Pyeloplasty may reverse the effect of growth delay from ureteropelvic junction obstruction in infants.

PURPOSE: To determine if children with UPJO demonstrate a clinically significant change in somatic growth following pyeloplasty. METHODS: We retrospectively evaluated the growth chart data of infants with SFU grade 3 or 4 congenital hydronephrosis at our institution from 2015 to 2022. Of those, 35 patients underwent pyeloplasty and 66 had no surgical intervention. Patients met criteria if they had SFU 3 or 4 hydronephrosis and MAG3 renal scan. If patients underwent surgery, height and weight percentiles were recorded from the pre-op and 6-16-month follow-up visits. In non-surgery patients, measurements were taken near the median age of surgery in the intervention group and 6-16 months later. Interval changes in group height and weight percentiles are compared for significant changes. RESULTS: The surgery and non-surgery groups did not differ in terms of gender (71% vs 74% Male), starting age (296 vs 244 days), starting weight (58th vs 52nd percentile), or time between measurements (255 vs 260 days), though the surgery group had significantly less height in the pre-operative period (43rd vs 55th percentile, p = 0.050) and were more likely to have delayed drainage on renal scan (83% w/delay vs 35%). The surgery group showed a significant increase in height (18.9 percentiles; 95% CI 11-27) and weight (6.0 percentiles; 95% CI 0.50-12) after intervention. CONCLUSIONS: Patients with congenital hydronephrosis due to UPJO that underwent pyeloplasty showed a significant increase in weight and height at 6-16 months postoperatively compared to those that were managed with close observation. This suggests UPJO might lead to growth delay in infants.

Utilizing eye tracking to assess electronic health record use by pharmacists in the intensive care unit.

PURPOSE: A study was conducted using high-fidelity electronic health record (EHR)-based simulations with incorporated eye tracking to understand the workflow of critical care pharmacists within the EHR, with specific attention to the data elements most frequently viewed. METHODS: Eight critical care pharmacists were given 25 minutes to review 3 simulated intensive care unit (ICU) charts deployed in the simulation instance of the EHR. Using monitor-based eye trackers, time spent reviewing screens, clinical information accessed, and screens used to access specific information were reviewed and quantified to look for trends. RESULTS: Overall, pharmacists viewed 25.5 total and 15.1 unique EHR screens per case. The majority of time was spent looking at screens focused on medications, followed by screens displaying notes, laboratory values, and vital signs. With regard to medication data, the vast majority of screen visitations were to view information on opioids/sedatives and antibiotics. With regard to laboratory values, the majority of views were focused on basic chemistry and hematology data. While there was significant variance between pharmacists, individual navigation patterns remained constant across cases. CONCLUSION: The study results suggest that in addition to medication information, laboratory data and clinical notes are key focuses of ICU pharmacist review of patient records and that navigation to multiple screens is required in order to view these data with the EHR. New pharmacy-specific EHR interfaces should consolidate these elements within a primary interface.

Pilot Testing of Simulation in the Evaluation of a Novel, Rapidly Deployable Electronic Health Record for use in Disaster Intensive Care.

OBJECTIVES: The SARS-CoV-2 pandemic has highlighted the need for rapid creation and management of ICU field hospitals with effective remote monitoring which is dependent on the rapid deployment and integration of an Electronic Health Record (EHR). We describe the use of simulation to evaluate a rapidly scalable hub-and-spoke model for EHR deployment and monitoring using asynchronous training. METHODS: We adapted existing commercial EHR products to serve as the point of entry from a simulated hospital and a separate system for tele-ICU support and monitoring of the interfaced data. To train our users we created a modular video-based curriculum to facilitate asynchronous training. Effectiveness of the curriculum was assessed through completion of common ICU documentation tasks in a high-fidelity simulation. Additional endpoints include assessment of EHR navigation, user satisfaction (Net Promoter), system usability (System Usability Scale-SUS), and cognitive load (NASA-TLX). RESULTS: A total of 5 participants achieved a 100% task completion on all domains except ventilator data (91%). Systems demonstrated high degrees of satisfaction (Net Promoter = 65.2), acceptable usability (SUS = 66.5), and acceptable cognitive load (NASA-TLX = 41.5); with higher levels of cognitive load correlating with the number of screens employed. CONCLUSIONS: Clinical usability of a comprehensive and rapidly deployable EHR was acceptable in an intensive care simulation which was preceded by < 1 hour of video education about the EHR. This model should be considered in plans for integrated clinical response with remote and accessory facilities.

