PD10-05: Development of a microultrasound-Based Nomogram to predict Extracapsular Extension in Patients with Prostate Cancer Undergoing Robot-Assisted Radical Prostatectomy
Introduction: We aimed to build a nomogram including clinicopathological parameters and microultrasound (mUS) findings to predict non-organ confined disease. Methods: Data of consecutive patients undergoing RARP between September 2020 and October 2021 were prospectively collected. All patients underwent mUS the day prior to RARP and the operators were blinded for both mpMRI and biopsy results. ECE on definitive pathology was the primary outcome. Variables significantly associated with ECE at univariable analysis were used to build the logistic multivariable models, and the regression coefficients of the model with the higher Area Under the Receiver Operating Curve (AUC) were used to develop the nomogram. Calibration plot was used to assess the extent of over- and under-estimation of the model. The model was subjected to 1000 bootstrap resamples for internal validation, and the bootstrapped model was compared to the base model. Sensibility, specificity, and negative predictive value (NPV) for each cut-off was evaluated. Results: Overall, 91/232 patients showed signs of ECE at mUS assessment before RARP, and 101/232 (43.5%) cases had a diagnosis of ECE on definitive pathology. MicroUS correctly identify ECE in 69/101 cases showing a sensitivity and a specificity of 65.7% and 82.5%, with an AUC of 77%. The mUS-based nomogram for the prediction of ECE is showed in Figure 1 and provided an AUC of 83.5%. The calibration plot showed a satisfactory concordance between predicted probabilities and observed frequencies of ECE, with a slightly tendency to underestimation. After 1000 bootstrap resamples, the predictive accuracy of the model was 83.2% (p=0.3). Finally, sensitivity and specificity and NPV associated with a 2% cut-off were 93%, 51.6% and 90%, respectively. Conclusions: We developed a mUS-based nomogram for the prediction of ECE. Its high accuracy and NPV make it a promising tool in planning the surgical approach. External validation and a direct comparison with mpMRI-based nomogram is crucial to corroborate our results. SOURCE OF Funding: None.
