Evaluating the Use of ChatGPT to Accurately Simplify Patient-centered Information about Breast Cancer Prevention and Screening.

Purpose To evaluate the use of ChatGPT as a tool to simplify answers to common questions about breast cancer prevention and screening. Materials and Methods In this retrospective, exploratory study, ChatGPT was requested to simplify responses to 25 questions about breast cancer to a sixth-grade reading level in March and August 2023. Simplified responses were evaluated for clinical appropriateness. All original and simplified responses were assessed for reading ease on the Flesch Reading Ease Index and for readability on five scales: Flesch-Kincaid Grade Level, Gunning Fog Index, Coleman-Liau Index, Automated Readability Index, and the Simple Measure of Gobbledygook (ie, SMOG) Index. Mean reading ease, readability, and word count were compared between original and simplified responses using paired t tests. McNemar test was used to compare the proportion of responses with adequate reading ease (score of 60 or greater) and readability (sixth-grade level). Results ChatGPT improved mean reading ease (original responses, 46 vs simplified responses, 70; P < .001) and readability (original, grade 13 vs simplified, grade 8.9; P < .001) and decreased word count (original, 193 vs simplified, 173; P < .001). Ninety-two percent (23 of 25) of simplified responses were considered clinically appropriate. All 25 (100%) simplified responses met criteria for adequate reading ease, compared with only two of 25 original responses (P < .001). Two of the 25 simplified responses (8%) met criteria for adequate readability. Conclusion ChatGPT simplified answers to common breast cancer screening and prevention questions by improving the readability by four grade levels, though the potential to produce incorrect information necessitates physician oversight when using this tool. Keywords: Mammography, Screening, Informatics, Breast, Education, Health Policy and Practice, Oncology, Technology Assessment Supplemental material is available for this article. © RSNA, 2023.

Association of Lipoprotein (a) with peri-coronary inflammation in persons with and without HIV infection.

BACKGROUND: Persons with HIV (PWH) have an increased risk of developing cardiovascular disease (CVD) compared to persons without HIV (PWoH). Lipoprotein a (Lp(a)) is a known atherosclerotic risk factor in PWoH, but there are no studies investigating Lp(a) and peri-coronary inflammation. OBJECTIVE: To investigate whether Lp(a) is associated with peri-coronary inflammation as assessed by the fat attenuation index (FAI) and activated monocytes and T lymphocytes in PWH and PWoH. METHODS: We measured plasma levels of Lp(a) at study entry in 58 PWH and 21 PWoH without CVD and who had FAI measurements. Associations of Lp(a) with FAI values of the right coronary artery (RCA) and left anterior descending artery (LAD) were evaluated using multivariable regression models adjusted for potential confounders. Correlations between Lp(a) levels and systemic inflammatory markers and immune cell subsets were examined. RESULTS: Lp(a) was associated with greater peri-coronary inflammation among PWH compared to PWoH (ß=1.73, P=0.019) in the RCA, in adjusted models. Significant correlations were observed with certain inflammatory markers (TNFR-I, b=0.295, P<0.001; TNFR-II, b=0.270, P=0.002; hs-CRP, b=0.195, P=0.028). Significant correlations were found between Lp(a) levels and several markers of monocyte activation: CD16 -CD163+ (b= -0.199, P=0.024), and CD16 -DR+ MFI (b= -0.179, P=0.042) and T cell subset CD38+CD4+ TEMRA (b= 0.177, P= 0.044). CONCLUSIONS: Lp(a) was associated with greater peri-coronary inflammation in the RCA in PWH compared to PWoH, as well as with select systemic inflammatory markers and specific subsets of immune cells in peripheral circulation.

Risk factors and treatment interventions associated with incomplete thrombus resolution and pulmonary hypertension after pulmonary embolism.

