TA306 · STA

Pixantrone monotherapy for treating multiply relapsed or refractory aggressive non-Hodgkin's B-cell lymphoma

Recommended with restrictionsAppraisal CommitteeDecember 2013

Pixantrone is recommended only if: the person has previously been treated with rituximab; the person is receiving third- or fourth-line treatment; and the manufacturer provides pixantrone with the discount agreed in the patient access scheme.

Source documents

Intervention

pixantrone (Pixuvri)
aza-anthracenedione analogue and inhibitor of topoisomerase II · anthracycline with reduced cardiotoxicity · intravenous

Condition

multiply relapsed or refractory aggressive non-hodgkin's b-cell lymphomahaematology · relapsed_refractory

Comparators

NameType Established Committee preferred
physician's choice of single-agent comparatorsstandard of careYes
vinorelbineactive drug
oxaliplatinactive drug
ifosfamideactive drug
etoposideactive drug
mitoxantroneactive drug
gemcitabineactive drug
treatment of physician's choicestandard of careYesYes

Clinical trials

TrialDesignPhasePivotal
PIX301RCTphase IIIYes
PIX203RCTII

Economic model

partitioned survival (company)
Time horizon: lifetime (23 years)
Cycle length: not specified

ICER

£20,000–£30,000 (pixantrone vs treatment of physician's choice) · high uncertainty
Confidential (PAS-dependent) (pixantrone vs treatment of physician's choice) · very_high uncertainty
Below £20,000 (pixantrone vs treatment of physician's choice) · moderate uncertainty

Methodological decisions (28)

comparator selection

PIX301 used physician's choice of single-agent comparators (vinorelbine, oxaliplatin, ifosfamide, etoposide, mitoxantrone, gemcitabine) rather than a single standard comparator.

comparator selection

ERG noted lack of consensus on which chemotherapy regimens should be used after second-line treatment fails and concluded that choice of treatment in comparator arm was unlikely to be a key issue, despite small numbers per treatment.

Company: Used treatment of physician's choice

ERG: No consensus on appropriate comparators; small sample sizes preclude reliable analysis

ICER impact: uncertain_direction

cost assumption

Drug costs calculated based on average dose per administration from PIX301 trial. Manufacturer initially used incorrect pixantrone vial price (£343.80) which was corrected to £553.50, with minimal impact on cost-effectiveness estimates

Company: Costs calculated from BNF edition 62 and NHS reference costs based on actual doses used in trial

cost assumption

Drug costs for comparator treatments changed from BNF to NHS Commercial Medicines Unit eMIT database in line with NICE Guide to the methods of technology appraisal (2013).

Company: Updated to use eMIT database

Committee: eMIT database method

ICER impact: uncertain_direction

model structure

Semi-Markov model used with partitioned survival approach. Stable/no progression state divided into 2 subpopulations (on treatment vs off treatment). Adverse events modelled as events rather than health states with time-independent occurrence.

Company: Semi-Markov partition approach chosen as particularly suited to progressive conditions with varying risks over time

model structure

Model assumed overall survival and progression-free survival were independent, leading to survival curves crossing in probabilistic analysis (30% of simulations affected). Manufacturer provided alternative assuming dependencies using Cholesky decomposition.

Company: Independence assumption with manual adjustment; also provided dependent assumption scenario

ICER impact: increases

other

Central independent pathological review undertaken retrospectively (after trial), confirming aggressive disease in only 104 of 140 randomised patients. ERG considered this important for evaluating benefit in patients with confirmed aggressive B-cell lymphoma.

Company: Not practical to confirm aggressive disease at enrolment

ERG: Retrospective confirmation creates uncertainty; subgroup analyses should be evaluated separately as they are underpowered

ICER impact: uncertain_direction

population generalisability

Generalisability of PIX301 to NHS population: only 54 of 70 (77%) patients in pixantrone arm had disease confirmed by central independent pathological review; 36 patients had non-confirmed histology with various reasons for non-confirmation.

Company: Manufacturer considered post-hoc subgroup of patients with aggressive B-cell lymphoma confirmed by onsite pathological review to be similar to the population eligible for treatment according to pixantrone's European marketing authorisation

population generalisability

Base case included patients who had received 2 or 3 prior therapies and whose disease was sensitive to anthracyclines, consistent with pixantrone's European marketing authorisation. This focused on 3rd or 4th line treatment, with note that treatment benefit not established for 5th line or greater.

Company: Population restricted to patients consistent with marketing authorisation (2-3 prior therapies, anthracycline-sensitive)

population generalisability

ERG had concerns about generalisability of PIX301 population to English clinical practice, particularly regarding previous rituximab treatment (given as part of standard first-line treatment in UK). About 50% of PIX301 patients had received a biological agent previously.

