449. How to Fix the Incentives in Cancer Research

Episode Summary

Executive Summary: The episode argues that cancer is not one disease but thousands, which explains both the impressive progress in some areas and the stubborn failures in others. It highlights gains from molecular profiling, immunotherapy, and better screening, while focusing on pancreatic cancer as a model for how collaborative, data-driven adaptive trials could speed discovery. COVID-era collaboration and trial innovation may further reshape oncology.

Main Topics: Cancer as a collection of diseases (Priority: 5/5): Sharpless explains that the field shifted from viewing cancer as a handful of diseases to recognizing hundreds or thousands of distinct cancers with unique biology, causes, and treatment responses. Progress in oncology is real but uneven (Priority: 5/5): Overall cancer mortality has fallen substantially, yet some cancers are still lethal or worsening, showing that advances benefit certain patients and tumor types more than others. Immunotherapy transformed skepticism into success (Priority: 4/5): The transcript traces how immune-based treatments went from being widely dismissed to becoming one of the most important modern cancer therapies, especially after checkpoint inhibitors proved effective. Pancreatic cancer as the hardest problem (Priority: 5/5): Pancreatic cancer is used as the clearest example of a deadly cancer with poor detection, limited treatment success, and very low survival, motivating new research strategies. Adaptive platform trials and collaboration (Priority: 5/5): Diane Simeone’s Precision Promise project is presented as a new clinical trial model that shares data, tests multiple therapies efficiently, and aligns hospitals, nonprofits, regulators, and companies. Data fragmentation and the need for aggregation (Priority: 4/5): The episode emphasizes that medical data are siloed across institutions, limiting research; better aggregation could improve understanding, prediction, and treatment selection. COVID-19 as a catalyst for trial innovation (Priority: 3/5): Pandemic-era vaccine development demonstrated that fast collaboration, remote trials, and cross-sector coordination can accelerate biomedical progress and may influence cancer research.

Key Arguments: Cancer research progressed slowly because the field first needed decades of basic science to understand biology rather than simply engineering a cure. Molecular profiling revealed that cancers thought to be single entities often contain multiple subtypes requiring different treatments. Immunotherapy succeeded only after years of failure and skepticism, showing that persistence in unconventional ideas can pay off. The broad decline in cancer mortality is meaningful, but it masks major disparities across cancer types and age groups. Lung cancer mortality is falling fastest because of tobacco control, targeted kinase inhibitors, and immunotherapy. Pancreatic cancer remains especially lethal because it is often detected late, spreads early, and lacks effective screening and therapy. Traditional clinical trials are too slow and redundant for rare, heterogeneous cancers; adaptive platform trials can test more treatments with less waste. Data sharing and standardized imaging, blood tests, and wearable monitoring could make each patient’s experience more informative for research. Negative trial results should be published because they prevent duplicated failures and can strengthen future breakthroughs. COVID-era collaboration and mRNA manufacturing success may create lasting improvements in trial design, speed, and cross-company cooperation.

Data Points: Initial expectation for curing cancer: by 1976 - Early 1970s “war on cancer” optimism after the National Cancer Act NCI annual funding: about $6 billion a year - Budget level described as allowing some self-determination in research priorities Cancer mortality decline in the U.S.: roughly 30% over three decades - Overall long-term improvement in cancer death rates Largest single-year decline in cancer death rate: 2.4% in 2018 - Most recent year cited in the transcript Global cancer deaths per year: nearly 10 million - Cancer remains the second leading cause of death worldwide after cardiovascular disease Cancer subtype example in breast cancer: about five diseases - Expression profiling showed breast cancer was far more heterogeneous than previously believed Men’s cancer death rates (2012–2016): decreased for 10 of 19 common cancers; increased for 6 - Shows uneven progress across cancer types Women’s cancer death rates (2012–2016): decreased for 13 of 20 common cancers; increased for 5 - Shows uneven progress across cancer types Lung cancer share targeted by kinase inhibitors: 15% to 20% - Subset of lung cancer patients who benefit from targeted pills U.S. lung cancer patients who never smoked: 12% - Reminder that smoking is not the sole cause of lung cancer Pancreatic cancer survival rate: 9% - Single-digit survival rate cited as evidence of extreme lethality Pancreatic cancer cases in the U.S.: about 58,000 per year - Approximate annual incidence Pancreatic cancer cases worldwide: about half a million per year - Global incidence estimate Surgically resectable pancreatic tumors at presentation: about 15% - Most patients are diagnosed too late for surgery Survival after removing very small pancreatic tumors: 60% to 70% five-year survival - Data from Japan when tumors are less than one centimeter Clinical trial participation in pancreatic cancer: 4% of patients - Illustrates how few patients enter trials Trial structure in Precision Promise: 30% control arm, 70% experimental arms - Adaptive platform trial design Number of pharmaceutical companies involved: about 30 - Industry participation in the pancreatic cancer ecosystem Potential timing improvement for FDA approval: cut in half - Expectation if the adaptive platform works as intended FDA lead-in sample size: 20 to 30 patients - Small early cohort used for go/no-go decisions Maximum waiting period for lead-in decision: 60 days - Accelerated review feature in the platform trial Moderna/Pfizer vaccine development speed: about one-tenth the usual time - COVID-19 vaccine development cited as proof of accelerated collaboration

Pivotal Quotes: "cancer really isn't a small number of entities. It's hundreds or thousands of diseases" — Ned Sharpless: Explaining the central paradigm shift in oncology "if clinical trials aren't working like we need, how do we fix them?" — Diane Simeone: Motivating the Precision Promise adaptive trial model for pancreatic cancer "I hate it. I'm driven to change that conversation." — Diane Simeone: Describing the emotional burden of telling pancreatic cancer patients there is no cure

Implications: Cancer progress will likely come less from one universal cure than from better classification, earlier detection, data sharing, and adaptive trials. Pancreatic cancer may be the proving ground for a faster, more collaborative model of oncology.