Objective:

To investigate the value of magnetic resonance imaging (MRI) radiomics parameters of periprostatic fat (PPF) in predicting biochemical recurrence of prostate c

ancer after radical prostatectomy (RP).

Methods:

MRI images

clinical information and follow-up results of 114 patients (without biochemical recurrence 64 case) and (biochemical recurrence 50 case) who underwent radical prostatectomy (RP) in the First Affiliated Hospital of Soochow University from January 2016 to December 2020 were retrospectively analyzed. All patients underwent MRI scan before surgery. Transverse T2-weighted imaging (T2WI)

diffusion-weighted imaging (DWI)

apparent diffusion coefficient (ADC) and sagittal T2WI sequences were selected

and 3D manual separation of the fat area around the prostate was performed on the transverse T2WI images using 3D Slicer 4.13.0. The feature extraction module of open-source software FAE is used to extract image omics features through 10 image conversion types and 7 feature extraction methods. 1 647 features were extracted from T2WI sequence

3 290 features were extracted from DWI and ADC sequence

and 4 937 features were extracted from T2WI combined with DWI and ADC sequence. The area under curve (AUC)

accuracy

positive predictive value of the training set and the test set are determined by establishing the image omics model

machine learning and model verification. Positive predictive value and negative predictive value were used to evaluate the quantitative performance of the radiomics model.

Results:

The top 13 feature models in T2WI sequence were selected

with AUC of training set was 0.982 and test set was 0.912. The AUC of training set and test set was 0.916 and 0.814 for the top 12 feature models with T2WI combined with DWI and ADC sequences. The AUC of training set and test set was 0.94 and 0.86

respectively

for the feature model with the top 9 selected features in DWI and ADC sequences.

Conclusion:

The MRI PPF radiomics model around prostate has a high ability to predict biochemical recurrence after RP

especially the T2WI sequence model.