A Dual-Modal Machine Learning Framework Integrating Red Blood Indices and Smartphone-Captured Microscopic Images for β-Thalassemia Screening

A Dual-Modal Machine Learning Framework Integrating Red Blood Indices and Smartphone-Captured Microscopic Images for β-Thalassemia Screening | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article A Dual-Modal Machine Learning Framework Integrating Red Blood Indices and Smartphone-Captured Microscopic Images for β-Thalassemia Screening C. R. Wijesinghe, Liyanaarachchi DC, Rathnayake PM, Weerasinghe AR This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8281069/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Mar, 2026 Read the published version in BMC Medical Informatics and Decision Making → Version 1 posted 14 You are reading this latest preprint version Abstract Background Thalassemia is the most prevalent single-gene disorder in Sri Lanka, imposing a significant socioeconomic and healthcare burden. Early carrier detection is essential for genetic counselling and the prevention of thalassemia major births. Current screening programs rely on red cell indices, manual examination of blood smears by expert haematologists, and expensive confirmatory tests such as High-Performance Liquid Chromatography (HPLC) or genetic analysis, which limit large-scale applicability in low-resource settings. This study aims to develop a machine learning–based automated screening system that integrates Red Blood Cell (RBC) indices and blood smear image analysis to support cost-effective and scalable β-thalassemia trait detection. Methods In a cross-sectional study of 152 individuals (54 confirmed β-thalassemia trait, 98 negative), 30% of the dataset was allocated for independent testing. The remaining 70% was used for model training, with 5-fold cross-validation for the RBC analysis model and 10-fold cross-validation for the image analysis model. RBC features were classified using a Multi-Layer Perceptron (MLP; scikit-learn), while blood smear images, captured through a conventional microscope using a smartphone camera, were analysed using a transfer learning–based VGG-16 CNN (TensorFlow/Keras). Data balancing and image augmentation (rotation, flipping, brightness variation) were applied to address class imbalance and overfitting. A two-step screening pipeline was proposed, applying RBC analysis first, followed by smear image analysis for RBC-negative cases. Results The RBC indices analysis model achieved 88.2% sensitivity (95% CI: 63.6–98.5%) and 92.9% specificity (95% CI: 76.5–99.1%), while the image analysis model reached 88.2% sensitivity (95% CI: 63.6–98.5%) and 64.3% specificity (95% CI: 44.1–81.4%). When integrated sequentially, the combined pipeline achieved 100% sensitivity (95% CI: 80.5–100%) and 60.7% specificity (95% CI: 40.6–78.5%), reducing the overall need for smear preparation by 37.7%. Conclusions This dual-modal system provides a highly sensitive, cost-effective, and practical solution for β-thalassemia carrier detection, enabling mass screening and supporting sustainable prevention strategies in resource-limited settings. With larger, independent, multi-centre validation, it could be integrated into laboratory workflows to expand screening coverage, while ongoing evaluation will ensure generalisability and inform policy adoption. Clinical Trial Registration Not applicable Thalassemia screening β-thalassemia trait detection Machine learning in healthcare Clinical decision support Red blood cell indices Blood smear image analysis Convolutional neural networks (CNN) Transfer learning Low-resource settings Smartphone-based microscopy Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Background Thalassemia[ 1 ] represents a major global health challenge due to its high prevalence and the substantial medical, social and economic burden associated with its management, particularly in low- and middle-income countries (LMICs) such as Sri Lanka [ 2 ]. It is a heterogeneous group of genetic disorders characterised by defective synthesis of hemoglobin chains, leading to anemia, ineffective erythropoiesis, and hemolysis [ 3 ]. The two principal forms, alpha-thalassemia and beta-thalassemia, vary widely in clinical severity, ranging from asymptomatic microcytosis to life-threatening anemia requiring lifelong transfusion therapy. In Sri Lanka, approximately 3,000 patients [ 4 ] depend on regular transfusions, consuming a significant portion of the national healthcare budget [ 5 ] [ 6 ]. Early screening and diagnosis are critical for preventing thalassemia major births. Current screening methods in Sri Lanka involve assessing red blood cell indices from a full blood count (FBC), peripheral blood smear examination by haematologists, and confirmatory tests such as High-Performance Liquid Chromatography (HPLC). While HPLC and genetic analysis provide very high diagnostic accuracy, these methods are time-consuming, expensive, and dependent on specialised expertise, making widespread implementation challenging. Despite Sri Lanka’s existing prevention program, the birth rate of thalassemia major cases remains inadequately controlled, demonstrating the need for more effective screening strategies. A simple, cost-effective, yet highly sensitive screening approach is therefore needed—one that can efficiently identify individuals at risk and prioritise them for confirmatory testing without overburdening clinicians and laboratory personnel. Such a method should ideally function with basic RBC data and blood smear images, even in rural primary care settings lacking haematologists. Machine learning has emerged as a powerful tool in medical diagnostics, offering potential improvements in efficiency, accuracy, and cost reduction. Various machine learning approaches have been applied to haematological disorders, but research on thalassemia screening remains limited. Previous studies have either focused on RBC-based classification models or blood image analysis, but not both simultaneously. Integrating these two approaches could significantly enhance diagnostic reliability by combining morphological and haematological features. However, a critical gap persists: current digital tools do not provide an integrated ML pipeline capable of processing both numerical RBC indices and smear-image features, nor are they designed for practical deployment in LMICs. Many existing systems require high-quality laboratory imaging, powerful computing resources, or specialised environments, making them unsuitable for decentralised or primary-care screening. In LMICs, the broader deployment of AI tools is challenged by variability in smartphone imaging quality, limited computational capacity, intermittent internet connectivity, and difficulties integrating digital tools into routine health workflows. Addressing these context-specific barriers is essential for achieving real-world impact. A notable innovation of the present study is the development of a machine learning pipeline explicitly designed to operate on average-priced smartphones, without requiring expensive high-tech equipment or internet connectivity. The system features a simple user interface, enabling healthcare workers to quickly and accurately screen individuals, triage a small number for confirmatory testing, and facilitate decentralised testing with minimal infrastructure. Embedding such a tool within national screening pathways has significant potential to enhance prevention efforts, particularly as Sri Lanka prepares to implement premarital thalassemia screening policies. This study aims to develop a machine-learning–based screening system optimised for high sensitivity and practical implementation, leveraging convolutional neural networks, transfer learning, and complementary ML techniques. A notable innovation of the present study is the development of an integrated pipeline that fuses Red Blood Cell (RBC) indices with CNN-derived blood smear image features into a unified, deployable screening system. The pipeline is explicitly designed to operate on average-priced smartphones, without requiring expensive high-tech equipment or internet connectivity, and features a simple user interface that enables healthcare workers to quickly and accurately screen individuals, triage a small number for confirmatory testing, and facilitate decentralised testing with minimal infrastructure, enabling use even in rural primary care settings without haematologists. This approach is relevant not only for thalassemia prevention but also for broader digital health implementation, as it demonstrates how AI tools can be adapted for low-resource settings, embedded into existing care pathways, and scaled to support national public-health strategies. By offering a sensitive, efficient, and sustainable solution, the proposed system has significant potential to strengthen thalassemia screening, particularly as Sri Lanka prepares to implement premarital screening policies and expand population-level preventive interventions. Machine learning techniques have increasingly been applied to automate microscopic image analysis in haematological cytology, addressing critical challenges in blood image processing. The accuracy of these methods depends on multiple factors, including slide preparation, staining quality, and digitisation settings. Optimising algorithm performance often requires standardised image acquisition protocols to ensure consistency and reliability across datasets [ 7 ]. One major challenge in blood smear analysis is detecting rare cells in peripheral blood. While analysing common blood cells can be done with small sample sections, the same approach may not be effective for rare cell detection due to limited imaging, potentially leading to false negatives. To mitigate this, studies emphasise the need for analysing larger slide areas, where image acquisition speed becomes a crucial factor [ 8 ]. To transition from research to practical application, systematic validation strategies must be established in laboratory environments. Defining comprehensive screening protocols will be essential to maximise the potential of automated recognition methodologies, making them suitable for clinical deployment [ 9 ]. Several studies have focused on automated thalassemia detection. Sandanayake et al. [ 10 ] developed a system to identify thalassemia using RBC images and machine learning. The system extracted features such as cell shape, size, and colour using edge detection (Canny), Hough Transform, and ellipse fitting, alongside quantitative parameters including RBC count, mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC), red blood cell distribution width (RDW), and haemoglobin content. Classification was performed using Artificial Neural Networks (ANN) and Support Vector Machines (SVM), achieving high accuracy (89.4% for females and 97% for males). The approach included a web interface for patient interaction, improving accessibility. However, the system relied solely on RBC morphological features without integrating RBC indices or other haematological parameters, limiting potential diagnostic accuracy. Validation was performed on a relatively small and gender-specific dataset, and modern deep learning methods such as CNNs or transfer learning were not explored. Khan et al. [ 11 ] developed a deep learning approach using Hb electrophoresis images, achieving up to 95.8% accuracy with pre-trained CNNs and confirming the model's focus on relevant haemoglobin bands. Although HPLC remains faster, more accurate, and convenient for confirmation, this study provides insights into transfer learning, pre-processing techniques, and interpretability methods such as Score-CAM, which are valuable for developing image-based screening models like ours. ThalPred is a machine learning tool designed to discriminate thalassemia trait (TT) from iron deficiency anaemia (IDA) using routine RBC parameters. Differentiating these two causes of hypochromic microcytic anaemia is challenging, especially in Thailand, due to the variability of RBC indices across populations. Using 186 adult samples, multiple models, including decision trees, random forests, SVM, and neural networks, were tested. While traditional indices like KF2 and RI performed moderately well, machine learning models, particularly SVM, achieved higher accuracy and robustness even on independent validation sets. Additionally, interpretable rules extracted from random forests enabled clinicians to screen using simple thresholds of Hb, RBC, RDW, and MCHC. Compared to our model, ThalPred emphasises the value of SVM for imbalanced datasets and highlights the importance of interpretability in clinical tools, reinforcing the benefit of combining robust algorithms with transparent decision rules. Lin et al. [ 12 ] used a Mask R-CNN-based deep learning model to detect and segment thalassemic RBCs from quantitative phase images, achieving 97.8% classification accuracy. Their approach combined instance segmentation with feature-based classifiers to extract single-cell morphology and haemoglobin content, providing detailed interpretability. However, the reliance on specialised digital holographic microscopy limits accessibility and real-time applicability. In contrast, our model uses full blood count parameters and blood smear images captured through a mid-range smartphone via a microscope eyepiece, offering greater accessibility and feasibility for widespread use. Insights from Lin et al. underscore the value of combining morphological information with quantitative features, guiding our approach to integrate image- and parameter-based data for accurate and interpretable thalassemia screening. Sadiq et al. [ 13 ] proposed an ensemble model, SGR-VC, combining SVM, Gradient Boosting Machine (GBM), and Random Forest (RF) to classify carriers and non-carriers using a dataset of 5,066 patients from the Punjab Thalassemia Prevention Program. The model achieved 93% accuracy, precision, recall, and F1-score, outperforming individual classifiers. While earlier works primarily focused on distinguishing β-thalassemia from iron deficiency anaemia or on specific subgroups such as women of fertile age, SGR-VC demonstrates a more generalisable approach across genders and age groups. However, gaps remain in exploring deep learning methods, regression-based models, and datasets from diverse geographic populations, which could further improve predictive power. Insights from this study highlight the effectiveness of ensemble approaches in leveraging complementary strengths of tree-based and probability-based classifiers, and the importance of balanced, normalised datasets for robust carrier detection. Compared to our model, SGR-VC serves as a benchmark showing that combining multiple classifiers can reduce misclassification and improve precision, providing a cost-effective, rapid screening tool suitable for large-scale public health applications. Purwar et al. [ 14 ] proposed a novel approach for thalassemia detection by fusing deep image features from blood smear images with clinical features from RBC indices. Using AlexNet CNN to extract morphological features and ten