Dasatinib, nilotinib and high-dose imatinib for treating imatinib-resistant or intolerant chronic myeloid leukaemia

The committee discussed current clinical practice for treating CML. The committee heard from the clinical experts that standard-dose imatinib is given in line with NICE technology appraisal guidance 70 to people presenting with chronic-phase CML. The clinical experts stated that in approximately 60% of people there is a good response to standard-dose imatinib, and that these people will continue to receive the treatment for life and have a normal life expectancy. The committee recognised the innovative nature and major change in the treatment of CML that imatinib had provided. However, it heard that 40% of people develop intolerance or resistance to standard-dose imatinib.

The committee heard that high-dose imatinib, dasatinib and nilotinib are in widespread use and are a major advance over earlier therapies (that is, interferon alfa and hydroxycarbamide). The clinical experts suggested that if dasatinib, high-dose imatinib or nilotinib were not available, people would receive treatment with interferon alfa, hydroxycarbamide or best supportive care, and that for many people hydroxycarbamide or interferon alfa are considered to be little better than best supportive care. The committee also heard from the clinical experts that bone marrow stem cell transplantation could be used, although it carries high risks and is restricted to fit, younger people. The committee concluded that any one of these treatments could be considered a comparator with high-dose imatinib, nilotinib or dasatinib.

The committee noted that high-dose imatinib had been recommended only in the context of clinical research in NICE technology appraisal guidance 70. It heard from the clinical experts that high-dose imatinib is being used in clinical practice for people whose CML has previously had a good response to treatment with standard-dose imatinib. The committee acknowledged the clinical experts' view that for CML that is resistant to standard-dose imatinib, high-dose imatinib was unlikely to be as beneficial as dasatinib and nilotinib.

The committee heard from the clinical experts that, in clinical practice, treatment with dasatinib, high-dose imatinib and nilotinib is given in accordance with European guidelines, which specify time-dependent targets. If the CML is responding to treatment, the treatment will be continued until progression or until the person dies (from non-CML causes). If CML does not respond to dasatinib or nilotinib within 12 months, treatment may be stopped, or may be changed to hydroxycarbamide and/or, if suitable, stem cell transplantation.

The committee heard from the clinical experts that in more than 50% of people with imatinib-resistant CML who have dasatinib or nilotinib, there is a good response to treatment and that this response is usually as good as the initial response to standard-dose imatinib. The clinical experts expected that these people would receive dasatinib or nilotinib treatment for the rest of their lives, and possibly have a nearly normal life expectancy (that is, at least 10 more years). For people receiving interferon alfa or hydroxycarbamide in the chronic phase, the prognosis is poor, with a median life expectancy of around 5 years. It heard from the clinical experts that with modern therapy the accelerated phase is no longer considered to be a distinct disease phase, so in effect the disease progresses from a prolonged chronic-phase to blast-crisis phase.

The committee discussed the clinical-effectiveness evidence for dasatinib, high-dose imatinib and nilotinib for the treatment of chronic-phase CML that is resistant to standard-dose imatinib. It was aware of only 1 comparative trial, which compared dasatinib with high-dose imatinib, but noted the restricted comparison (only with high-dose imatinib) and the comments from the assessment groups on the interpretation of this trial.

The committee noted that the clinical trials available were non-comparative, of short duration and had used surrogate outcomes to predict overall survival. The committee noted the wide range of results across the interventions, with major cytogenetic response rates ranging from 33.3 to 58.9% with dasatinib, 32.7 to 42.5% with high-dose imatinib (but with 1 outlying result of 63.5%), and 35.3 to 56.1% with nilotinib. The committee discussed the clinical trial evidence in light of the views of the patient and clinical experts. The committee noted the poor quality of the evidence base. However, it heard from the clinical experts and patient experts that clinical benefits, particularly of dasatinib and nilotinib, have been demonstrated. In addition, the clinical experts argued that the people in the clinical trials did not reflect the population seen in clinical practice because the trials included people who had worse disease prognoses than would be seen in current clinical practice.

