Related Experiment Video
Updated: Sep 17, 2025
06:08
A Cognitive Fusion-guided Prostate Biopsy Using Multiparametric Magnetic Resonance Imaging and Transrectal Ultrasound
Published on: March 21, 2025
382
Combining multi-parametric MRI radiomics features with tumor abnormal protein to construct a machine learning-based
Chi Zhang1, Zewen Wang1, Peicheng Shang1
1Department of Urology, The Second Affiliated Hospital of Soochow University, Suzhou, China.
Scientific Reports
|July 2, 2025
Summary
Integrating multi-parametric MRI radiomic features with tumor abnormal protein (TAP) and clinical data significantly improves prostate cancer diagnosis. This combined approach enhances diagnostic accuracy, offering a more precise tool for identifying the disease.
Area of Science:
- Radiology
- Oncology
- Medical Imaging
Background:
- Prostate cancer diagnosis relies on various methods, but improved accuracy is crucial.
- Multi-parametric magnetic resonance imaging (mpMRI) offers detailed anatomical and functional information.
- Radiomic analysis extracts quantitative features from medical images, potentially aiding diagnosis.
Purpose of the Study:
- To evaluate the diagnostic value of combining mpMRI radiomic features with tumor abnormal protein (TAP) and clinical data for prostate cancer.
- To develop and assess a machine learning model for enhanced prostate cancer diagnosis.
Main Methods:
- Radiomic features were extracted from T2WI and ADC maps of 109 patients.
- Feature selection used t-tests and LASSO regression; models were built with random forest.
- Clinical factors (age, PSA, prostate volume) and TAP data were integrated with radiomic features.
Main Results:
- Individual mpMRI radiomic features showed significant correlation with prostate cancer.
- Random forest models using radiomic features achieved AUCs up to 0.87.
- Combining mpMRI radiomic features with TAP and clinical data yielded AUCs of 0.91-0.92, outperforming models using single data types.
Conclusions:
- The integration of mpMRI radiomic features, TAP, and clinical characteristics demonstrates high predictive efficiency for prostate cancer diagnosis.
- Machine learning models incorporating these multimodal data offer a promising approach for improving diagnostic accuracy.
- This comprehensive approach can aid clinicians in more precise prostate cancer detection.
