222 study, people with disease progression on 400 mg/day imatinib, received a higher dose of imatinib if they were described as having a good performance status (such as ECOG status <2). The Committee heard from the manufacturer that there are no ongoing trials that address the decision problem in this appraisal. The Committee heard from the clinical specialists that a trial comparing high-dose imatinib with sunitinib had been stopped after it failed to recruit a sufficient number of people. The Committee concluded that the Assessment Group had made a good effort to include all available data relevant to this appraisal in its report but was concerned about the lack of data and the nature of the evidence available.
After receiving comments from consultees and commentators on the appraisal consultation document, the Committee considered the UK guideline for the management of GISTs that was published in May 2009. The Committee noted that the guideline contains the same evidence that was identified for this appraisal by the Assessment Group and the manufacturer, and that the development of the guideline had been sponsored by the manufacturer of imatinib. The Committee was aware that the guideline did not consider the cost effectiveness of any treatments and therefore the recommendations in this appraisal would likely be different from the guideline.
The Committee considered the effectiveness of imatinib in slowing disease progression in unresectable and/or metastatic GISTs. The Committee heard from the clinical specialists that comparable measures of disease progression had been used in the different clinical trials. The Committee heard that the 3 studies in which the dose of imatinib was increased from 400 mg/day to 800 mg/day showed that approximately one-third of people had either a partial response or had stable disease after receiving the increased dose. The Committee concluded that imatinib treatment at higher doses may offer some benefit to people whose disease progresses on 400 mg/day imatinib; however, because of the biases inherent to the clinical-effectiveness evidence available, it was aware that this conclusion was uncertain.
The Committee heard from the patient experts that measuring plasma concentrations of imatinib could be a major advantage, because it might allow an individualised approach to the dosing of imatinib. However, the clinical specialists noted that this does not happen in routine UK clinical practice, and the Committee noted that no data had been presented to demonstrate an association between plasma concentrations and outcomes. The Committee concluded that while measuring plasma concentrations of imatinib might potentially be of benefit in the future, it could not base any recommendations on this because of the current lack of evidence and because it was not done in routine clinical practice.
The Committee discussed whether benefits from increased doses of imatinib might be greater in certain subgroups of people. The Committee heard from the clinical specialists that there is some evidence suggesting that GISTs with certain mutations in the KIT gene are likely to be more or less sensitive to imatinib treatment. The clinical specialists suggested that the presence of an exon 9 mutation may be associated with a better outcome in people whose dose is increased to 800 mg/day imatinib. In addition, the clinical specialists explained that, although outside the current marketing authorisation, clinicians might choose to begin treatment with 800 mg/day imatinib without having tried lower doses in people with confirmed exon 9 mutations. However, they explained that the clinical evidence supporting this practice is based on the experience of a small number of people. The Committee also noted that data from a meta-analysis (metaGIST), in which people with exon 9 mutations started treatment on 800 mg/day imatinib, showed that there was no statistically significant difference in overall survival between people with exon 9 mutations treated with 400 mg/day imatinib compared with 800 mg/day imatinib. Furthermore, in light of the limited data available, the Committee noted that any economic analyses for this subgroup would not be considered more robust than for the entire population. The Committee also understood that mutational analysis in people with progressive disease had a limited role, if any, in clinical decision-making about increasing imatinib doses. Therefore, the Committee concluded that there was not sufficient evidence to justify a separate recommendation for the use of 600 or 800 mg/day imatinib for people with exon 9 mutations whose disease had progressed on imatinib 400 mg/day.
The Committee considered the data reported by the Assessment Group for the comparator treatment, sunitinib. The Committee noted that this evidence was mainly from an 'expanded access programme', in which regulators allow investigational drugs to be used to treat people with serious or immediately life-threatening diseases who cannot participate in clinical trials and who have no alternative therapy. The Committee was aware that people in this study were treated with sunitinib after receiving higher (600 or 800 mg/day) rather than lower (400 mg/day) doses of imatinib and did not necessarily reflect the population of interest in this appraisal – that is, people whose disease progresses on 400 mg/day imatinib.
