Adjusting for the effect of IV contrast on automated CT body composition measures during the portal venous phase.

OBJECTIVE: Fully-automated CT-based algorithms for quantifying numerous biomarkers have been validated for unenhanced abdominal scans. There is great interest in optimizing the documentation and reporting of biophysical measures present on all CT scans for the purposes of opportunistic screening and risk profiling. The purpose of this study was to determine and adjust the effect of intravenous (IV) contrast on these automated body composition measures at routine portal venous phase post-contrast imaging. METHODS: Final study cohort consisted of 1,612 older adults (mean age, 68.0 years; 594 women) all imaged utilizing a uniform CT urothelial protocol consisting of pre-contrast, portal venous, and delayed excretory phases. Fully-automated CT-based algorithms for quantifying numerous biomarkers, including muscle and fat area and density, bone mineral density, and solid organ volume were applied to pre-contrast and portal venous phases. The effect of IV contrast upon these body composition measures was analyzed. Regression analyses, including square of the Pearson correlation coefficient (r2), were performed for each comparison. RESULTS: We found that simple, linear relationships can be derived to determine non-contrast equivalent values from the post-contrast CT biomeasures. Excellent positive linear correlation (r2 = 0.91-0.99) between pre- and post-contrast values was observed for all automated soft tissue measures, whereas moderate positive linear correlation was observed for bone attenuation (r2 = 0.58-0.76). In general, the area- and volume-based measurement require less adjustment than attenuation-based measures, as expected. CONCLUSION: Fully-automated quantitative CT-biomarker measures at portal venous phase abdominal CT can be adjusted to a non-contrast equivalent using simple, linear relationships.

The effects of prehospital TXA on mortality and neurologic outcomes in patients with traumatic intracranial hemorrhage: a subgroup analysis from the prehospital TXA for TBI trial.

BACKGROUND: In the Prehospital Tranexamic Acid (TXA) for TBI Trial, TXA administered within two hours of injury in the out-of-hospital setting did not reduce mortality in all patients with moderate/severe traumatic brain injury (TBI). We examined the association between TXA dosing arms, neurologic outcome, and mortality in patients with intracranial hemorrhage (ICH) on computed tomography (CT). METHODS: This was a secondary analysis of the Prehospital Tranexamic Acid for TBI Trial (ClinicalTrials.gov [NCT01990768]) that randomized adults with moderate/severe TBI (Glasgow Coma Scale<13) and systolic blood pressure > =90 mmHg within two hours of injury to a 2-gram out-of-hospital TXA bolus followed by an in-hospital saline infusion, a 1-gram out-of-hospital TXA bolus/1-gram in-hospital TXA infusion, or an out-of-hospital saline bolus/in-hospital saline infusion (placebo). This analysis included the subgroup with ICH on initial CT. Primary outcomes included 28-day mortality, 6-month Glasgow Outcome Scale-Extended (GOSE) < = 4, and 6-month Disability Rating Scale (DRS). Outcomes were modeled using linear regression with robust standard errors. RESULTS: The primary trial included 966 patients. Among 541 participants with ICH, 28-day mortality was lower in the 2-gram TXA bolus group (17%) compared to the other two groups (1-gram bolus/1-gram infusion 26%, placebo 27%). The estimated adjusted difference between the 2-gram bolus and placebo groups was -8·5 percentage points (95% CI, -15.9 to -1.0) and between the 2-gram bolus and 1-gram bolus/1-gram infusion groups was -10.2 percentage points (95% CI, -17.6 to -2.9). DRS at 6 months was lower in the 2-gram TXA bolus group than the 1-gram bolus/1-gram infusion (estimated difference -2.1 [95% CI, -4.2 to -0.02]) and placebo groups (-2.2 [95% CI, -4.3, -0.2]). Six-month GOSE did not differ among groups. CONCLUSIONS: A 2-gram out-of-hospital TXA bolus in patients with moderate/severe TBI and ICH resulted in lower 28-day mortality and lower 6-month DRS than placebo and standard TXA dosing. LEVEL OF EVIDENCE: Therapeutic/Care Management, Level II.