ERG: Significant concerns about generalisability, especially regarding rituximab-pretreated subgroup

ICER impact: uncertain_direction

population generalisability

Study population definition: whether the ITT population (including ~10% patients without confirmed aggressive B-cell lymphoma) or the subgroup with disease confirmed by central independent pathological review is most relevant to NHS clinical practice and marketing authorisation

Company: Manufacturer submitted multiple subgroup analyses based on onsite and central pathological review

ERG: ERG judged the subgroup with aggressive B-cell lymphoma confirmed by central independent pathological review to be most relevant because it excluded patients with irrelevant disease (e.g. indolent disease). ERG noted this subgroup was underpowered and based on post-hoc analysis.

Committee: Committee concluded it would be more appropriate to consider the population with aggressive B-cell lymphoma confirmed by central independent pathological review, particularly those who had received third- or fourth-line treatment, when assessing PIX301 results

ICER impact: increases

population generalisability

Histology determination method: onsite single pathologist review vs central multidisciplinary team pathological review

Company: PIX301 population had tumour histology determined by onsite review by single pathologist

ERG: ERG was advised this was not representative of NHS clinical practice where multidisciplinary team review is routine with examination by 2-3 pathologists. ERG considered population with central independent pathological review more relevant.

Committee: Committee agreed that central independent pathological review population was more generalisable because it reflected NHS practice and excluded considerable proportion of patients with confirmed indolent disease

ICER impact: increases

population generalisability

Committee appropriateness of using third- or fourth-line subgroup rather than intention-to-treat population

Company: Intention-to-treat analysis of all 305 enrolled patients

Committee: Use subgroup of patients receiving third- or fourth-line treatment and who had previously received rituximab

ICER impact: decreases

population generalisability

Relevance of tumour histology determination method (onsite single pathologist vs central independent review)

Company: Onsite review by single pathologist

ERG: Central independent pathological review more relevant to NHS practice

Committee: Population with tumour histology confirmed by retrospective central independent pathological review by consensus

ICER impact: decreases

population generalisability

Committee concluded that the subgroup with aggressive B-cell lymphoma confirmed by central independent pathological review receiving third- or fourth-line treatment and who had previously received rituximab was the most appropriate for decision-making, rather than the intention-to-treat population which included patients ineligible under marketing authorisation terms

Company: intention-to-treat population

Committee: subgroup with aggressive B-cell lymphoma confirmed by central independent pathological review, third- or fourth-line treatment, prior rituximab exposure

ICER impact: decreases

surrogate endpoint validity

Choice of primary endpoint in PIX301: complete or unconfirmed complete response vs overall survival/progression-free survival

Company: Manufacturer designed PIX301 with primary endpoint of complete or unconfirmed complete response

ERG: ERG concerned about statistical power. European regulators prefer overall survival or progression-free survival. Unconfirmed complete response is now considered obsolete in modern trials.

Committee: Committee noted this was a fundamental concern about trial design and concluded the endpoint choice meant considerable uncertainty in validity and robustness of results. Clinical specialists stated they prefer studies powered to detect differences in overall survival.

ICER impact: uncertain_direction

survival extrapolation

Manufacturer used log-normal distribution for both progression-free survival and overall survival to extrapolate beyond PIX301 data (around 2 years follow-up)

Company: Log-normal parametric curve fitted to patient-level data from PIX301

survival extrapolation

Treatment of survival modelling uncertainty in probabilistic analysis

Committee: Mean probabilistic ICER of £22,000 may overestimate uncertainty; true value likely lower; median probabilistic ICER £14,700; ICER reduced to £10,000 when assuming PFS and OS not independent

ICER impact: decreases

survival extrapolation

Committee was persuaded that the manufacturer's mean probabilistic ICER of £22,000 per QALY gained could overestimate the uncertainty associated with survival modelling and that the true ICER might be lower

Committee: true value of ICER likely lower than probabilistic estimate

ICER impact: decreases

treatment sequencing

Line of therapy eligibility: Pixantrone's European marketing authorisation notes reduced benefit when used as fifth line or greater, and benefit not established in patients refractory to last therapy. Economic model and manufacturer's analyses focus on third- or fourth-line treatment.

Committee: third- or fourth-line treatment

treatment sequencing

Which line of therapy (third, fourth, or fifth+) is appropriate for pixantrone use within its marketing authorisation

Company: Not explicitly stated in this section

ERG: ERG viewed patients with aggressive B-cell lymphoma confirmed by central independent pathological review who had previously received rituximab and were receiving third- or fourth-line treatment as most relevant population

Committee: Committee concluded pixantrone would most likely be used as third- or fourth-line treatment in clinical practice and noted that fifth-line patients would likely receive palliative care or clinical trials. Marketing authorisation does not restrict to specific lines but fifth-line benefit not established.

ICER impact: increases

utility source

Manufacturer used utility data from published sources for similar disease areas including diffuse large B-cell lymphoma, chronic myelogenous leukaemia, chronic lymphocytic leukaemia, follicular lymphoma, renal cell carcinoma and melanoma. Specifically selected self-reported quality of life in older patients with aggressive diffuse large B-cell lymphoma (pre-progression 0.81, post-progression 0.60) without providing rationale.

