RESUMEN
PURPOSE: Diagnosing myocarditis relies on multimodal data including magnetic resonance imaging (MRI), clinical symptoms, and blood values. The correct interpretation and integration of MRI findings requires radiological expertise and knowledge. We aimed to investigate the performance of Generative Pre-trained Transformer 4 (GPT-4), a large language model, for report-based medical decision-making in the context of cardiac MRI for suspected myocarditis. METHODS: This retrospective study includes MRI reports from 396 patients with suspected myocarditis and eight centers, respectively. MRI reports and patient data including blood values, age, and further clinical information were provided to GPT-4 and to radiologists with 1 (Resident 1), 2 (Resident 2), and 4 years (Resident 3) of experience in cardiovascular MRI and knowledge of the 2018 Lake Louise Criteria. The final impression of the report regarding the radiological assessment of whether myocarditis is present or not was not provided. The performance of GPT-4 and of the human readers were compared to a consensus reading (two board-certified radiologists with 8 and 10 years of experience in cardiovascular MRI). Sensitivity, specificity, and accuracy were calculated. RESULTS: GPT-4 yielded an accuracy of 83%, sensitivity of 90%, and specificity of 78%, which was comparable to the physician with 1 year of experience (R1: 86%, 90%, 84%, p=.14) and lower than that of more experienced physicians (R2: 89%, 86%, 91%, p=.007 and R3: 91%, 85%, 96%, p<.001). GPT-4 and human readers showed a higher diagnostic performance when results from T1- and T2-mapping sequences were part of the reports, for Residents 1 and Resident 3 with statistical significance (p=.004 and p=.02, respectively). CONCLUSION: GPT-4 yielded good accuracy for diagnosing myocarditis based on MRI reports in a large dataset from multiple centers and therefore holds the potential to serve as a diagnostic decision supporting tool in this capacity, particularly for less experienced physicians. Further studies are required to explore the full potential and elucidate educational aspects of the integration of large language models in medical decision-making.
RESUMEN
BACKGROUND: While computed tomography pulmonary angiography (CTPA) is an integral part of the work-up in patients with suspected pulmonary hypertension (PH), there is no established CTPA-derived prognostic marker. We aimed to assess whether quantitative readouts of lung vessel morphology correlate with established prognostic indicators in PH. METHODS: We applied a fully-automatic in-house developed algorithm for segmentation of arteries and veins to determine lung vessel morphology in patients with precapillary PH who underwent right heart catheterization and CTPA between May 2016 and May 2019. Primary endpoint of this retrospective study was the calculation of receiver operating characteristics for identifying low and high mortality risk according to the 3-strata risk assessment model presented in the current guidelines. RESULTS: We analyzed 73 patients, median age 65 years (interquartile range (IQR): 54-76), female/male ratio 35/38, median mean pulmonary arterial pressure 37 mm Hg (IQR: 30-46), and found significant correlations with important prognostic factors in pulmonary arterial hypertension. N-terminal pro-brain natriuretic peptide, cardiac index, mixed venous oxygen saturation, and 6-minute walking distance were correlated with the ratio of the number of arteries over veins with vessel diameters of 6-10 mm (Spearman correlation coefficients ρ = 0.64, p < 0.001; ρ = -0.60, p < 0.001; ρ = -0.47, p = 0.005; ρ = -0.45, p = 0.001, respectively). This ratio predicted a low- and high-risk score with an area under the curve of 0.73 (95% confidence interval (CI): 0.56-0.90) and 0.86 (95% CI: 0.74-0.97), respectively. CONCLUSIONS: The ratio of the number of arteries over veins with diameters between 6 and 10 mm is significantly correlated with prognostic markers in pulmonary hypertension and predicts low and high mortality risk.
Asunto(s)
Hipertensión Pulmonar , Humanos , Masculino , Femenino , Anciano , Hipertensión Pulmonar/diagnóstico por imagen , Pronóstico , Estudios Retrospectivos , Tomografía Computarizada por Rayos X/métodos , Arteria Pulmonar/diagnóstico por imagen , PulmónRESUMEN
Background: Strain analyses derived from cardiovascular magnetic resonance-feature tracking (CMR-FT) provide incremental prognostic benefit in patients sufferring from acute myocardial infarction (AMI). This study aims to evaluate and revalidate previously reported prognostic implications of comprehensive strain analyses in a large independent cohort of patients with ST-elevation myocardial infarction (STEMI). Methods: Overall, 566 STEMI patients enrolled in the CONDITIONING-LIPSIA trial including pre- and/or postconditioning treatment in addition to conventional percutaneous coronary intervention underwent CMR imaging in median 3 days after primary percutaneous coronary intervention. CMR-based left atrial (LA) reservoir (Es), conduit (Ee), and boosterpump (Ea) strain analyses, as well as left ventricular (LV) global longitudinal strain (GLS), circumferential strain (GCS), and radial strain (GRS) analyses were carried out. Previously identified cutoff values were revalidated for risk stratification. Major adverse cardiac events (MACE) comprising death, reinfarction, and new congestive heart failure were assessed within 12 months after the occurrence of the index event. Results: Both atrial and ventricular strain values were significantly reduced in patients with MACE (p < 0.01 for all). Predetermined LA and LV strain cutoffs enabled accurate risk assessment. All LA and LV strain values were associated with MACE on univariable regression modeling (p < 0.001 for all), with LA Es emerging as an independent predictor of MACE on multivariable regression modeling (HR 0.92, p = 0.033). Furthermore, LA Es provided an incremental prognostic value above LVEF (a c-index increase from 0.7 to 0.74, p = 0.03). Conclusion: External validation of CMR-FT-derived LA and LV strain evaluations confirmed the prognostic value of cardiac deformation assessment in STEMI patients. In the present study, LA strain parameters especially enabled further risk stratification and prognostic assessment over and above clinically established risk parameters. Clinical Trial Registration: ClinicalTrials.gov, identifier NCT02158468.
RESUMEN
BACKGROUND: Chronic thromboembolic pulmonary hypertension (CTEPH) is considered a complication of pulmonary embolism (PE). However, signs of CTEPH may exist in patients with a first symptomatic PE. RESEARCH QUESTION: Which radiologic findings on CT pulmonary angiography (CTPA) at the time of acute PE could indicate the presence of preexisting CTEPH? STUDY DESIGN AND METHODS: This study included unselected patients with acute PE who were prospectively followed up for 2 years with a structured visit schedule. Two expert radiologists independently assessed patients' baseline CTPAs for preexisting CTEPH; in case of disagreement, a decision was reached by a 2:1 majority with a third expert radiologist. In addition, the radiologists checked for predefined individual parameters suggesting chronic PE and pulmonary hypertension. RESULTS: Signs of chronic PE or CTEPH at baseline were identified in 46 of 303 included patients (15%). Intravascular webs, arterial narrowing or retraction, dilated bronchial arteries, and right ventricular hypertrophy were the main drivers of the assessment. Five (1.7%) patients were diagnosed with CTEPH during follow-up. All four patients diagnosed with CTEPH early (83-108 days following acute PE) were found in enriched subgroups based on the experts' overall assessment or fulfilling a minimum number of the predefined radiologic criteria at baseline. The specificity of preexisting CTEPH diagnosis and the level of radiologists' agreement improved as the number of required criteria increased. INTERPRETATION: Searching for predefined radiologic parameters suggesting preexisting CTEPH at the time of acute PE diagnosis may allow for targeted follow-up strategies and risk-adapted CTEPH screening, thus facilitating earlier CTEPH diagnosis.
Asunto(s)
Hipertensión Pulmonar , Embolia Pulmonar , Humanos , Hipertensión Pulmonar/complicaciones , Hipertensión Pulmonar/diagnóstico , Embolia Pulmonar/diagnóstico , Embolia Pulmonar/diagnóstico por imagen , Pulmón , Angiografía/efectos adversos , Angiografía por Tomografía Computarizada , Enfermedad CrónicaRESUMEN
BACKGROUND: Myocardial deformation analyses using cardiovascular magnetic resonance (CMR) feature tracking (CMR-FT) have incremental value in the assessment of cardiac function beyond volumetric analyses. Since guidelines do not recommend specific imaging parameters, we aimed to define optimal spatial and temporal resolutions for CMR cine images to enable reliable post-processing. METHODS: Intra- and inter-observer reproducibility was assessed in 12 healthy subjects and 9 heart failure (HF) patients. Cine images were acquired with different temporal (20, 30, 40 and 50 frames/cardiac cycle) and spatial resolutions (high in-plane 1.5 × 1.5 mm through-plane 5 mm, standard 1.8 × 1.8 x 8mm and low 3.0 × 3.0 x 10mm). CMR-FT comprised left ventricular (LV) global and segmental longitudinal/circumferential strain (GLS/GCS) and associated systolic strain rates (SR), and right ventricular (RV) GLS. RESULTS: Temporal but not spatial resolution did impact absolute strain and SR. Maximum absolute changes between lowest and highest temporal resolution were as follows: 1.8% and 0.3%/s for LV GLS and SR, 2.5% and 0.6%/s for GCS and SR as well as 1.4% for RV GLS. Changes of strain values occurred comparing 20 and 30 frames/cardiac cycle including LV and RV GLS and GCS (p < 0.001-0.046). In contrast, SR values (LV GLS/GCS SR) changed significantly comparing all successive temporal resolutions (p < 0.001-0.013). LV strain and SR reproducibility was not affected by either temporal or spatial resolution, whilst RV strain variability decreased with augmentation of temporal resolution. CONCLUSION: Temporal but not spatial resolution significantly affects strain and SR in CMR-FT deformation analyses. Strain analyses require lower temporal resolution and 30 frames/cardiac cycle offer consistent strain assessments, whilst SR measurements gain from further increases in temporal resolution.
Asunto(s)
Ventrículos Cardíacos , Imagen por Resonancia Cinemagnética , Ventrículos Cardíacos/diagnóstico por imagen , Humanos , Imagen por Resonancia Magnética , Valor Predictivo de las Pruebas , Reproducibilidad de los Resultados , Función Ventricular IzquierdaRESUMEN
BACKGROUND: Right heart catheterization using exercise stress is the reference standard for the diagnosis of heart failure with preserved ejection fraction (HFpEF) but carries the risk of the invasive procedure. We hypothesized that real-time cardiac magnetic resonance (RT-CMR) exercise imaging with pathophysiologic data at excellent temporal and spatial resolution may represent a contemporary noninvasive alternative for diagnosing HFpEF. METHODS: The HFpEF-Stress trial (CMR Exercise Stress Testing in HFpEF; URL: https://www.clinicaltrials.gov; Unique identifier: NCT03260621. URL: https://dzhk.de/; Unique identifier: DZHK-17) prospectively recruited 75 patients with echocardiographic signs of diastolic dysfunction and dyspnea on exertion (E/e'>8, New York Heart Association class ≥II) to undergo echocardiography, right heart catheterization, and RT-CMR at rest and during exercise stress. HFpEF was defined according to pulmonary capillary wedge pressure (≥15 mm Hg at rest or ≥25 mm Hg during exercise stress). RT-CMR functional assessments included time-volume curves for total and early (1/3) diastolic left ventricular filling, left atrial (LA) emptying, and left ventricular/LA long axis strain. RESULTS: Patients with HFpEF (n=34; median pulmonary capillary wedge pressure at rest, 13 mm Hg; at stress, 27 mm Hg) had higher E/e' (12.5 versus 9.15), NT-proBNP (N-terminal pro-B-type natriuretic peptide; 255 versus 75 ng/L), and LA volume index (43.8 versus 36.2 mL/m2) compared with patients with noncardiac dyspnea (n=34; rest, 8 mm Hg; stress, 18 mm Hg; P≤0.001 for all). Seven patients were excluded because of the presence of non-HFpEF cardiac disease causing dyspnea on imaging. There were no differences in RT-CMR left ventricular total and early diastolic filling at rest and during exercise stress (P≥0.164) between patients with HFpEF and noncardiac dyspnea. RT-CMR revealed significantly impaired LA total and early (P<0.001) diastolic emptying in patients with HFpEF during exercise stress. RT-CMR exercise stress LA long axis strain was independently associated with HFpEF (adjusted odds ratio, 0.657 [95% CI, 0.516-0.838]; P=0.001) after adjustment for clinical and imaging measures and emerged as the best predictor for HFpEF (area under the curve at rest 0.82 versus exercise stress 0.93; P=0.029). CONCLUSIONS: RT-CMR allows highly accurate identification of HFpEF during physiologic exercise and qualifies as a suitable noninvasive diagnostic alternative. These results will need to be confirmed in multicenter prospective research studies to establish widespread routine clinical use. Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03260621. URL: https://dzhk.de/; Unique identifier: DZHK-17.