clinical parameters, such as RBC count, haemoglobin, MCV, and haemoglobin A2, the study applied PCA to reduce redundancy and prevent overfitting in the high-dimensional fused feature set (1011 features) relative to a small sample size (20 patients). Classification using Naive Bayes, Random Forest, and K-NN achieved 99 ± 1% accuracy with 100% sensitivity and specificity. While demonstrating the power of multimodal feature fusion, gaps remain in scaling to larger, more diverse populations and in exploring end-to-end deep learning models without feature engineering. Compared to our model, this study highlights the performance gains achievable when combining clinical and morphological information. Insights suggest that integrating heterogeneous data sources with dimensionality reduction can substantially improve detection accuracy even with limited datasets, providing a practical framework for automated thalassemia screening. Collectively, these studies demonstrate the promise of machine learning and deep learning for automated thalassemia detection. While prior work often focuses either on blood smear image analysis or RBC-based classification, few approaches integrate both modalities. Combining morphological and haematological features enhances diagnostic reliability and supports scalable, patient-centred screening systems, particularly in resource-limited settings such as Sri Lanka. Methods Study Design and Sample Collection Following ethical clearance, data were collected from individuals attending the National Thalassemia Centre in Kurunegala, Sri Lanka. Participants were recruited through three main pathways: voluntary screening, which included individuals who independently sought thalassemia carrier testing; referrals from primary care physicians or hospital wards for those with clinically suspected anaemia or microcytic hypochromic red cell indices; and mass screening programs conducted in schools across the Kurunegala district, particularly targeting adolescent populations. The routine diagnostic workflow at the centre was used for the study without any alterations. In routine (non-programmatic) screening, a structured clinical history was first obtained, followed by a Full Blood Count (FBC). Individuals with microcytic, hypochromic red cell indices (Mean Corpuscular Volume [MCV] < 77 fL and Mean Corpuscular Haemoglobin [MCH] < 27 pg) were further investigated by peripheral blood smear examination. Smears were prepared by trained laboratory technicians from venous blood collected in Ethylenediaminetetraacetic acid (EDTA) tubes, stained using the Leishman, and examined by haematologists for morphological features suggestive of thalassemia, including target cells, pencil cells, and tear-drop cells. Suspected cases were subjected to quantitative haemoglobin analysis using High-Performance Liquid Chromatography (HPLC), which is the standard confirmatory test for β-thalassemia in Sri Lanka. Individuals with an HbA₂ level > 3.5% were classified as β-thalassemia carriers. Following completion of diagnostic confirmation, Full Blood Count (FBC) reports, blood smears, and HPLC results, materials that would otherwise be discarded, were collected for analysis. Sensitive personal identifiers were removed, and only anonymised report numbers were retained to link FBC, smear, and HPLC data. Written informed consent was obtained from all participants prior to data use. For minors and individuals unable to provide consent, proxy consent was obtained from parents or legal guardians. All data were stored securely accessible only to authorised research personnel, and documentation of consent and ethical approval will be made available to the ethics review committee or journal upon request. Exclusion criteria included incomplete FBC reports, inability to retrieve the corresponding HPLC result, or absence of usable blood smear images. In addition, individuals who were HPLC-negative for β-thalassaemia but subsequently underwent further investigations for other haemoglobinopathies or alternative causes of anaemia were excluded to avoid diagnostic heterogeneity. No exclusions were made on the basis of prior blood transfusions or iron therapy. Population Characteristics and Generalisability Considerations Although screening services are available in multiple tertiary hospitals and dedicated haematology clinics across the country, the National Thalassemia Centre at Teaching Hospital Kurunegala functions as the largest national referral and screening hub. It hosts the highest number of registered patients and conducts the most extensive screening programmes in Sri Lanka, reflecting the highest reported regional prevalence of thalassaemia in areas surrounding Kurunegala. As a result, the study cohort is likely representative of the population most affected by β-thalassaemia within the national context. Age groups were approximately equally represented, with a near-balanced sex distribution (79 females, 73 males), reducing the likelihood of sex- or age-related sampling bias within the cohort. However, the dataset was collected from a single geographical region, and demographic characteristics beyond age and sex, such as ethnicity, socioeconomic background, and place of residence, were not recorded. These factors may or may not influence red blood cell indices or peripheral smear morphology, and to date, no scientific evidence has demonstrated systematic variation in these haematological parameters attributable to such demographic variables in Sri Lankan populations. Nonetheless, the absence of these data limits the ability to fully assess potential demographic bias. Furthermore, while some sources of variability were incorporated into the dataset through differences in operators and acquisition devices, the model has not yet been evaluated in populations served by other screening centres, laboratories using different equipment, or regions with differing genetic backgrounds. Therefore, future work should aim to include a more demographically diverse cohort through a multi-centre, nationwide study to enhance external validity and support broader clinical deployment. Data Acquisition Two types of data were acquired: Red Blood Cell Indices and Blood Smear Imaging Red Blood Cell Indices The following red cell indices and haematological parameters were collected for each participant: Haemoglobin (Hb), Red Blood Cell count (RBC), Mean Corpuscular Volume (MCV), Mean Corpuscular Haemoglobin (MCH), Mean Corpuscular Haemoglobin Concentration (MCHC), and Red cell Distribution Width (RDW). In addition, patient age and sex were recorded as shown in Table 1 . The dataset comprised 152 participants, including 54 BTT-positive and 98 BTT-negative individuals. For each RBC parameter and age, the mean ± standard deviation (SD) and observed range are presented for the total cohort, and separately for male and female participants within each group. Table 1 Descriptive summary of the study dataset: Red Blood Cell indices and age stratified by β-thalassemia trait (BTT) status and sex. Parameter Total Dataset Positive (BTT) Negative (Normal) Male Female Total Male Female Total MCV 71.46 ± 10.63 (51.4–90.5) 59.57 ± 4.39 (51.4–70.2) 60.21 ± 4.06 (53.9–74.4) 59.89 ± 4.20 (51.4–74.4) 77.26 ± 8.14 (54.3–89.9) 77.71 ± 6.96 (56.1–90.5) 77.51 ± 7.47 (54.3–90.5) MCH 23.47 ± 4.22 (15.1–33.0) 19.01 ± 1.42 (16.4–21.6) 19.09 ± 1.34 (17.0–23.9) 19.05 ± 1.37 (16.4–23.9) 25.75 ± 3.56 (15.1–31.2) 25.80 ± 3.01 (17.7–33.0) 25.78 ± 3.25 (15.1–33.0) MCHC 32.59 ± 1.94 (14.3–35.7) 31.97 ± 0.64 (30.5–32.9) 31.83 ± 0.76 (29.6–33.2) 31.90 ± 0.70 (29.6–33.2) 33.25 ± 1.58 (27.9–35.7) 32.72 ± 2.67 (14.3–34.9) 32.96 ± 2.26 (14.3–35.7) Hb 12.00 ± 2.11 (5.7–19.9) 11.24 ± 1.63 (8.5–13.8) 10.58 ± 1.07 (8.5–12.4) 10.91 ± 1.40 (8.5–13.8) 13.30 ± 2.59 (5.7–19.9) 12.00 ± 1.63 (9.4–17.3) 12.57 ± 2.19 (5.7–19.9) RBC 5.16 ± 0.73 (3.2–7.1) 5.91 ± 0.77 (3.9–7.1) 5.52 ± 0.46 (4.6–6.1) 5.72 ± 0.65 (3.9–7.1) 5.12 ± 0.60 (3.2–6.3) 4.69 ± 0.52 (3.5–5.8) 4.88 ± 0.60 (3.2–6.3) RDW 15.56 ± 2.17 (11.4–24.7) 17.31 ± 1.24 (14.8–20.0) 16.45 ± 1.23 (12.5–19.0) 16.88 ± 1.30 (12.5–20.0) 14.39 ± 1.74 (11.4–21.0) 15.24 ± 2.48 (12.7–24.7) 14.86 ± 2.21 (11.4–24.7) Age 22.29 ± 12.85 (1.0–63.0) 24.02 ± 15.81 (1.0–63.0) 19.29 ± 12.46 (2.8–45.0) 21.65 ± 14.30 (1.0–63.0) 22.37 ± 13.21 (1.0–48.0) 22.84 ± 11.22 (1.0–59.0) 22.63 ± 12.07 (1.0–59.0) Key observations from the dataset indicate that BTT-positive individuals exhibit markedly lower MCV values and MCH values compared to BTT-negative participants, who show higher averages for both parameters. Additionally, BTT-positive individuals demonstrate a higher RBC count than the negative group reflecting the compensatory erythropoiesis characteristic of thalassemia carriers. RDW and MCHC values showed additional distinctions between groups. BTT-positive individuals had higher RDW values compared to BTT-negative participants, indicating greater variability in red cell size. Differences in MCHC were more modest, although slightly lower values were observed among BTT-positive cases compared with the negative group. The age distribution of participants ranged from 1 to 63 years, with similar mean ages across groups. BTT-positive individuals had a mean age of 21.65 ± 14.30 years, closely aligned with the mean age of 22.63 ± 12.07 years observed in the negative group, suggesting no notable age-related sampling bias. Sex distribution was also balanced across the cohort, with male and female participants represented in approximately equal proportions in both BTT-positive and BTT-negative groups. Minor sex-related differences were noted Hb and RBC, with males showing slightly higher values. However, key screening parameters such as MCV and MCH demonstrated minimal variation by sex, indicating that these indices are largely independent of sex within this dataset. The dataset provides a comprehensive representation of RBC indices for both BTT-positive and BTT-negative individuals, making it suitable for training and validating the dual-modal machine learning pipeline. The distribution of values also highlights the discriminative potential of CBC indices for initial β-thalassemia screening. Blood Smear Imaging Blood smear images were captured using two approaches - Smartphone-assisted microscopy and Standard microscope camera. Smartphone-assisted microscopy – Three different smartphones were used: Samsung Galaxy M21, Apple iPhone 15, and Xiaomi Redmi Note 13. Three operators independently captured images, with slides randomly assigned in approximately equal proportions to simulate real-world variability in community and laboratory settings. Each phone camera was manually aligned with the microscope eyepiece without customised adapters or camera settings, and the smears were focused under the ×100 oil immersion objective. The microscope lamp served as the sole illumination source, and no exposure, colour, or white-balance calibration was performed prior to acquisition. Images were saved in the native formats of the devices (HEIC or PNG) and subsequently converted to JPEG during preprocessing. This method emulates routine haematological practice, with the smartphone effectively substituting for the human eye. Standard microscope camera – An AxioCam ERc5s camera attached to a standard light microscope was used to acquire reference-quality images at the same ×100 magnification. Relevant camera metadata, including pixel accuracy (0.0039 mm), image size (up to 5712 × 4284 pixels), light path, and lens information, were retained for reproducibility. For each smear, 25–30 non-overlapping images were obtained from the morphological “zone of interest,” i.e., the area between the body and tail of the smear where red cells are evenly distributed, non-overlapping, and morphologically intact. The head and tail regions, where cells are either too crowded or distorted, were excluded following standard haematology practice. Smear examination was performed systematically using a Z-pattern scanning approach to ensure comprehensive coverage. Representative blood smear images used in the study is shown in Fig. 1 . Captured images were reviewed, and any blurred or artefact-containing images were excluded. The final dataset comprised 5,198 high-quality images from 152 blood smear samples, each sample labelled as thalassemia carrier or non-carrier according to the corresponding HPLC results. Data Preprocessing Haematological Data (Red Blood Cell indices) The FBC reports obtained from the laboratory information system were exported into a structured format and underwent several preprocessing steps. Personal identifiers were removed to ensure anonymisation, with only the report ID retained to match smear images and HPLC results. Reports containing incomplete or missing red cell indices were excluded, and any outliers falling outside physiologically plausible limits were flagged and manually verified against the original records. Numerical variables, including MCV, MCH, MCHC, haemoglobin (Hb), RDW, and RBC count, were normalised using z-score standardization. Patient sex was encoded as a binary variable (0 = female, 1 = male). HPLC-confirmed carrier status served as the ground truth label, with carriers coded as 1 and non-carriers as 0. These steps resulted in a structured tabular dataset of RBC indices aligned with each participant’s corresponding smear images. Blood Smear Images The raw smear images underwent several preprocessing steps to ensure consistency and suitability for deep learning analysis. Initial quality control filtering was performed to remove images affected by motion blur, poor focus, or uneven staining, using both visual inspection and automated quality checks. To minimise background artefacts, a rectangular Region of Interest (ROI) corresponding to the largest inscribed area within the circular microscope field was extracted using the Hough Circle Transform in OpenCV, allowing automatic cropping of the maximal area containing red blood cells. All images were then resized to 224 × 224 pixels to meet the input requirements of the convolutional neural network (CNN) and maintain computational efficiency. To address class imbalance and improve model generalisation, data augmentation techniques—including random rotations (± 15°), horizontal and vertical flips, minor translations and zooming, and controlled brightness and contrast variations—were applied (Fig. 2 ). Following preprocessing, the dataset was divided into training (70%) and independent testing (30%) subsets, with patient-level stratification to prevent data leakage by ensuring that images from the same individual did not appear across different partitions. Model Architecture and Training The proposed model architecture was designed with careful consideration of the requirements of a population-level screening tool. In such settings, maximising sensitivity is paramount to ensure that no β-thalassemia carrier is missed, even at the expense of specificity. A false-negative result could delay diagnosis and increase the risk of undetected carrier marriages, whereas false positives can be resolved at the confirmatory testing stage. In addition to achieving high diagnostic