The Committee discussed the cost effectiveness of imatinib 600 and 800 mg/day after disease progression on 400 mg/day imatinib. The Committee noted the Assessment Group's view that it had great difficulty undertaking an assessment of cost effectiveness in the absence of robust comparative clinical evidence for increased doses of imatinib after disease progression on 400 mg/day imatinib. The Committee also noted that the manufacturer stated that they would not submit an economic model because of the lack of robust data comparing increased doses of imatinib with sunitinib and best supportive care.
The Committee then considered the monthly mortality rates used in the Assessment Group's economic evaluation and noted that they were key drivers of the outcomes in the model. The Committee noted that a higher monthly mortality rate for people receiving sunitinib treatment was used in the Assessment Group's model than for people receiving best supportive care, which the Committee considered was implausible. The Committee heard from the Assessment Group that the limited evidence available reported that the mortality rate for 600 mg/day imatinib was lower than the rate for 800 mg/day imatinib, and that treatment with 800 mg/day imatinib generated fewer life years and fewer QALYs than treatment with 600 mg/day imatinib. The Committee considered these data implausible and agreed that this difference was unlikely to reflect the true effect of 600 and 800 mg/day imatinib, but that this highlighted the limitations in the clinical evidence for the 2 doses. The Committee also noted that the monthly mortality rate applied for best supportive care appeared to be very low, despite these data being pooled from 2 studies that had been carried out before imatinib was introduced into clinical practice. The Committee also considered that when the studies were carried out, advanced diagnostic methods for GISTs did not exist, and that people may have had other tumours (for example, leiomyosarcoma) which were associated with a different mortality rate than for GISTs. The Committee was also aware that best supportive care was likely to have improved since the studies were carried out, leading to better outcomes, including lower mortality rates. The Committee concluded that, taking into account the limitations of the data used to derive the monthly mortality rates, the results presented by the Assessment Group may not reflect the true value of cost effectiveness of high doses of imatinib after disease progression on 400 mg/day imatinib and therefore should be interpreted with caution.
The Committee discussed the options for collecting data to establish outcomes in people receiving increased doses of imatinib or 1 of the comparator treatments defined in the scope. The Committee noted that despite the research recommendation in NICE's technology appraisal guidance TA86 on imatinib (published in 2004) suggesting that a national register for people receiving imatinib treatment for GISTs be maintained, such a register had not been established. The Committee heard from the manufacturer that a national register for people with GISTs is currently being set up, with pilot testing expected to begin by the end of 2010. The Committee heard from the clinical specialists that a small register has also been set up in Scotland to collect long-term treatment outcomes for people with GISTs, and that observational data from specialist cancer centres in the USA may also be available. The Committee welcomed the initiative of the manufacturer to establish a register for people with GISTs and suggested that it is important that the register collects data on outcomes specific to unresectable and metastatic GISTs.
The Committee discussed the health-related quality of life of people with unresectable and/or metastatic GISTs. It noted that the Assessment Group did not identify any data to use in its economic model that specifically measured the quality of life of people with GISTs who received imatinib treatment. The Committee heard from patient experts that the health measures defined in the NICE reference case, such as the EQ-5D, might not capture the benefits that people gain from imatinib treatment. The Committee considered the utility value used in the Assessment Group's economic model for imatinib and sunitinib (0.935). The Committee considered that this value was implausibly high and noted that this value had been derived from 3 out of 9 clinicians who had responded to a questionnaire. The Committee also noted that this utility value was higher than the value used in NICE's technology appraisal guidance on sunitinib, after disease progression on imatinib treatment. Although the Assessment Group carried out some sensitivity analyses that varied the utility value, the Committee was not convinced that the most plausible value had been used and considered that this added further uncertainty to the model. The Committee also considered that using a more appropriate utility value would probably increase the ICER because the difference between the utility values for active treatment and the comparator would be smaller. Therefore, the Committee concluded that collecting utility data is important for any future informed decision-making for this population.
