Marco Donia: Evaluating generalizability of oncology trials with machine learning
Marco Donia, Professor at the University of Copenhagen and Research Group Leader (TIL group) at CCIT – Center for Cancer Immune Therapy, shared a post on LinkedIn:
“Evaluating Generalizability of Oncology Trials with Machine Learning.
Nature Medicine, Orcutt et al., Jan 2025
Randomized Controlled Trials (RCTs) vs. Real-World Patients: Oncology trials often fail to represent real-world patients due to restrictive eligibility criteria (melanoma; not just in oncology)
Key Insights:
Trial Translator: A Machine Learning(ML)-Based Framework: Uses electronic health record data to emulate RCTs across prognostic risk groups.
Categorizes patient phenotypes in:
- Low and Medium risk: survival times similar to those observed in RCTs
- High risk: significantly lower survival times and treatment-associated benefits.
Take-Home Message:
- Highlights the gap in RCT representativeness of real-world patients
- ML-based risk stratification may inform on treatment expectations.
Clinical Implications:
This tool could enable more realistic expectations for treatment outcomes in diverse patient populations, and help planning individualized care
Initiatives to Broaden Eligibility Criteria of Clinical Trials to increase representativeness: learn more.
Outstanding work from the team of Ravi B. Parikh and Qi Long.”
More posts featuring Marco Donia.
-
Challenging the Status Quo in Colorectal Cancer 2024
December 6-8, 2024
-
ESMO 2024 Congress
September 13-17, 2024
-
ASCO Annual Meeting
May 30 - June 4, 2024
-
Yvonne Award 2024
May 31, 2024
-
OncoThon 2024, Online
Feb. 15, 2024
-
Global Summit on War & Cancer 2023, Online
Dec. 14-16, 2023