accuracy, the architecture was intentionally designed to meet the practical challenges of real-world healthcare delivery, especially in resource-limited and rural settings where trained haematologists and advanced laboratory infrastructure may be scarce. The system therefore aimed to minimise the workload on clinicians and laboratory staff by reducing the need for unnecessary smear preparation and microscopic examination, while also lowering overall costs and reagent consumption to support large-scale screening initiatives. Furthermore, portability and scalability were prioritised to enable deployment across multiple platforms—from mobile devices to desktop systems—ensuring broad accessibility. Finally, the pipeline was optimised for rapid turnaround, delivering near real-time results suitable for screening during outreach activities and school-based screening programmes. Overview of the Dual-Modal Framework A two-step dual-modal pipeline was implemented. Step 1 of the screening pipeline involves haematological analysis based RBC indices together with patient age and sex. This stage was optimised for high specificity to ensure that individuals flagged as positive could be confidently referred directly for confirmatory HPLC testing without requiring additional intermediate assessments. Only those classified as negative at this stage proceed to the secondary evaluation. Step 2 of the pipeline focuses on morphological screening using blood smear images. For individuals not identified as positive in the first stage, peripheral blood smears are prepared, digitised, and analysed using a CNN-based model trained to detect subtle morphological features associated with thalassemia. This stage was intentionally designed to maximise sensitivity, ensuring that borderline or atypical cases are captured and the risk of missed carriers is minimised. By integrating the two modalities in a sequential architecture, the pipeline ensures high diagnostic reliability while maintaining operational efficiency. At the same time, the overall workload for laboratory staff is significantly reduced, as only a small subset of individuals who screen negative initially require smear preparation and further microscopic analysis. This design also optimises time, cost, and reagent consumption, making the workflow well suited for mass-screening programmes where preserving resources without compromising clinical robustness is essential. The dual-modal strategy is shown in Fig. 3 . Processing of Red Blood Cell Indices and Blood Smear Images Septs followed in processing the Blood Cell Indices and Blood Smear Images are discussed below. Processing Red Blood Cell Indices The first component of the dual-modal model processes Red Blood Cell (RBC) indices, along with patient age and sex. The preprocessing and model selection workflow included the following steps: I. Dimensionality Reduction (PCA): Principal Component Analysis (PCA) was applied to the RBC dataset to identify the most informative combinations of red cell indices. PCA helped reduce redundancy between correlated parameters while retaining the variance most relevant for discriminating carriers from non-carriers. The resulting principal components were analysed to guide feature selection for downstream model training. II. Algorithm Evaluation: To identify the optimal classifier for RBC-based screening, nine machine learning algorithm s were trained and compared using different permutations of RBC parameters: Logistic Regression (LR), k-Nearest Neighbors (k-NN), Decision Tree (DT), Random Forest (RF), Gradient Boosting Machine (GBM), AdaBoost, Support Vector Machine (SVM), Naive Bayes (NB), Multilayer Perceptron (MLP). Each model was evaluated on training and validation sets using cross-validation to ensure robustness, and performance metrics, including accuracy, precision, recall, and area under the ROC curve (AUC), were compared. III. Feature Selection & Optimisation: Based on performance, the most informative subset of RBC parameters was selected to maximise specificity in this first step, ensuring that predicted positives could be confidently sent for confirmatory HPLC without unnecessary smear preparation. Hyperparameters for each algorithm were tuned using grid search or Bayesian optimisation to achieve optimal predictive performance. Processing Blood Smear Images For model development, blood smears were prepared for 152 out of 163 participants, comprising 98 negative and 54 positive cases. All images underwent the preprocessing steps described previously, including ROI extraction using Hough Circle Transform, resolution standardisation (224 × 224 pixels), and data augmentation to increase variability and reduce class imbalance. The image component employed transfer learning to leverage pretrained convolutional neural networks (CNNs) for feature extraction. Multiple architectures were evaluated: VGG16, ResNet50, DenseNet121, MobileNetV2 and ConvNeXtTiny. Each architecture was fine-tuned and hyperparameters were systematically varied using grid search or random search to identify the optimal configuration for each model. Models were trained using training and validation subsets, with patient-level stratification to prevent data leakage. Performance metrics, including accuracy, sensitivity, specificity, F1-score, and area under the ROC curve (AUC), were calculated to identify the model with the best sensitivity, ensuring that borderline and atypical positive cases were reliably detected. The selected image model forms the second-stage screening component. Model Training and Optimisation Training and optimization of each model is described below. RBC Indices Model The processed RBC dataset was used to train a machine learning model for β-thalassemia carrier classification. The model was implemented in Python 3.9, using scikit-learn v1.2, along with NumPy, Pandas, Matplotlib, and Seaborn for evaluation and visualisation. The MLP architecture comprised hidden layers with ReLU activation, fine-tuned to optimise performance for binary classification. To ensure balanced representation of carriers and non-carriers and to prevent data leakage, stratified 5-fold cross-validation was employed throughout the optimisation process. Performance of the model was evaluated using a comprehensive set of custom metrics, including sensitivity, specificity, accuracy, F1-score, ROC–AUC, and the Matthews correlation coefficient. Blood Smear Image Model The second component of the dual-modal model processes blood smear images and was implemented in Python 3.9 using TensorFlow/Keras 2.x alongside standard scientific libraries such as NumPy, Pandas, and Matplotlib. This component was designed to ensure patient-level stratified training and evaluation, preventing data leakage across folds or between images belonging to the same individual. Pretrained convolutional neural networks were used as feature extractors. Their base layers were initially frozen, and a custom classifier head was added, consisting of a global average pooling layer, a dense layer with 256 ReLU-activated units, a dropout layer with a rate of 0.5, and a final softmax output layer with two units for binary classification. After initial training, the top four layers of each CNN backbone were unfrozen for fine-tuning. Training followed a two-stage strategy: 50 epochs with a frozen backbone using the Adam optimiser (learning rate = 0.001), followed by 50 fine-tuning epochs with a reduced learning rate of 1e-5. Images were processed at 224 × 224 resolution with a batch size of 32, and extensive data augmentation—such as random flips, brightness and contrast adjustments, saturation shifts, and rotations—was applied to improve generalisation. To maintain rigorous reproducibility, patient-level splitting was enforced using train_test_split to ensure that no images from the same patient appeared in both training and validation sets. A custom PatientFolderDataGenerator, implemented as a subclass of tf.keras.utils.Sequence, handled batch generation, preprocessing, augmentation, and GPU-accelerated loading. Training incorporated callbacks including EarlyStopping (patience = 10, monitoring validation accuracy), ReduceLROnPlateau (factor = 0.5, patience = 5), and ModelCheckpoint for saving the best-performing model. Implementation and Deployment The proposed dual-modal pipeline was designed with practical deployment in low-resource and primary care settings in mind. A key feature of the system is that it does not require specialised imaging hardware; smartphone-based microscopy was performed using standard laboratory microscopes without custom adapters, modified optics, or controlled imaging environments. This approach reflects the conditions under which peripheral smear evaluation is routinely carried out in many regional hospitals and rural clinics in Sri Lanka, enabling potential adoption without additional capital cost. In a real-world workflow, peripheral smear images captured using a smartphone or existing microscope camera could be uploaded to a local workstation or mobile application interface, where automated image analysis and FBC feature processing would be performed. The computational requirements of the final model allow inference to be executed either on-device for newer smartphones or via a low-resource edge or cloud-based server, making the system adaptable to settings with variable technical infrastructure. This supports potential deployment in rural primary care clinics that lack on-site haematologists, providing decision support for early carrier screening and referral. The intended operational use-case involves assisting healthcare workers by flagging individuals with a high likelihood of β-thalassaemia trait, thereby prioritising confirmatory testing and improving screening throughput. However, the current implementation remains a research prototype, and several constraints must be addressed prior to clinical integration. These include regulatory approval, standardisation of smear preparation and staining quality, operator training for consistent image capture, and integration with electronic laboratory information systems. Connectivity limitations in remote settings may also affect workflows reliant on cloud-based processing. Future development will focus on a full mobile application interface, streamlined clinician-facing reporting, and incorporation into existing national thalassaemia screening programmes. Prospective field evaluation and usability testing will be required to ensure that the system is safe, scalable, and operationally feasible in real-world clinical environments. Results RBC Indices Data Classification The PCA results revealed that MCV and MCH contributed most strongly to the first two principal components, indicating that these parameters alone capture the majority of the variance distinguishing β-thalassemia trait from normal individuals. This observation aligns with clinical knowledge, as microcytosis (low MCV) and hypochromia (low MCH) are hallmark features of thalassemia carriers. To further illustrate this, all nine algorithms were initially trained using only MCV and MCH as input features. The resulting classification performance is shown in Fig. 4 , which displays the separation of positive and negative cases by each algorithm in the test set. The graph demonstrates that even with just these two parameters, the models are able to distinguish BTT-positive from normal cases with reasonable accuracy, highlighting the predictive power of MCV and MCH in preliminary screening. This also informed subsequent model development, where additional parameters were incorporated to further improve classification metrics such as sensitivity, specificity, and overall accuracy. Table 2 shows the performance metrics of various machine learning models trained on RBC parameters for differentiating BTT-positive (BTT) and normal cases. Overall, the models demonstrated high performance, with Classification Accuracies exceeding 92% across most configurations and AUC values consistently above 95%, reflecting strong discrimination between BTT-positive and normal cases. Sensitivity and specificity were well-balanced, with specificity ranging from approximately 93–94% and sensitivity from 88–94%, which is critical for a screening approach where minimising false negatives is essential. True positive counts (BTT → BTT) ranged from 48 to 50, and true negative counts (Normal → Normal) ranged from 84 to 85, indicating a low rate of misclassification. Models trained on slightly different subsets of RBC parameters, including age, sex, or history of iron supplementation, showed minor variations in metrics, suggesting robustness to feature selection. Among the algorithms, the MLP (100) model achieved the highest AUC (~ 96.88%) with well-balanced sensitivity and specificity, whereas Naive Bayes models exhibited the highest sensitivity (~ 92.45%) while maintaining competitive overall accuracy. Logistic Regression and SVM models also performed strongly and offer the additional advantage of interpretability. Table 2 Performance metrics of various machine learning models trained on RBC parameters for differentiating BTT-positive (BTT) and normal cases. Blood Smear Image Analysis Table 3 summarises the performance of each CNN model on the independent test set, with all metrics reported along with their 95% confidence intervals. Table 3 Performance of each CNN model on the independent test set, with all metrics reported along with their 95% confidence intervals. Model CA (%) F1 (%) Precision (%) Recall (%) Specificity (%) MCC (%) MobileNetV2 37.2% (23.0–53.3) 52.6% (35.3–67.7) 35.7% (21.6–52.0) 100.0% (78.2–100.0) 3.6% (0.1–18.3) 0.113 (0.000–0.216) ConvNeXtTiny 37.8% (23.8–53.5) 51.7% (34.6–65.7) 34.9% (21.0–50.9) 100.0% (78.2–100.0) 6.7% (0.8–22.1) 0.152 (0.000–0.264) InceptionV3 60.5% (44.4–75.0) 63.8% (45.5–77.8) 46.9% (29.1–65.3) 100.0% (78.2–100.0) 39.3% (21.5–59.4) 0.429 (0.283–0.586) VGG16 72.1% (56.3–84.7) 68.4% (48.5–83.3) 56.5% (34.5–76.8) 86.7% (59.5–98.3) 64.3% (44.1–81.4) 0.487 (0.229–0.715) ResNet50 65.1% (49.1–79.0) 63.4% (43.9–79.1) 50.0% (29.9–70.1) 86.7% (59.5–98.3) 53.6% (33.9–72.5) 0.392 (0.131–0.624) DenseNet121 62.8% (46.7–77.0) 61.9% (42.4–77.8) 48.1% (28.7–68.1) 86.7% (59.5–98.3) 50.0% (30.6–69.4) 0.362 (0.106–0.590) VGG16 achieved the highest classification accuracy (72.1%) and the highest specificity (64.3%) among the models, while maintaining a strong recall (86.67%), indicating a balanced ability to identify both positive and negative samples. Models such as MobileNetV2, ConvNeXtTiny, and InceptionV3 reached perfect recall (100%), but their specificity was extremely low, leading to a high number of false positives and a risk of unnecessary follow-up testing. ResNet50 and DenseNet121 showed moderate performance across all metrics, offering more balanced but overall lower accuracy compared to VGG16. Overall, the results demonstrate a clear trade-off between recall and specificity among the CNN models. Since the primary objective of this step is screening, avoiding missed positive cases is critical. Therefore, VGG16 was selected for blood smear analysis