Evaluation of GPT Large Language Model Performance on RSNA 2023 Case of the Day Questions.

Background GPT-4V (GPT-4 with vision, ChatGPT; OpenAI) has shown impressive performance in several medical assessments. However, few studies have assessed its performance in interpreting radiologic images. Purpose To assess and compare the accuracy of GPT-4V in assessing radiologic cases with both images and textual context to that of radiologists and residents, to assess if GPT-4V assistance improves human accuracy, and to assess and compare the accuracy of GPT-4V with that of image-only or text-only inputs. Materials and Methods Seventy-two Case of the Day questions at the RSNA 2023 Annual Meeting were curated in this observer study. Answers from GPT-4V were obtained between November 26 and December 10, 2023, with the following inputs for each question: image only, text only, and both text and images. Five radiologists and three residents also answered the questions in an "open book" setting. For the artificial intelligence (AI)-assisted portion, the radiologists and residents were provided with the outputs of GPT-4V. The accuracy of radiologists and residents, both with and without AI assistance, was analyzed using a mixed-effects linear model. The accuracies of GPT-4V with different input combinations were compared by using the McNemar test. P < .05 was considered to indicate a significant difference. Results The accuracy of GPT-4V was 43% (31 of 72; 95% CI: 32, 55). Radiologists and residents did not significantly outperform GPT-4V in either imaging-dependent (59% and 56% vs 39%; P = .31 and .52, respectively) or imaging-independent (76% and 63% vs 70%; both P = .99) cases. With access to GPT-4V responses, there was no evidence of improvement in the average accuracy of the readers. The accuracy obtained by GPT-4V with text-only and image-only inputs was 50% (35 of 70; 95% CI: 39, 61) and 38% (26 of 69; 95% CI: 27, 49), respectively. Conclusion The radiologists and residents did not significantly outperform GPT-4V. Assistance from GPT-4V did not help human raters. GPT-4V relied on the textual context for its outputs. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Katz in this issue.

Partial facial paralysis induced by sialolithiasis of the parotid gland: a case report.

BACKGROUND: Facial paralysis due to parotid sialolithiasis-induced parotitis is a unusual clinical phenomenon that has not been reported in prior literature. This scenario can present a diagnostic challenge due to its rarity and complex symptomatology, particularly if a patient has other potential contributing factors such as facial trauma or bilateral forehead botox injections as in this patient. This case report elucidates such a complex presentation, aiming to increase awareness and promote timely recognition among clinicians. CASE PRESENTATION: A 56-year-old male, with a medical history significant for hyperlipidemia, recurrent parotitis secondary to parotid sialolithiasis, and recent bilateral forehead cosmetic Botox injections presented to the emergency department with right lower facial drooping. This onset was about an hour after waking up and was of 4 h duration. The patient also had a history of a recent ground level fall four days prior that resulted in facial trauma to his right eyebrow without any evident neurological deficits in the region of the injury. A thorough neurological exam revealed sensory and motor deficits across the entirety of the right face, indicating a potential lesion affecting the buccal and marginal mandibular branches of the facial nerve (CN VII). Several differential diagnoses were considered for the lower motor neuron lesion, including soft tissue trauma or swelling from the recent fall, compression due to the known parotid stone, stroke, and complex migraines. An MRI of the brain was conducted to rule out a stroke, with no significant findings. A subsequent CT scan of the neck revealed an obstructed and dilated right Stensen's duct with a noticeably larger and anteriorly displaced sialolith and evidence of parotid gland inflammation. A final diagnosis of facial palsy due to parotitis secondary to sialolithiasis was made. The patient was discharged and later scheduled for a procedure to remove the sialolith which resolved his facial paralysis. CONCLUSIONS: This case emphasizes the need for a comprehensive approach to the differential diagnosis in presentations of facial palsy. It underscores the potential involvement of parotid sialolithiasis, particularly in patients with a history of recurrent parotitis or facial trauma. Prompt recognition of such uncommon presentations can prevent undue interventions, aid in timely appropriate management, and significantly contribute to the patient's recovery and prevention of long-term complications.