Fully automated CT imaging biomarkers for opportunistic prediction of future hip fractures.

OBJECTIVE: Assess automated CT imaging biomarkers in patients who went on to hip fracture, compared with controls. METHODS: In this retrospective case-control study, 6926 total patients underwent initial abdominal CT over a 20-year interval at one institution. A total of 1308 patients (mean age at initial CT, 70.5 ± 12.0 years; 64.4% female) went on to hip fracture (mean time to fracture, 5.2 years); 5618 were controls (mean age 70.3 ± 12.0 years; 61.2% female; mean follow-up interval 7.6 years). Validated fully automated quantitative CT algorithms for trabecular bone attenuation (at L1), skeletal muscle attenuation (at L3), and subcutaneous adipose tissue area (SAT) (at L3) were applied to all scans. Hazard ratios (HRs) comparing highest to lowest risk quartiles and receiver operating characteristic (ROC) curve analysis including area under the curve (AUC) were derived. RESULTS: Hip fracture HRs (95% CI) were 3.18 (2.69-3.76) for low trabecular bone HU, 1.50 (1.28-1.75) for low muscle HU, and 2.18 (1.86-2.56) for low SAT. 10-year ROC AUC values for predicting hip fracture were 0.702, 0.603, and 0.603 for these CT-based biomarkers, respectively. Multivariate combinations of these biomarkers further improved predictive value; the 10-year ROC AUC combining bone/muscle/SAT was 0.733, while combining muscle/SAT was 0.686. CONCLUSION: Opportunistic use of automated CT bone, muscle, and fat measures can identify patients at higher risk for future hip fracture, regardless of the indication for CT imaging. ADVANCES IN KNOWLEDGE: CT data can be leveraged opportunistically for further patient evaluation, with early intervention as needed. These novel AI tools analyse CT data to determine a patient's future hip fracture risk.

Personalized Impression Generation for PET Reports Using Large Language Models.

Large language models (LLMs) have shown promise in accelerating radiology reporting by summarizing clinical findings into impressions. However, automatic impression generation for whole-body PET reports presents unique challenges and has received little attention. Our study aimed to evaluate whether LLMs can create clinically useful impressions for PET reporting. To this end, we fine-tuned twelve open-source language models on a corpus of 37,370 retrospective PET reports collected from our institution. All models were trained using the teacher-forcing algorithm, with the report findings and patient information as input and the original clinical impressions as reference. An extra input token encoded the reading physician's identity, allowing models to learn physician-specific reporting styles. To compare the performances of different models, we computed various automatic evaluation metrics and benchmarked them against physician preferences, ultimately selecting PEGASUS as the top LLM. To evaluate its clinical utility, three nuclear medicine physicians assessed the PEGASUS-generated impressions and original clinical impressions across 6 quality dimensions (3-point scales) and an overall utility score (5-point scale). Each physician reviewed 12 of their own reports and 12 reports from other physicians. When physicians assessed LLM impressions generated in their own style, 89% were considered clinically acceptable, with a mean utility score of 4.08/5. On average, physicians rated these personalized impressions as comparable in overall utility to the impressions dictated by other physicians (4.03, P = 0.41). In summary, our study demonstrated that personalized impressions generated by PEGASUS were clinically useful in most cases, highlighting its potential to expedite PET reporting by automatically drafting impressions.

AI-generated CT body composition biomarkers associated with increased mortality risk in socioeconomically disadvantaged individuals.

