{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,27]],"date-time":"2026-02-27T05:08:57Z","timestamp":1772168937448,"version":"3.50.1"},"update-to":[{"DOI":"10.12688\/f1000research.166307.1","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2025,11,19]],"date-time":"2025-11-19T00:00:00Z","timestamp":1763510400000}},{"DOI":"10.12688\/f1000research.166307.2","type":"new_version","label":"New version","source":"publisher","updated":{"date-parts":[[2025,11,19]],"date-time":"2025-11-19T00:00:00Z","timestamp":1763510400000}}],"reference-count":35,"publisher":"F1000 Research Ltd","license":[{"start":{"date-parts":[[2025,11,19]],"date-time":"2025-11-19T00:00:00Z","timestamp":1763510400000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["f1000research.com"],"crossmark-restriction":false},"short-container-title":["F1000Res"],"abstract":"<ns3:p>Background Early and accurate prediction of Ischemic Heart Disease (IHD) is critical to reducing cardiovascular mortality through timely intervention. While deep learning (DL) models have shown promise in disease prediction, many lack interpretability, generalizability, and fairness\u2014particularly when deployed across demographically diverse populations. These shortcomings limit clinical adoption and risk reinforcing healthcare disparities. Methods This study proposes a novel model: X-TLRABiLSTM (Explainable Transfer Learning\u2013based Residual Attention Bidirectional LSTM). The architecture integrates transfer learning from pre-trained cardiovascular models into a BiLSTM framework with residual attention layers to improve temporal feature extraction and convergence. To ensure transparency, the model incorporates SHAP (SHapley Additive exPlanations) to quantify the contribution of each clinical feature to the final prediction. Additionally, a demographic reweighting strategy is applied to the training process to reduce bias across subgroups defined by age, gender, and ethnicity. The model was evaluated on the UCI Heart Disease dataset using 10-fold cross-validation. Results The X-TLRABiLSTM model achieved a classification accuracy of 98.2%, with an F1-score of 98.1% and an AUC of 99.1%, outperforming standard ML classifiers and state-of-the-art DL baselines. SHAP-based interpretability analysis highlighted clinically relevant predictors such as chest pain type, ST depression, and thalassemia. A fairness-aware reweighting strategy was applied during training, and fairness evaluation revealed minimal performance disparity across demographic subgroups, with F1-score gaps \u2264 0.6% and error rate gaps \u2264 0.4%. Confusion matrix analysis demonstrated low false-positive and false-negative rates, reinforcing the model\u2019s reliability for clinical deployment. Conclusions X-TLRABiLSTM offers a highly accurate, interpretable, and demographically fair framework for IHD prognosis. By combining transfer learning, residual attention, explainable AI, and fairness-aware optimization, this model advances trustworthy AI in healthcare. Its successful performance on benchmark clinical data supports its potential for real-world integration in ethical, AI-assisted cardiovascular diagnostics.<\/ns3:p>","DOI":"10.12688\/f1000research.166307.3","type":"journal-article","created":{"date-parts":[[2025,11,19]],"date-time":"2025-11-19T14:40:07Z","timestamp":1763563207000},"page":"651","update-policy":"https:\/\/doi.org\/10.12688\/f1000research.crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Explainable Transfer Learning with Residual Attention BiLSTM for Prognosis of Ischemic Heart Disease."],"prefix":"10.12688","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3715-6941","authenticated-orcid":false,"given":"Cenitta","family":"D","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Arul","family":"N","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Praveen Pai","family":"T","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1402-6573","authenticated-orcid":false,"given":"VIijaya