Asunto(s)
Prueba de Esfuerzo/métodos , Insuficiencia Cardíaca/diagnóstico por imagen , Insuficiencia Cardíaca/diagnóstico , Imagen por Resonancia Magnética/métodos , Volumen Sistólico/fisiología , Anciano , Femenino , Humanos , Masculino , Persona de Mediana EdadRESUMEN
AIMS: The exact pathophysiology of Takotsubo syndrome (TTS) remains not fully understood with most studies focussing on ventricular pathology. Since atrial involvement may have a significant role, we assessed the diagnostic and prognostic potential of atrial cardiovascular magnetic resonance feature tracking (CMR-FT) in TTS. METHODS AND RESULTS: This multicentre study recruited 152 TTS patients who underwent CMR on average within 3 days after hospitalization. Reservoir [total strain εs and peak positive strain rate (SR) SRs], conduit (passive strain εe and peak early negative SRe), and booster pump function (active strain εa and peak late negative SRa) were assessed in a core laboratory. Results were compared with 21 control patients with normal biventricular function. A total of 20 patients underwent follow-up CMR (median 3.5 months, interquartile range 3-5). All patients were approached for general follow-up. Left atrial (LA) but not right atrial (RA) reservoir and conduit function were impaired during the acute phase (εs: P = 0.043, εe: P < 0.001, SRe: P = 0.047 vs. controls) and recovered until follow-up (εs: P < 0.001, SRs: P = 0.04, εe: P = 0.001, SRe: P = 0.04). LA and RA booster pump function were increased in the acute setting (LA-εa: P = 0.045, SRa: P = 0.002 and RA-εa: P = 0.004, SRa: P = 0.002 vs. controls). LA-εs predicted mortality [hazard ratio 1.10, 95% confidence interval (CI) 1.01-1.20; P = 0.037] irrespectively of established cardiovascular risk factors (P = 0.019, multivariate analysis) including left ventricular ejection fraction (LVEF) (area under the curve 0.71, 95% CI 0.55-0.86, P = 0.048). CONCLUSION: TTS pathophysiology comprises transient impairments in LA reservoir and conduit functions and enhanced bi-atrial active booster pump functions. Atrial CMR-FT may evolve as a superior marker of adverse events over and above established parameters such as LVEF and atrial volume.
Asunto(s)
Atrios Cardíacos/diagnóstico por imagen , Atrios Cardíacos/fisiopatología , Imagen por Resonancia Cinemagnética , Cardiomiopatía de Takotsubo/diagnóstico por imagen , Cardiomiopatía de Takotsubo/fisiopatología , Anciano , Estudios de Casos y Controles , Electrocardiografía , Femenino , Humanos , Masculino , Persona de Mediana Edad , Pronóstico , Estudios RetrospectivosRESUMEN
BACKGROUND: Cardiovascular magnetic resonance feature tracking (CMR-FT) is increasingly used for myocardial deformation assessment including ventricular strain, showing prognostic value beyond established risk markers if used in experienced centres. Little is known about the impact of appropriate training on CMR-FT performance. Consequently, this study aimed to evaluate the impact of training on observer variance using different commercially available CMR-FT software. METHODS: Intra- and inter-observer reproducibility was assessed prior to and after dedicated one-hour observer training. Employed FT software included 3 different commercially available platforms (TomTec, Medis, Circle). Left (LV) and right (RV) ventricular global longitudinal as well as LV circumferential and radial strains (GLS, GCS and GRS) were studied in 12 heart failure patients and 12 healthy volunteers. RESULTS: Training improved intra- and inter-observer reproducibility. GCS and LV GLS showed the highest reproducibility before (ICC >0.86 and >0.81) and after training (ICC >0.91 and >0.92). RV GLS and GRS were more susceptible to tracking inaccuracies and reproducibility was lower. Inter-observer reproducibility was lower than intra-observer reproducibility prior to training with more pronounced improvements after training. Before training, LV strain reproducibility was lower in healthy volunteers as compared to patients with no differences after training. Whilst LV strain reproducibility was sufficient within individual software solutions inter-software comparisons revealed considerable software related variance. CONCLUSION: Observer experience is an important source of variance in CMR-FT derived strain assessment. Dedicated observer training significantly improves reproducibility with most profound benefits in states of high myocardial contractility and potential to facilitate widespread clinical implementation due to optimized robustness and diagnostic performance.