Development of a microultrasound-based nomogram to predict extra-prostatic extension in patients with prostate cancer undergoing robot-assisted radical prostatectomy

OBJECTIVES

To develop a microultrasound-based nomogram including clinicopathological parameters and microultrasound findings to predict the presence of extra-prostatic extension and guide the grade of nerve-sparing.

MATERIAL AND METHODS

All patients underwent microultrasound the day before robot-assisted radical prostatectomy. Variables significantly associated with extra-prostatic extension at univariable analysis were used to build the multivariable logistic model, and the regression coefficients were used to develop the nomogram. The model was subjected to 1000 bootstrap resamples for internal validation. The performance of the microultrasound-based model was evaluated using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve, calibration plot, and decision curve analysis (DCA).

RESULTS

Overall, 122/295 (41.4%) patients had a diagnosis of extra-prostatic extension on definitive pathology. Microultrasound correctly identify extra-prostatic extension in 84/122 (68.9%) cases showing a sensitivity and a specificity of 68.9% and 84.4%, with an AUC of 76.6%. After 1000 bootstrap resamples, the predictive accuracy of the microultrasound-based model was 85.9%. The calibration plot showed a satisfactory concordance between predicted probabilities and observed frequencies of extra-prostatic extension. The DCA showed a higher clinical net-benefit compared to the model including only clinical parameters. Considering a 4% cut-off, nerve-sparing was recommended in 173 (58.6%) patients and extra-prostatic extension was detected in 32 (18.5%) of them.

CONCLUSION

We developed a microultrasound-based nomogram for the prediction of extra-prostatic extension that could aid in the decision whether to preserve or not neurovascular bundles. External validation and a direct comparison with mpMRI-based nomogram is crucial to corroborate our results.

Quantitative MRI in the Local Staging of Prostate Cancer: A Systematic Review and Meta-Analysis

BACKGROUND

Local staging of prostate cancer (PCa) is important for treatment planning. Radiologist interpretation using qualitative criteria is variable with high specificity but low sensitivity. Quantitative methods may be useful in the diagnosis of extracapsular extension (ECE).

PURPOSE

To assess the performance of quantitative MRI markers for detecting ECE.

STUDY TYPE

Systematic review and meta-analysis.

SUBJECTS

4800 patients from 28 studies with histopathologically confirmed PCa on radical prostatectomy were pooled for meta-analysis. Patients from 46 studies were included for systematic review.

FIELD STRENGTH/SEQUENCE

Diffusion-weighted, T2-weighted, and dynamic contrast-enhanced MRI at 1.5 T or 3 T.

ASSESSMENT

PubMed, Embase, Web of Science, Scopus, and Cochrane databases were searched to identify studies on diagnostic test accuracy or association of any quantitative MRI markers with ECE. Results extracted by two independent reviewers for tumor contact length (TCL) and mean apparent diffusion coefficient (ADC-mean) were pooled for meta-analysis, but not for other quantitative markers including radiomics due to low number of studies available.

STATISTICAL TESTS

Hierarchical summary receiver operating characteristic (HSROC) curves were computed for both TCL and ADC-mean, but summary operating points were computed for TCL only. Heterogeneity was investigated by meta-regression. Results were significant if P ≤ 0.05.

RESULTS

At the 10 mm threshold for TCL, summary sensitivity and specificity were 0.76 [95% confidence interval (CI) 0.71-0.81] and 0.68 [95% CI 0.63-0.73], respectively. At the 15 mm threshold, summary sensitivity and specificity were 0.70 [95% CI 0.53-0.83] and 0.74 [95% CI 0.60-0.84] respectively. The area under the HSROC curves for TCL and ADC-mean were 0.79 and 0.78, respectively. Significant sources of heterogeneity for TCL included timing of MRI relative to biopsy.

DATA CONCLUSION

Both 10 mm and 15 mm thresholds for TCL may be reasonable for clinical use. From comparison of the HSROC curves, ADC-mean may be superior to TCL at higher sensitivities.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2.

Pre-operative prediction of extracapsular extension of prostate cancer: first external validation of the PRECE model on an independent dataset

INTRODUCTION

The PRECE is a model predicting the risk of extracapsular extension (ECE) of prostate cancer: it has been developed on more than 6000 patients who underwent robotic radical prostatectomy (RARP) at the Global Robotic Institute, FL, USA. Up to now, it is the single tool predicting either the side and the amount of ECE. The model has a free user-friendly interface and is made up from simple and available covariates, namely age, PSA, cT, GS and percent of positive core, the latter topographically distributed within the prostate gland. Despite the successful performance at internal validation, the model is still lacking an external validation (EV). The aim of the paper is to externally validate the PRECE model on an Italian cohort of patients elected to RARP.