Arjunan","family":"Ranganathan","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tanuja","family":"Shailesh","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Andrew","family":"J","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"2560","published-online":{"date-parts":[[2025,11,19]]},"reference":[{"key":"ref1","article-title":"Cardiovascular Diseases (CVDs).","year":"2023"},{"key":"ref2","doi-asserted-by":"crossref","first-page":"4600","DOI":"10.3390\/jcm13164600","article-title":"Clinical Updates in Coronary Artery Disease: A Comprehensive Review.","volume":"13","author":"A Bottardi","year":"2024","journal-title":"J. Clin. Med."},{"key":"ref3","article-title":"Ischemic Heart Disease Prognosis: A Hybrid Residual Attention-Enhanced LSTM Model.","author":"D Cenitta","year":"2024","journal-title":"IEEE Access."},{"key":"ref4","first-page":"1781-1789","article-title":"A hybrid model for heart disease prediction using recurrent neural network and long short term memory.","volume":"14","author":"G Bhavekar","year":"2022","journal-title":"Int. J. Inf. Technol."},{"key":"ref6","doi-asserted-by":"crossref","first-page":"9816142","DOI":"10.1155\/2020\/9816142","article-title":"Improving an intelligent detection system for coronary heart disease using a two-tier classifier ensemble.","volume":"2020","author":"B Tama","year":"2020","journal-title":"Biomed. Res. Int."},{"key":"ref7","doi-asserted-by":"crossref","first-page":"263-275","DOI":"10.1007\/s40860-021-00133-6","article-title":"A decision support system for heart disease prediction based upon machine learning.","volume":"7","author":"P Rani","year":"2021","journal-title":"J. Reliab. Intell. Environ."},{"key":"ref39","doi-asserted-by":"crossref","first-page":"1763-1795","DOI":"10.1007\/s11831-024-10194-4","article-title":"A comprehensive review on heart disease risk prediction using machine learning and deep learning algorithms.","volume":"32","author":"V Karna","year":"2025","journal-title":"Arch. Comput. Method Eng."},{"key":"ref9","doi-asserted-by":"crossref","first-page":"106628","DOI":"10.1016\/j.compeleceng.2020.106628","article-title":"Support vector machines-based heart disease diagnosis using feature subset, wrapping selection and extraction methods.","volume":"84","author":"S Shah","year":"2020","journal-title":"Comput. Electr. Eng."},{"key":"ref10","doi-asserted-by":"crossref","first-page":"102573","DOI":"10.1016\/j.jksus.2023.102573","article-title":"Predicting the outcome of heart failure against chronic-ischemic heart disease in elderly population\u2013Machine learning approach based on logistic regression, case to Villa Scassi hospital Genoa, Italy.","volume":"35","author":"D Stojanov","year":"2023","journal-title":"J. King Saud Univ. Sci."},{"key":"ref11","first-page":"1781-1789","article-title":"A hybrid model for heart disease prediction using recurrent neural network and long short term memory.","volume":"14","author":"G Bhavekar","year":"2022","journal-title":"Int. J. Inf. Technol."},{"key":"ref12","first-page":"2585235","article-title":"A reliable machine intelligence model for accurate identification of cardiovascular diseases using ensemble techniques.","volume":"1","author":"B Doppala","year":"2022","journal-title":"J. Healthc. Eng."},{"key":"ref13","doi-asserted-by":"crossref","first-page":"1831-1838","DOI":"10.11591\/ijece.v12i2.pp1831-1838","article-title":"A hybrid approach to medical decision-making: diagnosis of heart disease with machine-learning model.","volume":"12","author":"T Suresh","year":"2022","journal-title":"Int. J. Elec. Comp. Eng."},{"key":"ref14","doi-asserted-by":"crossref","first-page":"1146-1168","DOI":"10.1080\/10255842.2020.1869726","article-title":"Multi disease-prediction framework using hybrid deep learning: an optimal prediction model.","volume":"24","author":"A Ampavathi","year":"2021","journal-title":"Comput. Methods Biomech. Biomed. Engin."},{"key":"ref16","article-title":"Ischemic Heart Disease Prognosis: A Hybrid Residual Attention-Enhanced LSTM Model.","author":"D Cenitta","year":"2024","journal-title":"IEEE Access."