The committee concluded that it is clear that dasatinib, high-dose imatinib and nilotinib provide clinical benefit for people with imatinib-resistant CML. However, the committee agreed that the limited evidence base means that the magnitude of the benefit is uncertain. The committee also agreed that there was no good evidence to distinguish between dasatinib and nilotinib; a conclusion supported by the clinical experts.

The committee was aware that continued use of imatinib is not an option for people with imatinib intolerance. It noted that most of the clinical-effectiveness evidence came from trials that included a mixed population of people with imatinib-resistant CML and people with imatinib intolerance. The committee noted that in the trials that reported response rates separately, CML in people with imatinib intolerance generally had a higher response rate to dasatinib and nilotinib than people with imatinib-resistant CML, and that this was reflected in the estimates of overall survival used in the economic analyses. The committee agreed that this was a reasonable assumption given that people with imatinib intolerance generally have had a shorter duration of prior treatment than those whose CML develops resistance to imatinib over time.

The committee discussed the side effects of treatment for imatinib-resistant CML and for people with CML who have imatinib intolerance. It noted the adverse effects reported in the trials with dasatinib, high-dose imatinib and nilotinib in imatinib-resistant CML. The committee concluded that dasatinib and nilotinib are better tolerated than imatinib, and that older treatments, particularly interferon alfa, can be poorly tolerated.

The committee considered the treatment of the blast-crisis phase of CML in clinical practice. The committee heard from the clinical experts that at the blast-crisis stage of the disease, life expectancy is about 3 to 6 months. The committee also heard from the clinical experts that the treatment strategy in the blast-crisis phase of the disease is different from that in the accelerated or chronic phases, with dasatinib and high-dose imatinib given as adjuvant treatment with intensive chemotherapy for acute leukaemia. The committee was aware that no evidence was presented on the use of dasatinib or high-dose imatinib in this way and that the evidence base for the blast-crisis phase of the disease is very limited.

The committee then considered the economic models provided by the companies and the assessment groups for chronic-phase CML that is resistant to standard-dose imatinib. In each it took particular note of the incremental cost-effectiveness ratio (ICER) for the comparison between the most cost effective of the tyrosine kinase inhibitors (given that dasatinib and nilotinib are considered equal), and the most cost effective of the older treatments (given that none were definitively favoured). In all the comparisons, the committee also took particular note of the relationship between treatment duration and overall survival; because these are the main influences on costs and benefits and the clinical experts stated that these were closely related.

From the model developed by Bristol-Myers Squibb, the committee particularly noted the comparison between dasatinib and interferon alfa, which generated an ICER of £38,900 per quality-adjusted life year (QALY) gained. The estimated treatment duration with interferon alfa was 0.65 years (at a total estimated cost of £129,000), resulting in 3.56 years of overall survival, and the estimated treatment duration with dasatinib was 7.46 years (at a cost of £314,000), resulting in 11.76 years of overall survival. The committee considered that the model had a number of limitations, of which the most important were that it estimated the cost for people receiving interferon alfa to be higher than (in some cases double) that of all the other economic models, and it did not include a comparison with hydroxycarbamide. After consultation on the appraisal consultation document, Bristol-Myers Squibb provided an additional economic analysis. The committee noted that the additional analysis included hydroxycarbamide as a comparator and bone marrow stem cell transplantation as a third-line treatment. It noted that Bristol-Myers Squibb calculated the ICER for dasatinib to be £28,000 per QALY gained compared with hydroxycarbamide, and the total QALYs and costs associated with treatment with dasatinib in the additional economic analysis were more favourable to dasatinib than those in the company's original economic analysis.

The committee compared these findings with those of the other economic models, and examined the assumptions that had been used in the additional analysis. Bristol-Myers Squibb's estimates for comparator costs were higher than had been used in other economic models. The committee considered that the assumption that 30.8% of people who stopped treatment would receive bone marrow stem cell transplantation was likely to be an overestimate given contraindications to bone marrow stem cell transplantation and the lack of availability of a matched donor for many people. Secondly, the committee considered that the assumed ongoing monthly cost of £2,400 after bone marrow stem cell transplantation (at £80,000) was an unreasonably high estimate, given that only a minority of people who survive transplantation develop complications that incur high ongoing costs. Thirdly, the committee considered the utility value estimate of 0.6 for the health state associated with successful transplantation to be unreasonable, in view of the utility value of 0.85 for successful dasatinib treatment, and the utility value of 0.68 for failed dasatinib treatment. The committee noted that these utility values were not derived from a common source. The committee therefore concluded that the ICER from this analysis was not reliable and could not form a suitable basis for a recommendation.

