Estimating Long-Term Survival Outcomes for Tumor-Agnostic Therapies: Larotrectinib Case Study.

BACKGROUND: Larotrectinib is a precision oncology treatment for solid tumors with neurotrophic tyrosine receptor kinase (NTRK) gene fusions. Larotrectinib efficacy has been evaluated in single-arm basket trials with limited follow-up and sample sizes at the initial regulatory approval due to the rarity of solid tumors with NTRK gene fusion. OBJECTIVES: We aim to demonstrate that trends in progression-free survival (PFS) and overall survival (OS) in survival data with longer follow-up may be predicted from long-term survival estimates from survival data with shorter follow-up, including predictions for median survival when it is not observed in the trial. METHODS: Patient-level data were pooled from 3 clinical trials (NCT02122913, NCT02576431, and NCT02637687) using the 2018 and 2020 data cuts for the same subset of pediatric and adult patients. The Weibull distribution was selected for survival models. Survival predictions using 2018 data were compared to 2020 Kaplan-Meier (KM) curves. RESULTS: A total of 102 patients representing 15 tumor types were included in the analysis, with a mean age of 37 years. When comparing PFS from the 2018 survival prediction to observed 2020 KM data, the 12-month PFS rate was identical (66.6%). The 36-month PFS rate was lower for the 2018 prediction (35.3%) compared to 2020 KM data (44.4%). The median OS had not yet been reached in either data cut but was predicted to be 90 months using the 2018 data. When comparing OS from the 2018 survival prediction to the observed 2020 KM data, the 12-month OS rate was 89.0% and 86.6% and the 48-month OS rate was 67.2% and 63.0%, respectively. CONCLUSION: Long-term PFS predictions deviated from observed PFS rates due to response differences across tumor types and heavy censoring towards the end of the survival curve. However, for OS, the 48-month survival prediction was consistent with the observed 2020 KM estimate.

Developing a Framework for the Health Technology Assessment of Histology-independent Precision Oncology Therapies.

The arrival of precision oncology is challenging the evidence standards under which technologies are evaluated for regulatory approval as well as for health technology assessment (HTA) purposes. Several key concepts are discussed to highlight the source of the challenges in evaluating these products, particularly those impacting the HTA of histology-independent therapies. These include the basket trial design, high uncertainty in (potentially substantial) benefits for histology-independent therapies, and the inability to identify and quantify benefits of standard of care in daily practice when the biomarker is not currently used in practice. There is little precedent for a technology with the unique mixture of challenges for HTA of histology-independent therapies and they will be evaluated using standard HTA, as there currently is no evidence suggesting the standard HTA framework is not appropriate. A number of questions proposed to help guide HTA bodies when assessing the appropriateness of local processes to optimally evaluate histology-independent therapies. Pragmatic solutions are further proposed to decrease uncertainty in the benefits of histology independent therapies as well as fill gaps in comparative evidence. The proposed solutions ensure a consistent and streamlined approach to evaluation across histology-independent products, although with varying strengths and limitations. Alongside these solutions, sponsors should engage early with HTA bodies/payers and regulatory agencies through parallel/joint scientific advice to facilitate the integration of both regulatory and HTA perspectives into one clinical development programme, potentially reconciling evidence requirements.