Survival Extrapolation Research Articles

P45 The Use of Historical Clinical Trial Data to Inform Survival Extrapolation

Standard parametric distributions are commonly used for the extrapolation of survival data in cost-effectiveness analyses. However, survival data is often immature and uncertainty remains around the survival extrapolations. Mature historical data can be used to better predict survival beyond trial data. This study assessed two methods to incorporate historical data in the extrapolation of immature survival data.

P37 The Cure Myth: Experience from Recent Health Technology Assessment (HTA) Submissions in Relapsed/Refractory Diffuse Large B-Cell Lymphoma (R/R DLBCL)

Recent clinical trials testing novel cancer therapies such as CAR-Ts are detecting a plateau in the survival curves, conveying a potential opportunity for cure. Long-term survival data are limited, and clinicians are often cautious to make conclusions on the basis of the trial outcomes. Nevertheless, cure assumptions have appeared in oncology HTA submissions such as those in R/R DLBCL. This study aims to investigate how cure assumptions are applied across different HTA submissions for therapies in R/R DLBCL. Publicly available HTA dossiers in R/R DLBCL submitted to NICE, SMC, CADTH, ICER, PBAC and HAS in the past five years were reviewed. Information related to cure assumptions (methodology, cure time point, cure proportion) and the critiques received by the HTA bodies were reviewed and extracted. Of the twelve HTA documents identified, ten submissions applied cure assumption in overall survival extrapolation. These submissions focused on two CAR-T agents (axicabtagene and tisagenlecleucel) and polatuzumab vedotin. From the ten submissions that reported cure assumptions, six used cure-mixture models and four used simple cure assumptions. Eight models set mortality of cured patients to that of the general population, while two considered an adjustment factor. Cure was considered only for the intervention arm in seven models, while the rest considered it for all treatments. The timepoint and proportion of cure varied across the submissions. Critiques were mainly around the issues caused by immaturity of survival data, such as uncertainty in cure assumptions and the choice of cure timepoint. While cure modelling is frequently used in R/R DLBCL and in oncology, there seems to be no consensus on the implication of cure and its implementation in the economic analyses. Clear guidance must be provided by health authorities on how to approach the cure concept in economic modelling and how to validate the assumptions.

P49 Accuracy of Life Year Gains Predictions for CAR-T Therapy in the Long Term: An Analysis for Axicabtagene Ciloleucel in Refractory Large B-Cell Lymphoma

Survival profiles for chimeric antigen receptor T cell (CAR-T) therapies commonly exhibit a plateau that may indicate a cure. This leads to challenges predicting long-term outcomes for novel CAR-T therapies when trial data are immature. This study retrospectively analysed the accuracy of overall survival (OS) extrapolations from interim data cuts in predicting realised long-term life years (LYs) for axicabtagene ciloleucel in patients with refractory large B-cell lymphoma. Published OS data for axicabtagene ciloleucel from successive data cuts of ZUMA-1 (median follow-ups of 15.4, 27.1 and 51.1 months) were digitised. Standard parametric, spline (1–2 knots; normal, odds and hazard) and mixture cure models (MCMs) were fitted to the first two data cuts. Statistical fit was tested using Akaike and Bayesian information criteria (AIC and BIC). Cumulative LYs were estimated for each model over a 58-month time horizon, corresponding to the longest duration of published OS data. These projected LYs were then compared to realised LYs over this period. At the earliest data cut, MCMs provided the best predictions of realised LYs, with a mean absolute difference of 6.8% between predicted and realised LYs across MCM models (range: 0.5%–10.7%). Standard parametric extrapolations considerably underestimated realised LYs (mean absolute difference: 19.2%; range: 9.9%–28.0%), whilst spline models offered a mean absolute difference of 8.3% (range: 1.4%–13.2%). Similar findings were observed for extrapolations based on the second data cut (representing 11.7 months additional follow-up), but differences between model classes were less pronounced. MCMs may offer the best predictions of long-term survival for CAR-T therapies, particularly when only short-term data are available. Standard parametric models may be inappropriate to predict survival when extrapolating immature data, failing to capture the plateau in survival typical of CAR-T therapies. Further research is required to determine whether these findings are generalisable across CAR-T therapies and indications.

Key questions in marine mammal bioenergetics.