because it provides high sensitivity (recall 86.67%), ensuring that most potential β-thalassemia carriers are flagged for confirmation testing. Although this increases false positives, the pipeline tolerates this, as all flagged cases will undergo HPLC confirmation. This approach maximises patient safety and supports effective population screening within the dual-modal workflow. Overall, the results demonstrate a clear trade-off between recall and specificity among the CNN models. Models such as ConvNeXtTiny and MobileNetV2 achieved very high recall, correctly identifying nearly all positive cases, but their extremely low specificity led to many false positives and unnecessary follow-up testing. In contrast, VGG16 provided the best balance, with the highest classification accuracy and specificity while maintaining high sensitivity (86.67%), reducing the risk of missed positive cases. Since the primary objective of this second step is screening, avoiding false negatives is critical. Therefore, VGG16 was selected as the preferred model for blood smear analysis, as its high sensitivity ensures that most potential β-thalassemia carriers are flagged for confirmation testing. Although this increases false positives, the pipeline is designed to tolerate this, as all flagged cases will undergo HPLC confirmation. This approach maximises patient safety, aligns with clinical screening practice, and supports the overall efficiency of the dual-modal pipeline. Error Analysis To further interpret the performance of the proposed thalassemia screening pipeline, an error analysis was conducted to examine the characteristics and potential causes of model misclassifications. This analysis aimed to identify patterns among false positives and false negatives across both the RBC-based and image-based models, evaluate the complementary effects of the two-step approach, and reveal factors contributing to residual errors. Understanding these limitations provides valuable insight into model behaviour and guides future improvements in data preprocessing, feature selection, and threshold optimisation. Error Analysis of Red Blood Cell Indices-based MLP model To further investigate the misclassifications of the Red Blood Cell Indices-based MLP model, scatter plots were generated for multiple pairs of haematological parameters. Each plot highlights true positives (TP), true negatives (TN), false positives (FP), and false negatives (FN), allowing visual assessment of the model’s decision boundaries and error patterns. The plots also indicate patient sex and adjust point transparency to emphasise misclassified cases. This multi-pair approach provides an intuitive way to explore the distribution of misclassified samples across different parameter combinations, identify regions where the model struggles, and guide potential improvements in feature selection or thresholding. Figure 5 shows Multi-pair scatter plots showing MLP model predictions on the RBC dataset. Visualisations illustrate the distribution of errors, reveal parameter regions contributing to misclassifications, and provide insights for refining the screening pipeline. Analysis of the MLP misclassifications revealed that false negatives (FNs) cluster in regions of the MCV vs MCH scatter plot that largely overlap with true negatives (TNs). These samples exhibited relatively higher MCV and MCH values compared to true positives, while their RBC counts were lower than those of typical positive cases but similar to those of negative samples. Biologically, this pattern may reflect borderline or atypical β-thalassemia carriers whose red blood cell indices do not fall within classical microcytic and hypochromic ranges. Such cases are known to occur in carriers with mild phenotypic expression, coexisting iron deficiency, or recent blood transfusions, leading to partial normalisation of haematological parameters. In the present dataset, the observed false negatives did not have recent blood transfusions or iron treatment, suggesting that their borderline haematological profiles may be due to mild genetic variants, heterozygosity with less severe mutation expression, or individual physiological compensation mechanisms that partially normalise RBC indices. Consequently, the model may classify these borderline carriers as negative, illustrating the inherent limitation of using Red Blood Cell indices alone for precise detection in all patient subpopulations. Conversely, false positives (FPs) were primarily located within regions characteristic of true positives, exhibiting lower MCV and MCH values alongside higher RBC counts. This pattern suggests that while these individuals do not carry the β-thalassemia trait, their haematological parameters resemble those of carriers, possibly due to benign microcytosis, variations in haemoglobin concentration, or other haematological conditions. Such overlap underscores the trade-off between sensitivity and specificity in screening approaches and highlights the necessity of confirmatory testing to accurately distinguish true carriers from individuals with overlapping but non-pathological profiles. Error Analysis of Image-Based Model (VGG-16) To further interpret the performance of the image-based screening step, misclassifications from the VGG-16 model were examined. False negatives (FNs) and false positives (FPs) were analysed in relation to morphological characteristics observed in the blood smear images, providing insight into patterns that contribute to residual errors. This analysis aimed to identify visual features associated with borderline cases, evaluate limitations in feature extraction by the CNN, and highlight scenarios where morphological ambiguity affects predictive accuracy. Grad-CAM visualisations were generated for representative true negative (TN), true positive (TP), false positive (FP), and false negative (FN) samples to investigate the image-based model’s attention patterns (Fig. 6 ). For the TP sample (ID: 63542, Positive Probability: 99.29%), attention was strongly focused on the central regions of RBCs, particularly highlighting morphological features consistent with β-thalassemia. Similarly, the TN sample (ID: 63769, Positive Probability: 1.71%) showed minimal attention on RBCs, with the model correctly ignoring normal cell morphology. In contrast, the FP sample (ID: 63627, Positive Probability: 91.32%) exhibited concentrated attention on RBC clusters, likely misinterpreting normal variations or overlapping cells as thalassemia-like features, resulting in high predicted probability despite the sample being negative. The FN sample (ID: 4303, Positive Probability: 11.72%) displayed diffuse and less distinct attention across RBCs, indicating that subtle or borderline morphological abnormalities were insufficient for confident detection. Collectively, these patterns suggest that misclassifications arise from either over-interpretation of normal variations (FPs) or under-recognition of mild, atypical morphological features (FNs), highlighting inherent limitations of relying solely on image morphology for screening in borderline cases. Evaluation of Alternative Screening Pipelines To further evaluate the robustness and optimal configuration of the dual-modal screening system, two alternative pipeline architectures were tested by altering the order of Red Blood Cell Indices and image-based analysis. This served as an ablation-style assessment to determine how sequencing the analytical stages influences diagnostic accuracy, sensitivity, and specificity. Pipeline 1 (Red Blood Cell Indices Analysis → Image Analysis): In this configuration, the Multi-Layer Perceptron (MLP) model based on full blood count parameters was applied first. Samples classified as negative were subsequently analysed using the VGG-16 CNN model. This approach aimed to minimise the number of smear analyses required while ensuring that no β-thalassemia carriers were missed. The combined pipeline achieved 100% sensitivity and 60.7% specificity, indicating that all positive cases were successfully detected but at the cost of increased false positives. Importantly, this configuration reduced the overall need for smear preparation and image processing by 37.7%, representing a practical advantage in large-scale screening workflows. Table 4 Performance metrics of the Red Blood Cell Indices → Image Analysis pipeline for β-thalassemia screening, with all metrics reported along with their 95% confidence intervals. The MLP model was applied to Red Blood Cell parameters first, followed by VGG-16–based image analysis for Red Blood Cell indices analysis-negative cases. Step Model Sensitivity (%) Specificity (%) Classification Accuracy (%) First step Red Blood Cell Indices Analysis (MLP) 88.2% (63.6–98.5%) 92.9% (76.5–99.1%) 91.1% (78.8–97.5%) Second step Image Analysis (VGG-16) 100.0% (15.8–100%) 57.7% (36.9–76.6%) 60.7% (40.6–78.5%) Overall Pipeline 100% (80.5–100%) 60.7% (40.6–78.5%) 75.6% (60.5–87.1%) Pipeline 2 (Image Analysis → Red Blood Cell Indices Analysis): The second configuration reversed the order, applying VGG-16–based image analysis first and subsequently re-evaluating negative cases with the MLP model. This sequence also achieved 100% sensitivity , confirming that no positive samples were missed. Although specificity increased slightly to 64.3% , all samples required blood smear preparation and image processing. This introduced significant manual effort, time consumption, and computational workload . Consequently, despite maintaining perfect sensitivity, the operational overhead makes this architecture less practical for large-scale or resource-limited screening programs. Table 5 Performance metrics of the Image Analysis → Red Blood Cell Indices analysis pipeline for β-thalassemia screening. VGG-16–based image analysis was applied first, followed by MLP analysis for image-negative cases. Step Model Sensitivity (%) Specificity (%) Classification Accuracy (%) First step Image Analysis (VGG-16) 88.2% (63.6–98.5%) 64.3% (44.1–81.4%) 73.3% (58.1–85.4%) Second step Red Blood Cell Indices Analysis (MLP) 100.0% (15.8–100.0%) 94.4% (72.7–99.9%) 95% (75.1–99.9%) Overall Pipeline 100.0% (80.5–100.0%) 64.3% (44.1–81.4) 77.8% (62.9–88.8%) Comparative Interpretation: Both pipelines maintained perfect sensitivity, confirming the value of combining haematological and morphological features for comprehensive β-thalassemia screening. However, starting with Red Blood Cell indices analysis followed by image evaluation of negatives provided greater operational efficiency by reducing smear preparation and processing requirements. This indicates that the Red Blood Cell Indices → Image pipeline is more suitable for scalable, resource-conscious screening in clinical and public health settings. Overall Dual-Modal Pipeline Performance The proposed dual-modal screening pipeline combines Red Blood Cell indices analysis and blood smear image classification to maximise sensitivity while reducing unnecessary laboratory workload. In this study, all positive cases (BTT carriers) were successfully identified, with zero missed cases, demonstrating that the pipeline is highly reliable for screening purposes. By first using the Red Blood Cell indices analysis model to classify cases, only the negative cases from the first step were forwarded to blood smear analysis. Consequently, the number of blood smear preparations was reduced from 152 total samples to 97, effectively saving 37% of laboratory time, reagents, and personnel effort. This two-step approach ensures that almost all positive cases are flagged for confirmation while minimising unnecessary laboratory procedures, making the pipeline both practical and cost-effective, particularly in settings with limited access to haematologists or specialised laboratory infrastructure. Additionally, by using a high-sensitivity VGG16 model for the second step, the pipeline preserves patient safety without compromising screening efficiency. Discussion This study demonstrates the feasibility and potential utility of a dual-modal machine learning pipeline for β-thalassemia screening in Sri Lanka, combining Red Blood Cell (RBC) indices and peripheral blood smear images. By integrating these two complementary modalities, the system leverages both quantitative haematological features and morphological patterns, providing a sensitive, cost-effective, and scalable approach suitable for deployment in low-resource settings. Diagnostic Performance and Key Findings Analysis of the RBC dataset revealed clear microcytic and hypochromic patterns among β-thalassemia trait (BTT) carriers, characterised by lower MCV (59.89 ± 4.20 fL) and MCH (19.05 ± 1.37 pg) and elevated RBC counts (5.72 ± 0.65 ×10¹²/L), compared to non-carriers. RDW was also higher among carriers, reflecting increased variability in red cell size. These differences provided a strong discriminatory signal for the initial RBC-based screening step. The MLP machine learning classifier, optimised using these parameters, achieved high specificity, allowing positive cases to be confidently referred for confirmatory HPLC testing while minimising unnecessary smear preparation for individuals screened as negative. The blood smear image analysis component further enhanced screening sensitivity, effectively detecting subtle morphological changes, including target cells, pencil cells, and other characteristic features of thalassemia. Using the VGG16 CNN architecture with transfer learning, the system reliably identified borderline and atypical carriers that could be missed by RBC indices alone. The integration of structured and image-based data in a sequential pipeline ensured that no true positive cases were overlooked, while maintaining operational efficiency by limiting smear analysis to individuals not flagged in the first step. Practical Implications This dual-modal approach addresses several challenges faced by national screening programmes in low- and middle-income countries. Firstly, it reduces reliance on specialised laboratory equipment, as smartphone-assisted microscopy provided sufficiently high-quality images. Secondly, the sequential workflow optimises resource utilisation by minimising the number of smears requiring manual examination, thereby saving time, reagents, and labour. Thirdly, given the limited number of haematologists in Sri Lankan hospitals, conducting mass screening is extremely tiring and exhausting for available staff; the model’s portability and compatibility with widely available smartphones enable decentralised screening in rural and primary care settings. This aligns with ongoing efforts to expand thalassaemia prevention programmes in Sri Lanka, including premarital screening initiatives. By combining RBC indices and smear morphology, the system also improves diagnostic confidence, reducing false positives and negatives compared to single-modality approaches. This is particularly relevant in LMICs, where conventional confirmatory tests like HPLC or genetic analysis may be limited by cost, infrastructure, or accessibility. The proposed pipeline could serve as a triage tool, prioritising individuals for confirmatory