Researchers have developed a novel approach using 'digital twins' of cancer patients to accurately recreate clinical trials and predict treatment responses. The technology, named FarrSight-Twin, leverages algorithms originally designed for astrophysical black hole discovery to model individual therapeutic responses to chemotherapy. This innovation promises to accelerate drug development and personalize cancer treatment by simulating trials and identifying optimal therapies for patients.
FarrSight-Twin: Recreating Clinical Trials Virtually
FarrSight-Twin creates digital twins from biological data of thousands of cancer patients, incorporating molecular data on their tumors. This enables the prediction of how a patient is likely to respond to a specific treatment. In a study presented at the 36th EORTC-NCI-AACR Symposium on Molecular Targets and Cancer Therapeutics, the technology was used to recreate published clinical trials involving patients with breast, pancreatic, or ovarian cancer. These trials compared different drug therapies, including anthracyclines, taxanes, platinum-based drugs, capecitabine, and hormone treatments.
Accurate Prediction of Treatment Outcomes
The digital trials accurately predicted the outcomes of the actual clinical trials across all simulated studies. Further testing revealed that patients who received the treatment predicted by FarrSight-Twin to be the most effective had a 75% response rate, compared to a 53.5% response rate in those who received a different treatment. Response rate is defined as the proportion of patients whose tumors shrank following treatment.
Potential to Transform Cancer Care
Dr. Uzma Asghar, Co-founder and Chief Scientific Officer at Concr, stated, "We can use digital twins to represent individual patients, build clinical trial cohorts and compare treatments to see if they are likely to be successful before testing them out with real patients." This technology allows researchers to simulate patient trials earlier in drug development, rerun simulations to test different scenarios, and maximize the likelihood of success. It is also being used to simulate patients as controls for comparing the effect of new treatments with existing standards of care.
Clinical Applications and Future Directions
Currently, the technology is being developed to predict treatment responses for individual patients in the clinic, helping doctors understand which chemotherapies will be most effective. Dr. Asghar and her colleagues are also testing whether the technology can predict the best available treatments for patients with triple-negative breast cancer, in collaboration with The Institute of Cancer Research, Durham University, and The Royal Marsden Hospital.
Professor Timothy A. Yap from the University of Texas MD Anderson Cancer Center, commented, "If we can exploit digital tools to make this process quicker and easier, that should help us find better treatments for patients more efficiently in the future."
The use of digital twins enhances the interpretation of single-arm studies when a novel agent is combined with a standard of care, facilitating combination studies of novel agents in earlier lines of therapy. The predicted log odds ratios for overall response rates (ORR) in each treatment were consistent across the eight selected clinical trials. Drug response rates showed significant improvement (ORR = 2.55, Fisher exact P < .001), and overall survival (OS) increased with cohort enrichment (log rank P < .0001). The model accurately predicted OS across 23 solid tumors (AUC: 0.78). Patients in The Cancer Genome Atlas Program cohort who received treatments recommended by the digital twin model had noticeably better therapeutic responses (Fisher Exact Test P < .0001) and survival outcomes (log rank P < .0001) compared with those receiving alternative standard of care therapies.