The Committee considered comments from consultees and commentators on the appraisal consultation document. The Committee heard that recommendations in NICE's technology appraisal guidance TA86 on imatinib for stopping imatinib 400 mg/day were not supported by clinical specialists. The Committee heard that consultees noted that the lack of clinical evidence for this appraisal was directly attributable to the rarity of GISTs. The Committee acknowledged that, although the rarity of GISTs did contribute to the lack of evidence, more could have been done to describe the clinical experience that exists. The consultees and commentators repeated the Committee's concerns that a disease register had not been established since the publication of NICE's technology appraisal guidance TA86 on imatinib. The Committee also noted that the Assessment Group had highlighted issues contributing to the uncertainty about the clinical and cost effectiveness of imatinib at increased doses after disease progression on 400 mg/day. These issues include differing study populations, a lack of clinical evidence, and particularly sparse data on the clinical pathway of treatment with sunitinib after disease progression on 400 mg/day imatinib. The Committee appreciated the point made by consultees and commentators that, as this appraisal affected only a small group of people, giving clinicians the discretion to prescribe imatinib at doses higher than 400 mg/day would have little overall financial impact on the NHS. However, the Committee emphasised that (in line with NICE's guide to the methods of technology appraisal) the potential budget impact of the adoption of a new technology does not determine its decision.
The Committee explored whether it was possible to estimate a most plausible ICER. The Committee noted that the lowest ICER calculated by the Assessment Group was £27,300 per QALY gained for 600 mg/day imatinib after disease progression on 400 mg/day imatinib compared with sunitinib. However, in light of the inconsistencies in the model inputs and in the results, it recognised that this value was associated with considerable uncertainty, and should be interpreted with caution. The Committee discussed whether making any changes to the major assumptions made by the Assessment Group or further modelling might reduce the uncertainty in the estimates of cost effectiveness for high doses of imatinib. The Committee noted that any further modelling would need estimates of disease progression and mortality rates to be plausible as well as comparable across different treatment arms. The Committee agreed that all of the following changes in the assumptions would be likely to increase the ICER associated with imatinib treatment at increased doses:
Decreasing the utility value for imatinib and sunitinib from 0.935 to a more plausible value.
Assuming that only 50%, rather than 100%, of people receive 400 mg/day imatinib in addition to best supportive care after progression of disease at higher doses of imatinib, which is in line with current clinical practice.
Assuming that no one receives 400 mg/day imatinib in addition to best supportive care after progression of disease at higher doses of imatinib, which is in line with current NICE recommendations.
Using more up-to-date estimates of the effectiveness of best supportive care, which is assumed to be more effective now than when the data used in the modelling were collected.
Accounting for utility values for additional adverse events associated with higher doses of imatinib.
Using a more realistic effectiveness estimate for sunitinib treatment – this would be likely to increase the ICER for imatinib compared with sunitinib.
Because these assumptions would increase rather than decrease the ICERs for increased doses of imatinib, the Committee concluded that it was highly likely that the ICERs for 600 or 800 mg/day imatinib after disease progression on 400 mg/day imatinib, compared with best supportive care or with sunitinib were above £30,000 per QALY gained.
Because of the lack of robust clinical effectiveness evidence available, the Committee explored if there were any other approaches for exploring a most plausible ICER. The Committee acknowledged that when imatinib was first appraised in 2004 in NICE's technology appraisal guidance TA86 on imatinib, the results of the model from the Decision Support Unit (DSU) suggested that 400 mg/day imatinib had an ICER of approximately £32,000 per QALY gained compared with best supportive care. In addition, the ICER for 600 mg/day imatinib after disease progression on 400 mg/day compared with best supportive care was estimated to be £39,000 per QALY gained. The Committee noted that at the time these results were considered to be uncertain and that since then no clinical evidence has been published that would improve the robustness of the modelling results or reduce the uncertainty of the cost effectiveness of imatinib at increased doses. The Committee also considered that, given that the acquisition cost of 600 and 800 mg/day imatinib is obviously much higher than the cost of 400 mg/day imatinib, higher doses would need to be substantively more effective to be considered cost effective. The Committee concluded that the available evidence does not suggest that higher doses of imatinib lead to a substantive increase in effectiveness for the treatment of unresectable and/or metastatic GISTs after disease progression on 400 mg/day imatinib and that using this approach corroborates the conclusion that the most plausible ICER for 600 or 800 mg/day imatinib compared with best supportive care or with sunitinib was above £30,000 per QALY gained.