Bioenergetic approaches are increasingly used to understand how marine mammal populations could be affected by a changing and disturbed aquatic environment. There remain considerable gaps in our knowledge of marine mammal bioenergetics, which hinder the application of bioenergetic studies to inform policy decisions. We conducted a priority-setting exercise to identify high-priority unanswered questions in marine mammal bioenergetics, with an emphasis on questions relevant to conservation and management. Electronic communication and a virtual workshop were used to solicit and collate potential research questions from the marine mammal bioenergetic community. From a final list of 39 questions, 11 were identified as 'key' questions because they received votes from at least 50% of survey participants. Key questions included those related to energy intake (prey landscapes, exposure to human activities) and expenditure (field metabolic rate, exposure to human activities, lactation, time-activity budgets), energy allocation priorities, metrics of body condition and relationships with survival and reproductive success and extrapolation of data from one species to another. Existing tools to address key questions include labelled water, animal-borne sensors, mark-resight data from long-term research programs, environmental DNA and unmanned vehicles. Further validation of existing approaches and development of new methodologies are needed to comprehensively address some key questions, particularly for cetaceans. The identification of these key questions can provide a guiding framework to set research priorities, which ultimately may yield more accurate information to inform policies and better conserve marine mammal populations.

POSA303 Influence of Treatment Effect on Parameters in Fractional Polynomials for Network Meta-Analyses

Traditionally, 2nd order fractional polynomials (FP) have three parameters (β0, β1, β2) for each treatment in the trial plus power parameters p1 and p2. NICE DSU 21 provides guidance on flexible survival extrapolation models and the future research based on this DSU, involves parametrization of treatment effects, however, FPs are not included within the analysis.

POSB269 The Use of External Data to Inform Survival Extrapolation in Oncology Technology Appraisals: A Comprehensive Review

POSB269 The Use of External Data to Inform Survival Extrapolation in Oncology Technology Appraisals: A Comprehensive Review

POSB310 The Impact of Small Sample Size on Selecting Appropriate Distributions for Parametric Survival Extrapolations Models: A Methodological Commentary and Simulation Study

This simulation study explored the impact of sample size on parametric survival extrapolation model selection guided using statistical criteria and the potential effect of erroneous extrapolation on economic evaluations. Weibull survival data were simulated using ‘simsurv’ package in R, applying scale parameter 1.0 and consecutive shape parameters 0.5, 1.5, and 5.0. Datasets were randomly sampled with sample sizes increasing from 25 to 250 and extrapolated using seven functional forms 1000 times per sample size. The concordance between the actual and selected parametric survival distribution was evaluated at each sample size. Akaike information criterion (AIC) and Bayesian information criterion (BIC) were used to identify the best fitting distributions. Differences compared using the restricted mean survival time (RMST). The Weibull distribution was the most probable distribution selected at all sample sizes using BIC but was selected only 31.3%, 51.9%, and 35.6% during 1000 simulations at n=25, for shape parameters 0.5, 1.5, and 5.0, respectively. Weibull selection increased linearly with each n=25 increment, by 6.15% (p=0.01), 6.38% (p<0.01), and 4.49% (p=0.01), for shape parameters 0.5, 1.5, and 5.0, respectively. AIC demonstrated similar trends to BIC apart from increasing generalized gamma selection at sample sizes greater than n=125 (p>0.05). The relative differences in the RMST to the sampled data were greatest in the log-normal and log-logistic distributions. These distributions were selected more frequently when the sample size was small and the shape parameter less than 1.0 resulting in an overestimation of the RMST by 151%. Sample size influences the probability of selecting the appropriate parametric survival extrapolation distribution to model survival outcomes when informed by statistical criteria. An incorrect extrapolation of the sampled data can lead to under- and overestimation of the RMST, directly impacting the calculation of the benefits and costs in an economic evaluation.

POSB259 Immuno-Oncology Treatments: Key Factors Influencing NICE Approval, 2011-2020

