SCORPIO machine learning system can predict certain outcomes across diverse cancer types and health care settings
By Elana Gotkine HealthDay Reporter
FRIDAY, Jan. 10, 2025 (HealthDay News) — The machine learning system SCORPIO can predict patient outcomes with immune checkpoint inhibitor (ICI) therapy across diverse cancer types and health care settings, according to a study published online Jan. 6 in Nature Medicine.
Seong-Keun Yoo, Ph.D., from the Icahn School of Medicine at Mount Sinai in New York City, and colleagues developed and evaluated SCORPIO utilizing routine blood tests and clinical characteristics from 9,745 ICI-treated patients across 21 cancer types. SCORPIO was trained on data from 1,628 patients across 17 cancer types and tested in internal test sets made up of 2,511 patients across 19 cancer types. External validation was performed on 10 global phase 3 trials with 4,447 patients across six cancer types and in a real-world cohort from Mount Sinai Health System with 1,159 patients across 18 cancer types.
The researchers found that SCORPIO achieved median time-dependent area under the receiver operating characteristic curve (AUC[t]) values of 0.763 and 0.759 in the two internal test sets for predicting overall survival at six, 12, 18, 24, and 30 months, outperforming median AUC(t) values of 0.503 and 0.543 for tumor mutational burden (TMB). For predicting clinical benefit (tumor response or prolonged stability), SCORPIO demonstrated superior predictive performance, with AUC values of 0.714 and 0.641, compared with 0.546 and 0.573 for TMB. Performance in predicting ICI outcomes was maintained in external cohorts, surpassing programmed death-ligand 1 immunostaining.
“Despite its limitations, SCORPIO remains a highly accessible model for predicting ICI efficacy and can aid clinical decision-making when used alongside other assessments,” the authors write.
Several authors have a provisional patent application, and three are co-inventors on a patent filed for using TMB to predict immunotherapy response. Several authors also disclosed ties to the pharmaceutical industry.
Copyright © 2025 HealthDay. All rights reserved.