BACKGROUND: Residual pulmonary vascular occlusion (RPVO) affects one half of patients after a pulmonary embolism (PE). The relationship between the risk factors and therapeutic interventions for the development of RPVO and chronic thromboembolic pulmonary hypertension is unknown. METHODS: This retrospective review included PE patients within a 26-month period who had baseline and follow-up imaging studies (ie, computed tomography [CT], ventilation/perfusion scans, transthoracic echocardiography) available. We collected the incidence of RPVO, percentage of pulmonary artery occlusion (%PAO), baseline CT %PAO, most recent CT %PAO, and difference between the baseline and most recent %PAO on CT (Δ%PAO). RESULTS: A total of 354 patients had imaging reports available; 197 with CT and 315 with transthoracic echocardiography. The median follow-up time was 144 days (interquartile range [IQR], 102-186 days). RPVO was present in 38.9% of the 354 patients. The median Δ%PAO was -10.0% (IQR, -32% to -1.2%). Fewer patients with a provoked PE developed RPVO (P ≤ .01), and the initial troponin level was lower in patients who developed RPVO (P = .03). The initial thrombus was larger in the patients who received advanced intervention vs anticoagulation (baseline CT %PAO: median, 61.2%; [IQR, 27.5%-75.0%] vs median, 12.5% [IQR, 2.5%-40.0%]; P ≤ .0001). Catheter-directed thrombolysis (CDT; median Δ%PAO, -47.5%; IQR, -63.7% to -8.7%) and surgical pulmonary embolectomy (SPE; median Δ%PAO, -42.5; IQR, -68.1% to -18.7%) had the largest thrombus reduction compared with anticoagulation (P = .01). Of the 354 patients, 76 developed pulmonary hypertension; however, only 14 received pulmonary hypertension medications and 12 underwent pulmonary thromboendarterectomy. Cancer (odds ratio [OR], 1.7) and planned prolonged anticoagulation (>1 year; OR, 2.20) increased the risk of RPVO. In contrast, the risk was lower for men (OR, 0.61), patients with recent surgery (OR, 0.33), and patients treated with SPE (OR, 0.42). A larger Δ%PAO was found in men (coefficient, -8.94), patients with a lower body mass index (coefficient, -0.66), patients treated with CDT (coefficient, -18.12), and patients treated with SPE (coefficient, -21.69). A lower Δ%PAO was found in African-American patients (coefficient, 7.31). CONCLUSIONS: The use of CDT and SPE showed long-term benefit in thrombus reduction.

Metastatic Pulmonary Calcification: Single-Center Review of Typical and Atypical Imaging Features.

PURPOSE: The purpose of this study is to bring attention to an atypical form of metastatic pulmonary calcification, which is conventionally described as a metabolic process with upper lobe predominance in patients with a specific clinical history, which has not been reported as a distinct entity. METHODS: Patients with metastatic pulmonary calcification (MPC) were first identified with mPower keyword search, including MPC or metastatic calcifications on computed tomography chest radiological reports. Patients were then filtered on likelihood of MPC based off imaging reports. Images were then reviewed by three senior radiologists for pertinent characteristics such as location of MPC, degree of calcifications and pleural effusions. Based on the predominant location of MPC, cases were labeled as either typical or atypical. Clinical and imaging characteristics relevant to MPC were noted and compared across typical and atypical cases. RESULTS: In our study, we describe 25 patients with MPC, 13 defined as typical MPC and 12 with atypical MPC. Through consensus of senior radiologists, MPC was deemed to be mild (52%), moderate (44%), or severe (4%). Twenty-three patients (92%) had underlying renal disease including 21 requiring dialysis at the time of diagnosis. Outside of age at diagnosis, there was no significant clinical difference between the two groups. Evaluation of imaging characteristics (average HU attenuation, 267; range, 186-295), pattern and distribution of calcification, and clinical history strongly supported a diagnosis of atypical MPC. CONCLUSION: This study presents several cases of lower lobe subpleural MPC associated with pleural effusions, which has not been reported as a distinct entity, despite comprising a significant portion of MPC cases at our institution.

Lung-Reporting and Data System 2.0: Impact of the Updated Approach to Juxtapleural Nodules During Lung Cancer Screening Using the National Lung Cancer Screening Trial Data Set.