testing while reducing the burden on central laboratories. Limitations Despite its promising performance, the study has several limitations. The dataset was derived from a single tertiary referral centre in Kurunegala, limiting geographical and ethnic diversity. While age and sex distributions were balanced, other demographic factors, such as socioeconomic background, ethnicity, and regional prevalence variations, were not captured. Consequently, external validity remains to be established, and multi-centre studies are necessary to confirm generalisability across diverse populations. The sample size is modest, comprising 152 participants, which may affect the robustness of the findings and the ability to capture less common presentations of β-thalassaemia. Although smartphone-assisted image acquisition reflects real-world conditions, image quality may vary depending on operator experience, lighting, and device specifications. Data augmentation and CNN training mitigated some of these effects; however, further field validation is necessary to assess robustness under variable community-level conditions. The current pipeline remains a research prototype and has not yet undergone regulatory review, integration with laboratory information systems, or prospective clinical evaluation. Future Directions Future work should focus on expanding the dataset to include multi-centre, nationwide samples and additional demographic variables, enhancing model generalisability and robustness. Prospective studies are also required to evaluate real-world usability, workflow integration, and the impact on screening coverage, particularly in the context of mass thalassaemia prevention programs. Engagement with policymakers, healthcare administrators, and frontline health workers will be essential to understand operational constraints, gather feedback, and secure support for implementation. The development of a clinician-friendly mobile application with streamlined reporting could facilitate adoption, improve operational efficiency, and enable decentralised screening in rural and primary care settings. The dual-modal framework could be adapted for other haemoglobinopathies, supporting broader public health screening initiatives and contributing to scalable, sustainable, and equitable approaches to genetic disorder prevention. Implementation and Ethical Considerations While the dual-modal machine learning pipeline demonstrates high sensitivity and practical feasibility, several considerations must be addressed prior to deployment. Data Security and Privacy The system relies on patient data and blood smear images, which are sensitive health information. Secure storage, anonymisation, and controlled access are essential to prevent unauthorised use or breaches. Any mobile or cloud-based deployment must comply with national and international data protection regulations, including secure data transmission and storage protocols. Explainability and AI Reliability Supervised machine learning models, including deep learning, can produce incorrect or unexpected predictions, particularly when presented with inputs that differ from the training data. Ensuring explainability of these predictions is essential when they inform clinical decisions. Providing interpretable outputs, such as feature importance for RBC indices or saliency maps for blood smear images, can enhance clinician trust, facilitate oversight, and support safe integration into screening workflows. Operator and Workflow Variability The quality of smartphone-assisted images may vary with operator skill, lighting, and device type. Standardised training, clear imaging protocols, and continuous quality control are necessary to maintain model reliability across diverse settings. Ethical and Regulatory Oversight Clinical deployment must align with ethical guidelines, obtain regulatory approvals, and integrate seamlessly with existing laboratory and screening workflows. Continuous monitoring for potential biases, misclassifications, or inequitable outcomes is essential, particularly when expanding to new populations. Stakeholder Engagement Successful implementation requires input from healthcare workers, administrators, and policymakers to ensure the system is feasible, acceptable, and sustainable within national screening programmes. Addressing practical concerns, such as workload, cost, and training, will be critical for uptake and long-term adoption. Proactively addressing these considerations will enable safe, effective integration of the system into national screening pathways while maintaining patient trust and clinical reliability. Conclusion This study demonstrates that a dual-modal machine learning pipeline combining RBC indices and peripheral smear morphology can provide a sensitive, cost-effective, and practical approach for β-thalassemia screening. By enabling early detection and prioritisation of at-risk individuals, the system has potential to strengthen national prevention programs, reduce healthcare burden, and support decentralised screening in low-resource settings. It’s design, optimised for smartphones and minimal infrastructure, illustrates a scalable model for implementing AI-driven diagnostic support in LMICs. Declarations Ethics approval and consent to participate Ethical approval for this study was obtained from the Institutional Ethical Review Committee of the Teaching Hospital, Kurunegala. Written informed consent was obtained from all participants prior to enrolment using a certificate of consent. Consent for publication Not applicable Availability of data and materials The datasets used during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests Funding This study was funded by University of Colombo School of Computing, Sri Lanka. Authors' contributions AR, CR and DL work collaboratively in conceptualizing the study. DL collected data, performed data analysis, and was a major contributor in writing the manuscript. CR led data collection and preparation, model development, data analysis and reviewed and approved the final manuscript. PM assisted with data collection and reviewed the final manuscript. AR contributed in model development, data analysis and interpretation of the results. Acknowledgements The research team expresses sincere gratitude to Teaching Hospital Kurunegala for granting permission for data collection. A very special thanks goes to the medical laboratory technologists at the National Thalassemia Centre, Kurunegala, for their invaluable support in data collection. We also extend our appreciation to the University of Colombo School of Computing for providing necessary guidance and funding for this research. Finally, we acknowledge all individuals who contributed their time, expertise, and resources to make this study possible. References Aksu T, Ünal Ş, Thalassemia. Trends Pediatr. 2023;2(1):1–7. 10.5222/TP.2021.10820 . Perera O, Ranathunga T. Socio-economic impact of thalassemia: a case study of Kurunegala district, Sri Lanka. Fac J Humanit Social Sci (Univ Ruhuna). 2020;9(1):69–81. (hss.ruh.ac.lk). Marengo-Rowe AJ. The thalassemias and related disorders. Proc (Bayl Univ Med Cent). 2007;20(1):27–30. 10.1080/08998280.2007.11928230 . Amarasinghe N, et al. Redesigning new policy options for thalassemia prevention in Sri Lanka. Thalassemia Rep. 2022;12(4):135–42. 10.3390/thalassrep12040018 . de Silva S, Fisher CA, Premawardhena A, Lamabadusuriya SP, Peto TE, Perera G, Old JM, Clegg JB, Olivieri NF, Weatherall DJ, Sri Lanka Thalassaemia Study Group. Thalassaemia in Sri Lanka: implications for the future health burden of Asian populations. Lancet. 2000;355(9206):786 – 91. 10.1016/S0140-6736(99)08246-X . PMID: 10711926. Reed-Embleton H, Wijesiriwardana B, Perera S, Fernando M, Pathiraja R, Padeniya A. A cost-of-illness analysis of β-thalassaemia major in children in Sri Lanka: experience from a tertiary level teaching hospital. BMC Pediatr. 2020;20(1):160. 10.1186/s12887-020-02160-3 . Díaz G, Manzanera A. Automatic analysis of microscopic images in hematological cytology applications. In: González FA, Romero E, editors. Biomedical Image Analysis and Machine Learning Technologies: Applications and Techniques. Hershey (PA): IGI Global; 2009. Chapter 8. 10.4018/978-1-60566-956-4.ch008 . Lee SY, Chen CME, Lim EYP, Shen L, Sathe A, Singh A, Sauer J, Taghipour K, Yip CYC. Image analysis using machine learning for automated detection of Hemoglobin H inclusions in blood smears – a method for morphologic detection of rare cells. J Pathol Inf. 2021;12:18. 10.4103/jpi.jpi_110_20 . Rodellar J, Alférez S, Acevedo A, Molina A, Merino A. Image processing and machine learning in the morphological analysis of blood cells. Int J Lab Hematol. 2018;40(Suppl 1):46–53. 10.1111/ijlh.12818 . Sandanayake TC, Thalewela ATPMN, Thilakesooriya HP, Rathnayake RMAU, Wimalasooriya SAYA. Automated thalassemia identifier using image processing. IJIRT. 2016;3(7):97–105. Khan MS, Khan M, Khan SA, Imran M, Ullah F, Khan H, Irfan M, Nawaz M, Mehmood I. Deep learning assisted automated assessment of thalassemia from haemoglobin electrophoresis images. Diagnostics (Basel). 2022;12(10):2405. 10.3390/diagnostics12102405 . Lin YH, Liao KY-K, Sung K-B. Automatic detection and characterization of quantitative phase images of thalassemic red blood cells using a mask region-based convolutional neural network. J Biomed Opt. 2020;25(11):116502. 10.1117/1.JBO.25.11.116502 . Sadiq S, Khalid MU, Mui-Zzud-Din, Ullah S, Aslam W, Mehmood A, Choi GS, On B-W. Classification of β-Thalassemia carriers from red blood cell indices using ensemble classifier. IEEE Access. 2021;9:45528–38. 10.1109/ACCESS.2021.3066782 . Purwar S, Singh P, Singh A. Classification of thalassemia patients using a fusion of deep image and clinical features. Mater Today Proc. 2021;46:10418-22. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 21 Mar, 2026 Read the published version in BMC Medical Informatics and Decision Making → Version 1 posted Editorial decision: Revision requested 12 Jan, 2026 Reviews received at journal 12 Jan, 2026 Reviewers agreed at journal 12 Jan, 2026 Reviewers agreed at journal 11 Jan, 2026 Reviewers agreed at journal 10 Jan, 2026 Reviewers agreed at journal 10 Jan, 2026 Reviewers agreed at journal 10 Jan, 2026 Reviews received at journal 09 Jan, 2026 Reviewers agreed at journal 09 Jan, 2026 Reviewers invited by journal 12 Dec, 2025 Editor invited by journal 09 Dec, 2025 Editor assigned by journal 09 Dec, 2025 Submission checks completed at journal 09 Dec, 2025 First submitted to journal 04 Dec, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8281069","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":556210651,"identity":"14d16376-48dc-49b0-acb0-ad9092225ac5","order_by":0,"name":"C. R. Wijesinghe","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYHACxgdQBhuY5OMhrIXZAEULGxFa2CRI08I/+4xZNU9NnTwD++m0Bww1dkAtB/BrkTiXY3ab59hhwwae3O0GDMeSGdh4G/BrMeDh3Xabt+FAAoME7zYJBrYDDGz8BBwG0lLM21AH1fKPSC3MvA3MEC2MbQcIO0ziDP9nyTlAv7SB/JLYl8xD0Pv8PWyJH94AQ4yf/ey2Bx++2cnx8yQQcBkMgCMFqJiIiBwFo2AUjIJRQBAAAFYdMfkuGgP2AAAAAElFTkSuQmCC","orcid":"","institution":"University of Colombo School of Computing","correspondingAuthor":true,"prefix":"","firstName":"C.","middleName":"R.","lastName":"Wijesinghe","suffix":""},{"id":556210653,"identity":"03148c60-07be-4cef-ad52-6f7526383b69","order_by":1,"name":"Liyanaarachchi DC","email":"","orcid":"","institution":"University of Colombo School of Computing","correspondingAuthor":false,"prefix":"","firstName":"Liyanaarachchi","middleName":"","lastName":"DC","suffix":""},{"id":556210654,"identity":"dd0755b1-4d1a-400e-8874-6e5ac4cbe847","order_by":2,"name":"Rathnayake PM","email":"","orcid":"","institution":"University of Peradeniya","correspondingAuthor":false,"prefix":"","firstName":"Rathnayake","middleName":"","lastName":"PM","suffix":""},{"id":556210656,"identity":"c4243b6b-109e-44e2-86f9-9cd428385bdd","order_by":3,"name":"Weerasinghe AR","email":"","orcid":"","institution":"Informatics Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Weerasinghe","middleName":"","lastName":"AR","suffix":""}],"badges":[],"createdAt":"2025-12-04 15:53:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8281069/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8281069/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12911-026-03451-y","type":"published","date":"2026-03-21T15:58:38+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":97893956,"identity":"5b09c72c-eff4-4afb-b992-52d701d46098","added_by":"auto","created_at":"2025-12-10 15:31:41","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3265715,"visible":true,"origin":"","legend":"","description":"","filename":"ADualModalMachineLearningFrameworkforThalassemiascreening.docx","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/fd64337b9381f94f5dddf658.docx"},{"id":97695272,"identity":"d536ab55-3c74-4c6e-ac21-a1184680baca","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7378,"visible":true,"origin":"","legend":"","description":"","filename":"1391eadec05b401ea3b3242502833096.json","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/a9969ede978ba0c11bfe890d.json"},{"id":97695274,"identity":"b572f925-37d2-4c3e-914f-d0189a651e26","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128683,"visible":true,"origin":"","legend":"","description":"","filename":"1391eadec05b401ea3b32425028330961enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/f829bfeaa8578649c27b378a.xml"},{"id":97695288,"identity":"bedc2c37-ab61-4500-a98b-e4fa37344920","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":125396,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/4bd554298580c455d5e33abb.png"},{"id":97695280,"identity":"e5cae29a-350c-4583-b246-6e2d12070a2b","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":157895,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/a6c85c77254cbd6c0660c99b.png"},{"id":97695285,"identity":"e330df8e-8b2c-4503-bdd1-e09754440def","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":81339,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/eaef2d0c3f676e8de52cb540.png"},{"id":97894392,"identity":"6ab934a6-ec27-44cc-bb20-1d364e9126da","added_by":"auto","created_at":"2025-12-10 15:32:25","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13511,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/7cae576b4e24d0ed5f264dd6.png"},{"id":97894077,"identity":"16694e5a-a192-4f97-84f2-af6d34605e0f","added_by":"auto","created_at":"2025-12-10 15:31:54","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":65777,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/91230201246bb898267146df.png"},{"id":97894203,"identity":"53386703-e311-4269-8403-b0cd0ae6ead6","added_by":"auto","created_at":"2025-12-10 15:32:03","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40471,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/41141821b7c2cd3cf48dee35.png"},{"id":97893787,"identity":"249a5472-a81f-414f-8660-7beee901d88a","added_by":"auto","created_at":"2025-12-10 15:31:12","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59672,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/dcea0831b0a23fd237b09df3.png"},{"id":97894277,"identity":"3d4186c0-7f88-4fbd-81f9-d0169a30eb15","added_by":"auto","created_at":"2025-12-10 15:32:13","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":290332,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/db7c66d0d9d6c099a679ad4e.png"},{"id":97695283,"identity":"4661f65f-4835-4735-8c4f-92ca77b58006","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"xml","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":125967,"visible":true,"origin":"","legend":"","description":"","filename":"1391eadec05b401ea3b32425028330961structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/dc4198c6d907c69b87ee62de.xml"},{"id":97695291,"identity":"36697c38-4e85-4c98-aa73-5c1cdbf7f0fc","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"html","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":136453,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/dc4c645096f44c964a740643.html"},{"id":97893032,"identity":"790cd0cc-fd6b-44dc-9c62-731ba21a759f","added_by":"auto","created_at":"2025-12-10 15:26:03","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":571276,"visible":true,"origin":"","legend":"\u003cp\u003eRepresentative blood smear images used in the study. (A) Smartphone microscopy; (B) Standard microscope camera.