Over the last decade, the rapid development of immuno-oncology (I-O) therapies has transformed the cancer treatment landscape. However, a thorough analysis of the rationale behind I-O reimbursement is yet to be provided. We evaluated the I-O treatment appraisal process by the National Institute for Health and Care Excellence (NICE) in the UK, to quantify their success rate and highlight the main reimbursement barriers. We used the Cancer Research Institute classification to identify I-O therapies published between 2011-2020 from the NICE website (excluding terminated appraisals). We then extracted approval determinants, including all NICE critiques of I-O treatment clinical and cost-effectiveness evidence, from Final Appraisal Documents. Of the 75 I-O appraisals identified, 62.6% (47/75) were recommended for reimbursement by the National Health Service, 22.7% (17/75) within the Cancer Drugs Fund (CDF), and 14.7% (11/75) were not recommended. Most NHS-recommended treatments (59.6%, 28/47) demonstrated significantly improved overall survival (OS) during clinical trials, as opposed to 35.3% (6/17) of CDF-recommended treatments, for which OS uncertainties were frequent. The majority of non-recommended treatment also improved OS (63.6%, 7/11); the main barrier to their reimbursement was their incremental cost-effectiveness ratio (ICER) exceeding the NICE accepted threshold. Notably, due to modest clinical effectiveness only 27.3% (3/11) of non-recommended treatments met NICE's end-of-life criteria (for which the accepted ICER threshold is substantially higher than treatments that do not meet these criteria) as opposed to 53.2% (25/47) of recommended treatments. Additional barriers to recommendation were OS data immaturity, leading to uncertainties in long-term survival extrapolations (3/11), and lack of survival benefit (3/11). Trial treatment regimen and choice of efficacy outcomes were also frequently criticised by NICE. Most I-O treatments receive positive recommendations by NICE. However, manufacturers should beware that high ICERs, not meeting end-of-life criteria and long-term survival benefit uncertainties can undermine I-O approval.

A Review of Overall Survival Extrapolations of Immune-Checkpoint Inhibitors Used in Health Technology Assessments by the French Health Authorities.

Extrapolation is often required to inform cost-effectiveness (CE) evaluations of immune-checkpoint inhibitors (ICIs) since survival data from pivotal clinical trials are seldom complete. The objectives of this study were to evaluate the accuracy of estimates of long-term overall survival (OS) predicted in French CE assessment reports of ICIs, and to identify models presenting the best fit to the observed long-term survival data. A systematic review of French assessment reports of ICIs in the metastatic setting since inception until May 2020 was performed. A targeted literature review was conducted to collect associated extended follow-up of randomized controlled trials (RCTs) used in the CE assessment reports. Difference between projected and observed OS was calculated. A range of standard parametric and spline-based models were applied to the extended follow-up data from the RCT to determine the best-fitting survival models. Of the 121 CE assessment reports published, 11 reports met the inclusion criteria. OS was underestimated in 73percent of the CE assessment reports. The mean relative difference between each source was -13percent (median: -15percent; IQR: -0.4 to 26percent). Models providing the best fit were those that could reflect nonmonotonic hazards. Based on the available data at the time of submission, longer-term survival of ICIs was not fully captured by the extrapolation models used in CE assessments. Standard and flexible parametric models which can capture nonmonotonic hazard functions provided the best fit to the extended follow-up data. However, these models may still have performed poorly if fitted to survival data available at the time of submission to the French National Authority for Health.

Impact of limited sample size and follow-up on single event survival extrapolation for health technology assessment: a simulation study

IntroductionExtrapolation of time-to-event data from clinical trials is commonly used in decision models for health technology assessment (HTA). The objective of this study was to assess performance of standard parametric survival analysis techniques for extrapolation of time-to-event data for a single event from clinical trials with limited data due to small samples or short follow-up.MethodsSimulated populations with 50,000 individuals were generated with an exponential hazard rate for the event of interest. A scenario consisted of 5000 repetitions with six sample size groups (30–500 patients) artificially censored after every 10% of events observed. Goodness-of-fit statistics (AIC, BIC) were used to determine the best-fitting among standard parametric distributions (exponential, Weibull, log-normal, log-logistic, generalized gamma, Gompertz). Median survival, one-year survival probability, time horizon (1% survival time, or 99th percentile of survival distribution) and restricted mean survival time (RMST) were compared to population values to assess coverage and error (e.g., mean absolute percentage error).ResultsThe true exponential distribution was correctly identified using goodness-of-fit according to BIC more frequently compared to AIC (average 92% vs 68%). Under-coverage and large errors were observed for all outcomes when distributions were specified by AIC and for time horizon and RMST with BIC. Error in point estimates were found to be strongly associated with sample size and completeness of follow-up. Small samples produced larger average error, even with complete follow-up, than large samples with short follow-up. Correctly specifying the event distribution reduced magnitude of error in larger samples but not in smaller samples.ConclusionsLimited clinical data from small samples, or short follow-up of large samples, produce large error in estimates relevant to HTA regardless of whether the correct distribution is specified. The associated uncertainty in estimated parameters may not capture the true population values. Decision models that base lifetime time horizon on the model’s extrapolated output are not likely to reliably estimate mean survival or its uncertainty. For data with an exponential event distribution, BIC more reliably identified the true distribution than AIC. These findings have important implications for health decision modelling and HTA of novel therapies seeking approval with limited evidence.

