No abstract

Figure) on the first postoperative CT were assessed to determine risk factors for PH.RESULTS: The median age and median BMI were 70 years (IQR 63-74.75) and 23.09 kg/m2 (IQR 20.4-24.64), respectively. PH was observed in 20 patients (10.3%) after a median follow-up of 25.5 months. Of the 20 patients, 3 were symptomatic and 2 underwent repair of the PH. The cumulative incidences were 3.6%, 10.1% and 15.1 % at 1 year, 2 years and 5 years after operation, respectively. The sex-adjusted BMI and DPRAM were significant risk factors for PH (p[0.0377,p<0.0001,respectively). The odds ratios for developing PH were 3.42 and 24.6 for BMI !22.89 kg/m2 and DPRAM !2.4 cm, respectively.CONCLUSIONS: The radiography-based incidence of PH in the current Japanese series was lower than reported incidences in American series (13.535.2%). Although most patients in the current series had low BMI, higher BMI was still significantly associated with the incidence of PH. Moreover, DPRAM was also significantly associated with the incidence, suggesting that the operative procedure for creation of the passage is critical for future development of PH.

1557 Background: Patients with prostate cancer are diagnosed through a prostate needle biopsy (PNB). Information contained in PNB pathology reports is critical for informing clinical risk stratification and treatment; however, patient comprehension of PNB pathology reports is low, and formats vary widely by institution. Natural language processing (NLP) models trained to automatically extract key information from unstructured PNB pathology reports could be used to generate personalized educational materials for patients in a scalable fashion and expedite the process of collecting registry data or screening patients for clinical trials. As proof of concept, we trained and tested four NLP models for accuracy of information extraction. Methods: Using 403 positive PNB pathology reports from over 80 institutions, we converted portable document formats (PDFs) into text using the Tesseract optical character recognition (OCR) engine, removed protected health information using the Philter open-source tool, cleaned the text with rule-based methods, and annotated clinically relevant attributes as well as structural attributes relevant to information extraction using the Brat Rapid Annotation Tool. Text pre-processing for classification and extraction was done using Scispacy and rule-based methods. Using a 75:25 train:test split (N = 302, 101), we tested conditional random field (CRF), support vector machine (SVM), bidirectional long-short term memory network (Bi-LSTM), and Bi-LSTM-CRF models, reserving 46 training reports as a validation subset for the latter two models. Model-extracted variables were compared with values manually obtained from the unprocessed PDF reports for clinical accuracy. Results: Clinical accuracy of model-extracted variables is reported in the Table. CRF was the highest performing model, with accuracies of 97% for Gleason grade, 82% for percentage of positive cores ( < 50% vs. ≥50%), 90% for perineural or lymphovascular invasion, and 100% for presence of non-acinar carcinoma histology. On manual review of inaccurate results, model performance was limited by PDF image quality, errors in OCR processing of tables or columns, and practice variability in reporting number of biopsy cores. Conclusions: Our results demonstrate successful proof of concept for the use of NLP models in accurately extracting information from PNB pathology reports, though further optimization is needed before use in clinical practice.[Table: see text]

No abstract

scan. Outcomes of interest were hypertension, proteinuria and kidney disease (CKD stages II-IV or ESRD). These 3 outcomes were also considered as a composite outcome. Kaplan-Meier survival curves were constructed and association of initial RS with event-free survival was analyzed using Log-Rank test.RESULTS: Among 745 eligible patients, 393 (53%) had RS on DMSA scan and 352 (47%) did not. There was no significant difference in median age (4 vs 3 years) or follow up time (5.6 vs 5.5 years). The RS cohort had fewer females (68% vs 84%, p < 0.001) and a higher proportion of patients with VUR grades IV or V (28% vs 13%). RS patients were more likely to undergo surgical repair of VUR (59% vs 26%, p < 0.001) at a median of 83 days (IQR 35-370) after initial DMSA scan.Significantly more patients with RS developed proteinuria (5.1% vs 1.6%, p [ 0.005) and kidney disease (2% vs 0%, p [ 0.005; Table 1). Hypertension occurred in 2.8% of patients with RS and 1% of those without (p [ 0.09). Kaplan-Meier curves stratified by cohort demonstrate significantly increased probability of adverse renal outcomes for those with RS (6.2% vs 2.6%, p [ 0.02; Figure 1).CONCLUSIONS: In this cohort study, patients with RS were significantly more likely to develop proteinuria and kidney disease. Serial, longitudinal evaluation is essential for these at-risk children.