American College of Surgeons survival calculator for biliary tract cancers: using machine learning to individualize predictions

  • 0American College of Surgeons Cancer Programs, Chicago, IL; Department of Surgery, Northwestern University Feinberg School of Medicine, Chicago, IL. Electronic address: https://www.twitter.com/LJanczewskiMD.

|

|

Summary

This summary is machine-generated.

Machine learning created a Biliary Tract Cancer Survival Calculator for personalized prognosis. This tool, using patient, tumor, and treatment data, offers more accurate survival estimates than stage alone.

Area Of Science

  • Oncology
  • Medical Informatics
  • Machine Learning in Healthcare

Background

  • Cancer prognosis commonly relies on tumor stage, but survival is multifactorial.
  • Biliary tract cancers (bile duct, gallbladder) have complex survival dynamics.
  • Developing advanced prognostic tools is crucial for personalized cancer care.

Purpose Of The Study

  • To develop a prototype Biliary Tract Cancer Survival Calculator using machine learning.
  • To generate personalized survival estimates based on diverse patient, tumor, and treatment factors.
  • To improve upon traditional staging-based prognosis for biliary tract malignancies.

Main Methods

  • Utilized the National Cancer Database (2010-2017) for 62,877 biliary tract cancer patients.
  • Employed random forest algorithms to identify key prognostic variables.
  • Developed a survival model using extreme gradient boosting with survival embeddings for 3-year survival prediction.

Main Results

  • Identified metastatic disease, age, and lack of surgery as key factors for worse survival.
  • The machine learning model demonstrated superior performance (c-index: 0.74) compared to a stage-only model (c-index: 0.64).
  • The calculator integrates patient demographics, tumor characteristics (stage, grade, site), and treatment modalities (surgery, chemotherapy, radiation).

Conclusions

  • The Biliary Tract Cancer Survival Calculator is a highly accurate prognostic tool.
  • It provides individualized, real-time survival estimates.
  • This tool enhances personalized treatment planning and patient counseling in oncology.