Parametric Survival Extrapolation of Early Survival Data in Economic Analyses: A Comparison of Projected Versus Observed Updated Survival

ObjectivesTo establish the value of cancer drugs by cost-effectiveness analysis, lifetime parametric survival extrapolations are often fitted to early data. Recent literature suggests that the benefit of cancer agents in primary publications is often different compared with updated data. This study aimed to examine the projected survival based on parametric extrapolations compared with observed survival based on updated data. MethodsUS Food and Drug Administration oncology approvals from January 2006 to December 2015 were reviewed to identify randomized controlled trials, with updated overall survival (OS) or progression-free survival (PFS) data within 5 years. Individual patient data were reconstructed using established methods on initial and updated publications. Projected survival was calculated as the best-fit parametric restricted mean survival time (RMST) based on extrapolated initial Kaplan-Meier curves whereas observed survival was calculated as observed RMST based on updated Kaplan-Meier curves. Mean deviations, mean absolute error (MAE), mean absolute percentage error, and linear regressions were conducted to examine the relationship between projected and observed survival. ResultsIn total, 32 randomized controlled trials were included. The MAE between the projected RMST and observed RMST was 3.18 months (OS) and 2.84 months (PFS) and absolute percentage error of 100% (OS) and 23% (PFS), suggesting substantial imprecision of the projected RMST in predicting the updated RMST. The linear regression indicated MAE increased as time extrapolated and as the percentage of censored patients increased. ConclusionsThis study demonstrated substantial difference in projected survival between initial and updated publications. Health technology assessment committees need to be aware of the potential uncertainty of incremental effectiveness and resultant value-for-money assessment when making reimbursement decisions based on initial publications with immature survival data.

Approaches for Enhanced Extrapolation of Long-Term Survival Outcomes Using Electronic Health Records of Patients With Cancer

ObjectivesThis study aimed to demonstrate enhanced survival extrapolation methods using electronic health record-derived real-world data (RWD). MethodsThe study population included patients diagnosed of ER+/HER2− metastatic breast cancer who started first-line treatment with anastrozole or letrozole between November 18, 2014, and November 18, 2015. Two patient cohorts were constructed: a clinical trial cohort from digitized MONARCH-3 clinical trial results and a RWD cohort from a deidentified electronic health record-derived database. RWD patients were weighted to trial baseline covariate distributions. Standard parametric approaches were applied to trial data and a “best-fit” model was selected. We demonstrate traditional and enhanced hybrid (pooling with weighted RWD at start, 75%, or end of trial) extrapolation approaches. ResultsObserved and estimated 5-year progression-free survival (PFS) rates in extrapolating the trial control arm (n = 165) were comparable across all methods. Compared with the observed 5-year mean PFS in the RWD cohort (n = 118) of 20.4 months (95% confidence interval [CI] 16.9-23.8), there was some variation among studied methods. Best-fit standard parametric model (log-normal) had 5-year mean PFS of 21.3 months (95% CI 18.2-24.9), and for the hybrid methods in order of estimate conservativeness was start of trial (20.8 months; 95% CI 18.5-23.2), 75% of trial (21.3 months; 95% CI 18.1-24.5), and end of trial (21.8 months; 95% CI 18.8-25.2). ConclusionsOur study leverages RWD to enhance long-term survival extrapolation. Future use cases should include applying patient eligibility criteria, weighting on baseline characteristics, and choice of time window to add RWD to trial data.

Cost Effectiveness of Durvalumab in Unresectable Stage III NSCLC: 4-Year Survival Update and Model Validation from a UK Healthcare Perspective.

