Trastuzumab for the treatment of HER2-positive metastatic gastric cancer

The Committee discussed current clinical practice in the UK. It noted comments from consultation that no treatment options for metastatic gastric cancer currently target HER2 overexpression, and that trastuzumab offers a new option for this patient group. The Committee heard from clinical specialists that current clinical practice in the UK is normally a triple regimen that includes an anthracycline, a platinum agent, and a fluoropyrimidine. It understood that epirubicin, cisplatin and capecitabine are the standard anthracycline, platinum agent and fluoropyrimidine therapies used. It also understood that oxaliplatin is sometimes used in place of cisplatin and that 5-fluorouracil is sometimes used in place of capecitabine. The Committee considered this in the context of the comparator regimen given in the ToGA trial, which was a double regimen including a platinum agent and a fluoropyrimidine, but not an anthracycline. The Committee discussed the fact that double regimens were not often used in UK clinical practice, and concluded that the comparator in the ToGA trial did not represent current practice in the UK.

The Committee considered whether the population in the ToGA trial could be considered representative of the population of people with HER2-positive metastatic gastric cancer in England in Wales. It noted that most of the people in the trial were from Asia. The Committee acknowledged subgroup analyses that appeared to confirm a similar overall survival benefit for the group of European people in the trial.

The Committee discussed the clinical effectiveness of the trastuzumab regimens presented in the ToGA trial. It noted that, compared with cisplatin plus either capecitabine or 5-fluorouracil, trastuzumab plus cisplatin and either capecitabine or 5-fluorouracil provided a 4.2-month gain in overall survival and a 2.1-month gain in progression-free survival. The Committee concluded that the ToGA trial demonstrated trastuzumab plus cisplatin and capecitabine or 5-fluorouracil offers clinical benefit.

The Committee discussed the likely clinical effectiveness of the current UK triple regimens, that is, when epirubicin is added to cisplatin and fluoropyrimidine. It heard from clinical specialists that adding epirubicin to cisplatin and a fluoropyrimidine provides sufficient additional benefit to be standard practice. However, clinical specialists underlined that this treatment regimen had not been subject to rigorous evaluation. The Committee discussed the evidence identified by the manufacturer and the ERG, recognising that none of the studies were completed in a HER2-positive population. The Committee acknowledged the lack of ideal evidence for this comparison. However, it considered, based on the evidence and the views of the clinical specialists, that epirubicin provided some additional benefit when added to a cisplatin and fluoropyrimidine combination.

The Committee discussed whether the results of the ToGA trial could be applied to the comparison between a trastuzumab plus chemotherapy regimen and an epirubicin plus chemotherapy regimen. It noted that the ToGA trial had used a higher dose of cisplatin than would be used as part of a triple regimen in UK clinical practice, and recognised the manufacturer's view that adding epirubicin to high-dose cisplatin would offer less benefit than adding it to lower-dose cisplatin. The Committee also noted comments from consultation that dose intensity was an important factor in chemotherapy and that reduced doses over a longer number of cycles could not be considered equivalent to higher doses over a shorter number of cycles. However, based on clinical specialist opinion, the Committee was not persuaded that the outcomes for the chemotherapy comparator group in the ToGA trial were representative of the outcomes of triple regimens in UK clinical practice. It did, however, accept that trastuzumab plus cisplatin and either capecitabine or 5-fluorouracil was likely to offer a survival benefit over treatment with epirubicin plus cisplatin and either capecitabine or 5-fluorouracil.

The Committee then discussed the range of possible estimates submitted for the clinical effectiveness of the triple regimens over the double regimens. It first considered the hazard ratio for progression-free survival of 0.96 for epirubicin plus cisplatin and capecitabine compared with cisplatin plus capecitabine from the Yun study. The Committee heard from the ERG that it had chosen to use this estimate in its revised base-case analysis because it was taken from a study that compared cisplatin plus capecitabine with epirubicin plus cisplatin and capecitabine. In the absence of any evidence specifically for the HER2-positive group, it considered that this study best represented the decision problem in the appraisal.

The Committee also considered the hazard ratio for overall survival of 0.77 (indicating a more favourable benefit of treatment with epirubicin) from the Wagner meta-analysis. It discussed concerns raised by the ERG that this study was not directly applicable to the estimates of effectiveness of the regimens including capecitabine, because 2 of the studies used 5-fluorouracil regimens and the third study compared 2 triple therapies. The Committee heard from clinical specialists that they considered that the meta-analysis may have overestimated the survival benefit of adding epirubicin to cisplatin and a fluoropyrimidine. It further noted comments from consultees that the quality of the studies in the meta-analysis was poor. The evidence from the largest study (Ross) was from an unplanned subgroup, which provided a greater estimate of the effect of epirubicin than the full population. The Committee further noted comments from consultees that the low doses of cisplatin in the studies did not reflect the higher dose used in the ToGA trial. The Committee concluded that the survival benefit of a triple regimen including epirubicin compared with that of a double regimen without epirubicin was unlikely to be represented by a hazard ratio of 0.77, and that the estimate would be closer to 0.96. However, this was associated with considerable uncertainty.