The committee considered the economic model developed by Novartis for chronic-phase CML that is resistant to standard-dose imatinib. It noted that in the base-case analysis, nilotinib dominated (that is, it was less expensive and more effective than) high-dose imatinib and, in an exploratory analysis, nilotinib compared with a combination of hydroxycarbamide and stem cell transplantation resulted in an ICER of £44,000 per QALY gained. The estimated treatment duration with hydroxycarbamide and stem cell transplantation resulting in 4.21 years of overall survival (at a cost of £80,900) was not reported, and the estimated treatment duration with nilotinib was 2 years, resulting in 5.8 years of overall survival (at a cost of £139,000). The committee noted that if the treatment duration and overall survival seen in clinical practice were more accurately modelled and if hydroxycarbamide alone was a comparator, the base-case ICER of £44,000 per QALY gained would be likely to increase.

The committee considered the economic model developed by Peninsula Technology Assessment Group (PenTAG) and subsequently updated by Southampton Health Technology Assessments Centre (SHTAC) for chronic-phase CML that is resistant to standard-dose imatinib. The committee noted that the PenTAG model did not link treatment duration with overall survival and that some of the results were not plausible. In particular, it noted that the estimated overall survival for interferon alfa was implausible and the treatment duration for people receiving nilotinib was lower than would be seen in clinical practice, given the estimated overall survival.

The committee understood that the model updated by SHTAC attempted to correct PenTAG's overestimate of survival on interferon alfa and the discrepancy between the nilotinib and dasatinib treatment durations, but the SHTAC base-case treatment durations still did not reflect the fact that in clinical practice, people will receive treatment until progression or death (this was confirmed by the clinical experts; see section 4.5).

The committee did not consider that a conclusive ICER had been presented in any of the economic models, but agreed that, taking all the models' assumptions into account, the least implausible analysis was the SHTAC scenario in which the treatment durations of dasatinib, high-dose imatinib and nilotinib were set to 10 years with overall survival estimates of 12.4 to 13.4 years. It noted that in this analysis both high-dose imatinib and nilotinib were dominated (that is, more expensive and less effective) by dasatinib, and dasatinib compared with hydroxycarbamide resulted in an ICER of £43,800 per QALY gained. The committee noted its earlier conclusions that more than 50% of people receiving these treatments are likely to do so for more than 10 years, with many people receiving them until death. The committee agreed that if treatment is continued for most of the person's lifetime, then the ICERs would increase. The committee concluded that there was no evidence to distinguish between dasatinib and nilotinib and that the ICERs for these treatments compared with hydroxycarbamide were uncertain and likely to be higher than £43,800 per QALY gained.

The committee discussed the cost effectiveness of the technologies for the treatment of chronic-phase CML in people who have imatinib intolerance. It acknowledged the difficulties of undertaking an assessment of cost effectiveness without reasonable comparative evidence, relying on surrogate outcomes and uncertain treatment durations. However, it was aware that the effectiveness of dasatinib and nilotinib was likely to be greater in people with imatinib intolerance than in people with imatinib-resistant CML. Noting the uncertainties in these analyses, particularly about treatment duration, the committee concluded that dasatinib and nilotinib were likely to be at least as cost effective in people with imatinib intolerance as in people with imatinib-resistant CML and, as such, the cost effectiveness of dasatinib and nilotinib for people with imatinib intolerance could be inferred from the cost effectiveness in people with imatinib-resistant CML.

The committee noted that Novartis had agreed a patient access scheme with the Department of Health. The company had presented ICERs for the scheme based on an analysis reflecting the scenario considered most plausible by the committee, outlined in section 4.19.