PURPOSE: To evaluate the relationship between socioeconomic disadvantage using national area deprivation index (ADI) and CT-based body composition measures derived from fully automated artificial intelligence (AI) tools to identify body composition measures associated with increased risk for all-cause mortality and adverse cardiovascular events. METHODS: Fully automated AI body composition tools quantifying abdominal aortic calcium, abdominal fat (visceral [VAT], visceral-to-subcutaneous ratio [VSR]), and muscle attenuation (muscle HU) were applied to non-contrast CT examinations in adults undergoing screening CT colonography (CTC). Patients were partitioned into 5 socioeconomic groups based on the national ADI rank at the census block group level. Pearson correlation analysis was performed to determine the association between national ADI and body composition measures. One-way analysis of variance was used to compare means across groups. Odds ratios (ORs) were generated using high-risk, high specificity (90% specificity) body composition thresholds with the most disadvantaged groups being compared to the least disadvantaged group (ADI < 20). RESULTS: 7785 asymptomatic adults (mean age, 57 years; 4361:3424 F:M) underwent screening CTC from April 2004-December 2016. ADI rank data were available in 7644 patients. Median ADI was 31 (IQR 22-43). Aortic calcium, VAT, and VSR had positive correlation with ADI and muscle attenuation had a negative correlation with ADI (all p < .001). Compared with the least disadvantaged group, mean differences for the most disadvantaged group (ADI > 80) were: Aortic calcium (Agatston) = 567, VAT = 27 cm2, VSR = 0.1, and muscle HU = -6 HU (all p < .05). Compared with the least disadvantaged group, the most disadvantaged group had significantly higher odds of having high-risk body composition measures: Aortic calcium OR = 3.8, VAT OR = 2.5, VSR OR = 2.0, and muscle HU OR = 3.1(all p < .001). CONCLUSION: Fully automated CT body composition tools show that socioeconomic disadvantage is associated with high-risk body composition measures and can be used to identify individuals at increased risk for all-cause mortality and adverse cardiovascular events.

Automated detection of incidental abdominal aortic aneurysms on computed tomography.

PURPOSE: To detect and assess abdominal aortic aneurysms (AAAs) on CT in a large asymptomatic adult patient population using fully-automated deep learning software. MATERIALS AND METHODS: The abdominal aorta was segmented using a fully-automated deep learning model trained on 66 manually-segmented abdominal CT scans from two datasets. The axial diameters of the segmented aorta were extracted to detect the presence of AAAs-maximum axial aortic diameter greater than 3 cm were labeled as AAA positive. The trained system was then externally-validated on CT colonography scans of 9172 asymptomatic outpatients (mean age, 57 years) referred for colorectal cancer screening. Using a previously-validated automated calcified atherosclerotic plaque detector, we correlated abdominal aortic Agatston and volume scores with the presence of AAA. RESULTS: The deep learning software detected AAA on the external validation dataset with a sensitivity, specificity, and AUC of 96%, (95% CI 89%, 100%), 96% (96%, 97%), and 99% (98%, 99%) respectively. The Agatston and volume scores of reported AAA-positive cases were statistically significantly greater than those of reported AAA-negative cases (p < 0.0001). Using plaque alone as a AAA detector, at a threshold Agatston score of 2871, the sensitivity and specificity were 84% (73%, 94%) and 87% (86%, 87%), respectively. CONCLUSION: Fully-automated detection and assessment of AAA on CT is feasible and accurate. There was a strong statistical association between the presence of AAA and the quantity of abdominal aortic calcified atherosclerotic plaque.

Comparing fully automated AI body composition measures derived from thin and thick slice CT image data.