Disadvantaged groups have greater spatial access to pharmacies in New York state.

BACKGROUND: The accessibility of pharmacies has been associated with overall health and wellbeing. Past studies have suggested that low income and racial minority communities are underserved by pharmacies. However, the literature is inconsistent in finding links between area-level income or racial and ethnic composition and access to pharmacies. Here we aim to assess area-level spatial access to pharmacies across New York State (NYS), hypothesizing that Census Tracts with higher poverty rates and higher percentages of Black and Hispanic residents would have lower spatial access. METHODS: The population weighted mean shortest road network distance (PWMSD) to a pharmacy in 2018 was calculated for each Census Tract in NYS. This statistic was calculated from the shortest road network distance to a pharmacy from the centroid of each Census block within a tract, with the mean across census blocks weighted by the population of the census block. Cross-sectional analyses were conducted to assess links between Tract-level socio demographic characteristics and Tract-level PWMSD to a pharmacy. RESULTS: Overall the mean PWMSD to a pharmacy across Census tracts in NYS was 2.07 Km (SD = 3.35, median 0.85 Km). Shorter PWMSD to a pharmacy were associated with higher Tract-level % poverty, % Black/African American (AA) residents, and % Hispanic/Latino residents and with lower Tract-level % of residents with a college degree. Compared to tracts in the lowest quartile of % Black/AA residents, tracts in the highest quartile had a 70.7% (95% CI 68.3-72.9%) shorter PWMSD to a pharmacy. Similarly, tracts in the highest quartile of % poverty had a 61.3% (95% CI 58.0-64.4%) shorter PWMSD to a pharmacy than tracts in the lowest quartile. CONCLUSION: The analyses show that tracts in NYS with higher racial and ethnic minority populations and higher poverty rates have higher spatial access to pharmacies.

Machine learning-based mortality prediction in hip fracture patients using biomarkers.

The purpose of this retrospective study was to assess whether mortality following a hip fracture can be predicted by a machine learning model trained on basic blood and lab test data as well as basic demographic data. Additionally, the purpose was to identify the key variables most associated with 1-, 5-, and 10-year mortality and investigate their clinical significance. Input data included 3751 hip fracture patient records sourced from the Medical Information Mart for Intensive Care IV database, which provided records from in-hospital database systems at the Beth Israel Deaconess Medical Center. The 1-year mortality rate for all patients studied was 21% and for those aged 80+ was 29%. We assessed 10 different machine learning classification models, finding LightGBM to have the strongest 1-year mortality prediction performance, with accuracy of 81%, AUC of 0.79, sensitivity of 0.34, and specificity of 0.98 on the test set. The strongest-weighted features of the 1-year model included age, glucose, red blood cell distribution width, mean corpuscular hemoglobin concentration, white blood cells, urea nitrogen, prothrombin time, platelet count, calcium levels, and partial thromboplastin time. Most of these were also in the top 10 features of the LightGBM 5- and 10-year prediction models trained. Testing for these high-ranking biomarkers in new hip fracture patients can aid clinicians in assessing the likelihood of poor outcomes for hip fracture patients, and additional research can use these biomarkers to develop a mortality risk score.

A Comparison of CT-Based Pancreatic Segmentation Deep Learning Models.