The Committee discussed the clinical effectiveness of trastuzumab for the IHC3-positive subgroup, who in clinical practice would not require a confirmatory FISH test. It noted the efficacy in the trial was greater for the subgroup than for the whole population. The Committee discussed the biological plausibility of greater benefit in the IHC3-positive subgroup and considered that greater effectiveness may be experienced with higher levels of HER2. The Committee concluded it was an appropriate subgroup and discussed the clinical evidence. It noted that the hazard ratio of overall survival for the trastuzumab plus chemotherapy group compared with the chemotherapy alone group was 0.57, corresponding to a median survival for the trastuzumab plus chemotherapy group of 18 months compared with 12.4 months for the chemotherapy alone group (a 5.6-month gain in survival). The Committee concluded that trastuzumab plus cisplatin and capecitabine or 5-fluorouracil is clinically effective in the IHC3-positive subgroup.

The Committee discussed the quality-of-life data from the ToGA trial. It noted that there were no differences in quality-of-life scores between the 2 treatment groups and no statistical analysis of these data. It heard from clinical specialists that treatment could bring about some improvement in quality of life, because gastric cancer has serious disease symptoms including pain, vomiting and anaemia, which chemotherapy can reduce. The clinical specialists considered that reduced symptoms outweighed the side effects of chemotherapy. The Committee was persuaded that because of the disease symptoms associated with gastric cancer it was plausible that quality of life could increase during progression-free survival.

The Committee discussed the adverse event data provided by the manufacturer. The Committee heard from clinical specialists that they considered that the adverse effects of trastuzumab are known and manageable in clinical practice because of its use in breast cancer. It further heard that epirubicin was associated with adverse effects and that providing trastuzumab as an alternative to epirubicin may have benefits. The Committee understood that trastuzumab is associated with cardiotoxicity, but noted that the incidence of cardiac failure in the trial was low and was similar in both treatment groups in the ToGA trial (2 in the chemotherapy alone group and 1 in the trastuzumab plus chemotherapy group). It also noted commercial-in-confidence information relating to infusion-related reactions.

The Committee considered the base-case estimates of cost effectiveness in the manufacturer's model. It noted that after consultation on the ACD the manufacturer's alternative base-case ICER of £62,800 per QALY gained was approximately £11,000 per QALY gained higher than the original base-case ICER of £51,900 per QALY gained. It further noted that the alternative base case incorporated a number of assumptions considered equally plausible by the ERG in its report. The Committee discussed the analyses of cost effectiveness and specifically considered the following assumptions:

• The most plausible estimate of comparator effectiveness.

• The frequency of cardiac monitoring for people treated with epirubicin.

• A parallel or sequential HER2 testing strategy.

• The change in utility values during progression-free survival.

• The assumption that 80% of centres would share vials of trastuzumab.

The Committee discussed the most plausible estimate of clinical effectiveness for the triple comparators. It recognised that, in the manufacturer's base-case analysis, a hazard ratio of 1.00 was used to estimate overall survival and progression-free survival of people treated with epirubicin plus cisplatin and capecitabine compared with cisplatin plus capecitabine. It noted that the ERG modified this to 0.96 in its exploratory analysis, and that this was incorporated into the manufacturer's alternative base case. It noted that when this parameter change was applied to both overall survival and progression-free survival, the ICER increased by less than £1,000. The Committee also considered the estimate of the ICER using a hazard ratio of 0.77 (indicating a more favourable benefit of treatment with epirubicin) from the Wagner meta-analysis. The Committee noted that, although the effect of changing the hazard ratio to 0.96 had been relatively minor, the effect of changing the hazard ratio to 0.77 increased the ICER by up to £50,000 per QALY gained. The Committee concluded that it was appropriate to consider the ICERs that had used a hazard ratio of 0.96 (see section 4.8). However, it considered that there was considerable uncertainty in this estimate.