Automated digital TIL analysis (ADTA) adds prognostic value to standard assessment of depth and ulceration in primary melanoma.

Accurate prognostic biomarkers in early-stage melanoma are urgently needed to stratify patients for clinical trials of adjuvant therapy. We applied a previously developed open source deep learning algorithm to detect tumor-infiltrating lymphocytes (TILs) in hematoxylin and eosin (H&E) images of early-stage melanomas. We tested whether automated digital (TIL) analysis (ADTA) improved accuracy of prediction of disease specific survival (DSS) based on current pathology standards. ADTA was applied to a training cohort (n = 80) and a cutoff value was defined based on a Receiver Operating Curve. ADTA was then applied to a validation cohort (n = 145) and the previously determined cutoff value was used to stratify high and low risk patients, as demonstrated by Kaplan-Meier analysis (p ≤ 0.001). Multivariable Cox proportional hazards analysis was performed using ADTA, depth, and ulceration as co-variables and showed that ADTA contributed to DSS prediction (HR: 4.18, CI 1.51-11.58, p = 0.006). ADTA provides an effective and attainable assessment of TILs and should be further evaluated in larger studies for inclusion in staging algorithms.

Deep Learning Based on Standard H&E Images of Primary Melanoma Tumors Identifies Patients at Risk for Visceral Recurrence and Death.

PURPOSE: Biomarkers for disease-specific survival (DSS) in early-stage melanoma are needed to select patients for adjuvant immunotherapy and accelerate clinical trial design. We present a pathology-based computational method using a deep neural network architecture for DSS prediction. EXPERIMENTAL DESIGN: The model was trained on 108 patients from four institutions and tested on 104 patients from Yale School of Medicine (YSM, New Haven, CT). A receiver operating characteristic (ROC) curve was generated on the basis of vote aggregation of individual image sequences, an optimized cutoff was selected, and the computational model was tested on a third independent population of 51 patients from Geisinger Health Systems (GHS). RESULTS: Area under the curve (AUC) in the YSM patients was 0.905 (P < 0.0001). AUC in the GHS patients was 0.880 (P < 0.0001). Using the cutoff selected in the YSM cohort, the computational model predicted DSS in the GHS cohort based on Kaplan-Meier (KM) analysis (P < 0.0001). CONCLUSIONS: The novel method presented is applicable to digital images, obviating the need for sample shipment and manipulation and representing a practical advance over current genetic and IHC-based methods.

A deep learning approach for real-time detection of sleep spindles.

OBJECTIVE: Sleep spindles have been implicated in memory consolidation and synaptic plasticity during NREM sleep. Detection accuracy and latency in automatic spindle detection are critical for real-time applications. APPROACH: Here we propose a novel deep learning strategy (SpindleNet) to detect sleep spindles based on a single EEG channel. While the majority of spindle detection methods are used for off-line applications, our method is well suited for online applications. MAIN RESULTS: Compared with other spindle detection methods, SpindleNet achieves superior detection accuracy and speed, as demonstrated in two publicly available expert-validated EEG sleep spindle datasets. Our real-time detection of spindle onset achieves detection latencies of 150-350 ms (~two-three spindle cycles) and retains excellent performance under low EEG sampling frequencies and low signal-to-noise ratios. SpindleNet has good generalization across different sleep datasets from various subject groups of different ages and species. SIGNIFICANCE: SpindleNet is ultra-fast and scalable to multichannel EEG recordings, with an accuracy level comparable to human experts, making it appealing for long-term sleep monitoring and closed-loop neuroscience experiments.