PURPOSE: To compare fully automated artificial intelligence body composition measures derived from thin (1.25 mm) and thick (5 mm) slice abdominal CT data. METHODS: In this retrospective study, fully automated CT-based body composition algorithms for quantifying bone attenuation, muscle attenuation, muscle area, liver attenuation, liver volume, spleen volume, visceral-to-subcutaneous fat ratio (VSR) and aortic calcium were applied to both thin (1.25 × 0.625 mm) and thick (5 × 3 mm) abdominal CT series from two patient cohorts: unenhanced scans in asymptomatic adults undergoing colorectal cancer screening, and post-contrast scans in patients with colorectal cancer. Body composition measures derived from thin and thick slice data were compared, including correlation coefficients and Bland-Altman analysis. RESULTS: A total of 9882 CT scans (mean age, 57.0 years; 4527 women, 5355 men) were evaluated, including 8947 non-contrast and 935 contrast-enhanced CT exams. Very strong positive correlation was observed for all soft tissue measures: muscle attenuation (r2 = 0.97), muscle area (r2 = 0.98), liver attenuation (r2 = 0.99), liver volume (r2 = 0.98) and spleen volume (r2 = 0.99), VSR (r2 = 0.98), and aortic calcium (r2 = 0.92); (p < 0.001 for all). Moderate positive correlation was observed for bone attenuation (r2 = 0.35). Bland-Altman analysis showed strong agreement for muscle attenuation, muscle area, liver attenuation, liver volume and spleen volume. Mean percentage differences amongst body composition measures were less than 5% for VSR (4.6%), muscle area (- 0.5%), liver attenuation (0.4%) and liver volume (2.7%) and less than 10% for muscle attenuation (- 5.5%) and spleen volume (5.1%). For aortic calcium, thick slice overestimated for Agatston scores between 0 and 100 and > 400 burden in 3.1% and 0.3% relative to thin slice, respectively, but underestimated scores between 100 and 400. CONCLUSION: Automated body composition measures derived from thin and thick abdominal CT data are strongly correlated and show agreement, particularly for soft tissue applications, making it feasible to use either series for these CT-based body composition algorithms.

CT image-based biomarkers acquired by AI-based algorithms for the opportunistic prediction of falls.

Objective: Evaluate whether biomarkers measured by automated artificial intelligence (AI)-based algorithms are suggestive of future fall risk. Methods: In this retrospective age- and sex-matched case-control study, 9029 total patients underwent initial abdominal CT for a variety of indications over a 20-year interval at one institution. 3535 case patients (mean age at initial CT, 66.5 ± 9.6 years; 63.4% female) who went on to fall (mean interval to fall, 6.5 years) and 5494 controls (mean age at initial CT, 66.7 ± 9.8 years; 63.4% females; mean follow-up interval, 6.6 years) were included. Falls were identified by electronic health record review. Validated and fully automated quantitative CT algorithms for skeletal muscle, adipose tissue, and trabecular bone attenuation at the level of L1 were applied to all scans. Uni- and multivariate assessment included hazard ratios (HRs) and area under the receiver operating characteristic (AUROC) curve. Results: Fall HRs (with 95% CI) for low muscle Hounsfield unit, high total adipose area, and low bone Hounsfield unit were 1.82 (1.65-2.00), 1.31 (1.19-1.44) and 1.91 (1.74-2.11), respectively, and the 10-year AUROC values for predicting falls were 0.619, 0.556, and 0.639, respectively. Combining all these CT biomarkers further improved the predictive value, including 10-year AUROC of 0.657. Conclusion: Automated abdominal CT-based opportunistic measures of muscle, fat, and bone offer a novel approach to risk stratification for future falls, potentially by identifying patients with osteosarcopenic obesity. Advances in knowledge: There are few well-established clinical tools to predict falls. We use novel AI-based body composition algorithms to leverage incidental CT data to help determine a patient's future fall risk.

Automatic Personalized Impression Generation for PET Reports Using Large Language Models.

Purpose: To determine if fine-tuned large language models (LLMs) can generate accurate, personalized impressions for whole-body PET reports. Materials and Methods: Twelve language models were trained on a corpus of PET reports using the teacher-forcing algorithm, with the report findings as input and the clinical impressions as reference. An extra input token encodes the reading physician's identity, allowing models to learn physician-specific reporting styles. Our corpus comprised 37,370 retrospective PET reports collected from our institution between 2010 and 2022. To identify the best LLM, 30 evaluation metrics were benchmarked against quality scores from two nuclear medicine (NM) physicians, with the most aligned metrics selecting the model for expert evaluation. In a subset of data, model-generated impressions and original clinical impressions were assessed by three NM physicians according to 6 quality dimensions (3-point scale) and an overall utility score (5-point scale). Each physician reviewed 12 of their own reports and 12 reports from other physicians. Bootstrap resampling was used for statistical analysis. Results: Of all evaluation metrics, domain-adapted BARTScore and PEGASUSScore showed the highest Spearman's ρ correlations (ρ=0.568 and 0.563) with physician preferences. Based on these metrics, the fine-tuned PEGASUS model was selected as the top LLM. When physicians reviewed PEGASUS-generated impressions in their own style, 89% were considered clinically acceptable, with a mean utility score of 4.08 out of 5. Physicians rated these personalized impressions as comparable in overall utility to the impressions dictated by other physicians (4.03, P=0.41). Conclusion: Personalized impressions generated by PEGASUS were clinically useful, highlighting its potential to expedite PET reporting.

