Key result
Random Forest models outperform deep learning ResNet-18 in classifying cardiomyocyte calcium imaging, achieving ~0.94 ROC-AUC.
Why the study?
No standardised framework existed for functional single-cell phenotyping of calcium dysregulation across cardiac diseases and species. Confocal calcium imaging analysis remained limited to a few manually defined parameters, creating a bottleneck for drug screening and genotype variant adjudication.
Does a finetuned ResNet-18 deep learning framework improve automated classification of cardiac phenotypes across species, pathologies, and arrhythmia genotypes from confocal calcium imaging compared to conventional machine learning models?
Does a finetuned ResNet-18 deep learning framework improve automated classification of cardiac phenotypes across species, pathologies, and arrhythmia genotypes from confocal calcium imaging compared to conventional machine learning models?
Machine learning and deep learning models can accurately classify cardiac pathologies and arrhythmia genotypes directly from confocal calcium imaging, establishing a foundation for automated functional biomarker platforms in drug screening and variant adjudication.
Classical ML may generalize better than CNNs for calcium imaging phenotyping; leaves open optimal models for cross-lab cardiac genotype classification.
Background/Introduction Despite the central role of calcium dysregulation across heart failure, ischaemic heart disease, and hereditary arrhythmia syndromes, no standardised framework exists for functional single-cell phenotyping across disease entities and species. Confocal calcium imaging captures these dynamics with high spatiotemporal resolution, yet analysis remains limited to few manually defined parameters, creating a bottleneck for drug screening and genotype variant adjudication. Purpose To establish and validate an automated machine-learning framework for classifying cardiac phenotypes solely from confocal calcium imaging alone across species, pathologies, and hereditary arrhythmia genotypes, and to assess cross-laboratory generalisability for translational deployment. Methods Recordings from 621 cardiomyocytes (5249 calcium transients; five species; heart failure, myocardial infarction, catecholaminergic polymorphic ventricular tachycardia [CPVT]) were analysed. Conventional ML models (Random Forest, SVM, XGBoost) were trained on engineered transient descriptors. A finetuned ResNet-18 was evaluated using three preprocessing strategies: whole-image downscaling, multi-crop aggregation, transient-centred cropping. Group-aware stratified cross-validation ensured strict leakage control. External validation used additional 191 scans from a separate laboratory and acquisition period. Results On training data, both conventional ML and ResNet-18 showed strong discriminative ability across all three tasks. Optimal preprocessing was task-dependent: transient-centred cropping was most effective for species and pathology, while multi-crop aggregation yielded strongest CPVT genotype discrimination, potentially reflecting marked distributed diastolic events benefiting from scan-wide sampling. Engineered features outperformed deep learning only for species while ResNet-18 was superior for pathology and genotype. On external validation, classical models generalised more robustly (Random Forest ROC-AUC 0.94±0.01 vs. best ResNet-18 0.86±0.01; p=0.013), confirming calcium signatures transfer across laboratories while image-based models show greater domain-shift vulnerability. Post-hoc calibration improved the reliability of CNN probability estimates (ECE 0.14 to 0.08). SHAP identified rise-time as most dominant; Grad-CAM confirmed biologically plausible attention. Conclusion(s) Calcium imaging contain sufficient information for automated classification across a biologically meaningful hierarchy. Deep learning surpassed conventional models for pathology and genotype; engineered features were more robust under domain shift. These findings establish the foundation for microscopy-based functional biomarker platforms, with next steps including multi-site validation for automated drug screening and genotype variant adjudication.
No takes yet. Share an insight, caveat, or question.
Kaas et al. (2026) studied Heart failure, myocardial infarction, and catecholaminergic polymorphic ventricular tachycardia (CPVT) (n=621). Finetuned ResNet-18 deep learning model vs. Conventional machine learning models (Random Forest, SVM, XGBoost) trained on engineered transient descriptors was evaluated on External validation classification performance (ROC-AUC) across cardiac phenotypes (ROC-AUC 0.86 vs 0.94, p=0.013). Random Forest models generalized more robustly than ResNet-18 on external validation of cardiomyocyte calcium imaging (ROC-AUC 0.94±0.01 vs 0.86±0.01; p=0.013).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: