Faculty Bibliography

Background/UNASSIGNED:Radical cystectomy for bladder cancer has one of the highest rates of morbidity among urologic surgery, but the ability to predict postoperative complications remains poor. Our study objective was to create machine learning models to predict complications and factors leading to extended length of hospital stay and discharge to a higher level of care after radical cystectomy. Methods/UNASSIGNED:Using the American College of Surgeons National Surgical Quality Improvement Program, peri-operative adverse outcome variables for patients undergoing elective radical cystectomy for bladder cancer from 2005 to 2016 were extracted. Variables assessed include occurrence of minor, infectious, serious, or any adverse events, extended length of hospital stay, and discharge to higher-level care. To develop predictive models of radical cystectomy complications, we fit generalized additive model (GAM), least absolute shrinkage and selection operator (LASSO) logistic, neural network, and random forest models to training data using various candidate predictor variables. Each model was evaluated on the test data using receiver operating characteristic curves. Results/UNASSIGNED:A total of 7557 patients were identified who met the inclusion criteria, and 2221 complications occurred. LASSO logistic models demonstrated the highest area under curve for predicting any complications (0.63), discharge to a higher level of care (0.75), extended length of stay (0.68), and infectious (0.62) adverse events. This was comparable with random forest in predicting minor (0.60) and serious (0.63) adverse events. Conclusions/UNASSIGNED:Our models perform modestly in predicting radical cystectomy complications, highlighting both the complex cystectomy process and the limitations of large healthcare datasets. Identifying the most important variable leading to each type of adverse event may allow for further strategies to model cystectomy complications and target optimization of modifiable variables pre-operative to reduce postoperative adverse events.

PURPOSE/OBJECTIVE:While MRI-Ultrasound Fusion-targeted biopsy (MRF-TB) allows for improved detection of clinically significant prostate cancer (csPCa), concerning numbers of clinically significant disease are still missed. We hypothesize that a number of these are due to the learning curve associated with MRF-TB. We report results of repeat MRF-TB in men with continued suspicion for cancer and the institutional learning curve in detection of csPCa over time. MATERIALS AND METHODS/METHODS:Analysis of 1813 prostate biopsies in a prospectively acquired cohort of men presenting for prostate biopsy over a 4-year period. All men were offered pre-biopsy MRI and assigned a maximum Prostate Imaging - Reporting and Data System version 2 (PI-RADS) score. Biopsy outcomes of men with suspicious region of interest (ROI) were compared. The relationship between time and csPCa detection was analyzed. RESULTS:csPCa detection rate increased 26% over time in men with PI-RADS 4 and 5 (4/5) ROI. On repeat MRF-TB in men with continued suspicion for cancer, 53% of men with PI-RADS 4/5 ROI demonstrated clinically significant discordance from initial MRF-TB, compared to only 23% of men with PI-RADS 1/2 ROI. Significantly less csPCa were missed or under-graded in the most recent biopsies as compared to the earliest biopsies. CONCLUSION/CONCLUSIONS:High upgrade rates on repeat MRF-TB and increasing cancer detection rate over time demonstrate the significant learning curve associated with MRF-TB. Men with low risk or negative biopsies with persistent concerning ROI should be promptly re-biopsied. Improved targeting accuracy with operator experience can help decrease the number of missed csPCa.

BACKGROUND:The number of prostate biopsy cores that need to be taken from each magnetic resonance imaging (MRI) region of interest (ROI) to optimize sampling while minimizing overdetection has not yet been clearly elucidated. OBJECTIVE:To characterize the incremental value of additional MRI-ultrasound (US) fusion targeted biopsy cores in defining the optimal number when planning biopsy and to predict men who might benefit from more than two targeted cores. DESIGN, SETTING, AND PARTICIPANTS/METHODS:This was a retrospective cohort study of MRI-US fusion targeted biopsies between 2015 and 2017. INTERVENTION/METHODS:MRI-US fusion targeted biopsy in which four biopsy cores were directed to each MRI-targeted ROI. OUTCOMES MEASUREMENTS AND STATISTICAL ANALYSIS/UNASSIGNED:The MRI-targeted cores representing the first highest Gleason core (FHGC) and first clinically significant cancer core (FCSC; GS≥3+4) were evaluated. We analyzed the frequency of FHGC and FCSC among cores 1-4 and created a logistic regression model to predict FHGC >2. The number of unnecessary cores avoided and the number of malignancies missed for each Gleason grade were calculated via clinical utility analysis. The level of agreement between biopsy and prostatectomy Gleason scores was evaluated using Cohen's κ. RESULTS AND LIMITATIONS/CONCLUSIONS:A total of 479 patients underwent fusion targeted biopsy with four individual cores, with 615 ROIs biopsied. Among those, FHGC was core 1 in 477 (76.8%), core 2 in 69 (11.6%), core 3 in 48 (7.6%), and core 4 in 24 men (4.0%) with any cancer. Among men with clinically significant cancer, FCSC was core 1 in 191 (77.8%), core 2 in 26 (11.1%), core 3 in 17 (6.2%), and core 4 in 11 samples (4.9%). In comparison to men with a Prostate Imaging-Reporting and Data System (PI-RADS) score of 5, patients were significantly less likely to have FHGS >2 if they had PI-RADS 4 (odds ratio [OR] 0.287; p=0.006), PI-RADS 3 (OR 0.284; p=0.006), or PI-RADS 2 (OR 0.343; p=0.015). Study limitations include a single-institution experience and the retrospective nature. CONCLUSIONS:Cores 1-2 represented FHGC 88.4% and FCSC 88.9% of the time. A PI-RADS score of 5 independently predicted FHGC >2. Although the majority of cancers in our study were appropriately characterized in the first two biopsy cores, there remains a proportion of men who would benefit from additional cores. PATIENT SUMMARY/UNASSIGNED:In men who undergo magnetic resonance imaging-ultrasound fusion targeted biopsy, the first two biopsy cores diagnose the majority of clinically significant cancers. However, there remains a proportion of men who would benefit from additional cores.