Training certified detectives to track down the intrinsic shortcuts in COVID-19 chest x-ray data sets.

Deep learning faces a significant challenge wherein the trained models often underperform when used with external test data sets. This issue has been attributed to spurious correlations between irrelevant features in the input data and corresponding labels. This study uses the classification of COVID-19 from chest x-ray radiographs as an example to demonstrate that the image contrast and sharpness, which are characteristics of a chest radiograph dependent on data acquisition systems and imaging parameters, can be intrinsic shortcuts that impair the model's generalizability. The study proposes training certified shortcut detective models that meet a set of qualification criteria which can then identify these intrinsic shortcuts in a curated data set.

Artificial intelligence tool detection of intravenous contrast enhancement using spleen attenuation.

PURPOSE: To assess the ability of an automated AI tool to detect intravenous contrast material (IVCM) in abdominal CT examinations using spleen attenuation. METHODS: A previously validated automated AI tool measuring the attenuation of the spleen was deployed on a sample of 32,994 adult (age ≥ 18) patients (mean age, 61.9 ± 14.7 years; 13,869 men, 19,125 women) undergoing 65,449 supine position CT examinations (41,020 with and 24,429 without IVCM by DICOM header) from January 1, 2000 to December 31, 2021. After exclusions, receiver operating characteristic (ROC) curve analysis was performed to determine the optimal threshold for binary classification of IVCM status (non-contrast vs IVCM enhanced), which was then applied to the sample. Discordant examinations (i.e., IVCM status determined by AI tool did not match DICOM header) were manually reviewed to establish ground truth. Repeat ROC curve and contingency table analysis were performed to assess AI tool performance. RESULTS: ROC analysis of the initial study sample of 61,783 CT examinations yielded AUC of 0.970 with Youden index suggesting an optimal spleen attenuation threshold of 65 Hounsfield units (HU). Manual review of 2094 discordant CT examinations revealed discordance due to DICOM header error in 1278 (61.0%) and AI tool misclassification in 410 (19.6%), with 406 (9.4%) meeting exclusion criteria. Analysis of 61,377 CT examinations in the final study sample yielded AUC of 0.999 with accuracy of 99.3% at the 65 HU threshold. Error rate for DICOM header information was 2.1% (1278/61,377) versus 0.7% (410/61,377) for the AI tool. CONCLUSION: The automated spleen attenuation AI tool was highly accurate for detection of IVCM at a threshold of 65 HU.

Opportunistic detection of type 2 diabetes using deep learning from frontal chest radiographs.

Deep learning (DL) models can harness electronic health records (EHRs) to predict diseases and extract radiologic findings for diagnosis. With ambulatory chest radiographs (CXRs) frequently ordered, we investigated detecting type 2 diabetes (T2D) by combining radiographic and EHR data using a DL model. Our model, developed from 271,065 CXRs and 160,244 patients, was tested on a prospective dataset of 9,943 CXRs. Here we show the model effectively detected T2D with a ROC AUC of 0.84 and a 16% prevalence. The algorithm flagged 1,381 cases (14%) as suspicious for T2D. External validation at a distinct institution yielded a ROC AUC of 0.77, with 5% of patients subsequently diagnosed with T2D. Explainable AI techniques revealed correlations between specific adiposity measures and high predictivity, suggesting CXRs' potential for enhanced T2D screening.

Automated Deep Learning Artificial Intelligence Tool for Spleen Segmentation on CT: Defining Volume-Based Thresholds for Splenomegaly.