The Committee discussed the frequency of cardiac monitoring for people treated with epirubicin in the UK. It noted that the manufacturer's alternative base-case analysis assumed that people would have cardiac monitoring every 3 months whether they were treated with epirubicin or trastuzumab. This assumption was the same as that proposed by the ERG. The Committee heard from the clinical specialists that people on epirubicin treatment were not necessarily given cardiac monitoring this often in the UK. It heard that in current practice people were tested before starting epirubicin treatment and this was only repeated when treatment levels made it necessary or if cardiac toxicity was suspected during treatment. The Committee therefore concluded that an alternative scenario assuming equal monitoring may still slightly overestimate the cost of the comparator strategies. However, the Committee noted that the ICER was not very sensitive to this parameter. It therefore agreed to consider the ICERs that assumed equal frequency of cardiac monitoring for trastuzumab and epirubicin.

The Committee discussed sequential and parallel HER2 testing. It heard from the clinical specialists that people with gastric cancer are not routinely tested for their HER2 status. However, for people with breast cancer in the UK, testing is sequential and only people with a score of IHC2 have a confirmatory FISH test. The clinical specialists considered that there were no reasons to assume a different testing strategy for metastatic gastric cancer. The Committee therefore concluded that sequential testing would be appropriate in assessing HER2 status in people with metastatic gastric cancer.

The Committee discussed the manufacturer's assumption of a daily increase in utility during progression-free survival that was included in both the base-case and alternative base-case analyses. It was aware that this assumption was based on data only for people in the clinical trial surviving without progression without any adjustments. The Committee discussed the ERG's alternative assumption of a very slow decrease in utility calculated from those of a similar age group based on UK general population norms for EQ-5D. It recognised comments from the clinical specialists (see section 4.10) and noted that after consultation the ERG suggested that an increase in utility may be appropriate, but not above that of the general population. The Committee was persuaded that an increase in utility was plausible. However, it accepted the ERG comments that such increases should be capped so that they did not go above those of the general population of a comparable age.

The Committee noted the manufacturer's assumption that sharing vials between patients to minimise wastage would occur in 80% of centres. It considered that this might be an overestimate and that in some centres, particularly smaller centres, sharing vials may not be possible, and therefore there was likely to be large variation in vial sharing. The Committee concluded that there was not enough evidence to estimate the proportion of centres that would vial share in clinical practice, but that 80% could be an overestimate.

The Committee considered the manufacturer's alternative base-case ICER of £62,800 per QALY gained. It noted that this incorporated a hazard ratio for overall survival and progression-free survival of 0.96, but assumed that quality of life increased in progression-free survival. It further noted the results of the manufacturer's probabilistic sensitivity analysis around the alternative base case, in which the mean value of the ICER increased by approximately £5,000. It concluded that the manufacturer's alternative base-case ICER was associated with considerable uncertainty and that the ICER would be greater than £62,800 per QALY gained.

The Committee discussed the results of the ERG's exploratory analyses, in which a cap on utility values during progression-free survival was applied (see section 3.40). It considered that this change to the assumptions was appropriate. The Committee agreed that the ICER for the population covered by the marketing authorisation in the manufacturer's base case probably lay between the ERG calculated deterministic ICER of £63,100 per QALY gained and the ERG calculated probabilistic ICER of £71,500 per QALY gained. The Committee concluded that the ICER for the population covered by the marketing authorisation was in excess of the range normally considered cost effective.

The Committee considered the manufacturer's estimate of the ICER based on the subgroup of people who tested IHC3 positive in ToGA. It noted that the assumptions in the alternative base case (about cardiac monitoring for epirubicin and trastuzumab, hazard ratios of comparator effectiveness and an increase in utility during progression-free survival) were applied. The Committee expressed the same concerns that the model allowed utility values during progression-free survival to increase more than those of the general population (see section 4.18). It further noted that the estimate of the ICER for the IHC3-positive subgroup was stratified for baseline imbalances in the characteristics of people in the treatment groups, and that the stratified hazard ratio was favourable to trastuzumab (see section 3.39). The Committee noted that the manufacturer's estimate of £43,000 per QALY gained for this population was based on a deterministic estimate (that is, a point estimate of the value of the ICER). The Committee concluded that the manufacturer's ICER for the IHC3-positive subgroup was probably an underestimate.

The Committee discussed the ERG's exploratory analyses for the IHC3-positive subgroup. It noted that these analyses imposed a cap on utility values during-progression free survival equal to general population utility estimates. Based on previous discussions (see section 4.16), it considered that this was a reasonable parameter change. The Committee noted that using probabilistic analysis increased the ICER by approximately £1,500 per QALY gained compared with deterministic analysis. It further noted that the effect of stratification accounted for £6,700 and £7,400 per QALY gained for the deterministic and probabilistic analyses respectively. The Committee noted that the ERG's ICERs for the IHC3 subgroup (deterministic and probabilistic, stratified and unstratified) were between £43,200 and £52,000 per QALY gained. The comparator for 3 of these estimates was epirubicin plus cisplatin and capecitabine; the comparator in the highest estimate (probabilistic unstratified analysis) was epirubicin plus oxaliplatin and capecitabine. The Committee agreed that the most plausible estimate of cost effectiveness of trastuzumab plus cisplatin and capecitabine lay between £45,000 and £50,000 per QALY gained. The Committee concluded that the ICER for the population covered by the marketing authorisation was higher than would normally be considered cost effective.