BackgroundIn the phase III PACIFIC study, durvalumab improved survival versus placebo in patients with unresectable stage III non-small-cell lung cancer (NSCLC) whose disease had not progressed after platinum-based concurrent chemoradiotherapy. The appraisal by the UK’s National Institute for Health and Care Excellence (NICE) included a cost-effectiveness analysis based on an early data readout from PACIFIC [March 2018 data cut-off (DCO); median follow-up duration 25.2 months; range 0.2–43.1]. Uncertainties regarding long-term survival outcomes with durvalumab led to some challenges in estimating the cost effectiveness of this therapy.ObjectiveHere, we validate the survival extrapolations used in the original company base-case analysis by benchmarking them against updated survival data from the 4-year follow-up analysis of PACIFIC (i.e. approximately 4 years after the last patient was randomised; March 2020 DCO; median follow-up duration 34.2 months; range 0.2–64.9). Moreover, we update the original analysis with these more mature survival data to examine the consistency of key economic outputs with the original analysis.MethodsThe original analysis used a semi-Markov (state-transition) approach and was based on patients whose tumours expressed programmed cell death-ligand 1 on ≥ 1% of cells (to reflect the European licence for durvalumab). We benchmarked the survival extrapolations used in the original company base-case analysis against survival data from the 4-year follow-up of PACIFIC and updated the cost-effectiveness analysis with these more mature survival data. Early deaths avoided by the adoption of durvalumab into the UK Cancer Drugs Fund (CDF) in March 2019 were estimated using the 4-year follow-up survival data and an assumed uptake of 125 patients/year (lower estimate) and 367 patients/year (higher estimate).ResultsThe original company base-case analysis had a good visual fit with the observed overall survival (OS) distribution for the durvalumab arm and accurately predicted the 48-month OS rate (predicted 55%; observed 55%); by comparison, the fit was less precise for the placebo arm, for which the analysis underestimated the 48-month OS rate (predicted 32%; observed 38%). In the updated company base-case analysis, durvalumab yielded 2.51 incremental quality-adjusted life-years (QALYs) (− 0.43 vs. the original company base-case analysis), corresponding to an incremental cost-effectiveness ratio of £22,665/QALY (+£3298 vs. the original analysis), which falls within the upper bound of NICE’s willingness-to-pay threshold (£30,000/QALY gained). We estimate that between 31 and 91 early patient deaths may have been avoided by the adoption of durvalumab into the CDF.ConclusionsThese findings reinforce the patient benefit observed with durvalumab in unresectable stage III NSCLC, support the routine use and cost effectiveness of this therapy, and demonstrate how appropriate modelling can inform the early adoption of therapies by payers to achieve patient benefit.Supplementary InformationThe online version contains supplementary material available at 10.1007/s41669-021-00301-7.

Applying State-of-the-Art Survival Extrapolation Techniques to the Evaluation of CAR-T Therapies: Evidence from a Systematic Literature Review.

IntroductionTraditional statistical techniques for extrapolating short-term survival data for anticancer therapies assume the same mortality rate for noncured and “cured” patients, which is appropriate for projecting survival of non-curative therapies but may lead to an underestimation of the treatment effectiveness for potentially curative therapies. Our objective was to ascertain research trends in survival extrapolation techniques used to project the survival benefits of chimeric antigen receptor T cell (CAR-T) therapies.MethodsA global systematic literature search produced a review of survival analyses of CAR-T therapies, published between January 1, 2015 and December 14, 2020, based on publications sourced from MEDLINE, scientific conferences, and health technology assessment agencies. Trends in survival extrapolation techniques used, and the rationale for selecting advanced techniques, are discussed.ResultsTwenty publications were included, the majority of which (65%, N = 13) accounted for curative intent of CAR-T therapies through the use of advanced extrapolation techniques, i.e., mixture cure models [MCMs] (N = 10) or spline-based models (N = 3). The authors’ rationale for using the MCM approach included (a) better statistical fits to the observed Kaplan–Meier curves (KMs) and (b) visual inspection of the KMs indicated that a proportion of patients experienced long-term remission and survival which is not inherently captured in standard parametric distributions.DiscussionOur findings suggest that an advanced extrapolation technique should be considered in base case survival analyses of CAR-T therapies when extrapolating short-term survival data to long-term horizons extending beyond the clinical trial duration.ConclusionAdvanced extrapolation techniques allow researchers to account for the proportion of patients with an observed plateau in survival from clinical trial data; by only using standard-partitioned modeling, researchers may risk underestimating the survival benefits for the subset of patients with long-term remission. Sensitivity analysis with an alternative advanced extrapolation technique should be implemented and re-assessment using clinical trial extension data and/or real-world data should be conducted as longer-term data become available.Supplementary InformationThe online version contains supplementary material available at 10.1007/s12325-021-01841-4.

Cost-effectiveness (CE) analysis of pembrolizumab as first-line (1L) treatment for recurrent or metastatic head and neck squamous cell carcinoma (R/M HNSCC) in the United States (US).