Company: Utility values selected from published sources for similar patient populations with similar expected survival, disease progression and quality of life characteristics

utility source

Original model used self-reported quality of life in older patients with aggressive DLBCL (0.81 pre-progression, 0.60 post-progression). Revised model changed to utilities for second- and subsequent-line treatment of renal cell carcinoma (0.76 pre-progression, 0.68 post-progression) in response to Committee concern that original values overestimated quality of life.

Company: Original utilities from DLBCL population

Committee: Utilities from renal cell carcinoma population

ICER impact: decreases

utility source

Source of health-state utility values for aggressive non-Hodgkin's lymphoma

Company: Manufacturer used utility values from a population of patients receiving first-line treatment for aggressive non-Hodgkin's lymphoma

ERG: ERG considered these utility values potentially inappropriate because they were from first-line patients (not relapsed/refractory) and derived from a study initially rejected by the manufacturer. ERG noted values were higher than those for healthy older UK patients. ERG explored alternative utility values from chronic lymphocytic leukaemia patients receiving third- or later-line treatment.

ICER impact: increases

utility source

Choice of utility values for health states

Company: Utility values for patients receiving second- and subsequent-line treatment for renal cell carcinoma (0.76 pre-progression, 0.68 post-progression)

ERG: Utility values for final-line treatment for chronic lymphocytic leukaemia (0.428 pre-progression, 0.279 post-progression)

Committee: Manufacturer's utility values acceptable; ERG's values likely to underestimate quality of life

ICER impact: uncertain_direction

utility source

Committee aware that utility value used by manufacturer for pre-progression health state was similar to that expected for an older UK population, whereas quality of life for patients receiving third- or fourth-line treatment could be lower. However, Committee concluded the manufacturer's utility values were appropriate for decision-making

Committee: manufacturer's revised model utility values are acceptable despite uncertainty

ICER impact: uncertain_direction

utility value choice

Manufacturer assumed no difference in baseline health-related quality of life between patients on treatment vs discontinued treatment within stable/no progression state, and no difference between complete response, partial response or stable disease

Company: All stable/no progression patients assumed to have similar quality of life regardless of treatment status or response type

utility value choice

ERG's exploratory adverse-event disutilities were incorporated into manufacturer's revised model. Any grade 1-4 adverse event occurring in less than 5% of trial population assumed to have no impact on quality of life.

Company: Conservative approach; manufacturer-determined disutilities from literature

ERG: Provided exploratory adverse-event disutilities

Committee: ERG's exploratory disutility values

ICER impact: uncertain_direction

Evidence gaps

no direct comparisonPIX301 compared pixantrone with physician's choice of multiple single-agent comparators rather than a single standard comparator
single arm evidence onlyNo randomised trials directly comparing pixantrone monotherapy with a single standard comparator regimen; trial used physician's choice of heterogeneous comparators
immature overall survivalNo statistically significant difference in overall survival between pixantrone and comparator group in intention-to-treat population
short follow upOnly 26 of 95 patients completed 18 months of follow-up after completing study treatment
short follow upCentral independent pathological review undertaken retrospectively after trial completion rather than at enrolment
no direct comparisonNo statistically significant differences between pixantrone and comparator arms for complete/unconfirmed complete response, progression-free survival or overall survival in subgroup with central pathological confirmation and prior rituximab treatment
single arm evidence onlySmall subgroup sample sizes in analyses by central pathological confirmation; statistical power of subgroup analyses likely less than intention-to-treat population
single arm evidence onlyPIX301 failed to recruit the planned number of patients, resulting in underpowered trial
short follow upUncertainty about utility data in patients with relapsed or refractory aggressive non-Hodgkin's lymphoma; current utility values may overestimate utility of patients receiving third- or fourth-line treatment
surrogate not validatedPrimary endpoint of complete or unconfirmed complete response is a surrogate endpoint; relationship to overall survival not clearly established. Unconfirmed complete response now considered obsolete.
immature overall survivalNo statistically significant difference in overall survival between treatment arms for any groups presented by manufacturer
single arm evidence onlyPIX301 underpowered - accrued less than half of planned 320 patients; failed to recruit planned number
short follow upLimited robustness of clinical effectiveness evidence in Committee's preferred subgroup
single arm evidence onlyPIX301 trial was underpowered and failed to recruit the planned number of patients
short follow upConcerns about the statistical non-significance of results for response rates, progression-free survival, and overall survival in the subgroup receiving third- or fourth-line treatment who had previously received rituximab
no direct comparisonOnly just over half of patients in PIX301 had previously received rituximab, whereas rituximab is integral to standard first-line and often second-line treatment in the NHS

Commercial arrangement

simple discount pas · confidential · critical for recommendation

Special considerations

End of life considered (not met) Innovation acknowledged

Cross-references

same condition — PIX306 study results expected 2015 - larger phase III study comparing pixantrone plus rituximab with gemcitabine plus rituximab; guidance should be reviewed when results available
TA65same condition — Rituximab for aggressive non-Hodgkin's lymphoma
cdf reconsideration — PIX306 trial results expected in 2015; guidance to be reviewed by NICE once results available