RATIONALE AND OBJECTIVES: Pancreas segmentation accuracy at CT is critical for the identification of pancreatic pathologies and is essential for the development of imaging biomarkers. Our objective was to benchmark the performance of five high-performing pancreas segmentation models across multiple metrics stratified by scan and patient/pancreatic characteristics that may affect segmentation performance. MATERIALS AND METHODS: In this retrospective study, PubMed and ArXiv searches were conducted to identify pancreas segmentation models which were then evaluated on a set of annotated imaging datasets. Results (Dice score, Hausdorff distance [HD], average surface distance [ASD]) were stratified by contrast status and quartiles of peri-pancreatic attenuation (5 mm region around pancreas). Multivariate regression was performed to identify imaging characteristics and biomarkers (n = 9) that were significantly associated with Dice score. RESULTS: Five pancreas segmentation models were identified: Abdomen Atlas [AAUNet, AASwin, trained on 8448 scans], TotalSegmentator [TS, 1204 scans], nnUNetv1 [MSD-nnUNet, 282 scans], and a U-Net based model for predicting diabetes [DM-UNet, 427 scans]. These were evaluated on 352 CT scans (30 females, 25 males, 297 sex unknown; age 58 ± 7 years [ ± 1 SD], 327 age unknown) from 2000-2023. Overall, TS, AAUNet, and AASwin were the best performers, Dice= 80 ± 11%, 79 ± 16%, and 77 ± 18%, respectively (pairwise Sidak test not-significantly different). AASwin and MSD-nnUNet performed worse (for all metrics) on non-contrast scans (vs contrast, P < .001). The worst performer was DM-UNet (Dice=67 ± 16%). All algorithms except TS showed lower Dice scores with increasing peri-pancreatic attenuation (P < .01). Multivariate regression showed non-contrast scans, (P < .001; MSD-nnUNet), smaller pancreatic length (P = .005, MSD-nnUNet), and height (P = .003, DM-UNet) were associated with lower Dice scores. CONCLUSION: The convolutional neural network-based models trained on a diverse set of scans performed best (TS, AAUnet, and AASwin). TS performed equivalently to AAUnet and AASwin with only 13% of the training set size (8488 vs 1204 scans). Though trained on the same dataset, a transformer network (AASwin) had poorer performance on non-contrast scans whereas its convolutional network counterpart (AAUNet) did not. This study highlights how aggregate assessment metrics of pancreatic segmentation algorithms seen in other literature are not enough to capture differential performance across common patient and scanning characteristics in clinical populations.

Machine Learning Predicts Cerebral Vasospasm in Subarachnoid Hemorrhage Patients.

Background: Cerebral vasospasm (CV) is a feared complication occurring in 20-40% of patients following subarachnoid hemorrhage (SAH) and is known to contribute to delayed cerebral ischemia. It is standard practice to admit SAH patients to intensive care for an extended period of vigilant, resource-intensive, clinical monitoring. We used machine learning to predict CV requiring verapamil (CVRV) in the largest and only multi-center study to date. Methods: SAH patients admitted to UCLA from 2013-2022 and a validation cohort from VUMC from 2018-2023 were included. For each patient, 172 unique intensive care unit (ICU) variables were extracted through the primary endpoint, namely first verapamil administration or ICU downgrade. At each institution, a light gradient boosting machine (LightGBM) was trained using five- fold cross validation to predict the primary endpoint at various timepoints during hospital admission. Receiver-operator curves (ROC) and precision-recall (PR) curves were generated. Results: A total of 1,750 patients were included from UCLA, 125 receiving verapamil. LightGBM achieved an area under the ROC (AUC) of 0.88 an average of over one week in advance, and successfully ruled out 8% of non-verapamil patients with zero false negatives. Minimum leukocyte count, maximum platelet count, and maximum intracranial pressure were the variables with highest predictive accuracy. Our models predicted "no CVRV" vs "CVRV within three days" vs "CVRV after three days" with AUCs=0.88, 0.83, and 0.88, respectively. For external validation at VUMC, 1,654 patients were included, 75 receiving verapamil. Predictive models at VUMC performed very similarly to those at UCLA, averaging 0.01 AUC points lower. Conclusions: We present an accurate (AUC=0.88) and early (>1 week prior) predictor of CVRV using machine learning over two large cohorts of subarachnoid hemorrhage patients at separate institutions. This represents a significant step towards optimized clinical management and improved resource allocation in the intensive care setting of subarachnoid hemorrhage patients.

Machine learning predicts cerebral vasospasm in patients with subarachnoid haemorrhage.

