A Pragmatic Randomized Controlled Trial Examining the Impact of the Retzius-sparing Approach on Early Urinary Continence Recovery After Robot-assisted Radical Prostatectomy.

BACKGROUND: Retzius-sparing (posterior) robot-assisted radical prostatectomy (RARP) may expedite postoperative urinary continence recovery. OBJECTIVE: To compare the short-term (≤3 mo) urinary continence (UC), urinary function (UF), and UF-related bother outcomes of posterior RARP compared with standard anterior approach RARP. DESIGN, SETTING, AND PARTICIPANTS: A total of 120 patients aged 40-75 yr with low-intermediate-risk prostate cancer (per the National Comprehensive Cancer Network guidelines) underwent primary RARP at a tertiary care institution. INTERVENTION: Eligible men were randomized to receive either posterior (n=60) or anterior (n=60) RARP. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSES: Primary outcome was UC (defined as 0 pads/one security liner per day) 1 week after catheter removal. Secondary outcomes were short-term (≤3 mo) UC recovery, and UF and UF-related bother scores (measured by the International Prostate Symptom Score [IPSS] and IPSS quality-of-life scores, respectively) assessed at 1 and 2 wk, and 1 and 3 mo following catheter removal. Continence outcomes were objectively verified using 24-hr pad weights. UC recovery was analyzed using Kaplan-Meier method and Cox proportional hazards regression; UF and UF-related bother outcomes were compared using linear generalized estimating equations (GEEs). Perioperative complications, positive surgical margin, and biochemical recurrence-free survival (BCRFS) represent secondary outcomes reported in the study. RESULTS AND LIMITATIONS: Compared with 48% in the anterior arm, 71% men undergoing posterior RARP were continent 1 wk after catheter removal (p=0.01); corresponding median 24-h pad weights were 25 and 5g (p=0.001). Median time to continence in posterior versus anterior RARP was 2 and 8 d postcatheter removal, respectively (log-rank p=0.02); results were confirmed on multivariable regression analyses. GEE analyses showed that UF-related bother (but not UF) scores were significantly lower in the posterior versus anterior RARP group at 1 wk, 2 wk, and 1 mo on GEE analyses. Incidence of postoperative complications (12% anterior vs 18% posterior) and probability of BCRFS (0.91 vs 0.91) were comparable in the two arms. CONCLUSIONS: In this single-center randomized study, the Retzius-sparing approach of RARP resulted in earlier recovery of UC and lower UF-related bother compared with standard RARP. These results require long-term validation and reproduction by other centers, as well as studies on men with high-risk localized disease. PATIENT SUMMARY: In our hands, men with low-intermediate-risk prostate cancer undergoing Retzius-sparing robot-assisted radical prostatectomy (RARP) had earlier recovery of urinary continence and lower urinary function-related bother than those undergoing standard RARP.

Development of CYP3A4 inhibition models: comparisons of machine-learning techniques and molecular descriptors.

Computational models of cytochrome P450 3A4 inhibition were developed based on high-throughput screening data for 4470 proprietary compounds. Multiple models differentiating inhibitors (IC(50) <3 microM) and noninhibitors were generated using various machine-learning algorithms (recursive partitioning [RP], Bayesian classifier, logistic regression, k-nearest-neighbor, and support vector machine [SVM]) with structural fingerprints and topological indices. Nineteen models were evaluated by internal 10-fold cross-validation and also by an independent test set. Three most predictive models, Barnard Chemical Information (BCI)-fingerprint/SVM, MDL-keyset/SVM, and topological indices/RP, correctly classified 249, 248, and 236 compounds of 291 noninhibitors and 135, 137, and 147 compounds of 179 inhibitors in the validation set. Their overall accuracies were 82%, 82%, and 81%, respectively. Investigating applicability of the BCI/SVM model found a strong correlation between the predictive performance and the structural similarity to the training set. Using Tanimoto similarity index as a confidence measurement for the predictions, the limitation of the extrapolation was 0.7 in the case of the BCI/SVM model. Taking consensus of the 3 best models yielded a further improvement in predictive capability, kappa = 0.65 and accuracy = 83%. The consensus model could also be tuned to minimize either false positives or false negatives depending on the emphasis of the screening.