Back to the list
Congress: ECR25
Poster Number: C-27030
Type: Poster: EPOS Radiologist (scientific)
Authorblock: Y. Yang, H. Huang, Z. Wen; Beijing/CN
Disclosures:
Yuming Yang: Nothing to disclose
Hao Huang: Nothing to disclose
Zhiyong Wen: Nothing to disclose
Keywords: Genital / Reproductive system male, Oncology, Pelvis, MR, Computer Applications-Detection, diagnosis, Diagnostic procedure, Cancer, Neoplasia
Results

Of the 193 patients included in total , CsPCa was detected in 64 patients (33.4%), and the other 129 patients were diagnosed with non-significant cancers or no malignant lesions. Logistic regression analysis demonstrated that f/tPSA, PSAD, bpMRI scores and mpMRI scores were independent predictors for CsPCa (P < 0.05).

The bpMRI nomogram was established using f/tPSA, PSAD and bpMRI scores.

Fig 1: Nomogram for predicting clinically significant prostate cancer based on f/tPSA, PSAD and bpMRI scores.
The ROC curve analysis indicated that the bpMRI nomogram model showed a favorable diagnostic performance (AUC = 0.946, 95% CI: 0.904-0.973), which had significant differences with f/tPSA (AUC = 0.720, 95% CI: 0.656-0.784, P < 0.001), PSAD (AUC = 0.829, 95% CI: 0.771-0.886, P < 0.001), bpMRI scores (AUC = 0.883, 95% CI: 0.835-0.932, P < 0.001), and mpMRI scores (AUC = 0.905, 95% CI: 0.855-0.943, P = 0.004), respectively.
Fig 2: Comparison of ROC curves for predicting clinically significant prostate cancer.

The model showed satisfying calibration on internal bootstrap validation (corrected C-index = 0.938). The calibration curves showed a great agreement between the apparent curve and the ideal curve. The Hosmer-Lemeshow test produced a non-significant result (P > 0.05), which supports the goodness-of-fit of the model.

Fig 3: Calibration curves of the nomogram, including f/tPSA, PSAD and bpMRI scores.
 
GALLERY