The committee noted that the Novartis adjusted analysis based on the SHTAC update of the PenTAG model resulted in an ICER of £30,800 per QALY gained. It also noted that when SHTAC replicated the analysis the ICER increased slightly to £31,300 per QALY gained. It also noted that the company argued that a number of further changes to the SHTAC analysis should be made, namely:

• a reduction in treatment duration from 10.0 to 6.5 years

• a lower dose intensity of nilotinib based on clinical trial data

• an assumption of survival benefit equal to that of dasatinib

• a lower utility value associated with hydroxycarbamide treatment in the chronic phase, and

• a lower estimate of overall survival for hydroxycarbamide treatment.
  
  The committee noted that when the modifications and the discount were applied, the ICERs for nilotinib compared with hydroxycarbamide decreased to £22,800 per QALY gained when a treatment duration of 6.5 years was assumed, and £25,000 per QALY gained when a treatment duration of 10 years was assumed. The committee agreed that some of these adjustments were plausible, but not all. The treatment duration could be less than 10 years but the estimate of 6.5 years, which was based on treatment being withdrawn in all people who did not have a complete cytogenetic response, was not plausible. Also the committee did not agree with Novartis that the utility value for people treated with hydroxycarbamide should be lower for the same health states achieved by other treatments. It accepted that health state durations were shorter with hydroxycarbamide but thought that this should not be compounded by utility value adjustments.

The committee therefore concluded that the Novartis adjusted ICER of £22,800 per QALY gained was too optimistic. However, the committee accepted that with the patient access scheme in place and its earlier conclusion that some of the adjustments to the model were plausible, the ICER for nilotinib is likely to be less than the SHTAC replicated ICER of £31,300 per QALY gained. The committee concluded that the use of nilotinib for the treatment of imatinib-resistant CML could be regarded as a cost-effective use of NHS resources. The committee therefore recommended the use of nilotinib for the treatment of adults with chronic- and accelerated-phase CML that is resistant to standard-dose imatinib or who have imatinib intolerance, if the company makes nilotinib available with the discount agreed as part of the patient access scheme.

The committee then reflected on all of the models and results presented for high-dose imatinib for the treatment of CML that is resistant to standard-dose imatinib, together with the clinical and patient experts' views on the use of the technologies. It noted that high-dose imatinib was dominated (that is, more expensive and less effective than another treatment) in all models. Therefore the committee agreed that high-dose imatinib could not be recommended as a cost-effective use of NHS resources for the treatment of CML that is resistant to standard-dose imatinib.

The committee then considered the cost effectiveness of dasatinib for the treatment of CML that is resistant to standard-dose imatinib. The committee noted its earlier conclusion that the updated economic analysis provided by Bristol-Myers Squibb could not form a suitable basis for a recommendation given the limitations described in section 4.15. It also noted that all other estimated ICERs were higher than those normally considered acceptable for the NHS, and were highly likely to be above the figures suggested. Therefore the committee concluded that dasatinib could not be recommended as a cost-effective use of NHS resources for the treatment of adults with chronic-phase CML that is resistant to standard-dose imatinib, or who have imatinib intolerance. Furthermore, the committee noted that, given the patient access scheme for nilotinib and the assumed equivalence of effectiveness of dasatinib and nilotinib, dasatinib is considerably more expensive but no more effective than nilotinib.

The committee then considered the cost-effectiveness evidence for dasatinib, high-dose imatinib and nilotinib for the treatment of the accelerated and blast-crisis phases of CML. The committee noted the clinical experts' view that there is no longer considered to be a distinguishable accelerated phase of CML. However, it acknowledged that this phase continues to be recognisable for some people, and saw no reason not to recommend nilotinib for treatment of CML in the accelerated phase. The committee noted that, as for the chronic phase, high-dose imatinib continued to be dominated (that is, it was more expensive and less effective than another treatment), and dasatinib continued to be as effective but more expensive, and concluded that neither drug could be recommended for the treatment of accelerated-phase CML.

