An ASER AI/ML expert panel formative user research study for an interpretable interactive splenic AAST grading graphical user interface prototype.

PURPOSE: The AAST Organ Injury Scale is widely adopted for splenic injury severity but suffers from only moderate inter-rater agreement. This work assesses SpleenPro, a prototype interactive explainable artificial intelligence/machine learning (AI/ML) diagnostic aid to support AAST grading, for effects on radiologist dwell time, agreement, clinical utility, and user acceptance. METHODS: Two trauma radiology ad hoc expert panelists independently performed timed AAST grading on 76 admission CT studies with blunt splenic injury, first without AI/ML assistance, and after a 2-month washout period and randomization, with AI/ML assistance. To evaluate user acceptance, three versions of the SpleenPro user interface with increasing explainability were presented to four independent expert panelists with four example cases each. A structured interview consisting of Likert scales and free responses was conducted, with specific questions regarding dimensions of diagnostic utility (DU); mental support (MS); effort, workload, and frustration (EWF); trust and reliability (TR); and likelihood of future use (LFU). RESULTS: SpleenPro significantly decreased interpretation times for both raters. Weighted Cohen's kappa increased from 0.53 to 0.70 with AI/ML assistance. During user acceptance interviews, increasing explainability was associated with improvement in Likert scores for MS, EWF, TR, and LFU. Expert panelists indicated the need for a combined early notification and grading functionality, PACS integration, and report autopopulation to improve DU. CONCLUSIONS: SpleenPro was useful for improving objectivity of AAST grading and increasing mental support. Formative user research identified generalizable concepts including the need for a combined detection and grading pipeline and integration with the clinical workflow.

Pulmonary contusion: automated deep learning-based quantitative visualization.

PURPOSE: Rapid automated CT volumetry of pulmonary contusion may predict progression to Acute Respiratory Distress Syndrome (ARDS) and help guide early clinical management in at-risk trauma patients. This study aims to train and validate state-of-the-art deep learning models to quantify pulmonary contusion as a percentage of total lung volume (Lung Contusion Index, or auto-LCI) and assess the relationship between auto-LCI and relevant clinical outcomes. METHODS: 302 adult patients (age ≥ 18) with pulmonary contusion were retrospectively identified from reports between 2016 and 2021. nnU-Net was trained on manual contusion and whole-lung segmentations. Point-of-care candidate variables for multivariate regression included oxygen saturation, heart rate, and systolic blood pressure on admission. Logistic regression was used to assess ARDS risk, and Cox proportional hazards models were used to determine differences in ICU length of stay and mechanical ventilation time. RESULTS: Mean Volume Similarity Index and mean Dice scores were 0.82 and 0.67. Interclass correlation coefficient and Pearson r between ground-truth and predicted volumes were 0.90 and 0.91. 38 (14%) patients developed ARDS. In bivariate analysis, auto-LCI was associated with ARDS (p < 0.001), ICU admission (p < 0.001), and need for mechanical ventilation (p < 0.001). In multivariate analyses, auto-LCI was associated with ARDS (p = 0.04), longer length of stay in the ICU (p = 0.02) and longer time on mechanical ventilation (p = 0.04). AUC of multivariate regression to predict ARDS using auto-LCI and clinical variables was 0.70 while AUC using auto-LCI alone was 0.68. CONCLUSION: Increasing auto-LCI values corresponded with increased risk of ARDS, longer ICU admissions, and longer periods of mechanical ventilation.

Artificial intelligence CAD tools in trauma imaging: a scoping review from the American Society of Emergency Radiology (ASER) AI/ML Expert Panel.