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/d18a8a50e7b391f05c1bca08.jpeg"},{"id":97894270,"identity":"4e6cab8a-5261-4714-9da3-eb0c38a876b5","added_by":"auto","created_at":"2025-12-10 15:32:09","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":363994,"visible":true,"origin":"","legend":"\u003cp\u003eBlood smear image after pre-processing.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/c87677ebfa70d599bdd42c42.png"},{"id":97893084,"identity":"c54cc756-8e16-4e11-a6b0-31d1b76ad46e","added_by":"auto","created_at":"2025-12-10 15:26:40","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":44141,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic overview of the dual-modal machine learning pipeline for β-thalassemia screening: MLP model and CNN Model\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/6e2e93c80468e44b3870cec0.png"},{"id":97894358,"identity":"959bdf9a-47d8-4ee5-bb69-592f7915b374","added_by":"auto","created_at":"2025-12-10 15:32:23","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":247394,"visible":true,"origin":"","legend":"\u003cp\u003eClassification of β-thalassemia trait (BTT) versus normal cases using only MCV and MCH as input features. The scatter plots illustrate the distribution of true positives (BTT) and true negatives (Normal) predicted by each model, demonstrating that MCV and MCH alone provide substantial discriminatory power for preliminary screening.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/426757ab7a5826c321280e7f.png"},{"id":97695273,"identity":"94585ca9-944d-4c03-94e7-e17a121164cc","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":65927,"visible":true,"origin":"","legend":"\u003cp\u003eMulti-pair scatter plots showing MLP model predictions on the RBC dataset. True positives (orange), true negatives (green), false positives (blue), and false negatives (red) are displayed for each parameter pair (e.g., MCV vs MCH, Hb vs RBC). Point shapes indicate patient sex, and alpha values highlight misclassified samples.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/b6c5fb0a2b759229d79c1cc5.png"},{"id":97695277,"identity":"07b7c610-32f7-4785-8816-72094a894763","added_by":"auto","created_at":"2025-12-08 11:34:12","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1162684,"visible":true,"origin":"","legend":"\u003cp\u003eGrad-CAM visualisations were generated for representative true negative (TN), true positive (TP), false positive (FP), and false negative (FN) samples\u003c/p\u003e","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/3a8f0c7a5fb7efc5f3e23657.jpeg"},{"id":105223530,"identity":"ad2c823b-fdb3-4b2d-815e-278ecb349605","added_by":"auto","created_at":"2026-03-23 16:08:09","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3853295,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8281069/v1/83c98826-7e37-4fe4-9fb3-a4d599533395.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Dual-Modal Machine Learning Framework Integrating Red Blood Indices and Smartphone-Captured Microscopic Images for β-Thalassemia Screening","fulltext":[{"header":"Background","content":"\u003cp\u003eThalassemia[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] represents a major global health challenge due to its high prevalence and the substantial medical, social and economic burden associated with its management, particularly in low- and middle-income countries (LMICs) such as Sri Lanka [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. It is a heterogeneous group of genetic disorders characterised by defective synthesis of hemoglobin chains, leading to anemia, ineffective erythropoiesis, and hemolysis [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The two principal forms, alpha-thalassemia and beta-thalassemia, vary widely in clinical severity, ranging from asymptomatic microcytosis to life-threatening anemia requiring lifelong transfusion therapy. In Sri Lanka, approximately 3,000 patients [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] depend on regular transfusions, consuming a significant portion of the national healthcare budget [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eEarly screening and diagnosis are critical for preventing thalassemia major births. Current screening methods in Sri Lanka involve assessing red blood cell indices from a full blood count (FBC), peripheral blood smear examination by haematologists, and confirmatory tests such as High-Performance Liquid Chromatography (HPLC). While HPLC and genetic analysis provide very high diagnostic accuracy, these methods are time-consuming, expensive, and dependent on specialised expertise, making widespread implementation challenging. Despite Sri Lanka\u0026rsquo;s existing prevention program, the birth rate of thalassemia major cases remains inadequately controlled, demonstrating the need for more effective screening strategies. A simple, cost-effective, yet highly sensitive screening approach is therefore needed\u0026mdash;one that can efficiently identify individuals at risk and prioritise them for confirmatory testing without overburdening clinicians and laboratory personnel. Such a method should ideally function with basic RBC data and blood smear images, even in rural primary care settings lacking haematologists.\u003c/p\u003e\u003cp\u003eMachine learning has emerged as a powerful tool in medical diagnostics, offering potential improvements in efficiency, accuracy, and cost reduction. Various machine learning approaches have been applied to haematological disorders, but research on thalassemia screening remains limited. Previous studies have either focused on RBC-based classification models or blood image analysis, but not both simultaneously. Integrating these two approaches could significantly enhance diagnostic reliability by combining morphological and haematological features.\u003c/p\u003e\u003cp\u003eHowever, a critical gap persists: current digital tools do not provide an integrated ML pipeline capable of processing both numerical RBC indices and smear-image features, nor are they designed for practical deployment in LMICs. Many existing systems require high-quality laboratory imaging, powerful computing resources, or specialised environments, making them unsuitable for decentralised or primary-care screening.\u003c/p\u003e\u003cp\u003eIn LMICs, the broader deployment of AI tools is challenged by variability in smartphone imaging quality, limited computational capacity, intermittent internet connectivity, and difficulties integrating digital tools into routine health workflows. Addressing these context-specific barriers is essential for achieving real-world impact. A notable innovation of the present study is the development of a machine learning pipeline explicitly designed to operate on average-priced smartphones, without requiring expensive high-tech equipment or internet connectivity. The system features a simple user interface, enabling healthcare workers to quickly and accurately screen individuals, triage a small number for confirmatory testing, and facilitate decentralised testing with minimal infrastructure. Embedding such a tool within national screening pathways has significant potential to enhance prevention efforts, particularly as Sri Lanka prepares to implement premarital thalassemia screening policies.\u003c/p\u003e\u003cp\u003eThis study aims to develop a machine-learning\u0026ndash;based screening system optimised for high sensitivity and practical implementation, leveraging convolutional neural networks, transfer learning, and complementary ML techniques. A notable innovation of the present study is the development of an integrated pipeline that fuses Red Blood Cell (RBC) indices with CNN-derived blood smear image features into a unified, deployable screening system. The pipeline is explicitly designed to operate on average-priced smartphones, without requiring expensive high-tech equipment or internet connectivity, and features a simple user interface that enables healthcare workers to quickly and accurately screen individuals, triage a small number for confirmatory testing, and facilitate decentralised testing with minimal infrastructure, enabling use even in rural primary care settings without haematologists.\u003c/p\u003e\u003cp\u003eThis approach is relevant not only for thalassemia prevention but also for broader digital health implementation, as it demonstrates how AI tools can be adapted for low-resource settings, embedded into existing care pathways, and scaled to support national public-health strategies. By offering a sensitive, efficient, and sustainable solution, the proposed system has significant potential to strengthen thalassemia screening, particularly as Sri Lanka prepares to implement premarital screening policies and expand population-level preventive interventions.\u003c/p\u003e\u003cp\u003eMachine learning techniques have increasingly been applied to automate microscopic image analysis in haematological cytology, addressing critical challenges in blood image processing. The accuracy of these methods depends on multiple factors, including slide preparation, staining quality, and digitisation settings. Optimising algorithm performance often requires standardised image acquisition protocols to ensure consistency and reliability across datasets [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. One major challenge in blood smear analysis is detecting rare cells in peripheral blood. While analysing common blood cells can be done with small sample sections, the same approach may not be effective for rare cell detection due to limited imaging, potentially leading to false negatives. To mitigate this, studies emphasise the need for analysing larger slide areas, where image acquisition speed becomes a crucial factor [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eTo transition from research to practical application, systematic validation strategies must be established in laboratory environments. Defining comprehensive screening protocols will be essential to maximise the potential of automated recognition methodologies, making them suitable for clinical deployment [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eSeveral studies have focused on automated thalassemia detection. Sandanayake et al. [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] developed a system to identify thalassemia using RBC images and machine learning. The system extracted features such as cell shape, size, and colour using edge detection (Canny), Hough Transform, and ellipse fitting, alongside quantitative parameters including RBC count, mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC), red blood cell distribution width (RDW), and haemoglobin content. Classification was performed using Artificial Neural Networks (ANN) and Support Vector Machines (SVM), achieving high accuracy (89.4% for females and 97% for males). The approach included a web interface for patient interaction, improving accessibility. However, the system relied solely on RBC morphological features without integrating RBC indices or other haematological parameters, limiting potential diagnostic accuracy. Validation was performed on a relatively small and gender-specific dataset, and modern deep learning methods such as CNNs or transfer learning were not explored.\u003c/p\u003e\u003cp\u003eKhan et al. [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e] developed a deep learning approach using Hb electrophoresis images, achieving up to 95.8% accuracy with pre-trained CNNs and confirming the model's focus on relevant haemoglobin bands. Although HPLC remains faster, more accurate, and convenient for confirmation, this study provides insights into transfer learning, pre-processing techniques, and interpretability methods such as Score-CAM, which are valuable for developing image-based screening models like ours.\u003c/p\u003e\u003cp\u003eThalPred is a machine learning tool designed to discriminate thalassemia trait (TT) from iron deficiency anaemia (IDA) using routine RBC parameters. Differentiating these two causes of hypochromic microcytic anaemia is challenging, especially in Thailand, due to the variability of RBC indices across populations. Using 186 adult samples, multiple models, including decision trees, random forests, SVM, and neural networks, were tested. While traditional indices like KF2 and RI performed moderately well, machine learning models, particularly SVM, achieved higher accuracy and robustness even on independent validation sets. Additionally, interpretable rules extracted from random forests enabled clinicians to screen using simple thresholds of Hb, RBC, RDW, and MCHC. Compared to our model, ThalPred emphasises the value of SVM for imbalanced datasets and highlights the importance of interpretability in clinical tools, reinforcing the benefit of combining robust algorithms with transparent decision rules.\u003c/p\u003e\u003cp\u003eLin et al. [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] used a Mask R-CNN-based deep learning model to detect and segment thalassemic RBCs from quantitative phase images, achieving 97.8% classification accuracy. Their approach combined instance segmentation with feature-based classifiers to extract single-cell morphology and haemoglobin content, providing detailed interpretability. However, the reliance on specialised digital holographic microscopy limits accessibility and real-time applicability. In contrast, our model uses full blood count parameters and blood smear images captured through a mid-range smartphone via a microscope eyepiece, offering greater accessibility and feasibility for widespread use. Insights from Lin et al. underscore the value of combining morphological information with quantitative features, guiding our approach to integrate image- and parameter-based data for accurate and interpretable thalassemia screening.