BACKGROUND. Splenomegaly historically has been assessed on imaging by use of potentially inaccurate linear measurements. Prior work tested a deep learning artificial intelligence (AI) tool that automatically segments the spleen to determine splenic volume. OBJECTIVE. The purpose of this study is to apply the deep learning AI tool in a large screening population to establish volume-based splenomegaly thresholds. METHODS. This retrospective study included a primary (screening) sample of 8901 patients (4235 men, 4666 women; mean age, 56 ± 10 [SD] years) who underwent CT colonoscopy (n = 7736) or renal donor CT (n = 1165) from April 2004 to January 2017 and a secondary sample of 104 patients (62 men, 42 women; mean age, 56 ± 8 years) with end-stage liver disease who underwent contrast-enhanced CT performed as part of evaluation for potential liver transplant from January 2011 to May 2013. The automated deep learning AI tool was used for spleen segmentation, to determine splenic volumes. Two radiologists independently reviewed a subset of segmentations. Weight-based volume thresholds for splenomegaly were derived using regression analysis. Performance of linear measurements was assessed. Frequency of splenomegaly in the secondary sample was determined using weight-based volumetric thresholds. RESULTS. In the primary sample, both observers confirmed splenectomy in 20 patients with an automated splenic volume of 0 mL; confirmed incomplete splenic coverage in 28 patients with a tool output error; and confirmed adequate segmentation in 21 patients with low volume (< 50 mL), 49 patients with high volume (> 600 mL), and 200 additional randomly selected patients. In 8853 patients included in analysis of splenic volumes (i.e., excluding a value of 0 mL or error values), the mean automated splenic volume was 216 ± 100 [SD] mL. The weight-based volumetric threshold (expressed in milliliters) for splenomegaly was calculated as (3.01 × weight [expressed as kilograms]) + 127; for weight greater than 125 kg, the splenomegaly threshold was constant (503 mL). Sensitivity and specificity for volume-defined splenomegaly were 13% and 100%, respectively, at a true craniocaudal length of 13 cm, and 78% and 88% for a maximum 3D length of 13 cm. In the secondary sample, both observers identified segmentation failure in one patient. The mean automated splenic volume in the 103 remaining patients was 796 ± 457 mL; 84% (87/103) of patients met the weight-based volume-defined splenomegaly threshold. CONCLUSION. We derived a weight-based volumetric threshold for splenomegaly using an automated AI-based tool. CLINICAL IMPACT. The AI tool could facilitate large-scale opportunistic screening for splenomegaly.

Knowledge and confidence gains after a COVID-19 vaccine continuing education program developed for nurse practitioners.

ABSTRACT: The COVID-19 pandemic has been marked by rapid innovation in vaccine development. Given that nurse practitioners (NPs) are often involved in vaccine counseling and administration, the American Association of Nurse Practitioners developed a continuing education (CE) series that covered COVID-19 vaccine development, recommendations, administration, and solutions for overcoming hesitancy. In 2020 and 2021, three separate live webinars were delivered; each webinar was updated with the latest vaccine recommendations and was then archived in an enduring format for up to 4 months. The goal of this study was to assess changes in preactivity and postactivity knowledge and confidence and to qualitatively report other learner outcomes. Across the three webinars, 3,580 unique learners who self-reported seeing patients eligible for COVID-19 vaccination completed at least one activity. Knowledge and competence improved from the preactivity to postactivity survey in all webinars, with the overall rates of correct answers increasing by 30% after webinar 1, 37% after webinar 2, and 28% after webinar 3 (all p < .001). Furthermore, mean confidence in learner's ability to address vaccine hesitancy improved across all three webinars (range, 31-32%; all p < .001). The majority of learners indicated that they planned to incorporate lessons from the activity into their clinical practice (range, 85-87%). In postactivity surveys, vaccine hesitancy was identified as an ongoing barrier by up to 33% of learners. In conclusion, this CE activity improved learner knowledge, competence, and confidence related to COVID-19 vaccination and underscores the importance of up-to-date CE targeted to NPs.

