WHEN THE TAIL WAGS THE DOG IV: A MARGINAL SURVIVAL BENEFIT WHERE THE TRANSPARENCY STATISTICS GIVE THE ALL-CLEAR AND A POPULATION FLOOR CORRECTLY DOES NOTHING (PANOVA-3)
Author(s)
Andrew Briggs, DPhil1, Alex Hill, PhD1, Anna Forsythe, MBA, MSc, PharmD2.
1London School of Hygiene & Tropical Medicine, London, United Kingdom, 2Founder, Oncoscope, Miami, FL, USA.
1London School of Hygiene & Tropical Medicine, London, United Kingdom, 2Founder, Oncoscope, Miami, FL, USA.
OBJECTIVES: Partitioned survival models can overstate benefit through optimistic extrapolation - but the same transparency statistics should equally flag when a benefit is mostly observed, and when a population-mortality floor has no work to do. Using PANOVA-3, the most mature trial in our series (TTFields added to chemotherapy in pancreatic cancer), we applied the Extrapolation Benefit Share (EBS: the share of incremental mean survival arising beyond the observed data) and the Survival Ratio Shift (SRS: whether the advantage is tail-concentrated), with a relative-survival floor for overall survival (OS).
METHODS: Pseudo-individual patient data were reconstructed (SurvdigitizeR) from the published OS and progression-free survival (PFS) Kaplan-Meier curves (TTFields + gemcitabine/nab-paclitaxel vs gemcitabine/nab-paclitaxel; locally advanced pancreatic cancer; 285 vs 286). Eight parametric distributions were fitted jointly. Mean survival was decomposed into observed and extrapolated (EMST) components over a 40-year horizon, with a boundary set from a reverse-Kaplan-Meier diagnostic. OS was additionally modelled as relative survival (excess hazard plus ONS UK general-population mortality).
RESULTS: The benefit was marginal and mostly observed. OS incremental survival was 2-3 months with EBS 13-40% - with mature follow-up most of the small gain lay within the data, not the extrapolation. PFS added 1-2 months, mostly too small for EBS to be meaningful. SRS was only modestly above 1 (≈1.2-1.95), so the advantage was barely tail-loaded. Relative survival barely changed OS - matching the all-cause result. Where disease-specific mortality dominates and data are mature, the population floor has nothing left to discipline.
CONCLUSIONS: PANOVA-3 is the mature counterpoint to extrapolation-dominated trials: statistics that elsewhere flag an assumption-driven benefit here return an all-clear, and the relative-survival floor correctly does nothing. A small survival gain that rests on observation locates the value question in the size of the benefit, not in how long it is presumed to persist.
METHODS: Pseudo-individual patient data were reconstructed (SurvdigitizeR) from the published OS and progression-free survival (PFS) Kaplan-Meier curves (TTFields + gemcitabine/nab-paclitaxel vs gemcitabine/nab-paclitaxel; locally advanced pancreatic cancer; 285 vs 286). Eight parametric distributions were fitted jointly. Mean survival was decomposed into observed and extrapolated (EMST) components over a 40-year horizon, with a boundary set from a reverse-Kaplan-Meier diagnostic. OS was additionally modelled as relative survival (excess hazard plus ONS UK general-population mortality).
RESULTS: The benefit was marginal and mostly observed. OS incremental survival was 2-3 months with EBS 13-40% - with mature follow-up most of the small gain lay within the data, not the extrapolation. PFS added 1-2 months, mostly too small for EBS to be meaningful. SRS was only modestly above 1 (≈1.2-1.95), so the advantage was barely tail-loaded. Relative survival barely changed OS - matching the all-cause result. Where disease-specific mortality dominates and data are mature, the population floor has nothing left to discipline.
CONCLUSIONS: PANOVA-3 is the mature counterpoint to extrapolation-dominated trials: statistics that elsewhere flag an assumption-driven benefit here return an all-clear, and the relative-survival floor correctly does nothing. A small survival gain that rests on observation locates the value question in the size of the benefit, not in how long it is presumed to persist.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE210
Topic
Clinical Outcomes, Economic Evaluation, Methodological & Statistical Research
Topic Subcategory
Trial-Based Economic Evaluation
Disease
Oncology