BACKGROUND: Cerebral vasospasm (CV) is a feared complication which occurs after 20-40% of subarachnoid haemorrhage (SAH). It is standard practice to admit patients with SAH to intensive care for an extended period of resource-intensive monitoring. We used machine learning to predict CV requiring verapamil (CVRV) in the largest and only multi-center study to date. METHODS: Patients with SAH admitted to UCLA from 2013 to 2022 and a validation cohort from VUMC from 2018 to 2023 were included. For each patient, 172 unique intensive care unit (ICU) variables were extracted through the primary endpoint, namely first verapamil administration or no verapamil. At each institution, a light gradient boosting machine (LightGBM) was trained using five-fold cross validation to predict the primary endpoint at various hospitalization timepoints. FINDINGS: A total of 1750 patients were included from UCLA, 125 receiving verapamil. LightGBM achieved an area under the ROC (AUC) of 0.88 > 1 week in advance and ruled out 8% of non-verapamil patients with zero false negatives. Our models predicted "no CVRV" vs "CVRV within three days" vs "CVRV after three days" with AUCs = 0.88, 0.83, and 0.88, respectively. From VUMC, 1654 patients were included, 75 receiving verapamil. VUMC predictions averaged within 0.01 AUC points of UCLA predictions. INTERPRETATION: We present an accurate and early predictor of CVRV using machine learning with multi-center validation. This represents a significant step towards optimized clinical management and resource allocation in patients with SAH. FUNDING: Robert E. Freundlich is supported by National Center for Advancing Translational Sciences federal grant UL1TR002243 and National Heart, Lung, and Blood Institute federal grant K23HL148640; these funders did not play any role in this study. The National Institutes of Health supports Vanderbilt University Medical Center which indirectly supported these research efforts. Neither this study nor any other authors personally received financial support for the research presented in this manuscript. No support from pharmaceutical companies was received.

MRISegmentator-Abdomen: A Fully Automated Multi-Organ and Structure Segmentation Tool for T1-weighted Abdominal MRI.

BACKGROUND: Segmentation of organs and structures in abdominal MRI is useful for many clinical applications, such as disease diagnosis and radiotherapy. Current approaches have focused on delineating a limited set of abdominal structures (13 types). To date, there is no publicly available abdominal MRI dataset with voxel-level annotations of multiple organs and structures. Consequently, a segmentation tool for multi-structure segmentation is also unavailable. METHODS: We curated a T1-weighted abdominal MRI dataset consisting of 195 patients who underwent imaging at National Institutes of Health (NIH) Clinical Center. The dataset comprises of axial pre-contrast T1, arterial, venous, and delayed phases for each patient, thereby amounting to a total of 780 series (69,248 2D slices). Each series contains voxel-level annotations of 62 abdominal organs and structures. A 3D nnUNet model, dubbed as MRISegmentator-Abdomen (MRISegmentator in short), was trained on this dataset, and evaluation was conducted on an internal test set and two large external datasets: AMOS22 and Duke Liver. The predicted segmentations were compared against the ground-truth using the Dice Similarity Coefficient (DSC) and Normalized Surface Distance (NSD). FINDINGS: MRISegmentator achieved an average DSC of 0.861$\pm$0.170 and a NSD of 0.924$\pm$0.163 in the internal test set. On the AMOS22 dataset, MRISegmentator attained an average DSC of 0.829$\pm$0.133 and a NSD of 0.908$\pm$0.067. For the Duke Liver dataset, an average DSC of 0.933$\pm$0.015 and a NSD of 0.929$\pm$0.021 was obtained. INTERPRETATION: The proposed MRISegmentator provides automatic, accurate, and robust segmentations of 62 organs and structures in T1-weighted abdominal MRI sequences. The tool has the potential to accelerate research on various clinical topics, such as abnormality detection, radiotherapy, disease classification among others.

From delay to diagnosis: Chronic invasive fungal rhinosinusitis presenting with facial and orbital complications.

