Preliminary Evaluation of a Novel Artificial Intelligence-based Prediction Model for Surgical Site Infection in Colon
Yuki Ohno1, Junichi Mazaki1, Ryutaro Udo1
1Department of Gastrointestinal and Pediatric Surgery, Tokyo Medical University, Tokyo, Japan.
Background/Aim:
There are few studies on artificial intelligence-based prediction models for colon cancer built using clinicopathological factors. Here, we aimed to perform a preliminary evaluation of a novel artificial intelligence-based prediction model for surgical site infection (SSI) in patients with stage II-III colon cancer.
Patients And Methods:
The medical records of 730 patients who underwent radical surgery for stage II-III colon cancer between 2000 and 2018 at our institute were retrospectively analyzed. Kaplan-Meier curves were used to examine the association between SSI and oncological outcomes (recurrence-free survival time). Next, we used the machine learning software Prediction One to predict SSI. Receiver-operating characteristic curve analysis was used to evaluate the accuracy of the artificial intelligence model.
Results:
The prognosis in terms of recurrence-free survival time was poor in patients with SSI (p=0.005, 95% confidence interval=4892.061-5525.251). The area under the curve of the artificial intelligence model in predicting SSI was 0.731.
Conclusion:
As SSI is an important prognostic factor associated with oncological outcomes, the prediction of SSI occurrence is important. Based on our preliminary evaluation, the artificial intelligence model for predicting SSI in patients with stage II-III colon cancer was as accurate as the previously reported model derived through conventional statistical analysis.
