Validation of a Zero-shot Learning Natural Language Processing Tool to Facilitate Data Abstraction for Urologic Research.

BACKGROUND: Urologic research often requires data abstraction from unstructured text contained within the electronic health record. A number of natural language processing (NLP) tools have been developed to aid with this time-consuming task; however, the generalizability of these tools is typically limited by the need for task-specific training. OBJECTIVE: To describe the development and validation of a zero-shot learning NLP tool to facilitate data abstraction from unstructured text for use in downstream urologic research. DESIGN, SETTING, AND PARTICIPANTS: An NLP tool based on the GPT-3.5 model from OpenAI was developed and compared with three physicians for time to task completion and accuracy for abstracting 14 unique variables from a set of 199 deidentified radical prostatectomy pathology reports. The reports were processed in vectorized and scanned formats to establish the impact of optical character recognition on data abstraction. INTERVENTION: A zero-shot learning NLP tool for data abstraction. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: The tool was compared with the human abstractors in terms of superiority for data abstraction speed and noninferiority for accuracy. RESULTS AND LIMITATIONS: The human abstractors required a median (interquartile range) of 93 s (72-122 s) per report for data abstraction, whereas the software required a median of 12 s (10-15 s) for the vectorized reports and 15 s (13-17 s) for the scanned reports (p < 0.001 for all paired comparisons). The accuracies of the three human abstractors were 94.7% (95% confidence interval [CI], 93.8-95.5%), 97.8% (95% CI, 97.2-98.3%), and 96.4% (95% CI, 95.6-97%) for the combined set of 2786 data points. The tool had accuracy of 94.2% (95% CI, 93.3-94.9%) for the vectorized reports and was noninferior to the human abstractors at a margin of -10% (α = 0.025). The tool had slightly lower accuracy of 88.7% (95% CI 87.5-89.9%) for the scanned reports, making it noninferior to two of three human abstractors. CONCLUSIONS: The developed zero-shot learning NLP tool offers urologic researchers a highly generalizable and accurate method for data abstraction from unstructured text. An open access version of the tool is available for immediate use by the urologic community. PATIENT SUMMARY: In this report, we describe the design and validation of an artificial intelligence tool for abstracting discrete data from unstructured notes contained within the electronic medical record. This freely available tool, which is based on the GPT-3.5 technology from OpenAI, is intended to facilitate research and scientific discovery by the urologic community.

Applications of artificial intelligence in prostate cancer histopathology.

The diagnosis of prostate cancer (PCa) depends on the evaluation of core needle biopsies by trained pathologists. Artificial intelligence (AI) derived models have been created to address the challenges posed by pathologists' increasing workload, workforce shortages, and variability in histopathology assessment. These models with histopathological parameters integrated into sophisticated neural networks demonstrate remarkable ability to identify, grade, and predict outcomes for PCa. Though the fully autonomous diagnosis of PCa remains elusive, recently published data suggests that AI has begun to serve as an initial screening tool, an assistant in the form of a real-time interactive interface during histological analysis, and as a second read system to detect false negative diagnoses. Our article aims to describe recent advances and future opportunities for AI in PCa histopathology.

Factors that influence preference for male or female urologist among underserved patients in New York City.

Objective: To examine the prevalence of patient preference for male or female urologic provider and explore which patient characteristics influence this preference. Materials and Methods: After obtaining hospital Institutional Review Board approval, a 14-question survey in English and Spanish was administered across four general urology clinic sites in a single hospital system in New York City. The survey asked demographic questions and preference for a male or a female urologist. The survey included questions pertaining to the nature of the clinic visit and subsequent provider preference as well. Statistics were performed using Stata 16 (StataCorp, College Station, TX). Results: A total of 540 patients completed the 14-question survey. The vast majority of survey respondents identified as male (90%). The largest proportion demographic groups were those aged 41-60 (47%), Hispanic or Latino (43%), Catholic (47%), unemployed (40%) and those with a high school level of education (34%). Most patients (60%) did not have a preference for a specific gender provider, whereas 37% preferred a male provider, and 3% preferred a female provider. On univariate analysis, patient age 25-40, less than high school education level and lack of employment were significant predictors of provider gender preference (p < 0.05), with most patients indicating a male provider preference. On multivariate analysis of gender, age, education level and employment status, gender and education level were not significant predictors of preference, whereas age 25-40 and being unemployed were significant predictors (p < 0.05). Conclusion: Patient gender, race and religion do not appear to influence their preference to be seen by a male or a female urologist in the clinic setting. However, patient age, unemployment and potentially educational attainment were significantly associated with a provider gender preference.

Biopsy and Erectile Functional Outcomes of Partial Prostate Ablation: A Systematic Review and Meta-analysis of Prospective Studies.