PURPOSE: To determine the frequency of malignancy of nonperifissural juxtapleural nodules (JPNs) measuring 6 to < 10 mm in a subset of low-dose chest computed tomographies from the National Lung Cancer Screening Trial and the rate of down-classification of such nodules in Lung-Reporting and Data System (RADS) 2.0 compared with Lung-RADS 1.1. MATERIALS AND METHODS: A secondary analysis of a subset of the National Lung Screening Trial was performed. An exemption was granted by the Institutional Review Board. The dominant noncalcified nodule measuring 6 to <10 mm was identified on all available prevalence computed tomographies. Nodules were categorized as pleural or nonpleural. Benign or malignant morphology was recorded. Initial and updated categories based on Lung-RADS 1.1 and Lung-RADS 2.0 were assigned, respectively. The impact of the down-classification of JPN was assessed. Both classification schemes were compared using the McNemar test (P < 0.01). RESULTS: A total of 2813 patients (62 ± 5 y, 1717 men) with 4408 noncalcified nodules were studied. One thousand seventy-three dominant nodules measuring 6 to <10 mm were identified. Three hundred forty-eight (32.4%) were JPN. The updated scheme allowed down-classification of 310 JPN from categories 3 (n = 198) and 4A (n = 112) to category 2. We, therefore, estimate a 4.8% rate of down-classification to category 2 in the entire National Lung Screening Trial screening group. Two/348 (0.57%) JPN were malignant, both nonbenign in morphology. The false-positive rate decreased in the updated classification (P < 0.01). CONCLUSION: This study demonstrates the low malignant potential of benign morphology JPN measuring 6 mm to <10 mm. The Lung-RADS 2.0 approach to JPN is estimated to reduce short-term follow-ups and false-positive results.

Association of Soluble Markers of Inflammation With Peri-coronary Artery Inflammation in People With and Without HIV Infection and Without Cardiovascular Disease.

Background: Inflammation is linked to elevated cardiovascular disease (CVD) risk in people with HIV (PWH) on antiretroviral therapy (ART). Fat attenuation index (FAI) is a measure of peri-coronary inflammation that independently predicts CVD risk in HIV-uninfected persons. Whether FAI is associated with soluble inflammatory markers is unknown. Methods: Plasma levels of inflammatory markers were measured in 58 PWH and 16 controls without current symptoms or prior known CVD who underwent coronary computed tomography angiography and had FAI measurements. A cross-sectional analysis was performed, and associations of markers with FAI values of the right coronary artery (RCA) and left anterior descending artery (LAD) were assessed using multivariable regression models adjusted for the potential confounders age, sex, race, low-density lipoprotein cholesterol, body mass index, and use of lipid-lowering medication. Results: Several inflammatory markers had significant associations with RCA or LAD FAI in adjusted models, including sCD14, sCD163, TNFR-I, and TNFR-II, CCL5, CX3CL1, IP-10. Conclusions: The associations between indices of systemic and peri-coronary inflammation are novel and suggest that these systemic markers and FAI together are promising noninvasive biomarkers that can be applied to assess asymptomatic CVD in people with and without HIV; they also may be useful tools to evaluate effects of anti-inflammatory interventions.

Applying Artificial Intelligence to Pediatric Chest Imaging: Reliability of Leveraging Adult-Based Artificial Intelligence Models.

OBJECTIVE: The scarcity of artificial intelligence (AI) applications designed for use in pediatric patients can cause a significant health disparity in this vulnerable population. We investigated the performance of an adult-trained algorithm in detecting pneumonia in a pediatric population to explore the viability of leveraging adult-trained algorithms to accelerate pediatric AI research. METHODS: We analyzed a publicly available pediatric chest x-ray dataset using an AI algorithm from TorchXRayVision. A 60% threshold was used to make binary predictions for pneumonia presence. Predictions were compared with dataset labels. Performance measures including true-positive rate, false-positive rate, true-negative rate, false-negative rate, sensitivity, specificity, positive predictive value (PPV), negative predictive value, accuracy, and F1-score were calculated for the complete dataset and bacterial and viral pneumonia subsets. RESULTS: Overall (n = 5,856), the algorithm identified 3,923 cases with pneumonia (67.00%) and 1,933 (33.00%) normal cases. In comparison with the actual image labels, there were 3,411 (58.25%) true-positive cases, 512 (8.74%) false-positive cases, 1,071 (18.29%) true-negative cases, and 862 (14.72%) false-negative cases resulting in 79.83% sensitivity, 67.66% specificity, 86.95% PPV, 55.41% negative predictive value, and 76.54% accuracy and an F1-score of 0.83. Although the performance remained consistent in the bacterial pneumonia group, there was a significant decrease in PPV (69.9%) and F1-score (0.74) in the viral pneumonia group. CONCLUSION: An adult-trained model adequately detected pneumonia in pediatric patients aged 1 to 5 years. Though models trained exclusively on pediatric images performed better, leveraging adult-based algorithms and datasets can expedite pediatric AI research.

