Pretrained transformers applied to clinical studies improve predictions of treatment efficacy and associated
Gustavo Arango-Argoty1, Elly Kipkogei2, Ross Stewart3
1Oncology Data Science, Oncology R&D, AstraZeneca, Waltham, MA, USA. gustavo.arango@astrazeneca.com.
Abstract:
Cancer treatment has made significant advancements in recent decades, however many patients still experience treatment failure or resistance. Attempts to identify determinants of response have been hampered by a lack of tools that simultaneously accommodate smaller datasets, sparse or missing measurements, multimodal clinicogenomic data, and that can be interpreted to extract biological or clinical insights. We introduce the Clinical Transformer, an explainable transformer-based deep-learning framework that addresses these challenges. Our framework maximizes data via self-supervised, gradual, and transfer learning, and yields survival predictions surpassing performance of state-of-the-art methods across diverse, independent datasets. The framework's generative capability enables in silico perturbation experiments to test counterfactual hypotheses. By perturbing immune-associated features in immunotherapy-naive patients, we identify a patient subset that may benefit from immunotherapy, and we validate this finding across three independent immunotherapy-treated cohorts. We anticipate our work will empower the scientific community to further harness data for the benefit of patients.
Insights
A new AI tool, the Clinical Transformer, improves cancer treatment predictions by analyzing complex patient data. It identifies patients who may benefit from immunotherapy, offering new hope for personalized cancer care.
Area of Science:
- Computational biology
- Artificial intelligence in medicine
- Oncology
Background:
- Cancer treatment faces challenges with treatment failure and resistance.
- Existing tools struggle with complex, multimodal clinicogenomic data and small datasets.
- Identifying determinants of treatment response requires advanced analytical methods.
Purpose of the Study:
- To introduce an explainable deep-learning framework, the Clinical Transformer, for analyzing complex cancer data.
- To improve survival predictions and identify patient subsets for targeted therapies.
- To overcome limitations of existing tools in handling multimodal and sparse data.
Main Methods:
- Developed an explainable transformer-based deep-learning framework (Clinical Transformer).
- Employed self-supervised, gradual, and transfer learning to maximize data utilization.
- Utilized generative capabilities for in silico perturbation experiments to test hypotheses.
Main Results:
- Achieved superior survival prediction performance compared to state-of-the-art methods across diverse datasets.
- Identified a subset of immunotherapy-naive patients who may benefit from immunotherapy through in silico experiments.
- Validated the identified patient subset across three independent immunotherapy-treated cohorts.
Conclusions:
- The Clinical Transformer offers a powerful, interpretable tool for analyzing complex clinicogenomic data.
- This framework can enhance personalized cancer treatment strategies, particularly for immunotherapy.
- The approach has the potential to accelerate data-driven discoveries in cancer research.
Related Concept Videos
01:16Clinical Trials
There are four phases in a clinical trial. A phase one...
01:11Clinical Trials: Overview
01:23Types of Biopharmaceutical Studies: Controlled and Non-Controlled Approaches
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast, controlled...
01:11Pharmacogenetics of Drug Targets: β₂-Adrenergic Receptors, Apo E, Thymidylate Synthase