Related Experiment Video
Updated: Sep 5, 2025
08:05
Detection and Isolation of Cancer in Prostate Biopsies Using Stimulated Raman Histology and Artificial Intelligence
Published on: June 10, 2025
373
Radiomics-Based Machine Learning Models for Predicting P504s/P63 Immunohistochemical Expression: A Noninvasive
Yun-Fan Liu1, Xin Shu1, Xiao-Feng Qiao1
1Department of Radiology, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, China.
Frontiers in Oncology
|July 7, 2022
Summary
This study developed a noninvasive radiomics machine learning model to identify P504s/P63 status and diagnose prostate cancer (PCa). The random forest model achieved high accuracy, showing potential for presurgical evaluation.
Area of Science:
- Radiology and Medical Imaging
- Machine Learning in Healthcare
- Oncology Diagnostics
Background:
- Accurate prostate cancer (PCa) diagnosis relies on pathological markers like P504s/P63.
- Noninvasive methods for assessing P504s/P63 status pre-surgery are needed.
Purpose of the Study:
- To develop and validate a noninvasive radiomic-based machine learning (ML) model.
- To identify P504s/P63 status and diagnose PCa.
Main Methods:
- Retrospective analysis of 315 patients' prostate MRI (T2WI, DWI, ADC) and P504s/P63 pathology.
- Radiomic features extracted using AI Kit software.
- Random Forest (RF) and other ML algorithms used for prediction, evaluated by AUC and accuracy.
Main Results:
- The RF model demonstrated strong performance (microaverage AUC=0.920, macroaverage AUC=0.870).
- T2WI sequence was optimal as a single-sequence predictor (microaverage AUC=0.94).
- Combined T2WI, DWI, and ADC sequences yielded the best results (microaverage AUC=0.930).
Conclusions:
- Radiomic-based RF classifier shows potential for noninvasive presurgical P504s/P63 status evaluation.
- This approach offers accurate and noninvasive diagnosis of PCa.