Use of ChatGPT, GPT-4, and Bard to Improve Readability of ChatGPT's Answers to Common Questions About Lung Cancer and Lung Cancer Screening.

ChatGPT's responses to questions about lung cancer and LCS, although deemed clinically appropriate by cardiothoracic radiologists, were difficult to read. Simplified responses from three LLMs (ChatGPT, GPT-4, and Bard) had improved reading ease and readability (in terms of U.S. grade levels). However, some simplified responses were no longer clinically appropriate.

Systematic Review of Radiology Residency Artificial Intelligence Curricula: Preparing Future Radiologists for the Artificial Intelligence Era.

OBJECTIVE: Although educating radiology trainees about artificial intelligence (AI) has become increasingly emphasized, the types of AI educational curricula are not well understood. We performed a systematic review of original studies describing curricula used to teach AI concepts and practical applications for radiology residents and fellows. MATERIALS AND METHODS: We performed a PubMed search for original studies published as of July 22, 2022, describing AI curricula geared toward radiology residents or fellows. Studies meeting inclusion criteria were evaluated for curricula design, implementation details, and outcomes. Descriptive statistics were used to summarize these curricula. RESULTS: Five studies were included describing an AI curriculum, all geared toward radiology residents. All five curricula were led by radiologists, mostly by individual academic radiology departments (4; 80%) with one led by the ACR Resident and Fellow Section. Curricula design included didactic sessions (5; 100%), assigned readings (4; 80%), hands-on learning (3; 60%), and journal clubs (3; 60%); only one had individualized learning plans. All four studies that evaluated the impact of the curricula on participants' knowledge or attitudes showed positive effects. DISCUSSION: Amid increasing recognition of the importance of AI education for radiologists-in-training, several AI curricula for radiology residents have been implemented. Although curricula designs varied and it is unclear if one type is superior, they have had a positive impact on residents' knowledge and attitudes toward AI. As AI becomes increasingly adopted in radiology, these curricula serve as models for other departments and programs to develop AI educational initiatives to prepare the next generation of radiologists for the AI era.

Comprehensive Computed Tomography Imaging of Vessel-specific and Lesion-specific Myocardial Ischemia.

Coronary computed tomographic angiography (CCTA) has emerged as a fast and robust tool with high sensitivity and excellent negative predictive value for the evaluation of coronary artery disease, but is unable to estimate the hemodynamic significance of a lesion. Advances in computed tomography (CT)-based diagnostic techniques, for example, CT-derived fractional flow reserve and CT perfusion, have helped transform CCTA primarily from an anatomic assessment tool to a technique that is able to provide both anatomic and functional information for a stenosis. With the results of the ISCHEMIA trial published in 2019, these advanced techniques can elevate CCTA into the role of a better gatekeeper for decision-making and can help guide referral for invasive management. In this article, we review the principles, limitations, diagnostic performance, and clinical utility of these 2 functional CT-based techniques in the evaluation of vessel-specific and lesion-specific ischemia.

A pilot study of deep learning-based CT volumetry for traumatic hemothorax.

PURPOSE: We employ nnU-Net, a state-of-the-art self-configuring deep learning-based semantic segmentation method for quantitative visualization of hemothorax (HTX) in trauma patients, and assess performance using a combination of overlap and volume-based metrics. The accuracy of hemothorax volumes for predicting a composite of hemorrhage-related outcomes - massive transfusion (MT) and in-hospital mortality (IHM) not related to traumatic brain injury - is assessed and compared to subjective expert consensus grading by an experienced chest and emergency radiologist. MATERIALS AND METHODS: The study included manually labeled admission chest CTs from 77 consecutive adult patients with non-negligible (≥ 50 mL) traumatic HTX between 2016 and 2018 from one trauma center. DL results of ensembled nnU-Net were determined from fivefold cross-validation and compared to individual 2D, 3D, and cascaded 3D nnU-Net results using the Dice similarity coefficient (DSC) and volume similarity index. Pearson's r, intraclass correlation coefficient (ICC), and mean bias were also determined for the best performing model. Manual and automated hemothorax volumes and subjective hemothorax volume grades were analyzed as predictors of MT and IHM using AUC comparison. Volume cut-offs yielding sensitivity or specificity ≥ 90% were determined from ROC analysis. RESULTS: Ensembled nnU-Net achieved a mean DSC of 0.75 (SD: ± 0.12), and mean volume similarity of 0.91 (SD: ± 0.10), Pearson r of 0.93, and ICC of 0.92. Mean overmeasurement bias was only 1.7 mL despite a range of manual HTX volumes from 35 to 1503 mL (median: 178 mL). AUC of automated volumes for the composite outcome was 0.74 (95%CI: 0.58-0.91), compared to 0.76 (95%CI: 0.58-0.93) for manual volumes, and 0.76 (95%CI: 0.62-0.90) for consensus expert grading (p = 0.93). Automated volume cut-offs of 77 mL and 334 mL predicted the outcome with 93% sensitivity and 90% specificity respectively. CONCLUSION: Automated HTX volumetry had high method validity, yielded interpretable visual results, and had similar performance for the hemorrhage-related outcomes assessed compared to manual volumes and expert consensus grading. The results suggest promising avenues for automated HTX volumetry in research and clinical care.

Early pulmonary complications related to cancer treatment in children.

In this review, we summarize early pulmonary complications related to cancer therapy in children and highlight characteristic findings on imaging that should be familiar to a radiologist reviewing imaging from pediatric cancer patients.

The Dean Effect: An Aortic Arch Flow Artifact Mimicking Dissection.

The unique hemodynamics of the aortic arch create conditions for potential formation of a flow-related artifact that mimics disease on CT angiographic images. The hemodynamic basis for this artifact can be explained by fluid mechanics incorporating a mathematical principle known as the Dean number. Therefore, in this review, the artifact is referred to as the Dean effect. It is important for radiologists and other clinicians to recognize this artifact when encountered. It is also helpful for the interpreting radiologist to have a basic understanding of the relevant hemodynamic principles. This review provides an example of the artifact, reviews the basic underlying hemodynamics, and presents methods of how to prevent this artifact and distinguish it from pathologic mimics in clinical practice. Keywords: CT Angiography, Vascular, Thorax, Aorta, Artifacts, Blood, Dissection, Hemodynamics/Flow Dynamics © RSNA, 2022.

Prone Chest Radiographs: Distinguishing Features and Identification of Support Devices.

PURPOSE: Prone position is known to improve acute lung injury, and chest radiographs are often necessary to monitor disease and confirm support device placement. However, there is a paucity of literature regarding radiographs obtained in this position. We evaluated prone radiographs for distinguishing features and ability to identify support devices. METHODS: Pairs of prone and supine radiographs obtained during the COVID-19 pandemic were assessed retrospectively. IRB approval and waiver of informed consent were obtained. Radiographs were assessed for imaging adequacy, distinguishing features, and support device identification (endotracheal tube, enteric tube, or central line). Radiographs were reviewed by ≥ 2 cardiothoracic radiologists. RESULTS: Radiographs from 81 patients (63yo ± 13, 30% women) were reviewed. Prone and supine radiographs were comparable for imaging the lung bases (81% vs. 90%, p = 0.35) and apices (93% vs. 94%, p = 1); prone radiographs more frequently had significant rotation (36% vs. 19%, p = 0.021). To identify prone technique, scapula tip located beyond the rib border was 89% sensitive (95%CI 80-95%) and 85% specific (76-92%), and a fundal stomach bubble was 44% sensitive (33-56%) and 90% specific (81-96%). For women, displaced breast shadow was 46% sensitive (26-67%) and 92% specific (73-99%). Prone and supine radiographs each identified > 99% of support devices. Prone exams trended toward increased rate of malpositioned device (12% vs. 6%, p = 0.07). CONCLUSION: Scapula position reliably distinguishes prone from supine position; fundal stomach bubble or displaced breast shadow is specific for prone position. Prone radiographs reliably identify line and tube position, which is particularly important as prone patients appear at increased risk for malpositioned devices.

Assessment of coronary inflammation in antiretroviral treated people with HIV infection and active HIV/hepatitis C virus co-infection.

