Retrospective Study
Copyright ©The Author(s) 2024. Published by Baishideng Publishing Group Inc. All rights reserved.
World J Gastroenterol. Apr 28, 2024; 30(16): 2233-2248
Published online Apr 28, 2024. doi: 10.3748/wjg.v30.i16.2233
Preoperative prediction of perineural invasion of rectal cancer based on a magnetic resonance imaging radiomics model: A dual-center study
Yan Liu, Bai-Jin-Tao Sun, Chuan Zhang, Bing Li, Xiao-Xuan Yu, Yong Du
Yan Liu, Bai-Jin-Tao Sun, Chuan Zhang, Bing Li, Xiao-Xuan Yu, Department of Radiology, The Affiliated Hospital of North Sichuan Medical College, Nanchong 637000, Sichuan Province, China
Yong Du, Department of Radiology, The Affiliated Hospital of North Sichuan Medical College, Nanchong 637000, Sichuan Province, China. duyong@nsmc.edu.cn
Author contributions: Liu Y data acquisition and analysis, drafting and writing of the manuscript; Sun BJT data collection and data analysis; Zhang C, Li B and Yu XX language editing and revisions to the manuscript; Du Y work concept or design and important revisions to the manuscript; all authors have read and approve the final manuscript.
Institutional review board statement: This study was reviewed and approved by the Ethics Committee of the Affiliated Hospital of North Sichuan Medical College.
Informed consent statement: Patients were not required to give informed consent to the study because the analysis used anonymous clinical data that were obtained after each patient agreed to treatment by written consent.
Conflict-of-interest statement: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Corresponding author: Yong Du, MD, Professor, Department of Radiology, The Affiliated Hospital of North Sichuan Medical College, No. 1 Maoyuannan Road, Nanchong 637000, Sichuan Province, China. duyong@nsmc.edu.cn
Received: January 2, 2024
Peer-review started: January 2, 2024
First decision: January 31, 2024
Revised: February 8, 2024
Accepted: March 20, 2024
Article in press: March 20, 2024
Published online: April 28, 2024
ARTICLE HIGHLIGHTS
Research background

Perineural invasion (PNI), is a potential pathway for the metastatic spread of rectal cancer (RC), and has been used as an important pathological indicator and independent prognostic factor. Preoperative stratification of RC patients according to PNI status facilitates individualized treatment and improves the prognosis of RC patients.

Research motivation

Nowadays, the preoperative predicton of PNI status is still challenging and needs further study.

Research objectives

To evaluate the usefulness of a model based on preoperative magnetic resonance imaging (MRI) radiomics for predicting PNI status in patients with RC and establishing and validating an optimal nomogram model for predicting PNI status preoperatively in RC patients.

Research methods

We enrolled 244 RC patients from two independent centers from May 2019 to August 2022. The patients from Center 1 were randomly divided into a training group (n = 118) and an internal validation group (n = 52), whereas 74 patients from Center 2 served as an external validation group. Extracted and selected quantitative radiomics features and clinical risk factors to establish and validate the radiomics predictive model and clinical-radiomics (CR) model.

Research results

We extracted 944 radiomics features from T2-weighted imaging and contrast-enhanced T1-weighted imaging sequences, combined with PNI-related clinical features (clinical TNM and histological grade) to construct the final CR model, and used internal and external validation groups to evaluate the models. The final CR model showed good performance to predict PNI status, the area under the curve of the CR model in the training and internal and external validation groups were 0.889, 0.889 and 0.894, respectively.

Research conclusions

The CR model based on MRI radiomics features and clinical risk factors was able to predict the PNI status of RC noninvasively, showed stable performance, which can provide support for individualized treatment of RC patients.

Research perspectives

Further external verification is needed to optimize the model, and explore the feasibility of applying deep learning to automatically describe volume of interest, reduce the difference between observers, and improve the applicability of the model.