To assess ChatGPT's ability as a resource for educating patients on various aspects of cardiac imaging, including diagnosis, imaging modalities, indications, interpretation of radiology reports, and management. 30 questions were posed to ChatGPT-3.5 and ChatGPT-4 three times in three separate chat sessions. Responses were scored as correct, incorrect, or clinically misleading categories by three observers-two board certified cardiologists and one board certified radiologist with cardiac imaging subspecialization. Consistency of responses across the three sessions was also evaluated. Final categorization was based on majority vote between at least two of the three observers. ChatGPT-3.5 answered seventeen of twenty eight questions correctly (61 %) by majority vote. Twenty one of twenty eight questions were answered correctly (75 %) by ChatGPT-4 by majority vote. Majority vote for correctness was not achieved for two questions. Twenty six of thirty questions were answered consistently by ChatGPT-3.5 (87 %). Twenty nine of thirty questions were answered consistently by ChatGPT-4 (97 %). ChatGPT-3.5 had both consistent and correct responses to seventeen of twenty eight questions (61 %). ChatGPT-4 had both consistent and correct responses to twenty of twenty eight questions (71 %). ChatGPT-4 had overall better performance than ChatGTP-3.5 when answering cardiac imaging questions with regard to correctness and consistency of responses. While both ChatGPT-3.5 and ChatGPT-4 answers over half of cardiac imaging questions correctly, inaccurate, clinically misleading and inconsistent responses suggest the need for further refinement before its application for educating patients about cardiac imaging.
1 February 2025 in Circulation: Cardiovascular Imaging
BACKGROUND: Risk stratification is highly desirable in patients with uncomplicated Stanford type B aortic dissection but inadequately supported by evidence. We sought to validate externally a published prediction model for late adverse events (LAEs), consisting of 1 clinical (connective tissue disease) and 4 imaging variables: maximum aortic diameter, false lumen circumferential angle, false lumen outflow, and number of identifiable intercostal arteries. METHODS: We assembled a retrospective multicenter cohort (ROADMAP [Registry of Aortic Diseases to Model Adverse Events and Progression]) of 401 patients with uncomplicated Stanford type B aortic dissection presenting to 1 of 8 aortic centers between 2001 and 2013, followed until 2020. LAEs were defined as fatal or nonfatal aortic rupture, new refractory hypertension or pain, organ or limb ischemia, aortic aneurysm formation (≥6 cm), or rapid growth (≥1 cm per year). We applied the original model parameters to the validation cohort and examined the effect on risk categorization using LAE end points. RESULTS: One hundred and seventy-six patients (44%) with incomplete imaging or clinical data were excluded. Of 225 patients in the final cohort, 90 (40%) developed LAEs, predominantly driven by aneurysm formation. Baseline maximum aortic diameter was significantly larger in patients with (42.6 [95% CI, 39.1–45.8] mm) compared with patients without LAEs (39.9 [95% CI, 36.3–44.2] mm; P =0.001). A multivariable Cox regression model indicated that only maximum diameter was associated with LAEs (hazard ratio, 1.07 [95% CI, 1.03–1.11] per mm; P 0.05). Applying the original prediction model to the validation cohort resulted in a poor 5-year sensitivity (38%) and specificity (69%). CONCLUSIONS: A clinical and imaging-based prediction model performed poorly in the ROADMAP cohort. Maximum aortic diameter remains the strongest predictor of LAEs in uncomplicated Stanford type B aortic dissection.
1 December 2022 in Radiology: Cardiothoracic Imaging
This imaging essay reviews the clinical and CT angiography imaging findings in patients with limited aortic intimal tears to improve awareness and understanding of this rare aortic lesion. Limited aortic intimal tear is an uncommon lesion of the dissection spectrum. The lesion has several imaging features that are not well known, including asymmetric aortic contour abnormalities, fil...
Background: Lung herniation has been described in case reports or series. There are scare data in the form of original research studies to systematically evaluate this condition. Objective: Our aim was to evaluate lung hernias with a focus on their natural history. Methods: This is a retrospective study at our institution of patients who were found to have lung herniation on imaging between September 2010 and November 2022. Electronic medical record review was performed to extract clinical information regarding patients. Computed tomographic imaging was used to evaluate hernia size and size progression over time with a median follow-up time of 3.8 years. Results: Thirty-eight patients were eligible for analysis. Majority of patients were overweight or obese (31/38), smokers (31/38), had prior thoracic surgery (30/38), and were asymptomatic (33/38). Twenty of 38 patients had stability in hernia size, 12 of 38 patients had hernia size progression, and 6 of 38 patients showed hernia size regression. Younger age was found to be predictive of hernia size progression with age of 60 years being the cut-off for its prediction. Conclusion: Lung hernias typically either remain stable in size or show size progression. Younger age (60 years cut-off) was found to be predictive of size progression. This is the largest systematic investigation at a medical institution to the best of our knowledge of lung hernias which used computed tomographic imaging to follow up lung hernias. Further information could be given to patients with this condition and to clinicians for potential management guidance.
1 December 2025 in Radiology: Cardiothoracic Imaging
Allocation of key type B aortic dissection features to aortic zones, as defined by the 2020 Society for Vascular Surgery/Society of Thoracic Surgeons reporting standards, was poorly reproducible among experienced cardiovascular radiologists and surgeons.