A combined AI echo-clinical model demonstrated superior performance for cardiac amyloidosis detection compared with a TTE-only model (AUC 0.94 vs 0.89; P=0.006).
Observational (n=1,043)
Yes
Does a combined AI echo-clinical model improve diagnostic performance for cardiac amyloidosis compared to a TTE-only AI model in patients with suspected disease?
Integrating clinical, laboratory, and echocardiographic data into a multimodal AI model significantly improves the detection of cardiac amyloidosis compared to an echocardiography-only model.
Absolute Event Rate: 0.94% vs 0.89%
p-value: p=0.006
BACKGROUND: Cardiac amyloidosis (CA) is an underdiagnosed yet treatable cause of heart failure in which timely diagnosis is essential to initiate life-prolonging therapies. While artificial intelligence (AI)-based tools using transthoracic echocardiography (TTE), electrocardiography, or electronic health records have demonstrated promise for CA detection, most rely on single data sources. We aimed to evaluate whether integrating clinical, laboratory, and TTE biomarkers improves the performance of an existing TTE-based AI model for CA detection. METHODS: We developed and tested a combined AI echo-clinical model (AI-ECM) incorporating demographics, laboratory biomarkers, and TTE parameters into a previously validated TTE-only AI model (Us2.Ca). Model training and internal validation were performed using the Amyloidosis Imaging International Consortium, a global multiethnic registry comprised of 727 patients with CA and 316 controls, including 202 with suspected transthyretin-CA with negative diagnostic evaluation and 114 patients with biopsy-proven extracardiac light chain amyloidosis without cardiac involvement. Ground truth CA diagnosis was adjudicated per consensus criteria. AI-ECM and Us2.Ca performance was assessed using area under the curve, accuracy, sensitivity, and specificity. RESULTS: In building the AI-ECM, feature importance analysis showed that having the Us2.Ca prediction scores, relative wall thickness, gender, and estimated glomerular filtration rate contributed most to performance. The AI-ECM demonstrated superior performance (area under the curve, 0.94; accuracy, 90%; sensitivity, 93%; specificity, 85%) compared with the Us2.Ca (area under the curve, 0.89; accuracy, 80%; sensitivity, 76%; specificity, 91%; P =0.006). While the Us2.Ca model classification was indeterminate in 9% of the cases, the AI-ECM allowed classification of all cases. The AI-ECM improved sensitivity for light chain-CA detection and maintained high accuracy across subtypes and control groups. CONCLUSIONS: A multiparametric AI model integrating basic clinical, laboratory, and TTE data with the deep learning Us2.Ca improved performance for CA detection over Us2.Ca alone. This approach represents a step toward scalable, AI-guided precision diagnostics for CA in diverse populations.
“The model shows particularly notable improvement in sensitivity for AL-CA detection, and while the specificity is modestly decreased, the increased sensitivity is more meaningful for a screening tool aiming to reduce missed and delayed diagnoses. The study population was notably diverse, with substantial Black and Hispanic/Latino representation across 9 centers in 4 countries. Though the model has not been externally validated, the multiethnic cohort partially offsets this limitation.”
Published in Circulation: Cardiovascular Imaging, this study introduces a new AI model that integrates clinical, lab, and echocardiography data to improve the accuracy of diagnosing cardiac amyloidosis. This is a significant advance in the application of AI to cardiac imaging, with the potential to facilitate earlier diagnosis and treatment of this complex disease.
Slivnick et al. (Wed,) conducted a observational in Cardiac amyloidosis (n=1,043). Combined AI echo-clinical model (AI-ECM) vs. TTE-only AI model (Us2.Ca) was evaluated on Area under the curve for cardiac amyloidosis detection (p=0.006). A combined AI echo-clinical model demonstrated superior performance for cardiac amyloidosis detection compared with a TTE-only model (AUC 0.94 vs 0.89; P=0.006).