The committee noted that nilotinib does not have a marketing authorisation for the treatment of blast-crisis phase CML. It noted that treatment for the blast-crisis phase is different from that used in the other phases, with interventions generally used as adjuvant treatment to intensive chemotherapy for acute leukaemia. The committee was aware that no evidence using the interventions in this way had been submitted. To the extent that dasatinib could be considered a stand-alone treatment, the committee concluded that the evidence was particularly limited. The committee considered that all 3 of the estimates it saw, 1 from PenTAG and 2 from Bristol-Myers Squibb to be highly speculative. The PenTAG model comparing dasatinib with best supportive care included cost estimates of £88,000 and £80,000 for dasatinib and no treatment respectively. The committee considered that the small cost difference from which this was derived was unlikely to reflect reality, as the costs for best supportive care included in the no treatment arm would also be incurred in the dasatinib treatment arm after treatment with dasatinib is stopped. Neither of the Bristol-Myers Squibb models included best supportive care as a comparator and the committee was not convinced that high-dose imatinib and bone marrow stem cell transplantation were sufficient comparators. This compounded the very poor evidence base supporting the calculations and the committee concluded that dasatinib could not be considered a cost-effective use of NHS resources for the treatment of blast-crisis phase CML.

The committee recognised the innovative nature and major change in the treatment of CML that imatinib has provided since it has been introduced and recommended for use by NICE, and discussed whether dasatinib and nilotinib should be considered to be innovative treatments. The committee considered that the development of dasatinib and nilotinib was not a step change in the treatment of CML when standard-dose imatinib had failed because of resistance or intolerance and did not identify any potential significant and substantial health-related benefits that had not been included in the economic models.

The committee noted the importance of registries in gathering data on CML, particularly when treatment with standard-dose imatinib has failed. It supported collecting information in a suitable registry about treatments, long-term outcomes (particularly overall survival) and treatment-related adverse events in CML that is resistant to standard-dose imatinib.

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 that the assumptions used in the reference case of the economic modelling are plausible, objective and robust.

The committee discussed the possibility that the end-of-life criteria defined by NICE in its supplementary advice might be met by dasatinib or high-dose imatinib for people with blast-crisis phase CML. The committee noted that in the blast-crisis phase of CML, life expectancy is short (about 3 to 6 months). The committee also agreed that this is a very small population, because fewer than 10% of all people with CML will present at the blast-crisis stage. However, the committee agreed that the available evidence on life extension in the blast-crisis phase was too weak and was not considered to be robust. In addition, no data were presented for the interventions as used in clinical practice. The committee concluded that dasatinib and high-dose imatinib do not fulfil the end-of-life criteria for people with blast-crisis phase CML.

The committee discussed whether NICE's duties under the equalities legislation required it to alter or add to its preliminary recommendations in any way. It noted that the submission from Bristol-Myers Squibb highlighted that if dasatinib, high-dose imatinib or nilotinib are not recommended for the treatment of imatinib-resistant CML, then allogeneic stem cell transplantation is the only treatment that may deliver clinical efficacy. Because only a small number of people who have imatinib-resistant CML are eligible for allogeneic stem cell transplantation, this could raise equality issues in relation to race, age (older people), and comorbidities. However, the committee concluded that allowing for clinical decisions relating to a range of possible treatments based on individual assessment of risk and benefit does not limit access to the technology for any specific protected group compared with other people.

This appraisal was a Cancer Drugs Fund partial reconsideration of the published NICE technology appraisal guidance on dasatinib, high-dose imatinib and nilotinib for the treatment of imatinib-resistant chronic myeloid leukaemia (CML) (part review of NICE technology appraisal guidance 70), and dasatinib and nilotinib for people with CML for whom treatment with imatinib has failed because of intolerance. The committee considered the company's (Bristol-Myers Squibb) submission for the Cancer Drugs Fund reconsideration that included:

• a revised patient access scheme that provides a simple discount to the list price of dasatinib

• an updated systematic literature review, which provided a naive comparison of clinical outcomes of dasatinib compared with nilotinib

• a cost-minimisation analysis of dasatinib compared with nilotinib and high-dose imatinib.