\u003c/p\u003e\u003cp\u003eSadiq et al. [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] proposed an ensemble model, SGR-VC, combining SVM, Gradient Boosting Machine (GBM), and Random Forest (RF) to classify carriers and non-carriers using a dataset of 5,066 patients from the Punjab Thalassemia Prevention Program. The model achieved 93% accuracy, precision, recall, and F1-score, outperforming individual classifiers. While earlier works primarily focused on distinguishing β-thalassemia from iron deficiency anaemia or on specific subgroups such as women of fertile age, SGR-VC demonstrates a more generalisable approach across genders and age groups. However, gaps remain in exploring deep learning methods, regression-based models, and datasets from diverse geographic populations, which could further improve predictive power. Insights from this study highlight the effectiveness of ensemble approaches in leveraging complementary strengths of tree-based and probability-based classifiers, and the importance of balanced, normalised datasets for robust carrier detection. Compared to our model, SGR-VC serves as a benchmark showing that combining multiple classifiers can reduce misclassification and improve precision, providing a cost-effective, rapid screening tool suitable for large-scale public health applications.\u003c/p\u003e\u003cp\u003ePurwar et al. [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] proposed a novel approach for thalassemia detection by fusing deep image features from blood smear images with clinical features from RBC indices. Using AlexNet CNN to extract morphological features and ten clinical parameters, such as RBC count, haemoglobin, MCV, and haemoglobin A2, the study applied PCA to reduce redundancy and prevent overfitting in the high-dimensional fused feature set (1011 features) relative to a small sample size (20 patients). Classification using Naive Bayes, Random Forest, and K-NN achieved 99\u0026thinsp;\u0026plusmn;\u0026thinsp;1% accuracy with 100% sensitivity and specificity. While demonstrating the power of multimodal feature fusion, gaps remain in scaling to larger, more diverse populations and in exploring end-to-end deep learning models without feature engineering. Compared to our model, this study highlights the performance gains achievable when combining clinical and morphological information. Insights suggest that integrating heterogeneous data sources with dimensionality reduction can substantially improve detection accuracy even with limited datasets, providing a practical framework for automated thalassemia screening.\u003c/p\u003e\u003cp\u003eCollectively, these studies demonstrate the promise of machine learning and deep learning for automated thalassemia detection. While prior work often focuses either on blood smear image analysis or RBC-based classification, few approaches integrate both modalities. Combining morphological and haematological features enhances diagnostic reliability and supports scalable, patient-centred screening systems, particularly in resource-limited settings such as Sri Lanka.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStudy Design and Sample Collection\u003c/h2\u003e\u003cp\u003e Following ethical clearance, data were collected from individuals attending the National Thalassemia Centre in Kurunegala, Sri Lanka. Participants were recruited through three main pathways: voluntary screening, which included individuals who independently sought thalassemia carrier testing; referrals from primary care physicians or hospital wards for those with clinically suspected anaemia or microcytic hypochromic red cell indices; and mass screening programs conducted in schools across the Kurunegala district, particularly targeting adolescent populations. The routine diagnostic workflow at the centre was used for the study without any alterations.\u003c/p\u003e\u003cp\u003eIn routine (non-programmatic) screening, a structured clinical history was first obtained, followed by a Full Blood Count (FBC). Individuals with microcytic, hypochromic red cell indices (Mean Corpuscular Volume [MCV]\u0026thinsp;\u0026lt;\u0026thinsp;77 fL and Mean Corpuscular Haemoglobin [MCH]\u0026thinsp;\u0026lt;\u0026thinsp;27 pg) were further investigated by peripheral blood smear examination. Smears were prepared by trained laboratory technicians from venous blood collected in Ethylenediaminetetraacetic acid (EDTA) tubes, stained using the Leishman, and examined by haematologists for morphological features suggestive of thalassemia, including target cells, pencil cells, and tear-drop cells. Suspected cases were subjected to quantitative haemoglobin analysis using High-Performance Liquid Chromatography (HPLC), which is the standard confirmatory test for β-thalassemia in Sri Lanka. Individuals with an HbA₂ level\u0026thinsp;\u0026gt;\u0026thinsp;3.5% were classified as β-thalassemia carriers.\u003c/p\u003e\u003cp\u003eFollowing completion of diagnostic confirmation, Full Blood Count (FBC) reports, blood smears, and HPLC results, materials that would otherwise be discarded, were collected for analysis. Sensitive personal identifiers were removed, and only anonymised report numbers were retained to link FBC, smear, and HPLC data.\u003c/p\u003e\u003cp\u003e Written informed consent was obtained from all participants prior to data use. For minors and individuals unable to provide consent, proxy consent was obtained from parents or legal guardians. All data were stored securely accessible only to authorised research personnel, and documentation of consent and ethical approval will be made available to the ethics review committee or journal upon request.\u003c/p\u003e\u003cp\u003eExclusion criteria included incomplete FBC reports, inability to retrieve the corresponding HPLC result, or absence of usable blood smear images. In addition, individuals who were HPLC-negative for β-thalassaemia but subsequently underwent further investigations for other haemoglobinopathies or alternative causes of anaemia were excluded to avoid diagnostic heterogeneity. No exclusions were made on the basis of prior blood transfusions or iron therapy.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003ePopulation Characteristics and Generalisability Considerations\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eAlthough screening services are available in multiple tertiary hospitals and dedicated haematology clinics across the country, the National Thalassemia Centre at Teaching Hospital Kurunegala functions as the largest national referral and screening hub. It hosts the highest number of registered patients and conducts the most extensive screening programmes in Sri Lanka, reflecting the highest reported regional prevalence of thalassaemia in areas surrounding Kurunegala. As a result, the study cohort is likely representative of the population most affected by β-thalassaemia within the national context.\u003c/p\u003e\u003cp\u003eAge groups were approximately equally represented, with a near-balanced sex distribution (79 females, 73 males), reducing the likelihood of sex- or age-related sampling bias within the cohort. However, the dataset was collected from a single geographical region, and demographic characteristics beyond age and sex, such as ethnicity, socioeconomic background, and place of residence, were not recorded. These factors may or may not influence red blood cell indices or peripheral smear morphology, and to date, no scientific evidence has demonstrated systematic variation in these haematological parameters attributable to such demographic variables in Sri Lankan populations. Nonetheless, the absence of these data limits the ability to fully assess potential demographic bias.\u003c/p\u003e\u003cp\u003eFurthermore, while some sources of variability were incorporated into the dataset through differences in operators and acquisition devices, the model has not yet been evaluated in populations served by other screening centres, laboratories using different equipment, or regions with differing genetic backgrounds. Therefore, future work should aim to include a more demographically diverse cohort through a multi-centre, nationwide study to enhance external validity and support broader clinical deployment.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eData Acquisition\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eTwo types of data were acquired: Red Blood Cell Indices and Blood Smear Imaging\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eRed Blood Cell Indices\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe following red cell indices and haematological parameters were collected for each participant: Haemoglobin (Hb), Red Blood Cell count (RBC), Mean Corpuscular Volume (MCV), Mean Corpuscular Haemoglobin (MCH), Mean Corpuscular Haemoglobin Concentration (MCHC), and Red cell Distribution Width (RDW). In addition, patient age and sex were recorded as shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The dataset comprised 152 participants, including 54 BTT-positive and 98 BTT-negative individuals. For each RBC parameter and age, the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD) and observed range are presented for the total cohort, and separately for male and female participants within each group.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDescriptive summary of the study dataset: Red Blood Cell indices and age stratified by β-thalassemia trait (BTT) status and sex.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eParameter\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eTotal Dataset\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c5\" namest=\"c3\"\u003e\u003cp\u003ePositive (BTT)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c8\" namest=\"c6\"\u003e\u003cp\u003eNegative (Normal)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTotal\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eTotal\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMCV\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e71.46\u0026thinsp;\u0026plusmn;\u0026thinsp;10.63\u003c/p\u003e\u003cp\u003e(51.4\u0026ndash;90.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e59.57\u0026thinsp;\u0026plusmn;\u0026thinsp;4.39\u003c/p\u003e\u003cp\u003e(51.4\u0026ndash;70.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e60.21\u0026thinsp;\u0026plusmn;\u0026thinsp;4.06\u003c/p\u003e\u003cp\u003e(53.9\u0026ndash;74.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e59.89\u0026thinsp;\u0026plusmn;\u0026thinsp;4.20\u003c/p\u003e\u003cp\u003e(51.4\u0026ndash;74.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e77.26\u0026thinsp;\u0026plusmn;\u0026thinsp;8.14\u003c/p\u003e\u003cp\u003e(54.3\u0026ndash;89.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e77.71\u0026thinsp;\u0026plusmn;\u0026thinsp;6.96\u003c/p\u003e\u003cp\u003e(56.1\u0026ndash;90.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e77.51\u0026thinsp;\u0026plusmn;\u0026thinsp;7.47\u003c/p\u003e\u003cp\u003e(54.3\u0026ndash;90.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMCH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e23.47\u0026thinsp;\u0026plusmn;\u0026thinsp;4.22\u003c/p\u003e\u003cp\u003e(15.1\u0026ndash;33.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19.01\u0026thinsp;\u0026plusmn;\u0026thinsp;1.42\u003c/p\u003e\u003cp\u003e(16.4\u0026ndash;21.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e19.09\u0026thinsp;\u0026plusmn;\u0026thinsp;1.34\u003c/p\u003e\u003cp\u003e(17.0\u0026ndash;23.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e19.05\u0026thinsp;\u0026plusmn;\u0026thinsp;1.37\u003c/p\u003e\u003cp\u003e(16.4\u0026ndash;23.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e25.75\u0026thinsp;\u0026plusmn;\u0026thinsp;3.56\u003c/p\u003e\u003cp\u003e(15.1\u0026ndash;31.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e25.80\u0026thinsp;\u0026plusmn;\u0026thinsp;3.01\u003c/p\u003e\u003cp\u003e(17.7\u0026ndash;33.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e25.78\u0026thinsp;\u0026plusmn;\u0026thinsp;3.25\u003c/p\u003e\u003cp\u003e(15.1\u0026ndash;33.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMCHC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e32.59\u0026thinsp;\u0026plusmn;\u0026thinsp;1.94\u003c/p\u003e\u003cp\u003e(14.3\u0026ndash;35.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31.97\u0026thinsp;\u0026plusmn;\u0026thinsp;0.64\u003c/p\u003e\u003cp\u003e(30.5\u0026ndash;32.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e31.83\u0026thinsp;\u0026plusmn;\u0026thinsp;0.76\u003c/p\u003e\u003cp\u003e(29.6\u0026ndash;33.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e31.90\u0026thinsp;\u0026plusmn;\u0026thinsp;0.70\u003c/p\u003e\u003cp\u003e(29.6\u0026ndash;33.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e33.25\u0026thinsp;\u0026plusmn;\u0026thinsp;1.58\u003c/p\u003e\u003cp\u003e(27.9\u0026ndash;35.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e32.72\u0026thinsp;\u0026plusmn;\u0026thinsp;2.67\u003c/p\u003e\u003cp\u003e(14.3\u0026ndash;34.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e32.96\u0026thinsp;\u0026plusmn;\u0026thinsp;2.26\u003c/p\u003e\u003cp\u003e(14.3\u0026ndash;35.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHb\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e12.00\u0026thinsp;\u0026plusmn;\u0026thinsp;2.11\u003c/p\u003e\u003cp\u003e(5.7\u0026ndash;19.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e11.24\u0026thinsp;\u0026plusmn;\u0026thinsp;1.63\u003c/p\u003e\u003cp\u003e(8.5\u0026ndash;13.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e10.58\u0026thinsp;\u0026plusmn;\u0026thinsp;1.07\u003c/p\u003e\u003cp\u003e(8.5\u0026ndash;12.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e10.91\u0026thinsp;\u0026plusmn;\u0026thinsp;1.40\u003c/p\u003e\u003cp\u003e(8.5\u0026ndash;13.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e13.30\u0026thinsp;\u0026plusmn;\u0026thinsp;2.59\u003c/p\u003e\u003cp\u003e(5.7\u0026ndash;19.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e12.00\u0026thinsp;\u0026plusmn;\u0026thinsp;1.63\u003c/p\u003e\u003cp\u003e(9.4\u0026ndash;17.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e12.57\u0026thinsp;\u0026plusmn;\u0026thinsp;2.19\u003c/p\u003e\u003cp\u003e(5.7\u0026ndash;19.9)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRBC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e5.16\u0026thinsp;\u0026plusmn;\u0026thinsp;0.73\u003c/p\u003e\u003cp\u003e(3.2\u0026ndash;7.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5.91\u0026thinsp;\u0026plusmn;\u0026thinsp;0.77\u003c/p\u003e\u003cp\u003e(3.9\u0026ndash;7.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e5.52\u0026thinsp;\u0026plusmn;\u0026thinsp;0.46\u003c/p\u003e\u003cp\u003e(4.6\u0026ndash;6.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5.72\u0026thinsp;\u0026plusmn;\u0026thinsp;0.65\u003c/p\u003e\u003cp\u003e(3.9\u0026ndash;7.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e5.12\u0026thinsp;\u0026plusmn;\u0026thinsp;0.60\u003c/p\u003e\u003cp\u003e(3.2\u0026ndash;6.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e4.69\u0026thinsp;\u0026plusmn;\u0026thinsp;0.52\u003c/p\u003e\u003cp\u003e(3.5\u0026ndash;5.