Key Clinical Message: Early identification and management of chronic invasive fungal rhinosinusitis (CIFRS) is key to optimizing outcomes. A missed diagnosis can result in permanent vision loss, chronic facial pain, or death. We present a case of CIFRS and literature review. Abstract: This case report presents a 56-year-old female with CIFRS involving orbital and facial complications. The patient experienced delayed diagnosis despite multiple ED visits for sinusitis with progressive facial pain and ocular deficits not alleviated with antibiotics, emphasizing the importance of early identification and maintaining high clinical suspicion for CIFRS. Prompt recognition, initiation of antifungal therapy, and aggressive surgical debridement were crucial for preventing disease progression and improving the patient's quality of life.

Machine Learning and Artificial Intelligence: A Web-Based Implant Failure and Peri-implantitis Prediction Model for Clinicians.

PURPOSE: To develop a machine learning model that can predict dental implant failure and peri-implantitis as a tool for maximizing implant success. MATERIALS AND METHODS: This study used a supervised learning model to retrospectively analyze 398 unique patients receiving a total of 942 dental implants presenting at the Philadelphia Veterans Affairs Medical Center from 2006 to 2013. Logistic regression, random forest classifiers, support vector machines, and ensemble techniques were employed to analyze this dataset. RESULTS: The random forest model possessed the highest predictive performance on test sets, with receiver operating characteristic area under curves (ROC AUC) of 0.872 and 0.840 for dental implant failures and peri-implantitis, respectively. The five most important features correlating with implant failure were amount of local anesthetic, implant length, implant diameter, use of preoperative antibiotics, and frequency of hygiene visits. The five most important features correlating with peri-implantitis were implant length, implant diameter, use of preoperative antibiotics, frequency of hygiene visits, and presence of diabetes mellitus. CONCLUSION: This study demonstrated the ability of machine learning models to assess demographics, medical history, and surgical plans, as well as the influence of these factors on dental implant failure and peri-implantitis. This model may serve as a resource for clinicians in the treatment of dental implants. Int J Oral Maxillofac Implants 2023;38:576-582. doi: 10.11607/jomi.9852.

Impacts of the New York State COVID-19 disaster emergency orders on prescription dispensing for opioids and medication for opioid use disorder.

AIMS: The aim of this study is to examine whether the March 2020 New York State (NYS) SARS-CoV-2 emergency orders were associated with an initial surge in opioid dispensing and a longer-term reduction in access to medications for opioid use disorder (MOUD). DESIGN: Time-series analyses of the dispensing of non-MOUD opioid and MOUD prescriptions using IQVIA's longitudinal prescription claims database (n = 16 087 429) in NYS by week, from 1 January 2018 to 31 July 2020. IQVIA is a multi-national company that provides biopharmaceutical development and commercial outsourcing services. SETTING AND PARTICIPANTS: NYS Zone Improvement Plan (ZIP) codes (n = 1218) in which prescriptions were dispensed. MEASUREMENT: For each ZIP code, for each week, the following dispensing measures were calculated: total weekly morphine milligram equivalents/day (MME/day), total weekly MME/day dispensed via prescriptions for ≤ 7 days and the count of MOUD prescriptions dispensed. Differences in dispensing metrics, comparing each week in 2020 with corresponding weeks in 2019, were calculated for each ZIP code. RESULTS: During the study period, weekly MME/day per ZIP code of dispensed non-MOUD opioids steadily declined. Compared with the difference in dispensing between 2019 and 2020 during the first week in 2020, there was a significantly larger drop in dispensed weekly total MME/day beginning 21 March 2020, and lasting until the week of 17 April (P < 0.05 for each week). Mean weekly total MME/day dispensed from 21 March to 17 April 2020 was 17.07% lower [95% confidence interval (CI) = 13.97%, 20.17%] than in the 4 weeks before 21 March almost entirely due to a drop in MME/day dispensed for prescriptions of ≤ 7 days. There was not a discernable drop in MOUD dispensing associated with the period of the emergency orders. CONCLUSIONS: New York State emergency orders in March 2020 to reduce SARS-CoV-2 transmission and preserve hospital capacity appeared to be associated with a decline in dispensing of opioids not used as MOUD. Access to MOUD appeared to be unaffected by the orders, probably because of policy initiatives by the Substance Abuse and Mental Health Services Administration.