},{"key":"ref17","doi-asserted-by":"crossref","first-page":"9816142","DOI":"10.1155\/2020\/9816142","article-title":"Improving an intelligent detection system for coronary heart disease using a two-tier classifier ensemble.","volume":"2020","author":"B Tama","year":"2020","journal-title":"Biomed. Res. Int."},{"key":"ref18","first-page":"50-67","article-title":"An efficient privacy-preserving deep learning scheme for medical image analysis.","volume":"12","author":"J Andrew Onesimu","year":"2020","journal-title":"J. Inf. Technol. Manag."},{"key":"ref19","first-page":"59247-59256","article-title":"Recursion enhanced random forest with an improved linear model (RERF-ILM) for heart disease detection on the internet of medical things platform.","volume":"8","author":"C Guo","year":"2020","journal-title":"IEEE Access."},{"key":"ref20","doi-asserted-by":"crossref","first-page":"105108","DOI":"10.1016\/j.compbiomed.2021.105108","article-title":"Automatic heartbeat classification using S-shaped reconstruction and a squeeze-and-excitation residual network.","volume":"140","author":"X Li","year":"2022","journal-title":"Comput. Biol. Med."},{"key":"ref21","doi-asserted-by":"crossref","first-page":"2668-2679","DOI":"10.1016\/j.jacc.2018.03.521","article-title":"Artificial intelligence in cardiology.","volume":"71","author":"K Johnson","year":"2018","journal-title":"J. Am. Coll. Cardiol."},{"key":"ref22","first-page":"884-892","article-title":"Heart disease prediction using hybrid machine learning: A brief review.","volume":"5","author":"M Ahmed","year":"2024","journal-title":"J. Robot. Control."},{"key":"ref23","article-title":"A decision support system for heart disease prediction based upon machine learning.","author":"P Rani","year":"2021","journal-title":"J. Reliab. Intell. Environ."},{"key":"ref24","article-title":"UCI Heart Disease Dataset."},{"key":"ref25","first-page":"262-272","article-title":"Ischemic heart disease multiple imputation technique using machine learning algorithm.","volume":"19","author":"D Cenitta","year":"2022","journal-title":"Eng. Sci."},{"key":"ref26","article-title":"Nips 2016 tutorial: Generative adversarial networks.","author":"I Goodfellow","year":"2016","journal-title":"arXiv preprint arXiv:1701.00160."},{"key":"ref27","doi-asserted-by":"crossref","first-page":"1347-1358","DOI":"10.1056\/NEJMra1814259","article-title":"Machine learning in medicine.","volume":"380","author":"A Rajkomar","year":"2019","journal-title":"N. Engl. J. Med."},{"key":"ref28","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-642-24797-2","article-title":"Supervised sequence labelling with recurrent neural networks.","author":"A Graves","year":"2012"},{"key":"ref30","article-title":"A unified approach to interpreting model predictions.","volume":"30","author":"S Lundberg","year":"2017","journal-title":"Adv. Neural Inf. Proces. Syst."},{"key":"ref31","doi-asserted-by":"crossref","first-page":"447-453","DOI":"10.1126\/science.aax2342","article-title":"Dissecting racial bias in an algorithm used to manage the health of populations.","volume":"366","author":"Z Obermeyer","year":"2019","journal-title":"Science."},{"key":"ref32","doi-asserted-by":"crossref","first-page":"105108","DOI":"10.1016\/j.compbiomed.2021.105108","article-title":"Automatic heartbeat classification using S-shaped reconstruction and a squeeze-and-excitation residual network.","volume":"140","author":"X Li","year":"2022","journal-title":"Comput. Biol. Med."},{"key":"ref33","first-page":"1-35","article-title":"A survey on bias and fairness in machine learning.","volume":"54","author":"N Mehrabi","year":"2021","journal-title":"ACM computing surveys (CSUR)."},{"key":"ref34","doi-asserted-by":"crossref","first-page":"866-872","DOI":"10.7326\/M18-1990","article-title":"Ensuring fairness in machine learning to advance health equity.","volume":"169","author":"A Rajkomar","year":"2018","journal-title":"Ann. Intern. Med."