Convolutional Neural Network-Based Deep Learning Engine for Mastoidectomy Instrument Recognition and Movement Tracking.

OBJECTIVE: To develop a convolutional neural network-based computer vision model to recognize and track 2 mastoidectomy surgical instruments-the drill and the suction-irrigator-from intraoperative video recordings of mastoidectomies. STUDY DESIGN: Technological development and model validation. SETTING: Academic center. METHODS: Ten 1-minute videos of mastoidectomies done for cochlear implantation by varying levels of resident surgeons were collected. For each video, containing 900 frames, an open-access computer vision annotation tool was used to annotate the drill and suction-irrigator class images with bounding boxes. A mastoidectomy instrument tracking module, which extracts the center coordinates of bounding boxes, was developed using a feature pyramid network and layered with DETECTRON, an open-access faster-region-based convolutional neural network. Eight videos were used to train the model, and 2 videos were used for testing. Outcome measures included Intersection over Union (IoU) ratio, accuracy, and average precision. RESULTS: For an IoU of 0.5, the mean average precision for the drill was 99% and 86% for the suction-irrigator. The model proved capable of generating maps of drill and suction-irrigator stroke direction and distance for the entirety of each video. CONCLUSIONS: This computer vision model can identify and track the drill and suction-irrigator from videos of intraoperative mastoidectomies performed by residents with excellent precision. It can now be employed to retrospectively study objective mastoidectomy measures of expert and resident surgeons, such as drill and suction-irrigator stroke concentration, economy of motion, speed, and coordination, setting the stage for characterization of objective expectations for safe and efficient mastoidectomies.

The placebo effect in randomized-controlled trials of medical treatments for chronic rhinosinusitis: A systematic review and meta-analysis.

BACKGROUND: The placebo effect observed in clinical trials evaluating medical treatments for chronic rhinosinusitis (CRS) is not well understood. This systematic review and meta-analysis sought to characterize the placebo effect present within CRS outcomes. METHODS: A systematic review of PubMed, Scopus, and Cumulated Index in Nursing and Allied Health Nursing (CINAHL) was performed. Randomized controlled trials (RCTs) evaluating medical treatments for CRS versus placebo were included. We assessed patient-reported (sino-nasal outcome test 22 [SNOT-22], nasal obstruction, sense of smell, nasal obstruction visual analogue score [VAS], sense of smell VAS, anterior rhinorrhea, and postnasal drip) and objective (Lund-Mackay Computed tomography (CT) score, peak nasal inspiratory flow [PNIF], nasal polyp scores, 40-item Smell Identification Test, serum IgE, and blood eosinophil levels) outcomes. RESULTS: Twenty-one RCTs were included, comprising 1437 patients (mean age 49.2 years). Biologics were the most common treatment investigated (n = 9). Eleven studies administered background steroids along with placebo. Following placebo administration, multiple patient-reported outcomes significantly decreased, including SNOT-22 (mean difference -9.49, 95% confidence interval [CI] [-11.26, -7.73]), nasal obstruction (-0.33 [-0.54, -0.13]), sense of smell (-0.22 [-0.33, -0.11]), nasal obstruction VAS (-2.47 [-2.87, -2.06]), and loss of smell VAS (-2.31 [-4.14, -0.47]) scores. For objective measures, significant changes occurred in Lund-Mackay CT score (-0.82, [-1.48, -0.16]) and PNIF (4.70, [4.76, 24.64]) with placebo. Placebo arms had the greatest impact when no background medications were used. CONCLUSIONS: Placebo treatments have a statistically and potentially clinically significant effect on patient-reported and some objective CRS outcomes. Further investigation is required to fully understand placebo effect, which could improve assessment of RCTs and impact patient care.

Systematic Review of Surgical Interventions for Inferior Turbinate Hypertrophy.

