Explainable Multimodal AI for In-Hospital Detection of Lung Cancer and Heart Failure Using Routine Data: Development, Internal Validation, and Simulated Federated Evaluation
- 1University of Dhaka
Abstract
ObjectiveTo determine whether combining four routine data sources improves in-hospital detection of lung cancer and heart failure, which sources matter, whether explanations are faithful, whether models are fair, and how much benefit federated training retains.
MethodsThis retrospective single-centre study linked MIMIC-IV admissions (73,637 patients) to chest radiographs, radiology reports, electrocardiograms (ECGs) and laboratory tests from the first 48 hours. Excluding prior cases, we compared single-source models, gradient boosting and early, late and attention-based fusion, tested explanations for faithfulness and weight dependence, audited subgroup performance, and simulated five hospitals for federated training.
ResultsLung cancer was coded in 1.7% and heart failure in 15.6%. Attention fusion reached an area under the receiver operating characteristic curve (AUROC) of 0.833 (95% CI 0.809-0.859) for lung cancer and 0.885 (0.878-0.892) for heart failure, against 0.753 and 0.794 for gradient boosting on laboratory data (both p < 0.001); early and late fusion performed similarly. Lung cancer relied most on laboratory data and reports, heart failure on the ECG. Explanations passed a deletion test; weight randomisation of absolute attributions passed for ECG but not laboratory explanations. Heart-failure discrimination fell with age. Federated training recovered 95% of the centralised gain over local training for heart failure; for lung cancer, federated averaging (FedAvg) recovered 51% and FedProx 82%.
ConclusionFusing routine data improved in-hospital detection of both diseases over laboratory data alone, and which data matter depends on the disease. Explanations should be checked per source, and performance in older patients monitored. External validation is needed.
Related
- Development and internal temporal validation of an explainable XGBoost model for predicting clinical pregnancy after frozen embryo transfer
- Development and validation of an explainable machine learning-based risk prediction model for macrosomia in han Chinese women with singleton cephalic pregnancies
- An explainable artificial intelligence system for prognostic assessment and decision support in diabetic foot ulcer care: a multicentre derivation and validation study with prospective randomised evaluation
- HMOX2-driven crosstalk between vascular aging and heart failure: A multimodal bioinformatics and explainable machine learning approach with experimental validation
- A multimodal optimisation framework for cardiac electrophysiology model personalisation using computed tomography and 12-lead electrocardiogram: Algorithm development and validation
BibTeX
@misc{ahmed2026explainable,
title = {{Explainable Multimodal AI for In-Hospital Detection of Lung Cancer and Heart Failure Using Routine Data: Development, Internal Validation, and Simulated Federated Evaluation}},
author = {M. Ahmed and M. M. Rahman and S. T. Wahid and G. Rabbi and A. Rafy and G. S. Khondoker and S. Dutta and M. A. Saadat and M. M. Haque},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.22.26366034},
url = {https://www.medrxiv.org/content/10.64898/2026.10.22.26366034v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Explainable Multimodal AI for In-Hospital Detection of Lung Cancer and Heart Failure Using Routine Data: Development, Internal Validation, and Simulated Federated Evaluation
AU - M. Ahmed
AU - M. M. Rahman
AU - S. T. Wahid
AU - G. Rabbi
AU - A. Rafy
AU - G. S. Khondoker
AU - S. Dutta
AU - M. A. Saadat
AU - M. M. Haque
PY - 2026
DA - 2026/10/22
PB - medRxiv
DO - 10.64898/2026.10.22.26366034
UR - https://www.medrxiv.org/content/10.64898/2026.10.22.26366034v1
AB - ObjectiveTo determine whether combining four routine data sources improves in-hospital detection of lung cancer and heart failure, which sources matter, whether explanations are faithful, whether models are fair, and how much benefit federated training retains. MethodsThis retrospective single-centre study linked MIMIC-IV admissions (73,637 patients) to chest radiographs, radiology reports, electrocardiograms (ECGs) and laboratory tests from the first 48 hours. Excluding prior cases, we compared single-source models, gradient boosting and early, late and attention-based fusion, tested explanations for faithfulness and weight dependence, audited subgroup performance, and simulated five hospitals for federated training. ResultsLung cancer was coded in 1.7% and heart failure in 15.6%. Attention fusion reached an area under the receiver operating characteristic curve (AUROC) of 0.833 (95% CI 0.809-0.859) for lung cancer and 0.885 (0.878-0.892) for heart failure, against 0.753 and 0.794 for gradient boosting on laboratory data (both p < 0.001); early and late fusion performed similarly. Lung cancer relied most on laboratory data and reports, heart failure on the ECG. Explanations passed a deletion test; weight randomisation of absolute attributions passed for ECG but not laboratory explanations. Heart-failure discrimination fell with age. Federated training recovered 95% of the centralised gain over local training for heart failure; for lung cancer, federated averaging (FedAvg) recovered 51% and FedProx 82%. ConclusionFusing routine data improved in-hospital detection of both diseases over laboratory data alone, and which data matter depends on the disease. Explanations should be checked per source, and performance in older patients monitored. External validation is needed.
ER -