e18013 Background: Pembrolizumab received FDA approval in 1L R/M HNSCC as monotherapy (P) in patients with combined positive score (CPS) ≥1, and in combination with platinum + 5-Fluorouracil (5-FU) chemotherapy (P + C) in the total population, based on KEYNOTE 048 (KN048) trial. Taking a US payer perspective, we evaluated the CE of P and P + C versus KN048 trial comparator cetuximab + platinum + 5-FU (EXTREME regimen), and other comparators commonly used in the US (i.e. cisplatin+docetaxel+cetuximab [TPEx regimen], platinum + paclitaxel/taxane [Pt + T] and platinum+5-FU [Pt + F]). Methods: A 3-state partitioned-survival model was used to project costs and outcomes over 20 years with a 3% annual discount rate. For P, P + C and EXTREME, health state occupancy was based on KN048 trial results with long-term parametric extrapolation of progression-free survival (PFS) and overall survival (OS) guided by statistical criteria. A network meta-analysis (NMA) was applied to the P and P + C interventions in KN048 to project the PFS and OS outcomes for other comparators. The time on treatment (ToT) for P, P+C and EXTREME was derived from the KN048 trial, whereas for other comparators, ToT was approximated by PFS. Costs of 1L and subsequent treatments, disease management, adverse event, and terminal care were included. Utilities were derived using the EuroQoL five-dimension data from KN048 and a US value set. Results: CE ratios, total and incremental costs and quality-adjusted life years (QALYs) from the base case analysis are summarized in the table below. P was more effective and less costly (i.e. dominant) versus EXTREME and TPEx. P + C was highly cost-effective versus EXTREME ($1,789/QALY) and TPEx regimen (dominant). CE ratios for both P and P + C versus all comparators were below a $100,000/QALY willingness-to-pay threshold. Probabilistic sensitivity analysis also showed a 100% CE probability for P and P + C at the $100,000/QALY threshold. Conclusions: P and P + C are cost-effective treatment options in the US with CE ratios below $100,000/QALY gained versus commonly used therapies in 1L R/M HNSCC.[Table: see text]

Comparison of Parametric Survival Extrapolation Approaches Incorporating General Population Mortality for Adequate Health Technology Assessment of New Oncology Drugs

ObjectivesSurvival extrapolation of trial outcomes is required for health economic evaluation. Generally, all-cause mortality (ACM) is modeled using standard parametric distributions, often without distinguishing disease-specific/excess mortality and general population background mortality (GPM). Recent National Institute for Health and Care Excellence guidance (Technical Support Document 21) recommends adding GPM hazards to disease-specific/excess mortality hazards in the log-likelihood function (“internal additive hazards”). This article compares alternative extrapolation approaches with and without GPM adjustment. MethodsSurvival extrapolations using the internal additive hazards approach (1) are compared to no GPM adjustment (2), applying GPM hazards once ACM hazards drop below GPM hazards (3), adding GPM hazards to ACM hazards (4), and proportional hazards for ACM versus GPM hazards (5). The fit, face validity, mean predicted life-years, and corresponding uncertainty measures are assessed for the active versus control arms of immature and mature (30- and 75-month follow-up) multiple myeloma data and mature (64-month follow-up) breast cancer data. ResultsThe 5 approaches yielded considerably different outcomes. Incremental mean predicted life-years vary most in the immature multiple myeloma data set. The lognormal distribution (best statistical fit for approaches 1-4) produces survival increments of 3.5 (95% credible interval: 1.4-5.3), 8.5 (3.1-13.0), 3.5 (1.3-5.4), 2.9 (1.1-4.5), and 1.6 (0.4-2.8) years for approaches 1 to 5, respectively. Approach 1 had the highest face validity for all data sets. Uncertainty over parametric distributions was comparable for GPM-adjusted approaches 1, 3, and 4, and much larger for approach 2. ConclusionThis study highlights the importance of GPM adjustment, and particularly of incorporating GPM hazards in the log-likelihood function of standard parametric distributions.

Abstract PS9-33: Cost-effectiveness of neratinib for the extended adjuvant treatment of adult patients with early-stage, HR+, HER2-overexpressed/amplified breast cancer who initiated neratinib within 1 year of completing trastuzumab in the US

Abstract Background: Neratinib, an oral, irreversible tyrosine kinase inhibitor of multiple human epidermal growth factor receptors (HERs), is indicated for the extended adjuvant treatment of adult patients with, HER2-positive early-stage breast cancer (eBC) to follow adjuvant trastuzumab-based therapy in the United States (US). However, greater efficacy has been seen in patients with hormone receptor-positive (HR+) breast cancer and who initiated neratinib within 1 year of completing trastuzumab. Objective: This study explores the cost-effectiveness of neratinib versus placebo from a US third-party payer perspective in adult patients with HR+ HER2-overexpressed/amplified eBC and who are less than 1 year from the completion of prior adjuvant trastuzumab-based therapy. Methods: A Markov model was constructed to model costs and health outcomes of neratinib and placebo over a lifetime horizon. The model consisted of five health states representing the primary stages of disease in eBC—disease-free, local recurrence, remission, distant recurrence, and dead—and corresponds to the primary and secondary endpoints in the neratinib ExteNET trial. Overall survival was modeled based on a combination of post-distant recurrence survival and general population mortality, assuming all cancer-related mortality would occur through the distant recurrence health state. Statistical extrapolation of invasive disease-free survival and post-distant recurrence survival were derived from the ExteNET clinical trial data as well as the proportion of local and distant recurrence and number of adverse events. Non-breast cancer-related mortality was derived from US life tables. Costs per treatment arm were calculated based on drug acquisition, administration, and monitoring costs as well as treatment costs for adverse events and health state-related medical resource use. Quality-adjusted life-years (QALYs) were estimated per health state, and disutilities were applied per adverse event independent of treatment. Utility values for the invasive disease-free survival health state and diarrhoea were estimated using the EQ-5D-3L data collected in ExteNET. Additional health state utilities and adverse events disutilities were identified from the literature. Costs were adjusted to 2020 US dollars. One-way (OWSA) and probabilistic sensitivity analyses (PSA) and scenario analyses were performed to investigate the robustness of the results. Further, analyses were performed to investigate the cost-effectiveness for the subgroup of patients who did not achieve a pathologic complete response (pCR) after neoadjuvant therapy. Results: The results of the base case analysis showed that, compared with placebo, neratinib treatment generated an additional 0.88 incremental QALYs, 0.99 incremental life-years, and a resulting cost per QALY gained of $56,367. The results were robust across multiple scenario analyses with incremental cost-effectiveness ratios (ICERs) below $70,000. OWSA showed that the parameters that most influenced the results were variations in treatment-related costs, efficacy, and health state utility values. The PSA indicated that the probability of neratinib being cost-effective at a $100,000 per QALY threshold was above 75%. The analysis for the subgroup of patients who did not achieve a pCR after neoadjuvant therapy resulted in an improved ICER. Conclusions: The current cost-effectiveness analysis shows that neratinib is a cost-effective therapy for treating adult patients with HR+ HER2-overexpressed/amplified eBC and who are less than 1 year from the completion of prior adjuvant trastuzumab-based therapy in the US and in patients who did not achieve a pCR after neoadjuvant treatment. Citation Format: Thor-Henrik Brodtkorb, Cheryl McDade, Deepa Lalla, Frankie Ann Holmes. Cost-effectiveness of neratinib for the extended adjuvant treatment of adult patients with early-stage, HR+, HER2-overexpressed/amplified breast cancer who initiated neratinib within 1 year of completing trastuzumab in the US [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS9-33.

Cost-Effectiveness of Durvalumab After Chemoradiotherapy in Unresectable Stage III NSCLC: A US Healthcare Perspective.

Durvalumab was approved by the FDA in February 2018 for patients with unresectable stage III NSCLC that has not progressed after platinum-based concurrent chemoradiotherapy (cCRT), and this regimen is the current standard of care. The objective of this study was to examine the cost-effectiveness of durvalumab following cCRT versus cCRT alone in patients with locally advanced, unresectable stage III NSCLC. A 3-state semi-Markov model was used. Modeling was performed in a US healthcare setting from Medicare and commercial payer perspectives over a 30-year time horizon. Clinical efficacy (progression-free and post progression survival) and utility inputs were based on PACIFIC study data (ClinicalTrials.gov identifier: NCT02125461; data cutoff March 22, 2018). Overall survival extrapolation was validated using overall survival data from a later data cutoff (January 31, 2019). The main outcome was the incremental cost-effectiveness ratio (ICER) of durvalumab following cCRT versus cCRT alone, calculated as the difference in total costs between treatment strategies per quality-adjusted life-year (QALY) gained. In the base-case analysis, durvalumab following cCRT was cost-effective versus cCRT alone from Medicare and commercial insurance perspectives, with ICERs of $55,285 and $61,111, respectively, per QALY gained. Durvalumab was thus considered cost-effective at the $100,000 willingness-to-pay (WTP) threshold. Sensitivity analyses revealed the model was particularly affected by variables associated with subsequent treatment, although no tested variable increased the ICER above the WTP threshold. Scenario analyses showed the model was most sensitive to assumptions regarding time horizon, treatment effect duration, choice of fitted progression-free survival curve, subsequent immunotherapy treatment duration, and use of a partitioned survival model structure. In a US healthcare setting, durvalumab was cost-effective compared with cCRT alone, further supporting the adoption of durvalumab following cCRT as the new standard of care in patients with unresectable stage III NSCLC.