METHODS

269 prostatic lobes from 141 patients represented the validation dataset. The EV was performed with the receiver operating characteristics (ROC) curves and calibration, to address the ability of PRECE to discriminate between patients with or without ECE.

RESULTS

Overall, an ECE was found in 91 out of the 269 prostatic lobes (34%). Twenty-five patients out of pT3 had a bilateral ECE. The ROC curve showed an AUC of 0.80 (95% CI 0.74-0.85). Sensitivity and specificity were 77% and 69%, respectively. The model showed an acceptable calibration with tendency towards overestimation.

CONCLUSIONS

From the current EV, the PRECE displays a good predictive performance to discriminate between cases with and without ECE; despite preliminary, outcomes may support the generalizability of the model in dataset other than the development one.

Explainable artificial intelligence to predict the risk of side-specific extraprostatic extension in pre-prostatectomy patients

INTRODUCTION

We aimed to develop an explainable machine learning (ML) model to predict side-specific extraprostatic extension (ssEPE) to identify patients who can safely undergo nerve-sparing radical prostatectomy using preoperative clinicopathological variables.

METHODS

A retrospective sample of clinicopathological data from 900 prostatic lobes at our institution was used as the training cohort. Primary outcome was the presence of ssEPE. The baseline model for comparison had the highest performance out of current biopsy-derived predictive models for ssEPE. A separate logistic regression (LR) model was built using the same variables as the ML model. All models were externally validated using a testing cohort of 122 lobes from another institution. Models were assessed by area under receiver-operating-characteristic curve (AUROC), precision-recall curve (AUPRC), calibration, and decision curve analysis. Model predictions were explained using SHapley Additive exPlanations. This tool was deployed as a publicly available web application.

RESULTS

Incidence of ssEPE in the training and testing cohorts were 30.7 and 41.8%, respectively. The ML model achieved AUROC 0.81 (LR 0.78, baseline 0.74) and AUPRC 0.69 (LR 0.64, baseline 0.59) on the training cohort. On the testing cohort, the ML model achieved AUROC 0.81 (LR 0.76, baseline 0.75) and AUPRC 0.78 (LR 0.75, baseline 0.70). The ML model was explainable, well-calibrated, and achieved the highest net benefit for clinically relevant cutoffs of 10-30%.

CONCLUSIONS

We developed a user-friendly application that enables physicians without prior ML experience to assess ssEPE risk and understand factors driving these predictions to aid surgical planning and patient counselling (https://share.streamlit.io/jcckwong/ssepe/main/ssEPE_V2.py).

Selection of patients for nerve sparing surgery in robot‐assisted radical prostatectomy

Context

Robot‐assisted radical prostatectomy (RARP) has become the standard surgical procedure for localized prostate‐cancer (PCa). Nerve‐sparing surgery (NSS) during RARP has been associated with improved erectile function and continence rates after surgery. However, it remains unclear what are the most appropriate indications for NSS.

Objective

The objective of this study is to systematically review the available parameters for selection of patients for NSS. The weight of different clinical variables, multiparametric magnetic‐resonance‐imaging (mpMRI) findings, and the impact of multiparametric‐nomograms in the decision‐making process on (side‐specific) NSS were assessed.

Evidence acquisition

This systematic review searched relevant databases and included studies performed from January 2000 until December 2020 and recruited a total of 15 840 PCa patients. Studies were assessed that defined criteria for (side‐specific) NSS and associated them with oncological safety and/or functional outcomes. Risk of bias assessment was performed.

Evidence synthesis

Nineteen articles were eligible for full‐text review. NSS is primarily recommended in men with adequate erectile function, and with low‐risk of extracapsular extension (ECE) on the side‐of NSS. Separate clinical and radiological variables have low accuracy for predicting ECE, whereas nomograms optimize the risk‐stratification and decision‐making process to perform or to refrain from NSS when oncological safety (organ‐confined disease, PSM rates) and functional outcomes (erectile function and continence rates) were assessed.

Conclusions

Consensus exists that patients who are at high risk of ECE should refrain from NSS. Several multiparametric preoperative nomograms were developed to predict ECE with increased accuracy compared with single clinical, pathological, or radiological variables, but controversy exists on risk thresholds and decision rules on a conservative versus a less‐conservative surgical approach. An individual clinical judgment on the possibilities of NSS set against the risks of ECE is warranted.

Patient summary

NSS is aimed at sparing the nerves responsible for erection. NSS may lead to unfavorable tumor control if the risk of capsule penetration is high. Nomograms predicting extraprostatic tumor‐growth are probably most helpful.

Preoperative Prediction of Extracapsular Extension: Radiomics Signature Based on Magnetic Resonance Imaging to Stage Prostate Cancer

PURPOSE

To investigate and validate the potential role of a radiomics signature in predicting the side-specific probability of extracapsular extension (ECE) of prostate cancer (PCa).