},{"key":"ref35","doi-asserted-by":"crossref","first-page":"447-453","DOI":"10.1126\/science.aax2342","article-title":"Dissecting racial bias in an algorithm used to manage the health of populations.","volume":"366","author":"Z Obermeyer","year":"2019","journal-title":"Science."},{"key":"ref36","doi-asserted-by":"crossref","first-page":"153","DOI":"10.3390\/fi14050153","article-title":"Medical internet-of-things based breast cancer diagnosis using hyperparameter-optimized neural networks.","volume":"14","author":"R Ogundokun","year":"2022","journal-title":"Future Internet."},{"key":"ref37","doi-asserted-by":"crossref","first-page":"24-29","DOI":"10.1038\/s41591-018-0316-z","article-title":"A guide to deep learning in healthcare.","volume":"25","author":"A Esteva","year":"2019","journal-title":"Nat. Med."},{"key":"ref38","first-page":"73","article-title":"A framework for understanding unintended consequences of machine learning.","volume":"2","author":"H Suresh","year":"2019","journal-title":"arXiv preprint arXiv:1901.10002."}],"container-title":["F1000Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/f1000research.com\/articles\/14-651\/v3\/xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/f1000research.com\/articles\/14-651\/v3\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/f1000research.com\/articles\/14-651\/v3\/iparadigms","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,19]],"date-time":"2025-11-19T14:40:13Z","timestamp":1763563213000},"score":1,"resource":{"primary":{"URL":"https:\/\/f1000research.com\/articles\/14-651\/v3"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,19]]},"references-count":35,"URL":"https:\/\/doi.org\/10.12688\/f1000research.166307.3","relation":{"has-version":[{"id-type":"doi","id":"10.12688\/f1000research.166307.2","asserted-by":"subject"},{"id-type":"doi","id":"10.12688\/f1000research.166307.1","asserted-by":"subject"}],"has-review":[{"id-type":"doi","id":"10.5256\/f1000research.191512.r434284","asserted-by":"object"}]},"ISSN":["2046-1402"],"issn-type":[{"value":"2046-1402","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,19]]},"assertion":[{"value":"Indexed","URL":"https:\/\/f1000research.com\/articles\/14-651\/v3#article-reports","order":0,"name":"referee-status","label":"Referee status","group":{"name":"current-referee-status","label":"Current Referee Status"}},{"value":"10.5256\/f1000research.183275.r411315, Krishna Kumar Joshi, Neelam Joshi, Rajiv Gandhi Proudyogiki Vishwavidyalaya, MP, 16 Sep 2025, version 1, 1 approved, 1 approved with reservations","URL":"https:\/\/f1000research.com\/articles\/14-651\/v1#referee-response-411315","order":0,"name":"referee-response-411315","label":"Referee Report","group":{"name":"article-reports","label":"Article Reports"}},{"value":"<b>VIJAYA ARJUNAN RANGANATHAN<\/b>; \n<i>Posted: 16 Sep 2025<\/i>; We sincerely thank the reviewer for taking the time to evaluate our manuscript and approving the work. We are grateful for the approval and encouragement, which supports the value of our work.","URL":"https:\/\/f1000research.com\/articles\/14-651\/v3#referee-comment-14597","order":1,"name":"referee-comment-14597","label":"Referee Comment","group":{"name":"article-reports","label":"Article Reports"}},{"value":"10.5256\/f1000research.183275.r411313, Ramesh Chandra Poonia, CHRIST, Delhi, 17 Sep 2025, version 1, 1 approved, 1 approved with reservations","URL":"https:\/\/f1000research.com\/articles\/14-651\/v1#referee-response-411313","order":2,"name":"referee-response-411313","label":"Referee Report","group":{"name":"article-reports","label":"Article Reports"}},{"value":"<b>VIJAYA ARJUNAN RANGANATHAN<\/b>; \n<i>Posted: 21 Sep 2025<\/i>; \n<b>Reviewer#2, Concern # 1:<\/b> The phrase \u201cFair and Accurate Prognosis\u201d could be shortened to just \u201cAccurate Prognosis\u201d unless fairness is explicitly addressed with a defined framework or metrics. \n<b>Author response:<\/b>&nbsp; We thank the reviewer for this observation. In our study, fairness is not only a conceptual point but is explicitly addressed through a demographic