BACKGROUND: AI/ML CAD tools can potentially improve outcomes in the high-stakes, high-volume model of trauma radiology. No prior scoping review has been undertaken to comprehensively assess tools in this subspecialty. PURPOSE: To map the evolution and current state of trauma radiology CAD tools along key dimensions of technology readiness. METHODS: Following a search of databases, abstract screening, and full-text document review, CAD tool maturity was charted using elements of data curation, performance validation, outcomes research, explainability, user acceptance, and funding patterns. Descriptive statistics were used to illustrate key trends. RESULTS: A total of 4052 records were screened, and 233 full-text articles were selected for content analysis. Twenty-one papers described FDA-approved commercial tools, and 212 reported algorithm prototypes. Works ranged from foundational research to multi-reader multi-case trials with heterogeneous external data. Scalable convolutional neural network-based implementations increased steeply after 2016 and were used in all commercial products; however, options for explainability were narrow. Of FDA-approved tools, 9/10 performed detection tasks. Dataset sizes ranged from < 100 to > 500,000 patients, and commercialization coincided with public dataset availability. Cross-sectional torso datasets were uniformly small. Data curation methods with ground truth labeling by independent readers were uncommon. No papers assessed user acceptance, and no method included human-computer interaction. The USA and China had the highest research output and frequency of research funding. CONCLUSIONS: Trauma imaging CAD tools are likely to improve patient care but are currently in an early stage of maturity, with few FDA-approved products for a limited number of uses. The scarcity of high-quality annotated data remains a major barrier.

Improving the efficacy of exome sequencing at a quaternary care referral centre: novel mutations, clinical presentations and diagnostic challenges in rare neurogenetic diseases.

BACKGROUND: We used a multimodal approach including detailed phenotyping, whole exome sequencing (WES) and candidate gene filters to diagnose rare neurological diseases in individuals referred by tertiary neurology centres. METHODS: WES was performed on 66 individuals with neurogenetic diseases using candidate gene filters and stringent algorithms for assessing sequence variants. Pathogenic or likely pathogenic missense variants were interpreted using in silico prediction tools, family segregation analysis, previous publications of disease association and relevant biological assays. RESULTS: Molecular diagnosis was achieved in 39% (n=26) including 59% of childhood-onset cases and 27% of late-onset cases. Overall, 37% (10/27) of myopathy, 41% (9/22) of neuropathy, 22% (2/9) of MND and 63% (5/8) of complex phenotypes were given genetic diagnosis. Twenty-seven disease-associated variants were identified including ten novel variants in FBXO38, LAMA2, MFN2, MYH7, PNPLA6, SH3TC2 and SPTLC1. Single-nucleotide variants (n=10) affected conserved residues within functional domains and previously identified mutation hot-spots. Established pathogenic variants (n=16) presented with atypical features, such as optic neuropathy in adult polyglucosan body disease, facial dysmorphism and skeletal anomalies in cerebrotendinous xanthomatosis, steroid-responsive weakness in congenital myasthenia syndrome 10. Potentially treatable rare diseases were diagnosed, improving the quality of life in some patients. CONCLUSIONS: Integrating deep phenotyping, gene filter algorithms and biological assays increased diagnostic yield of exome sequencing, identified novel pathogenic variants and extended phenotypes of difficult to diagnose rare neurogenetic disorders in an outpatient clinic setting.

Accelerating voxelwise annotation of cross-sectional imaging through AI collaborative labeling with quality assurance and bias mitigation.

