A deep-learning model to predict the completeness of cytoreductive surgery in colorectal cancer with peritoneal metastasis☆

Qingfeng Lin ¹, Can Chen ², Kangshun Li ³

¹Department of Colorectal Surgery and Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Diseases, The Sixth Affiliated Hospital, Sun Yat-Sen University, Guangzhou, Guangdong Province, China.
²College of Mathematics and Informatics, South China Agricultural University, Guangzhou, China; College of Computers, Central South University, Changsha, China.
³College of Mathematics and Informatics, South China Agricultural University, Guangzhou, China.
⁴Department of Radiology, The Sixth Affiliated Hospital, Sun Yat-Sen University, Guangzhou, China.
⁵Department of Colorectal Surgery, Fudan University Shanghai Cancer Center, Shanghai, China.
⁶Colon and Rectal Surgery, Baylor University Medical Center, Dallas, TX, USA.
⁷General Surgery Center, Department of Gastrointestinal Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou, China.
⁸College of Intelligent Manufacturing and Modern Industry (School of Mechanical Engineering), Xinjiang University, Urumqi, China.
⁹Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong, China.
¹⁰Department of Gastrointestinal Surgical Oncology, Fujian Cancer Hospital and Fujian Medical University Cancer Hospital, Fuzhou, China.

⁰Department of Colorectal Surgery and Guangdong Provincial Key Laboratory of Colorectal and Pelvic Floor Diseases, The Sixth Affiliated Hospital, Sun Yat-Sen University, Guangzhou, Guangdong Province, China.

European Journal of Surgical Oncology : the Journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology +

|

April 2, 2025

View abstract on PubMed

Summary

A new AI framework, DeAF, accurately predicts the completeness of cytoreductive surgery (CRS) for colorectal cancer with peritoneal metastasis. This tool aids in selecting suitable patients for CRS, improving surgical decision-making and patient outcomes.

Area Of Science

Oncology
Artificial Intelligence
Medical Imaging

Background

Colorectal cancer (CRC) with peritoneal metastasis (PM) presents a poor prognosis.
The Peritoneal Cancer Index (PCI) is currently used for patient selection for cytoreductive surgery (CRS), but its accuracy is limited.
Accurate patient selection is crucial for optimizing CRS outcomes in PM patients.

Purpose Of The Study

To develop and validate a novel AI framework, named DeAF (decoupling feature alignment and fusion), for improved patient selection for CRS.
To predict the completeness of CRS in patients with PM using deep learning and CT imaging.
To enhance surgical decision-making for PM patients undergoing CRS.

Main Methods

A deep learning model (DeAF) was trained using contrast CT images and clinicopathological parameters from 186 CRC patients with PM.
The DeAF model utilized Simsiam algorithms for feature alignment and fusion.
Model performance was rigorously evaluated using accuracy, sensitivity, specificity, and ROC AUC in internal and three external validation cohorts.

Main Results

The DeAF model demonstrated high accuracy in predicting CRS completeness, achieving an AUC of 0.9 in the internal validation cohort.
The model's predictive performance was consistently validated across three external cohorts, with AUC values ranging from 0.906 to 0.960.
The DeAF framework effectively aids in selecting suitable PM patients and predicting the potential benefits of CRS.

Conclusions

The DeAF AI framework offers a novel and effective approach to aid surgeons in selecting appropriate patients for CRS.
The model accurately predicts the completeness of CRS, potentially transforming surgical decision-making for PM patients.
This AI tool holds promise for improving outcomes for colorectal cancer patients with peritoneal metastasis.

Keywords:

Colorectal cancer Cytoreductive surgery DeAF model Deep learning Peritoneal metastasis