Conquering the Cobb Angle: A Deep Learning Algorithm for Automated, Hardware-Invariant Measurement of Cobb Angle on Radiographs in Patients with Scoliosis.

Scoliosis is a disease estimated to affect more than 8% of adults in the United States. It is diagnosed with use of radiography by means of manual measurement of the angle between maximally tilted vertebrae on a radiograph (ie, the Cobb angle). However, these measurements are time-consuming, limiting their use in scoliosis surgical planning and postoperative monitoring. In this retrospective study, a pipeline (using the SpineTK architecture) was developed that was trained, validated, and tested on 1310 anterior-posterior images obtained with a low-dose stereoradiographic scanning system and radiographs obtained in patients with suspected scoliosis to automatically measure Cobb angles. The images were obtained at six centers (2005-2020). The algorithm measured Cobb angles on hold-out internal (n = 460) and external (n = 161) test sets with less than 2° error (intraclass correlation coefficient, 0.96) compared with ground truth measurements by two experienced radiologists. Measurements, produced in less than 0.5 second, did not differ significantly (P = .05 cutoff) from ground truth measurements, regardless of the presence or absence of surgical hardware (P = .80), age (P = .58), sex (P = .83), body mass index (P = .63), scoliosis severity (P = .44), or image type (low-dose stereoradiographic image vs radiograph; P = .51) in the patient. These findings suggest that the algorithm is highly robust across different clinical characteristics. Given its automated, rapid, and accurate measurements, this network may be used for monitoring scoliosis progression in patients. Keywords: Cobb Angle, Convolutional Neural Network, Deep Learning Algorithms, Pediatrics, Machine Learning Algorithms, Scoliosis, Spine Supplemental material is available for this article. © RSNA, 2023.

A real-time COVID-19 surveillance dashboard to support epidemic response in Connecticut: lessons from an academic-health department partnership.

In response to the coronavirus disease-19 (COVID-19) pandemic, numerous institutions published COVID-19 dashboards for reporting epidemiological statistics at the county, state, or national level. However, statistics for smaller cities were often not reported, requiring these areas to develop their own data processing pipelines. For under-resourced departments of health, the development of these pipelines was challenging, leading them to rely on nonspecific and often delayed infection statistics during the pandemic. To avoid this issue, the Stamford, Connecticut Department of Health (SDH) contracted with the Columbia Mailman School of Public Health to develop an online dashboard that displays real-time case, death, test, vaccination, hospitalization, and forecast data for their city, allowing SDH to monitor trends for specific demographic and geographic groups. Insights from the dashboard allowed SDH to initiate timely and targeted testing/vaccination campaigns. The dashboard is widely used and highlights the benefit of public-academic partnerships in public health, especially during the COVID-19 pandemic.

Vertebral Deformity Measurements at MRI, CT, and Radiography Using Deep Learning.