Background: precision-medicine quantitative tools for cross-sectional imaging require painstaking labeling of targets that vary considerably in volume, prohibiting scaling of data annotation efforts and supervised training to large datasets for robust and generalizable clinical performance. A straight-forward time-saving strategy involves manual editing of AI-generated labels, which we call AI-collaborative labeling (AICL). Factors affecting the efficacy and utility of such an approach are unknown. Reduction in time effort is not well documented. Further, edited AI labels may be prone to automation bias. Purpose: In this pilot, using a cohort of CTs with intracavitary hemorrhage, we evaluate both time savings and AICL label quality and propose criteria that must be met for using AICL annotations as a high-throughput, high-quality ground truth. Methods: 57 CT scans of patients with traumatic intracavitary hemorrhage were included. No participant recruited for this study had previously interpreted the scans. nnU-net models trained on small existing datasets for each feature (hemothorax/hemoperitoneum/pelvic hematoma; n = 77-253) were used in inference. Two common scenarios served as baseline comparison- de novo expert manual labeling, and expert edits of trained staff labels. Parameters included time effort and image quality graded by a blinded independent expert using a 9-point scale. The observer also attempted to discriminate AICL and expert labels in a random subset (n = 18). Data were compared with ANOVA and post-hoc paired signed rank tests with Bonferroni correction. Results: AICL reduced time effort 2.8-fold compared to staff label editing, and 8.7-fold compared to expert labeling (corrected p < 0.0006). Mean Likert grades for AICL (8.4, SD:0.6) were significantly higher than for expert labels (7.8, SD:0.9) and edited staff labels (7.7, SD:0.8) (corrected p < 0.0006). The independent observer failed to correctly discriminate AI and human labels. Conclusion: For our use case and annotators, AICL facilitates rapid large-scale curation of high-quality ground truth. The proposed quality control regime can be employed by other investigators prior to embarking on AICL for segmentation tasks in large datasets.

Simple and economical HandClench Relaxometer device for reliable and sensitive measurement of grip myotonia in myotonic dystrophy.

Grip myotonia and weakness are attractive treatment response biomarkers in clinical trials of myotonic dystrophy type 1 (DM1). There is a need to develop simple, patient-friendly and reproducible methods of quantifying grip myotonia in multisite trial settings. We designed a HandClench Relaxometer (HCR) that measures grip myotonia and strength. In contrast with the existing quantitative myometry (QMA) setup, the HCR is portable, economical, can be used with any laptop and generates automated command prompts. We demonstrate the feasibility and reliability of HCR device in twenty DM1 individuals and ten age-matched controls; patients returned for follow up within two months. The device showed excellent day to day reproducibility (ICC >0.80) in patients. The HCR device detected myotonia in milder muscle disease and measured longer myotonia duration than QMA indicating enhanced sensitivity for quantifying myotonia in DM1. The reaction time to the relax but not squeeze command was delayed and showed warm up similar to myotonia in DM1. HCR outcomes were correlated with key pinch strength, hand dexterity test, and fat replacement in the MRI of the long finger flexor muscles. Use of the HCR is warranted for grip myotonia and strength measurements in longitudinal observational and interventional studies of DM1.

Association between initial morphine intake and body weight change, acoustic startle reflex and drug seeking in rats.

RATIONALE: Although chronic use of opiates can induce physical dependence and addiction, individual differences contributing to these symptoms are largely unknown. OBJECTIVES: Using intravenous morphine self-administration (MSA), we investigated whether individual differences in drug intake are associated with weight change, acoustic startle reflex (ASR), pre-pulse inhibition (PPI), and drug seeking during spontaneous withdrawal. METHODS: Male Sprague-Dawley rats self-administered morphine (0.5 mg/kg/infusion) or saline for 3 weeks (4-6 h/day, 5 days/week) and drug intake and body weight were monitored daily. The ASR and the PPI (baseline, 1 day and 1 week) and drug seeking (1 week) were measured during spontaneous withdrawal. RESULTS: Morphine animals did not gain weight (101 % ± 0.69), while the control animals did (115 % ± 1.06) after 3 weeks of self-administration. The ASR and the PPI were not significantly different between morphine and saline animals in 1-day or 1-week withdrawal. However, individual differences in initial (first 10 min), but not total (4-6 h), morphine intake of the daily sessions were positively correlated with weight change (r = 0.437, p = 0.037) and drug seeking (r = 0.424, p = 0.035) while inversely correlated with the ASR (r = -0.544, p = 0.005) in 1-week withdrawal from chronic morphine. CONCLUSIONS: A subgroup of animals that self-administered a larger amount of morphine at the beginning of the daily sessions exhibited subsequent weight gain, reduced ASR, and enhanced drug seeking in morphine withdrawal. Thus, individual differences in initial morphine intake may reveal a novel behavioral phenotype in opioid addiction.