Developing a Multimodal Deep Learning Framework for Classifying Thalassemia and Iron Deficiency Anemia: A Conceptual Step towards Digital-Assisted Decision-Making in Hematology
Abstract
Objectives: Anemia is one of the most common hematological disorders and is recognized by the WHO as a major global health concern. Differentiating between anemia subtypes, such as thalassemia and iron deficiency anemia, is crucial for guiding the appropriate treatment. This study aimed to develop a multimodal deep learning framework to support digital-assisted decision-making in hematological disorders.
Material and Methods: Patient characteristics and corresponding peripheral blood smear (PBS) images were retrieved from electronic medical records under ethical approval. The proposed framework consisted of two components. The first component comprised three predictive layers: image analysis, patient characteristics, and patient profile similarity. The second component integrated these predictions through a majority voting mechanism, whereby concordance of two out of three layers determined the final classification.
Results: A total of 166 patients and 764 high-resolution PBS images were included, comprising 425 from thalassemia and 339 from iron deficiency anemia. Within the first component, EfficientNetB0 yielded the best performance for image analysis, the decision tree performed best for patient characteristics, and k-nearest neighbor provided optimal results for profile similarity. Combining these three models produced the most reliable predictive performance for the dataset.
Conclusion: The multimodal deep learning framework developed in this study demonstrated promising predictive performance. Further refinement under clinical supervision may advance the integration of image analysis into hematology practice.
Keywords
Full Text:
PDF (Early Version)References
WHO. Trends in anemia in women and children: 1995 to 2016; WHO 2017. Available from: http://www.WHO.int/data/ gho/data/indicators/indicator-details/GHO/prevalence-of- anemia-in-women-of-reprodutive-age.
Brimson S, Suwanwong Y, Brimson JM. Nutritional anemia predominant form of anemia in educated young Thai women. Ethn Health 2019;24:405-14.
Khorwanichakij N, Kungwankiattichai S, Owattanapanich W. Validation of several formulas to differentiate thalassemia from iron deficiency anemia and proposal of a thalassemia–iron deficiency discrimination (TID) predictive score. Siriraj Med J 2022;74:256-65.
Akers AS, Howard D, Ford JC, editors. Distinguishing iron deficiency anaemia from thalassemia trait in clinical obstetric practice 2018.
Jameel T, Baig M, Ahmed I, Hussain MB, Alkhamaly MBD. Differentiation of beta thalassemia trait from iron deficiency anemia by hematological indices. Pak J Med Sci 2017;33:665-9.
Ferih K, Elsayed B, Elshoeibi AM, Elsabagh AA, Elhadary M, Soliman A, et al. Applications of artificial intelligence in thalassemia: a comprehensive review. Diagnostics (Basel) 2023;13.
Anderson PG, Tarder-Stoll H, Alpaslan M, Keathley N, Levin DL, Venkatesh S, et al. Deep learning improves physician accuracy in the comprehensive detection of abnormalities on chest X-rays. Sci Rep 2024;14:25151.
Arora L, Singh SK, Kumar S, Gupta H, Alhalabi W, Arya V, et al. Ensemble deep learning and EfficientNet for accurate diagnosis of diabetic retinopathy. Sci Rep 2024;14:30554.
Brinker TJ, Hekler A, Enk AH, Berking C, Haferkamp S, Hauschild A, et al. Deep neural networks are superior to dermatologists in melanoma image classification. Eur J Cancer 2019;119:11-7.
Schömig-Markiefka B, Pryalukhin A, Hulla W, Bychkov A, Fukuoka J, Madabhushi A, et al. Quality control stress test for deep learning-based diagnostic model in digital pathology. Mod Pathol 2021;34:2098-108.
Babker AM, Suliman RS, Ghazwani AAM, Alharbi W, Lyashenko VV. Combined use of edge selection operators in individual color channels for the analysis of cytological images presented in RGB format. Biomed Pharmacol J 2025;18:121-38.
Babker AM, Suliman RS, Elshaikh RH, Boboyorov S, Lyashenko VV. Sequence of simple digital technologies for detection of platelets in medical images. Biomed Pharmacol J 2024;17:141-52.
Kansal K, Chandra TB, Singh A. ResNet-50 vs. EfficientNet-B0: Multi-Centric Classification of Various Lung Abnormalities Using Deep Learning. Procedia Computer Science. 2024;235:70-80.
Integration of EfficientNetB0 and Machine Learning for Fingerprint Classification. Informatica 2023;47.
Chen X, Pu X, Chen Z, Li L, Zhao KN, Liu H, et al. Application of EfficientNet-B0 and GRU-based deep learning on classifying the colposcopy diagnosis of precancerous cervical lesions. Cancer Med 2023;12:8690-9.
Ashames MMA, Demir A, Gerek ON, Fidan M, Gulmezoglu MB, Ergin S, et al. Are deep learning classification results obtained on CT scans fair and interpretable? Phys Eng Sci Med 2024;47:967-79.
Uddin S, Khan A, Hossain ME, Moni MA. Comparing different supervised machine learning algorithms for disease prediction. BMC Med Inform Decis Mak 2019;19:281.
Shehab M, Abualigah L, Shambour Q, Abu-Hashem MA, Shambour MKY, Alsalibi AI, et al. Machine learning in medical applications: a review of state-of-the-art methods. Comput Biol Med 2022;145:105458.
Uddin S, Lu H. Confirming the statistically significant superiority of tree-based machine learning algorithms over their counterparts for tabular data. PLoS One 2024;19:e0301541.
Zhang R, Xu P, Guo L, Zhang Y, Li P, Yao D. Z-score linear discriminant analysis for eeg based brain-computer interfaces. PLoS One 2013;8:e74433.
Stephens CR, Huerta HF, Linares AR. When is the naive bayes approximation not so naive? Machine Learning 2018;107:397-441.
Wang N, Wang M, Zhou Y, Liu H, Wei L, Fei X, et al. Sequential Data-Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development. J Med Internet Res 2022;24:e30720.
Mavaie P, Holder L, Skinner MK. Hybrid deep learning approach to improve classification of low-volume high-dimensional data. BMC Bioinformatics 2023;24:419.
Qaid TS, Mazaar H, Al-Shamri MYH, Alqahtani MS, Raweh AA, Alakwaa W. Hybrid deep-learning and machine-learning models for predicting COVID-19. Comput Intell Neurosci 2021;2021:9996737.
Elshaikh RH, Amir R, Ahmeide A, Mohamedahmed KA, Alfeel AH, Higazi H, et al. Evaluation of the discrimination between beta-thalassemia trait and iron deficiency anemia using different indexes. Int J Biomed 2022;12:375-9.
Refbacks
- There are currently no refbacks.

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.