PURPOSE: To construct and evaluate the efficacy of a deep learning system to rapidly and automatically locate six vertebral landmarks, which are used to measure vertebral body heights, and to output spine angle measurements (lumbar lordosis angles [LLAs]) across multiple modalities. MATERIALS AND METHODS: In this retrospective study, MR (n = 1123), CT (n = 137), and radiographic (n = 484) images were used from a wide variety of patient populations, ages, disease stages, bone densities, and interventions (n = 1744 total patients, 64 years ± 8, 76.8% women; images acquired 2005-2020). Trained annotators assessed images and generated data necessary for deformity analysis and for model development. A neural network model was then trained to output vertebral body landmarks for vertebral height measurement. The network was trained and validated on 898 MR, 110 CT, and 387 radiographic images and was then evaluated or tested on the remaining images for measuring deformities and LLAs. The Pearson correlation coefficient was used in reporting LLA measurements. RESULTS: On the holdout testing dataset (225 MR, 27 CT, and 97 radiographic images), the network was able to measure vertebral heights (mean height percentage of error ± 1 standard deviation: MR images, 1.5% ± 0.3; CT scans, 1.9% ± 0.2; radiographs, 1.7% ± 0.4) and produce other measures such as the LLA (mean absolute error: MR images, 2.90°; CT scans, 2.26°; radiographs, 3.60°) in less than 1.7 seconds across MR, CT, and radiographic imaging studies. CONCLUSION: The developed network was able to rapidly measure morphometric quantities in vertebral bodies and output LLAs across multiple modalities.Keywords: Computer Aided Diagnosis (CAD), MRI, CT, Spine, Demineralization-Bone, Feature Detection Supplemental material is available for this article. © RSNA, 2021.

A deep learning system for automated, multi-modality 2D segmentation of vertebral bodies and intervertebral discs.

PURPOSE: Fractures in vertebral bodies are among the most common complications of osteoporosis and other bone diseases. However, studies that aim to predict future fractures and assess general spine health must manually delineate vertebral bodies and intervertebral discs in imaging studies for further radiomic analysis. This study aims to develop a deep learning system that can automatically and rapidly segment (delineate) vertebrae and discs in MR, CT, and X-ray imaging studies. RESULTS: We constructed a neural network to output 2D segmentations for MR, CT, and X-ray imaging studies. We trained the network on 4490 MR, 550 CT, and 1935 X-ray imaging studies (post-data augmentation) spanning a wide variety of patient populations, bone disease statuses, and ages from 2005 to 2020. Evaluated using 5-fold cross validation, the network was able to produce median Dice scores > 0.95 across all modalities for vertebral bodies and intervertebral discs (on the most central slice for MR/CT and on image for X-ray). Furthermore, radiomic features (skewness, kurtosis, mean of positive value pixels, and entropy) calculated from predicted segmentation masks were highly accurate (r ≥ 0.96 across all radiomic features when compared to ground truth). Mean time to produce outputs was <1.7 s across all modalities. CONCLUSIONS: Our network was able to rapidly produce segmentations for vertebral bodies and intervertebral discs for MR, CT, and X-ray imaging studies. Furthermore, radiomic quantities derived from these segmentations were highly accurate. Since this network produced outputs rapidly for these modalities which are commonly used, it can be put to immediate use for radiomic and clinical imaging studies assessing spine health.

The effects of perturbed cerebral blood flow and cerebrovascular reactivity on structural MRI and behavioral readouts in mild traumatic brain injury.

This study investigated the effects of perturbed cerebral blood flow (CBF) and cerebrovascular reactivity (CR) on relaxation time constant (T2), apparent diffusion coefficient (ADC), fractional anisotropy (FA), and behavioral scores at 1 and 3 hours, 2, 7, and 14 days after traumatic brain injury (TBI) in rats. Open-skull TBI was induced over the left primary forelimb somatosensory cortex (N=8 and 3 sham). We found the abnormal areas of CBF and CR on days 0 and 2 were larger than those of the T2, ADC, and FA abnormalities. In the impact core, CBF was reduced on day 0, increased to 2.5 times of normal on day 2, and returned toward normal by day 14, whereas in the tissue surrounding the impact, hypoperfusion was observed on days 0 and 2. CR in the impact core was negative, most severe on day 2 but gradually returned toward normal. T2, ADC, and FA abnormalities in the impact core were detected on day 0, peaked on day 2, and pseudonormalized by day 14. Lesion volumes peaked on day 2 and were temporally correlated with forelimb asymmetry and foot-fault scores. This study quantified the effects of perturbed CBF and CR on structural magnetic resonance imaging and behavioral readouts.