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, 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 economic modelling are plausible, objective and robust.

The Committee considered the criteria that needed to be met to consider trastuzumab as a life-extending, end-of-life treatment. First, the Committee considered the life expectancy of people with HER2-positive metastatic gastric cancer. It understood that the ToGA trial reported a median 11.8 months' overall survival for people receiving cisplatin plus either capecitabine or 5-fluorouracil. Therefore, the Committee was persuaded that trastuzumab plus cisplatin and either capecitabine or 5-fluorouracil met the criterion for short life expectancy.

The Committee then considered whether trastuzumab offers an extension to life of at least an additional 3 months. It noted that the median overall survival gain for the licensed population from the ToGA trial was 4.2 months for trastuzumab plus cisplatin and capecitabine or 5-fluorouracil compared with cisplatin and capecitabine or 5-fluorouracil alone. This produced a modelled mean overall survival gain of 4.8 months for people treated with trastuzumab plus cisplatin and capecitabine compared with people treated with epirubicin plus cisplatin and capecitabine. The Committee further noted that the median overall survival gain for the subgroup of people whose tumours were IHC3-positive in the ToGA trial was 5.6 months for trastuzumab plus cisplatin and capecitabine or 5-fluorouracil compared with cisplatin and capecitabine or 5-fluorouracil alone. This produced a modelled mean overall survival gain of 7.4 months for trastuzumab plus cisplatin and capecitabine compared with epirubicin plus cisplatin and capecitabine. The Committee considered that this was subject to uncertainty because of the absence of appropriate UK practice comparator data. However, on balance the Committee was persuaded that the addition of trastuzumab to chemotherapy would result in an extension to life of more than 3 months. The Committee therefore considered that trastuzumab plus cisplatin and capecitabine or 5-fluorouracil met the criterion for life extension.

The Committee considered the size of the patient population. Treatment with trastuzumab would be suitable for approximately 7000 people who have 1 of the diseases for which trastuzumab is licensed (that is, HER2-positive metastatic gastric cancer, HER2-positive early and locally advanced breast cancer and HER2-positive metastatic breast cancer). The Committee considered that 7000 was at the upper end of the population size for which it understood the supplementary advice to apply. However, the Committee concluded overall that applying the supplementary advice on end-of-life was appropriate.

The Committee discussed the cost effectiveness of trastuzumab plus cisplatin and capecitabine for the population covered by the marketing authorisation, taking into account the end-of-life criteria. It agreed that the most plausible estimate was between £63,100 per QALY gained (using the deterministic estimate from the ERG's alternative base-case analysis) and £71,500 per QALY gained (using the probabilistic estimate from the ERG's alternative base-case analysis). The Committee therefore considered that, even when taking the end-of-life considerations into account, the magnitude of weight required for the ICER to be in a range normally considered cost effective in the NHS was too high. The estimates exceed what can be considered a reasonable use of NHS resources. Therefore, the Committee concluded that trastuzumab plus cisplatin and capecitabine or 5-fluorouracil could not be recommended as an appropriate use of NHS resources for the whole population covered by the marketing authorisation.

The Committee then discussed the cost effectiveness of trastuzumab plus cisplatin and capecitabine for the subgroup of people whose tumours are IHC3 positive, taking into account the end-of-life criteria. It agreed that the ICER was between £45,000 and £50,000 per QALY gained. The Committee considered that the magnitude of weight required for the ICER to be in a range normally considered cost effective in the NHS was acceptable. The Committee recognised that the ICER range was specifically for trastuzumab plus cisplatin and capecitabine, and not trastuzumab plus cisplatin and 5-fluorouracil. However, accepting the uncertainties around comparator effectiveness estimates for HER2-positive metastatic gastric cancer, the Committee considered it was appropriate to recommend 5-fluorouracil as an alternative to capecitabine for those people who required it. The Committee therefore concluded that trastuzumab plus cisplatin and capecitabine or 5-fluorouracil should be recommended as an option for the treatment of people with HER2-positive, metastatic adenocarcinoma of the stomach or gastro-oesophageal junction who have not received prior treatment for their metastatic disease andwhose tumours express high levels of HER2, as defined by a positive immunohistochemistry score of 3 (IHC3 positive).