Prediction of acute postoperative protein depletion risk in colon cancer using an in-context learning foundation model: a retrospective cohort study - Takeaways - MDSpire

Prediction of acute postoperative protein depletion risk in colon cancer using an in-context learning foundation model: a retrospective cohort study

  • By

  • Xinke Cao

  • Linrui Han

  • Xinquan Zan

  • Yinchao Zhang

  • Zhiqiang Tian

  • Wei Shen

  • July 8, 2026

  • 0 min

Share

  • 1

    The study developed a predictive model for acute postoperative protein depletion in colon cancer patients using a tabular foundation model.

  • 2

    A total of 812 colon cancer patients were evaluated, with data collected from 2020 to 2025 for model training and validation.

  • 3

    The TabICLv2 model achieved a validation AUC of 0.766, indicating moderate discrimination for predicting protein depletion.

  • 4

    Key predictors identified included age, prealbumin, and globulin, as revealed by SHAP analysis.

  • 5

    A web-based calculator was created to facilitate clinical translation of the model for individual risk assessment.

Original Source(s)

Prediction of acute postoperative protein depletion risk in colon cancer using an in-context learning foundation model: a retrospective cohort study

  • by Xinke Cao, Linrui Han, Xinquan Zan, Yinchao Zhang, Zhiqiang Tian, Wei Shen

  • July 8, 2026

Related Content

Correction: Resistance mechanisms and overcoming strategies for CAR T-cell therapy in B-cell hematologic malignancies

Clinicopathological, proliferative, molecular, and prognostic characteristics of differentiated high-grade thyroid carcinoma: a multicenter retrospective study

Baptist Health Cancer Care to Present Research on Precision Radiation Therapy Across Complex Cancers at ASTRO 2026

Experts across Baptist Health Herbert Wertheim Cancer Institute and Eugene M. & Christine E. Lynn Cancer Institute will present new radiation oncology research and clinical expertise at the 2026 Annual Meeting of the American Society for Radiation Oncology (ASTRO), September 26-30, 2026, in Boston.

  • by

  • July 7, 2026

  • 8 min