ASCO 2024: Study shows artificial Intelligence accurately predicts short-, long-term survival in ovarian cancer

May 30, 2024

Positive results from Predict Oncology Inc.’s retrospective study, completed in collaboration with UPMC Magee-Womens Hospital, will be presented at the 2024 American Society of Clinical Oncology (ASCO) Annual Meeting in Chicago, IL from May 31-June 4. The study aimed to determine if Predictive Oncology could leverage its artificial intelligence and other capabilities to develop machine learning (ML) models that could more accurately predict both short-term and long-term survival outcomes among ovarian cancer patients. The authors found that superior prediction of short- and long-term survival was achieved as compared to clinical data alone. The specific drivers of the top performing models were different for the short- and long-term cohorts, identifying future research opportunities as well as development potential of a clinical decision tool.

• Researchers analyzed clinical data and tumor specimens from 2010-2016. Patient data, whole exome sequencing, whole transcriptome sequencing, drug response profile, and digital pathology profile were used as input feature sets for training the multi-omic ML models. Hypothesis-free training of the ML models was utilized to classify patient survival 2 yr and 5 yr threshold.
• 160 high grade serous ovarian cancer (HGSC) model builds were completed. ML models achieved high prediction accuracy for both short-term (2y) and long-term (5y) patient cohorts, identifying 7 models with ≥ 0.7 AUROC model performance for 2yr OS threshold and 13 models with ≥ 0.7 AUROC model performance for 5yr OS threshold.
• Addition of multi-omic feature sets to clinical profile improved the model’s ability to predict OS for both short-term and long-term thresholds. Multi-omic feature set inputs led to superior prediction and improved performance over clinical profile information alone, and top performing multi-omic models predicted better than any feature set in isolation.
• When comparing top features identified by 2y and 5y OS threshold models, researchers found that molecular features (WTS followed by WES feature sets) predominantly drove 2y cohort while digital pathology imaging features were the driver of the 5y cohort top performing models.

Sources:

(2024, April 24). Predictive Oncology. Predictive Oncology Announces Abstract Accepted for Presentation at the 2024 American Society of Clinical Oncology (ASCO) Annual Meeting. [Press release]. https://predictive-oncology.com/wp-content/uploads/2024/04/Predictive-Oncology-Announces-Abstract-Accepted-for-Presentation-at-the-2024-American-Society-of-Clinical-Oncology-ASCO-Annual-Meeting.pdf

Orr, B, et al. (2024, May). ASCO. Using artificial intelligence-powered evidence-based molecular decision-making for improved outcomes in ovarian cancer: abstract #5555. https://meetings.asco.org/abstracts-presentations/234582