OBJECTIVE: To provide a systematic summary of prospectively performed studies evaluating ablative therapies for the treatment of prostate cancer (PCa) that included protocol-mandated assessment of (1) residual disease by post-treatment biopsy and/or (2) erectile functional outcomes. MATERIALS AND METHODS: We performed a comprehensive literature search in September 2022. Studies were evaluated according to a predefined and registered plan in PROSPERO (CRD42022302777). Only prospective trials with protocol-mandated post-treatment prostate biopsies or functional assessments were included. Targeted focal therapy was the only ablation pattern with sufficient data to perform meta-analyses (29 studies, 1079 patients). RESULTS: At baseline, 65.0% of patients treated with targeted focal therapy harbored grade group (GG) ≥2 PCa. One year after treatment, in-field treatment failure with ≥GG1 and ≥GG2 PCa occurred in 25.7% (range 11.1%-66.7%) and 8.8% (range 0%-27.8%) of men, respectively. In patients that received whole-gland biopsies 1year after ablation, residual ≥GG1 and ≥GG2 PCa was detected anywhere in the prostate in 43.7% (range 19.4%-71.7%) and 13.0% (range 0%-35.9%) of men. Erectile function was negatively affected by treatment, but 78.7% were potent 1year after targeted focal therapy (7 studies, 197 patients), and the average decrease in erectile function scores was 8.8% at 1year (21 studies, 760 patients). CONCLUSION: Though long-term data after targeted focal therapy are limited, oncologic and treatment failure occurred in 13% and 9% (≥GG2 at 6-12months after treatment). Most men were able to maintain potency. This work can help benchmark new techniques and power future trials.

Robot-Assisted Repair of Ureteroenteric Strictures After Cystectomy with Urinary Diversion: Technique Description and Outcomes from the European Robotic Urology Section Scientific Working Group.

Background: Robot-assisted repair of benign ureteroenteric anastomotic strictures (UAS) provides an alternative to the open approach. We aimed to report short-, medium-, and long-term outcomes for robotic repair of benign UAS, and to provide a detailed video demonstration of critical operative techniques in performing this procedure robotically. Materials and Methods: Between January 2013 and September 2022, 31 patients from seven institutions who previously underwent radical cystectomy and subsequently developed UAS underwent robotic repair of UAS. Perioperative variables were prospectively collected, and postoperative outcomes were assessed. The surgery starts with a lysis of adhesions after previous surgery. Ureters are dissected, and the level of the stricture is identified. The ureter is then divided, and the stricture is resected. Finally, the ureter is spatulated and reimplanted with Nesbit technique after stenting with Double-J stents. In cases where both ureters show strictures, Wallace technique for reimplantation can be applied. Results: After robotic or open cystectomy, 31 patients had a total of 43 UAS at a median (interquartile range) follow-up of 21 (9-43) months. Median stricture length was 2.0 (1.0-3.25) cm, operative duration was 141 (121-232) minutes, estimated blood loss was 100 (50-150) mL, and length of hospital stay was 5 (3-9) days. One (3.2%) case was converted to open and one (3.2%) intraoperative complication occurred. Seven (22.6%) patients experienced postoperative complications, including four (12.9%) Clavien-Dindo grade 3 complications. No Clavien-Dindo grade 4 or 5 complications occurred. Stricture recurrence occurred in 2 (6.5%) patients. Conclusions: These results demonstrate that robotic repair of UAS is feasible and effective approach with outcomes in line with prior open series. Patient Consent Statement: Authors have received and archived patient consent for video recording and publication in advance of video recording of procedure.

Technical Tips in Managing Large Median Lobes During Robot-assisted Radical Prostatectomy.

Prostate cancer surgeons are commonly faced by a technically challenging situation dealing with prostate cancer having large median lobes. Patients with large median lobes often have larger prostates, which makes it difficult to visualize anatomical planes during robot-assisted radical prostatectomy (RARP). Herein, we described our experience in dealing with large median lobes during RARP. We have focused on technical tips to avoid complications and facilitate a smooth procedure in patients with large median lobes during RARP. A total of 2671 patients who underwent RARP were divided into two groups based on the presence or absence of a protruded median lobe (PML): group A (2411 patients without a PML) and group B (260 patients with a PML). All patients underwent preoperative magnetic resonance imaging and final intraoperative confirmation for the presence of a PML. Pre-, intra-, and postoperative parameters were compared in two groups using the Student t test and two-proportion t test as appropriate. Patients in group B have statistically significantly higher median prostate-specific antigen (PSA; 7.7 vs 5.8 ng/dl), PSA density (0.17 vs 0.09), and International Prostate Symptom Score (19.5 vs 7.2); longer median console time (114 vs 134 min) and surgery time (145 vs 170 min); and higher blood loss (150 vs 175 ml) than those in group A. There were no statistically significant differences in pathological stages (T2, T3; 87%, 13% vs 88%, 12%) and rates of positive surgical margins (7% vs 8.5%) between groups A and B. Single-center and retrospective design was the major limitation of our study. We conclude that understanding the key steps to facilitate bladder neck dissection is vital to avoid serious intraoperative events and to maximize outcomes. Patient summary: In this report, we looked at our robotic radical prostatectomy cohort with large median lobes. We found that surgery in these patients requires more time and blood loss, but similar cancer control. We conclude that following the key steps are important to avoid complications.