PROCEDURES

The preoperative magnetic resonance imaging data of 238 prostatic samples from 119 enrolled PCa patients were retrospectively assessed. The samples with were randomized in a two-to-one ratio into training (n = 74) and validation (n = 45) datasets. The radiomics features were derived from T2-weighted images (T2WIs). The optimal radiomics features were identified from the least absolute shrinkage and selection operator (LASSO) logistic regression model and were used to construct a predictive radiomics signature via dimension reduction and selection approaches. The association between the radiomics signatures and pathological ECE status was explored. Receiver operating characteristic (ROC) analysis was used to assess the discriminatory ability of the signature. The calibration performance and clinical usefulness of the radiomics signature were subsequently assessed by calibration curve and decision curve analyses.

RESULTS

The proposed radiomics signature that incorporated 17 selected radiomics features was significantly associated with pathological ECE outcomes (P < 0.001) in both the training and validation datasets. The constructed model displayed good discrimination, with areas under the curve (AUC) of 0.906 (95 % confidence interval (CI), 0.847, 0.948) and 0.821 (95 % CI, 0.726, 0.894) for the training and validation datasets, respectively, and had a good calibration performance. The clinical utility of this model was confirmed through decision curve analysis.

CONCLUSIONS

The radiomics signature based on T2WIs showed the potential to predict the side-specific probability of pathological ECE status and can facilitate the preoperative individualized predictions for PCa patients.

Nomograms in Urologic Oncology: Lights and Shadows

Decision-making in urologic oncology involves integrating multiple clinical data to provide an answer to the needs of a single patient. Although the practice of medicine has always been an “art” involving experience, clinical data, scientific evidence and judgment, the creation of specialties and subspecialties has multiplied the challenges faced every day by physicians. In the last decades, with the field of urologic oncology becoming more and more complex, there has been a rise in tools capable of compounding several pieces of information and supporting clinical judgment and experience when approaching a difficult decision. The vast majority of these tools provide a risk of a certain event based on various information integrated in a mathematical model. Specifically, most decision-making tools in the field of urologic focus on the preoperative or postoperative phase and provide a prognostic or predictive risk assessment based on the available clinical and pathological data. More recently, imaging and genomic features started to be incorporated in these models in order to improve their accuracy. Genomic classifiers, look-up tables, regression trees, risk-stratification tools and nomograms are all examples of this effort. Nomograms are by far the most frequently used in clinical practice, but are also among the most controversial of these tools. This critical, narrative review will focus on the use, diffusion and limitations of nomograms in the field of urologic oncology.

Development and External Validation of a Novel Nomogram to Predict Side-specific Extraprostatic Extension in Patients with Prostate Cancer Undergoing Radical Prostatectomy

BACKGROUND

Prediction of side-specific extraprostatic extension (EPE) is crucial in selecting patients for nerve-sparing radical prostatectomy (RP).

OBJECTIVE

To develop and externally validate nomograms including multiparametric magnetic resonance imaging (mpMRI) information to predict side-specific EPE.

DESIGN, SETTING, AND PARTICIPANTS

A retrospective analysis of 1870 consecutive prostate cancer patients who underwent robot-assisted RP from 2014 to 2018 at three institutions.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

Four multivariable logistic regression models were established, including combinations of patient-based and side-specific variables: prostate-specific antigen (PSA) density, highest ipsilateral International Society of Urological Pathology (ISUP) biopsy grade, ipsilateral percentage of positive cores on systematic biopsy, and side-specific clinical stage assessed by both digital rectal examination and mpMRI. Discrimination (area under the curve [AUC]), calibration, and net benefit of these models were assessed in the development cohort and two external validation cohorts.

RESULTS AND LIMITATIONS

On external validation, AUCs of the four models ranged from 0.80 (95% confidence interval [CI] 0.68-0.88) to 0.83 (95% CI 0.72-0.90) in cohort 1 and from 0.77 (95% CI 0.62-0.87) to 0.78 (95% CI 0.64-0.88) in cohort 2. The three models including mpMRI staging information resulted in relatively higher AUCs compared with the model without mpMRI information. No major differences between the four models regarding net benefit were established. The model based on PSA density, ISUP grade, and mpMRI T stage was superior in terms of calibration. Using this model with a cut-off of 20%, 1980/2908 (68%) prostatic lobes without EPE would be found eligible for nerve sparing, whereas non-nerve sparing would be advised in 642/832 (77%) lobes with EPE.

CONCLUSIONS

Our analysis resulted in a simple and robust nomogram for the prediction of side-specific EPE, which should be used to select patients for nerve-sparing RP.

PATIENT SUMMARY

We developed a prediction model that can be used to assess accurately the likelihood of tumour extension outside the prostate. This tool can guide patient selection for safe nerve-sparing surgery.