AI-Driven Chemotoxicity Prediction in Colorectal Cancer: Impact of Race, SDOH, and Biological Aging

Strategic Insights -

Research at Ohio State University College of Nursing has developed effective artificial intelligence (AI)/machine learning (ML) models to predict overall, gastrointestinal (GI), and hematological chemotoxicity in patients with colorectal cancer (CRC). The study, published in BMC Cancer, integrated racialized group, social determinants of health (SDOH), and biological aging to predict chemotoxicity risk. The findings suggest that higher Levine Phenotypic Age, elevated inflammatory markers, and poor SDOH are key predictors of overall and GI toxicities, while hematological toxicity is linked to lower inflammatory markers, higher Levine Phenotypic Age, and younger chronological age.

Key Takeaways:

  • The study developed AI/ML models to predict chemotoxicity risk in CRC patients, integrating racialized group, SDOH, and biological aging.
  • The most significant predictors of overall and GI toxicities were higher Levine Phenotypic Age, elevated inflammatory markers, and poor SDOH (e.g., higher ADI, unemployment).
  • Hematological toxicity was linked to lower inflammatory markers, higher Levine Phenotypic Age, and younger chronological age.
  • The Support Vector Machine and XGBoost models demonstrated high accuracy in predicting overall and GI chemotoxicity.
  • Race (non-Hispanic Black), body mass index, and lifestyle also influenced overall and GI toxicities.
  • The study suggests that integrating these factors into predictive models can help clinicians identify at-risk patients and tailor interventions to reduce chemotoxicity and improve survivorship outcomes.

Statistics:

  • 1,735 adult patients with CRC were analyzed in the study.
  • The AI/ML models were trained on 80% of cases (n = 1,388) and tested on 20% (n = 347) of the data.
  • Accuracy, area under the curve (AUC), and F1-score metrics were used to evaluate the performance of the models.
  • The Support Vector Machine model had an AUC of 0.988 in predicting overall chemotoxicity within the training dataset.
  • Higher Levine Phenotypic Age (beyond 80 years old) was associated with increased risk of overall and GI chemotoxicity.
  • Elevated inflammatory markers (e.g., C-reactive protein) were associated with increased risk of overall and GI toxicities.
  • Poor SDOH (e.g., higher ADI, unemployment) was associated with increased risk of overall and GI toxicities.

Sources:

  • Ohio State University College of Nursing, Center for Healthy Aging, Self-Management and Complex Care
  • BMC Cancer, BioMed Central - www.biomedcentral.com/; BMC Cancer - www.biomedcentral.com/bmccancer/
  • NewsRx LLC, Journal of Engineering - Journal of Engineering - www.journalofengineering.com/
  • Claire Han, Ohio State University College of Nursing, Ohio State University, Cancer Treatment and Research Center
  • Christin Burd, Jesse Plascak, Fode Tounkara, Ashley Rosko, Anne Noonan, Alai Tan, Diane Von Ah, and Xia Ning, Ohio State University College of Nursing, Center for Healthy Aging, Self-Management and Complex Care