Training certified detectives to track down the intrinsic shortcuts in COVID-19 chest x-ray data sets.

Deep learning faces a significant challenge wherein the trained models often underperform when used with external test data sets. This issue has been attributed to spurious correlations between irrelevant features in the input data and corresponding labels. This study uses the classification of COVID-19 from chest x-ray radiographs as an example to demonstrate that the image contrast and sharpness, which are characteristics of a chest radiograph dependent on data acquisition systems and imaging parameters, can be intrinsic shortcuts that impair the model's generalizability. The study proposes training certified shortcut detective models that meet a set of qualification criteria which can then identify these intrinsic shortcuts in a curated data set.

Opportunistic Screening: Radiology Scientific Expert Panel.

Radiologic tests often contain rich imaging data not relevant to the clinical indication. Opportunistic screening refers to the practice of systematically leveraging these incidental imaging findings. Although opportunistic screening can apply to imaging modalities such as conventional radiography, US, and MRI, most attention to date has focused on body CT by using artificial intelligence (AI)-assisted methods. Body CT represents an ideal high-volume modality whereby a quantitative assessment of tissue composition (eg, bone, muscle, fat, and vascular calcium) can provide valuable risk stratification and help detect unsuspected presymptomatic disease. The emergence of "explainable" AI algorithms that fully automate these measurements could eventually lead to their routine clinical use. Potential barriers to widespread implementation of opportunistic CT screening include the need for buy-in from radiologists, referring providers, and patients. Standardization of acquiring and reporting measures is needed, in addition to expanded normative data according to age, sex, and race and ethnicity. Regulatory and reimbursement hurdles are not insurmountable but pose substantial challenges to commercialization and clinical use. Through demonstration of improved population health outcomes and cost-effectiveness, these opportunistic CT-based measures should be attractive to both payers and health care systems as value-based reimbursement models mature. If highly successful, opportunistic screening could eventually justify a practice of standalone "intended" CT screening.

Technical Adequacy of Fully Automated Artificial Intelligence Body Composition Tools: Assessment in a Heterogeneous Sample of External CT Examinations.

BACKGROUND. Clinically usable artificial intelligence (AI) tools analyzing imaging studies should be robust to expected variations in study parameters. OBJECTIVE. The purposes of this study were to assess the technical adequacy of a set of automated AI abdominal CT body composition tools in a heterogeneous sample of external CT examinations performed outside of the authors' hospital system and to explore possible causes of tool failure. METHODS. This retrospective study included 8949 patients (4256 men, 4693 women; mean age, 55.5 ± 15.9 years) who underwent 11,699 abdominal CT examinations performed at 777 unique external institutions with 83 unique scanner models from six manufacturers with images subsequently transferred to the local PACS for clinical purposes. Three independent automated AI tools were deployed to assess body composition (bone attenuation, amount and attenuation of muscle, amount of visceral and sub-cutaneous fat). One axial series per examination was evaluated. Technical adequacy was defined as tool output values within empirically derived reference ranges. Failures (i.e., tool output outside of reference range) were reviewed to identify possible causes. RESULTS. All three tools were technically adequate in 11,431 of 11,699 (97.7%) examinations. At least one tool failed in 268 (2.3%) of the examinations. Individual adequacy rates were 97.8% for the bone tool, 99.1% for the muscle tool, and 98.9% for the fat tool. A single type of image processing error (anisometry error, due to incorrect DICOM header voxel dimension information) accounted for 81 of 92 (88.0%) examinations in which all three tools failed, and all three tools failed whenever this error occurred. Anisometry error was the most common specific cause of failure of all tools (bone, 31.6%; muscle, 81.0%; fat, 62.8%). A total of 79 of 81 (97.5%) anisometry errors occurred on scanners from a single manufacturer; 80 of 81 (98.8%) occurred on the same scanner model. No cause of failure was identified for 59.4% of failures of the bone tool, 16.0% of failures of the muscle tool, or 34.9% of failures of the fat tool. CONCLUSION. The automated AI body composition tools had high technical adequacy rates in a heterogeneous sample of external CT examinations, supporting the generalizability of the tools and their potential for broad use. CLINICAL IMPACT. Certain causes of AI tool failure related to technical factors may be largely preventable through use of proper acquisition and reconstruction protocols.