reweighting strategy and quantitative fairness evaluation. Specifically, we applied fairness-aware loss functions during training and reported subgroup-level performance (\u0394F1 \u2264 0.6, error rate gap \u2264 0.4) across gender and age categories. These results demonstrate that the model achieves equitable performance across demographic groups, thereby justifying the inclusion of the term \u201cFair\u201d in the title. \n<b>Author action: <\/b>We have clarified this more explicitly in the abstract, introduction, and results sections to highlight the fairness framework and metrics applied in the study. For example, we now state in the abstract that fairness evaluation revealed minimal disparity across demographic subgroups, and we emphasize in the methods\/results that fairness-aware optimization was implemented and quantitatively validated.  \n<b>Reviewer#2, Concern # 2:<\/b> Consider slightly restructuring for brevity, e.g., \n<i>\u201cExplainable Transfer Learning with Residual Attention BiLSTM for Prognosis of Ischemic Heart Disease\u201d<\/i> or \n<i>\u201cExplainable Residual Attention BiLSTM with Transfer Learning for Accurate Ischemic Heart Disease Prognosis\u201d<\/i> \n<b>Author response:<\/b>&nbsp; We thank the reviewer for this helpful suggestion and agree that a more concise title improves readability. \n<b>Author action: <\/b>The title has been revised to: \u201cExplainable Transfer Learning with Residual Attention BiLSTM for Prognosis of Ischemic Heart Disease.\u201d  \n<b>Reviewer#2, Concern # 3:<\/b> Ensure that the paper includes a strong justification for each component mentioned in the title (e.g., why BiLSTM over standard LSTM, why transfer learning is suitable for the dataset, and how fairness is ensured). \n<b>Author response:<\/b>&nbsp; We thank the reviewer for this constructive suggestion. We agree that the rationale for each component of the proposed framework should be clearly articulated. Our revised manuscript now emphasizes:  Why BiLSTM over standard LSTM: BiLSTM captures both forward and backward temporal dependencies, which is critical in modelling clinical progression patterns where both prior and subsequent states provide diagnostic cues. This leads to richer feature representations compared to standard LSTM. Why transfer learning is suitable: The UCI Heart Disease dataset is relatively small and limited in diversity. Transfer learning from pre-trained cardiovascular models improves generalisation, accelerates convergence, and reduces overfitting on small datasets. How fairness is ensured: A demographic reweighting strategy was integrated into the loss function, ensuring balanced performance across subgroups (age and gender). We quantitatively validated fairness with subgroup-specific F1-scores and error rates (\u0394F1 \u2264 0.6, error rate gap \u2264 0.4).  \n<b>Author action: <\/b>We revised the Introduction (Section 1) to include clearer motivation for BiLSTM, transfer learning, and fairness. Additionally, we expanded the Methods (Sections 3.3.1 and 3.4) to justify these components explicitly and highlighted the fairness evaluation more clearly in the Results (Section 4.4).","URL":"https:\/\/f1000research.com\/articles\/14-651\/v3#referee-comment-14635","order":3,"name":"referee-comment-14635","label":"Referee Comment","group":{"name":"article-reports","label":"Article Reports"}},{"value":"10.5256\/f1000research.188562.r424480, Karna Vishnu Vardhana Reddy, Aditya University, Surampalem, Andhra Pradesh, 08 Nov 2025, version 2, indexed","URL":"https:\/\/f1000research.com\/articles\/14-651\/v2#referee-response-424480","order":4,"name":"referee-response-424480","label":"Referee Report","group":{"name":"article-reports","label":"Article Reports"}},{"value":"The author(s) declared that no grants were involved in supporting this work","order":5,"name":"grant-information","label":"Grant Information"},{"value":"This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.","order":0,"name":"copyright-info","label":"Copyright"}]}}