The Committee considered supplementary advice from NICE that should be taken into account when appraising treatments that may extend the life of patients with a short life expectancy and that are licensed for indications that affect small numbers of people with incurable illnesses. For this advice to be applied, all the following criteria must be met:
The treatment is indicated for patients with a short life expectancy, normally less than 24 months.
There is sufficient evidence to indicate that the treatment offers an extension to life, normally of at least an additional 3 months, compared with current NHS treatment.
The treatment is licensed or otherwise indicated for small patient populations.
In addition, when taking these criteria into account, the Committee must be persuaded that the estimates of the extension to life are robust and the assumptions used in the reference case of the economic modelling are plausible, objective and robust.
The Committee was aware that the number of people newly diagnosed with GISTs in England and Wales ranges from approximately 200 to 2000 per year. It noted that imatinib has a marketing authorisation for a number of other indications in addition to the treatment of unresectable and/or metastatic GISTs. The Committee noted that for people with unresectable GISTs, prognosis is poor, with survival generally less than 2 years without further treatment. The Committee noted that in the economic model the survival benefit following treatment with imatinib 800 mg/day compared with best supportive care was 4.2 months. However, the Committee agreed that the evidence for this life extension could not be considered sufficiently robust, considering the uncertainty about the assumptions in the economic model, and the lack of comparative clinical effectiveness data. In addition, the Committee noted that the results of the meta-analysis (metaGIST) showed no overall survival benefit for people receiving 800 mg/day imatinib compared with people receiving 400 mg/day imatinib in people with exon 9 mutations. The Committee therefore concluded that increased doses of imatinib after disease progression on 400 mg/day imatinib did not meet the criteria for being a life-extending, end-of-life treatment.
The Committee considered whether its recommendation was associated with any potential issues related to equality, and noted comments made during consultation on the appraisal consultation document that not recommending 600 or 800mg/day of imatinib following disease progression with 400mg/day imatinib unfairly discriminates against people with rare diseases. The Committee also noted the respective consultees' acknowledgement that having a rare disease does not constitute 1 of the protected characteristics in the current equalities legislation or the Equality Act. However, the Committee was aware that it has general Public Law obligations of fairness and reasonableness in respect of the impact of its guidance on patients. The Committee was also aware that the Human Rights Act and article 14 of the European Convention on Human Rights (ECHR) can protect groups of people other than those covered by the UK equalities legislation. As regards article 14 ECHR, the Committee noted that it was not clear that patients affected by this appraisal or those with an exon 9 mutation would be regarded as a 'group' protected by article 14, nor that any of the substantive ECHR articles was engaged. In relation to both the ECHR obligations and public law requirements, the Committee considered that its recommendation did not unfairly disadvantage any groups within the remit of this appraisal. The Committee noted that its role was to appraise clinical and cost effectiveness and that in relation to each technology it is required to consider the robustness or otherwise of the available evidence. The Committee took into account the lack of robust clinical evidence for a survival benefit of higher doses of imatinib, specifically for the subgroup of people that have been reported to respond better, that is people with an exon 9 mutation. The Committee was also aware that an alternative treatment option is available for this group of people because NICE's technology appraisal guidance on sunitinib recommends that patients have the option to receive treatment with sunitinib after disease progression on 400 mg/day imatinib. The Committee was satisfied that its recommendation was consistent with NICE's legislative obligations under the equalities legislation and the requirement for fairness.
In summary, the Committee agreed that clinical opinion suggests that increased doses of imatinib after disease progression on 400 mg/day imatinib may offer benefit to some people. However, since the previous appraisal of imatinib (NICE's technology appraisal guidance TA86), there are no new good-quality data on the clinical effectiveness of increasing the dose of imatinib. The Committee concluded that the current available clinical and cost-effectiveness evidence does not justify a positive recommendation for the use of imatinib at increased doses of 600 mg/day and 800 mg/day as an appropriate use of NHS resources for the treatment of people with unresectable and/or metastatic GISTs whose disease has progressed on 400 mg/day imatinib.