Factors Contributing to Computed Tomography Trauma Scan Times at a Tertiary Center: Improving Emergency Department Trauma Imaging Workflow Through Targeted Interventions.

PURPOSES: The aims of the study are to identify factors contributing to computed tomography (CT) trauma scan turnaround time variation and to evaluate the effects of an automated intervention on time metrics. METHODS: Throughput metrics were captured via picture archiving and communication system from January 1, 2018, to December 16, 2019, and included 17,709 CT trauma scans from our institution. Initial data showed that imaging technologist variation played a significant role in trauma imaging turnaround time. In December 2019, we implemented a 2-pronged intervention: (1) educational intervention to techs and (2) modified trauma CT abdomen/pelvis to autogenerate and autosend reformats to picture archiving and communication system. A total of 13,169 trauma CT scans were evaluated from the postintervention period taking place from January 2020 to March 2021. Throughput metrics such as last image to first report interval and emergency department length of stay were captured and compared with performing technologist, time of day, and weekday versus weekend scans. RESULTS: Substantial variability among trauma CT scans was observed. For CT trauma abdomen/pelvis, the interval from last image to initial report decreased from 26.4 to 24.0 minutes ( P = 0.001) while the interval between first and last image time decreased from 11.4 to 4.2 minutes ( P < 0.001). Emergency department length of stay also decreased from 3.9 to 3.7 hours ( P < 0.0001) in the postintervention period. Variation among imaging technologist was statistically significant and became less significant after intervention ( P = 0.09, P = 0.54). CONCLUSIONS: Factors such as imaging technologist variability, time of day, and day of the week of trauma scans played a significant role in CT trauma turnaround time variability. Automation interventions can help with efficiency in image turnaround time.

Fully automated CT-based adiposity assessment: comparison of the L1 and L3 vertebral levels for opportunistic prediction.

PURPOSE: The purpose of this study is to compare fully automated CT-based measures of adipose tissue at the L1 level versus the standard L3 level for predicting mortality, which would allow for use at both chest (L1) and abdominal (L3) CT. METHODS: This retrospective study of 9066 asymptomatic adults (mean age, 57.1 ± 7.8 [SD] years; 4020 men, 5046 women) undergoing unenhanced low-dose abdominal CT for colorectal cancer screening. A previously validated artificial intelligence (AI) tool was used to assess cross-sectional visceral and subcutaneous adipose tissue areas (SAT and VAT), as well as their ratio (VSR) at the L1 and L3 levels. Post-CT survival prediction was compared using area under the ROC curve (ROC AUC) and hazard ratios (HRs). RESULTS: Median clinical follow-up interval after CT was 8.8 years (interquartile range, 5.2-11.6 years), during which 5.9% died (532/9066). No significant difference (p > 0.05) for mortality was observed between L1 and L3 VAT and SAT at 10-year ROC AUC. However, L3 measures were significantly better for VSR at 10-year AUC (p < 0.001). HRs comparing worst-to-best quartiles for mortality at L1 vs. L3 were 2.12 (95% CI, 1.65-2.72) and 2.22 (1.74-2.83) for VAT; 1.20 (0.95-1.52) and 1.16 (0.92-1.46) for SAT; and 2.26 (1.7-2.93) and 3.05 (2.32-4.01) for VSR. In women, the corresponding HRs for VSR were 2.58 (1.80-3.69) (L1) and 4.49 (2.98-6.78) (L3). CONCLUSION: Automated CT-based measures of visceral fat (VAT and VSR) at L1 are predictive of survival, although overall measures of adiposity at L1 level are somewhat inferior to the standard L3-level measures. Utilizing predictive L1-level fat measures could expand opportunistic screening to chest CT imaging.