OBJECTIVE: People with HIV (PWH) and co-infected with hepatitis C virus (PWH + HCV) have increased risk of cardiovascular disease (CVD). Peri-coronary inflammation, measured by fat attenuation index (FAI) on coronary computed tomography angiography (CCTA), independently predicts cardiovascular risk in the general population but has not been studied in the PWH + HCV population. We tested whether peri-coronary inflammation is increased in PWH or PWH + HCV, and whether inflammation changes over time. DESIGN: Cross-sectional analysis to determine FAI differences among groups. Longitudinal analysis in PWH to assess changes in inflammation over time. METHODS: Age-matched and sex-matched seropositive groups (PWH and PWH + HCV) virologically suppressed on antiretroviral therapy, HCV viremic, and without prior CVD and matched controls underwent CCTA. Peri-coronary FAI was measured around the proximal right coronary artery (RCA) and left anterior descending artery (LAD). Follow-up CCTA was performed in 22 PWH after 20.6-27.4 months. RESULTS: A total of 101 participants (48 women) were studied (60 PWH, 19 PWH + HCV and 22 controls). In adjusted analyses, peri-coronary FAI did not differ between seropositive groups and controls. Low attenuation coronary plaque was significantly less common in seropositive groups compared with controls (LAD, P = 0.035; and RCA, P = 0.017, respectively). Peri-coronary FAI values significantly progressed between baseline and follow-up in PWH (RCA: P = 0.001, LAD: P = <0.001). CONCLUSION: PWH and PWH + HCV without history of CVD do not have significantly worse peri-coronary inflammation, assessed by FAI, compared with matched controls. However, peri-coronary inflammation in mono-infected PWH significantly increased over approximately 22 months. FAI measures may be an important imaging biomarker for tracking asymptomatic CVD progression in PWH.

Cardiac MRI Findings of Myocarditis After COVID-19 mRNA Vaccination in Adolescents.

BACKGROUND. A possible association has been reported between COVID-19 messenger RNA (mRNA) vaccination and myocarditis. OBJECTIVE. The purpose of our study was to describe cardiac MRI findings in patients with myocarditis after COVID-19 mRNA vaccination. METHODS. This retrospective study included patients without known prior SARS-CoV-2 infection who underwent cardiac MRI between May 14, 2021, and June 14, 2021, for suspected myocarditis within 2 weeks of COVID-19 mRNA vaccination. Information regarding clinical presentation, hospital course, and events after hospital discharge were recorded. A cardiothoracic imaging fellow and cardiothoracic radiologist reviewed cardiac MRI examinations in consensus. Data were summarized descriptively. RESULTS. Of 52 patients without known prior SARS-CoV-2 infection who underwent cardiac MRI during the study period, five underwent MRI for suspected myocarditis after recent COVID-19 mRNA vaccination. All five patients were male patients ranging in age from 16 to 19 years (mean, 17.2 ± 1.0 [SD] years) who presented within 4 days of receiving the second dose of a COVID-19 mRNA vaccine. Troponin levels were elevated in all patients (mean peak troponin I value, 6.82 ± 4.13 ng/mL). Alternate possible causes of myocarditis were deemed clinically unlikely on the basis of medical history, physical examination findings, myocarditis viral panel, and toxicology screening. Cardiac MRI findings were consistent with myocarditis in all five patients on the basis of the Lake Louise criteria, including early gadolinium enhancement and late gadolinium enhancement (LGE) in all patients and corresponding myocardial edema in four patients. All five patients had a favorable hospital course and were discharged from the hospital in stable condition with improved or resolved symptoms after hospitalization (mean length of hospital stay, 4.8 days). Two patients underwent repeat cardiac MRI that showed persistent, although decreased, LGE. Three patients reported mild intermittent self-resolving chest pain after hospital discharge, and two patients had no recurrent symptoms after discharge. CONCLUSION. In this small case series, all patients with myocarditis after COVID-19 vaccination were male adolescents and had a favorable initial clinical course. All patients showed cardiac MRI findings typical of myocarditis from other causes. LGE persisted in two patients who underwent repeat MRI. These observations do not establish causality. CLINICAL IMPACT. Radiologists should be aware of a possible association of COVID-19 mRNA vaccination and myocarditis and recognize the role of cardiac MRI in the assessment of suspected myocarditis after COVID-19 vaccination.