Systems and methods to generate a surgical risk score and uses thereof
Inventors
Gaudilliere, Brice • Aghaeepour, Nima • Hedou, Julien • Rumer, Kristen • Angst, Martin S.
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Assignees
Leland Stanford Junior University
MemberStanford UniversityStanford UniversityTodd Coleman is an associate professor of bioengineering at Stanford University. His research combines applied probability, physiology, and bioelectronics to develop methods and devices for monitoring and modulating nervous-system and other physiological activity, including brain and visceral-organ signals.
Todd Coleman is an associate professor of bioengineering at Stanford University. His research combines applied probability, physiology, and bioelectronics to develop methods and devices for monitoring and modulating nervous-system and other physiological activity, including brain and visceral-organ signals.
Publication Number
US-12755850-B2
Publication Date
2026-10-06
Expiration Date
Abstract
Embodiments herein describe systems and methods to generate a surgical risk score. Various embodiments obtaining multi-omics data from an individual, such as genomics, transcriptomics, and proteomics. In certain embodiments, a machine algorithm is used to generate the surgical risk score based on the multi-omics data. In further embodiments, clinical data is further used in the determination of the surgical risk score.
Core Innovation
The claimed method assesses a patient's immune activation prior to surgery to calculate a surgical risk score indicating the probability of a surgical-site complication. It combines values of omic biological features, including immune-activation measurements of immune-related signaling proteins following ex-vivo stimulation, with a machine-learning model trained to output surgical risk scores. The background section is not explicitly described in the provided content.
The framework utilizes Multi-Omic Bootstrap model training across individual data layers containing biological and artificial features, applying feature selection based on statistical weights and occurrence frequencies. Reported performance demonstrates AUC values of 0.82 for preoperative multi-omic data, 0.92 upon adding clinical features, 0.77 in an independent cohort, and 0.86 using postoperative-day-one data.
Claims Coverage
The claims cover two inventive features embodied in two independent claims directed to preoperative immune-activation assessment and machine-learning calculation of a surgical risk score.
Preoperative immune activation assessment with bootstrap feature selection
The method determines a surgical risk score indicating the probability of a surgical-site complication from patient omic biological features, including immune-activation measurements obtained following ex-vivo stimulation [procedural detail omitted for safety]. Model training employs bootstrap procedures across data layers comprising biological and artificial features, selecting features based on the occurrence frequency of significant weights.
Risk assessment with hyperparameter-specific occurrence frequencies
The method calculates a surgical risk score from omic biological features and stimulated immune-activation measurements [procedural detail omitted for safety], wherein bootstrap-based model training computes initial weights for multiple hyperparameter values and calculates a unitary occurrence frequency for each feature and hyperparameter value from significant weights across repetitions.
The two independent claims cover preoperative risk scoring based on immune-activation and omic biological features using bootstrap-based machine-learning feature selection, with the second claim additionally specifying occurrence-frequency calculations across hyperparameter values.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Assessing a patient's probability of developing a surgical-site complication following surgery by determining a surgical risk score.
Guiding perioperative care by administering a treatment regimen configured to reduce the probability of a surgical-site complication when the surgical risk score exceeds a threshold.
Interested in licensing this patent?