BACKGROUND: Various surgical interventions exist for treatment of inferior turbinate hypertrophy (ITH). Though mucosal-sparing techniques are generally preferred, there is lack of consensus on the optimal technique. OBJECTIVE: This systematic review sought to evaluate the evidence for treatment of bilateral nasal obstruction via inferior turbinate reduction (ITR) and provide a meta-analysis of expected results of various techniques. METHODS: PubMed, Scopus, Cochrane Library databases were queried to include articles describing surgical treatment for ITH. Exclusion criteria were concurrent nasal procedures or non-mucosal ITH. Primary outcomes included visual analog scale for nasal obstruction, nasal cavity volume by acoustic rhinometry, and resistance by anterior rhinomanometry. Subgroup analyses assessed outcomes by rhinitis diagnosis and length of follow-up, and radiofrequency ablation (RFA) was compared to microdebrider-assisted turbinoplasty (MAIT). RESULTS: A total of 1870 studies were identified with 62 meeting inclusion criteria. Reported techniques included turbinectomy, submucosal resection, RFA, MAIT, laser, or electrocautery.All techniques demonstrated significant improvements in nasal obstruction using the visual analog scale. Further comprehensive physiologic data for RFA, MAIT, and laser was available and, compared to baseline, these techniques resulted in significant improvements in nasal resistance, nasal cavity volume, and nasal airflow. Six studies directly compared RFA and MAIT with statistically similar results on VAS, nasal cavity volume, and resistance with median follow-up time of 3.5 months. Assessment of VAS congestion over time reveals peak benefit is achieved between 3-6 months follow-up. CONCLUSIONS: All reviewed ITR techniques improve patient-reported nasal obstruction. RFA and MAIT provide comparable improvements in patient-reported and physiologic nasal airflow outcomes and while benefits are sustained long-term, the peak benefit for both techniques appears to be achieved within the first year.

Enhanced phenotypes for identifying opioid overdose in emergency department visit electronic health record data.

Background: Accurate identification of opioid overdose (OOD) cases in electronic healthcare record (EHR) data is an important element in surveillance, empirical research, and clinical intervention. We sought to improve existing OOD electronic phenotypes by incorporating new data types beyond diagnostic codes and by applying several statistical and machine learning methods. Materials and Methods: We developed an EHR dataset of emergency department visits involving OOD cases or patients considered at risk for an OOD and ascertained true OOD status through manual chart reviews. We developed and validated prediction models using Random Forest, Extreme Gradient Boost, and Elastic Net models that incorporated 717 features involving primary and second diagnoses, chief complaints, medications prescribed, vital signs, laboratory results, and procedural codes. We also developed models limited to single data types. Results: A total of 1718 records involving 1485 patients were manually reviewed; 541 (36.4%) patients had one or more OOD. Prediction performance was similar for all models; sensitivity varied from 94% to 97%; and area under the receiver operating characteristic curve (AUC) was 98% for all methods. The primary diagnosis and chief complaint were the most important contributors to AUC performance; primary diagnoses and medication class contributed most to sensitivity; chief complaint, primary diagnosis, and vital signs were most important for specificity. Models limited to decision support data types available in real time demonstrated robust prediction performance. Conclusions: Substantial prediction performance improvements were demonstrated for identifying OODs in EHR data. Our e-phenotypes could be applied in surveillance, retrospective empirical applications, or clinical decision support systems.

Video Analysis of Otologic Instrument Movement During Resident Mastoidectomies.

OBJECTIVE: To measure surgical instrument movement during resident mastoidectomies and identify metrics that correlate with experience. STUDY DESIGN: Retrospective case series. SETTING: Tertiary care center. SUBJECTS: Ten postgraduate year (PGY) 2, 6 PGY3, 7 PGY4, and 19 PGY5 recordings of mastoidectomy performed by otolaryngology residents. INTERVENTIONS: One-minute intraoperative recordings of mastoidectomies performed during cochlear implantation were collected. Drill and suction-irrigator motion were analyzed with sports motion tracking software. MAIN OUTCOME MEASURES: Mean instrument speed, angle, and angular velocity were calculated. Mann-Whitney U tests compared mean instrument metrics between PGY levels. Change in drill speed for seven residents between their PGY2 to PGY5 years was individually analyzed. RESULTS: Mean drill speed was significantly greater for PGY5 residents compared with PGY2s (2.9 versus 1.8 cm/s, p = 0.001). Compared with PGY2 residents, suction speed was greater as a PGY5 (1.2 versus 0.9 cm/s; p = 0.201) and significantly greater as a PGY4 (1.5 versus 0.9 cm/s, p = 0.039). Of the seven residents individually analyzed, group mean drill speed increased by 0.4 cm/s, yearly. CONCLUSIONS: Drill and suction-irrigator movement during the second minute of drilling of a cortical mastoidectomy seems to increase with resident level. Objective video analysis is a potential adjunct for differentiating novices from more experienced surgeons and monitoring surgical skills progress.