What Did Time Tell Us? A Comparison and Retrospective Validation of Different Survival Extrapolation Methods for Immuno-Oncologic Therapy in Advanced or Metastatic Renal Cell Carcinoma.

The immuno-oncologic (IO) mechanism of action may lead to an overall survival (OS) hazard that changes over time, producing shapes that standard parametric extrapolation methods may struggle to reflect. Furthermore, selection of the most appropriate extrapolation method for health technology assessment is often based on trial data with limited follow-up. To examine this problem, we fitted a range of extrapolation methods to patient-level survival data from CheckMate 025 (NCT01668784, CM-025), a phase III trial comparing nivolumab with everolimus for previously treated advanced renal cell carcinoma (aRCC), to assess their predictive accuracy over time. Six extrapolation methods were examined: standard parametric models, natural cubic splines, piecewise models combining Kaplan-Meier data with an exponential or non-exponential distribution, response-based landmark models, and parametric mixture models. We produced three database locks (DBLs) at minimum follow-ups of 15, 27, and 39 months to align with previously published CM-025 data. A three-step evaluation process was adopted: (1) selection of the distribution family for each method in each of the three DBLs, (2) internal validation comparing extrapolation-based landmark and mean survival with the latest CM-025 dataset (minimum follow-up, 64 months), and (3) external validation of survival projections using clinical expert opinion and long-term follow-up data from other nivolumab studies in aRCC (CheckMate 003 and CheckMate 010). All extrapolation methods, with the exception of mixture models, underestimated landmark and mean OS for nivolumab compared with CM-025 long-term follow-up data. OS estimates for everolimus tended to be more accurate, with four of the six methods providing landmark OS estimates within the 95% confidence interval of observed OS as per the latest dataset. The predictive accuracy of survival extrapolation methods fitted to nivolumab also showed greater variation than for everolimus. The proportional hazards assumption held for all DBLs, and a dependent log-logistic model provided reliable estimates of longer-term survival for both nivolumab and everolimus across the DBLs. Although mixture models and response-based landmark models provided reasonable estimates of OS based on the 39-month DBL, this was not the case for the two earlier DBLs. The piecewise exponential models consistently underestimated OS for both nivolumab and everolimus at clinically meaningful pre-specified landmark time points. This aRCC case study identified marked differences in the predictive accuracy of survival extrapolation methods for nivolumab but less so for everolimus. The dependent log-logistic model did not suffer from overfitting to early DBLs to the same extent as more complex methods. Methods that provide more degrees of freedom may accurately represent survival for IO therapy, particularly if data are more mature or external data are available to inform the long-term extrapolations.

EX1 How to Select the Appropriate Survival Extrapolation Method? a Proposed Framework

For HTA, apart from standard parametric models (SPMs), other extrapolation approaches are applied such as piecewise, splines, mixture cure and non-mixture cure, parametric-mixture, and fractional polynomial models. The objective is to validate a survival extrapolation decision framework suggesting cure models for curative therapies, parametric-mixture models in case of a subgroup of long-term survivors, piecewise, splines and fractional polynomials in case of fluctuating hazards, and SPMs for other data. All survival extrapolation approaches were fitted on immature data of three trials; in melanoma the framework recommends a cure model as cure seems clinically plausible, in breast cancer (BC) a parametric-mixture as a subgroup of long-term survivors is expected, and in multiple myeloma (MM) an SPM as no cure, long-term survivors or fluctuating hazards are expected. The statistical fit of the models was assessed using LOOIC. The extrapolation on immature data was compared to mature trial data in terms of ΔMean-Absolute-Deviation (ΔMAD) in months for all arms in the trial. Lower LOOIC and ΔMAD imply better predictions. Restricted mean survival time (RMST) of the mature data was determined to assess the relative size of ΔMAD. The best fitting approach on the melanoma dataset (RMST: 67.88 months) was the log-logistic parametric-mixture (LOOIC=2,870, ΔMAD=19.59). The best fitting cure model was the lognormal (LOOIC=3,249, ΔMAD=10.89). On the BC dataset (RMST: 123.42 months), the lognormal parametric-mixture fitted best (LOOIC=1,782, ΔMAD=9.86) and had one of the lowest ΔMAD. On the MM dataset (RMST: 92.55 months), the Weibull parametric-mixture fitted best (LOOIC=1,944, ΔMAD=11.24), and the best fitting SPM was the exponential (LOOIC=1,988, ΔMAD=7.21). This study shows that the model recommended by the framework generally leads to better long-term extrapolations compared to the statistically best fitting model. Therefore, additional to historical data and long-term comparative effectiveness expectations, this framework can be used for model selection.