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e4.88\u0026thinsp;\u0026plusmn;\u0026thinsp;0.60\u003c/p\u003e\u003cp\u003e(3.2\u0026ndash;6.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRDW\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e15.56\u0026thinsp;\u0026plusmn;\u0026thinsp;2.17\u003c/p\u003e\u003cp\u003e(11.4\u0026ndash;24.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e17.31\u0026thinsp;\u0026plusmn;\u0026thinsp;1.24\u003c/p\u003e\u003cp\u003e(14.8\u0026ndash;20.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16.45\u0026thinsp;\u0026plusmn;\u0026thinsp;1.23\u003c/p\u003e\u003cp\u003e(12.5\u0026ndash;19.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e16.88\u0026thinsp;\u0026plusmn;\u0026thinsp;1.30\u003c/p\u003e\u003cp\u003e(12.5\u0026ndash;20.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e14.39\u0026thinsp;\u0026plusmn;\u0026thinsp;1.74\u003c/p\u003e\u003cp\u003e(11.4\u0026ndash;21.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e15.24\u0026thinsp;\u0026plusmn;\u0026thinsp;2.48\u003c/p\u003e\u003cp\u003e(12.7\u0026ndash;24.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e14.86\u0026thinsp;\u0026plusmn;\u0026thinsp;2.21\u003c/p\u003e\u003cp\u003e(11.4\u0026ndash;24.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e22.29\u0026thinsp;\u0026plusmn;\u0026thinsp;12.85\u003c/p\u003e\u003cp\u003e(1.0\u0026ndash;63.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24.02\u0026thinsp;\u0026plusmn;\u0026thinsp;15.81\u003c/p\u003e\u003cp\u003e(1.0\u0026ndash;63.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e19.29\u0026thinsp;\u0026plusmn;\u0026thinsp;12.46\u003c/p\u003e\u003cp\u003e(2.8\u0026ndash;45.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e21.65\u0026thinsp;\u0026plusmn;\u0026thinsp;14.30\u003c/p\u003e\u003cp\u003e(1.0\u0026ndash;63.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e22.37\u0026thinsp;\u0026plusmn;\u0026thinsp;13.21\u003c/p\u003e\u003cp\u003e(1.0\u0026ndash;48.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e22.84\u0026thinsp;\u0026plusmn;\u0026thinsp;11.22\u003c/p\u003e\u003cp\u003e(1.0\u0026ndash;59.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e22.63\u0026thinsp;\u0026plusmn;\u0026thinsp;12.07\u003c/p\u003e\u003cp\u003e(1.0\u0026ndash;59.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eKey observations from the dataset indicate that BTT-positive individuals exhibit markedly lower MCV values and MCH values compared to BTT-negative participants, who show higher averages for both parameters. Additionally, BTT-positive individuals demonstrate a higher RBC count than the negative group reflecting the compensatory erythropoiesis characteristic of thalassemia carriers. RDW and MCHC values showed additional distinctions between groups. BTT-positive individuals had higher RDW values compared to BTT-negative participants, indicating greater variability in red cell size. Differences in MCHC were more modest, although slightly lower values were observed among BTT-positive cases compared with the negative group.\u003c/p\u003e\u003cp\u003eThe age distribution of participants ranged from 1 to 63 years, with similar mean ages across groups. BTT-positive individuals had a mean age of 21.65\u0026thinsp;\u0026plusmn;\u0026thinsp;14.30 years, closely aligned with the mean age of 22.63\u0026thinsp;\u0026plusmn;\u0026thinsp;12.07 years observed in the negative group, suggesting no notable age-related sampling bias. Sex distribution was also balanced across the cohort, with male and female participants represented in approximately equal proportions in both BTT-positive and BTT-negative groups. Minor sex-related differences were noted Hb and RBC, with males showing slightly higher values. However, key screening parameters such as MCV and MCH demonstrated minimal variation by sex, indicating that these indices are largely independent of sex within this dataset.\u003c/p\u003e\u003cp\u003eThe dataset provides a comprehensive representation of RBC indices for both BTT-positive and BTT-negative individuals, making it suitable for training and validating the dual-modal machine learning pipeline. The distribution of values also highlights the discriminative potential of CBC indices for initial β-thalassemia screening.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eBlood Smear Imaging\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eBlood smear images were captured using two approaches - Smartphone-assisted microscopy and Standard microscope camera.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSmartphone-assisted microscopy\u003c/b\u003e \u0026ndash; Three different smartphones were used: Samsung Galaxy M21, Apple iPhone 15, and Xiaomi Redmi Note 13. Three operators independently captured images, with slides randomly assigned in approximately equal proportions to simulate real-world variability in community and laboratory settings. Each phone camera was manually aligned with the microscope eyepiece without customised adapters or camera settings, and the smears were focused under the \u0026times;100 oil immersion objective. The microscope lamp served as the sole illumination source, and no exposure, colour, or white-balance calibration was performed prior to acquisition. Images were saved in the native formats of the devices (HEIC or PNG) and subsequently converted to JPEG during preprocessing. This method emulates routine haematological practice, with the smartphone effectively substituting for the human eye.\u003c/p\u003e\u003cp\u003e\u003cb\u003eStandard microscope camera\u003c/b\u003e \u0026ndash; An AxioCam ERc5s camera attached to a standard light microscope was used to acquire reference-quality images at the same \u0026times;100 magnification. Relevant camera metadata, including pixel accuracy (0.0039 mm), image size (up to 5712 \u0026times; 4284 pixels), light path, and lens information, were retained for reproducibility.\u003c/p\u003e\u003cp\u003eFor each smear, 25\u0026ndash;30 non-overlapping images were obtained from the morphological \u0026ldquo;zone of interest,\u0026rdquo; i.e., the area between the body and tail of the smear where red cells are evenly distributed, non-overlapping, and morphologically intact. The head and tail regions, where cells are either too crowded or distorted, were excluded following standard haematology practice. Smear examination was performed systematically using a Z-pattern scanning approach to ensure comprehensive coverage. Representative blood smear images used in the study is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eCaptured images were reviewed, and any blurred or artefact-containing images were excluded. The final dataset comprised 5,198 high-quality images from 152 blood smear samples, each sample labelled as thalassemia carrier or non-carrier according to the corresponding HPLC results.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eData Preprocessing\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eHaematological Data (Red Blood Cell indices)\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003e The FBC reports obtained from the laboratory information system were exported into a structured format and underwent several preprocessing steps. Personal identifiers were removed to ensure anonymisation, with only the report ID retained to match smear images and HPLC results. Reports containing incomplete or missing red cell indices were excluded, and any outliers falling outside physiologically plausible limits were flagged and manually verified against the original records. Numerical variables, including MCV, MCH, MCHC, haemoglobin (Hb), RDW, and RBC count, were normalised using z-score standardization. Patient sex was encoded as a binary variable (0\u0026thinsp;=\u0026thinsp;female, 1\u0026thinsp;=\u0026thinsp;male). HPLC-confirmed carrier status served as the ground truth label, with carriers coded as 1 and non-carriers as 0. These steps resulted in a structured tabular dataset of RBC indices aligned with each participant\u0026rsquo;s corresponding smear images.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eBlood Smear Images\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe raw smear images underwent several preprocessing steps to ensure consistency and suitability for deep learning analysis. Initial quality control filtering was performed to remove images affected by motion blur, poor focus, or uneven staining, using both visual inspection and automated quality checks. To minimise background artefacts, a rectangular Region of Interest (ROI) corresponding to the largest inscribed area within the circular microscope field was extracted using the Hough Circle Transform in OpenCV, allowing automatic cropping of the maximal area containing red blood cells. All images were then resized to 224 \u0026times; 224 pixels to meet the input requirements of the convolutional neural network (CNN) and maintain computational efficiency. To address class imbalance and improve model generalisation, data augmentation techniques\u0026mdash;including random rotations (\u0026plusmn;\u0026thinsp;15\u0026deg;), horizontal and vertical flips, minor translations and zooming, and controlled brightness and contrast variations\u0026mdash;were applied (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Following preprocessing, the dataset was divided into training (70%) and independent testing (30%) subsets, with patient-level stratification to prevent data leakage by ensuring that images from the same individual did not appear across different partitions.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eModel Architecture and Training\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe proposed model architecture was designed with careful consideration of the requirements of a population-level screening tool. In such settings, maximising sensitivity is paramount to ensure that no β-thalassemia carrier is missed, even at the expense of specificity. A false-negative result could delay diagnosis and increase the risk of undetected carrier marriages, whereas false positives can be resolved at the confirmatory testing stage.\u003c/p\u003e\u003cp\u003eIn addition to achieving high diagnostic accuracy, the architecture was intentionally designed to meet the practical challenges of real-world healthcare delivery, especially in resource-limited and rural settings where trained haematologists and advanced laboratory infrastructure may be scarce. The system therefore aimed to minimise the workload on clinicians and laboratory staff by reducing the need for unnecessary smear preparation and microscopic examination, while also lowering overall costs and reagent consumption to support large-scale screening initiatives. Furthermore, portability and scalability were prioritised to enable deployment across multiple platforms\u0026mdash;from mobile devices to desktop systems\u0026mdash;ensuring broad accessibility. Finally, the pipeline was optimised for rapid turnaround, delivering near real-time results suitable for screening during outreach activities and school-based screening programmes.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eOverview of the Dual-Modal Framework\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eA two-step dual-modal pipeline was implemented.\u003c/p\u003e\u003cp\u003e\u003cb\u003eStep 1\u003c/b\u003e of the screening pipeline involves haematological analysis based RBC indices together with patient age and sex. This stage was optimised for high specificity to ensure that individuals flagged as positive could be confidently referred directly for confirmatory HPLC testing without requiring additional intermediate assessments. Only those classified as negative at this stage proceed to the secondary evaluation.\u003c/p\u003e\u003cp\u003e\u003cb\u003eStep 2\u003c/b\u003e of the pipeline focuses on morphological screening using blood smear images. For individuals not identified as positive in the first stage, peripheral blood smears are prepared, digitised, and analysed using a CNN-based model trained to detect subtle morphological features associated with thalassemia. This stage was intentionally designed to maximise sensitivity, ensuring that borderline or atypical cases are captured and the risk of missed carriers is minimised.\u003c/p\u003e\u003cp\u003eBy integrating the two modalities in a sequential architecture, the pipeline ensures high diagnostic reliability while maintaining operational efficiency. At the same time, the overall workload for laboratory staff is significantly reduced, as only a small subset of individuals who screen negative initially require smear preparation and further microscopic analysis. This design also optimises time, cost, and reagent consumption, making the workflow well suited for mass-screening programmes where preserving resources without compromising clinical robustness is essential.\u003c/p\u003e\u003cp\u003eThe dual-modal strategy is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e\u003cp\u003eProcessing of Red Blood Cell Indices and Blood Smear Images\u003c/p\u003e\u003cp\u003eSepts followed in processing the Blood Cell Indices and Blood Smear Images are discussed below.\u003c/p\u003e\u003cp\u003eProcessing Red Blood Cell Indices\u003c/p\u003e\u003cp\u003eThe first component of the dual-modal model processes Red Blood Cell (RBC) indices, along with patient age and sex.\u003c/p\u003e\u003cp\u003eThe preprocessing and model selection workflow included the following steps:\u003c/p\u003e\u003cp\u003eI. Dimensionality Reduction (PCA):\u003c/p\u003e\u003cp\u003ePrincipal Component Analysis (PCA) was applied to the RBC dataset to identify the most informative combinations of red cell indices. PCA helped reduce redundancy between correlated parameters while retaining the variance most relevant for discriminating carriers from non-carriers. The resulting principal components were analysed to guide feature selection for downstream model training.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eII. Algorithm Evaluation:\u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTo identify the optimal classifier for RBC-based screening, nine machine learning algorithm\u003cb\u003es\u003c/b\u003e were trained and compared using different permutations of RBC parameters: Logistic Regression (LR), k-Nearest Neighbors (k-NN), Decision Tree (DT), Random Forest (RF), Gradient Boosting Machine (GBM), AdaBoost, Support Vector Machine (SVM), Naive Bayes (NB), Multilayer Perceptron (MLP). Each model was evaluated on training and validation sets using cross-validation to ensure robustness, and performance metrics, including accuracy, precision, recall, and area under the ROC curve (AUC), were compared.\u003c/p\u003e\u003c/div\u003e\u003cp\u003eIII. Feature Selection \u0026amp; Optimisation:\u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eBased on performance, the most informative subset of RBC parameters was selected to maximise specificity in this first step, ensuring that predicted positives could be confidently sent for confirmatory HPLC without unnecessary smear preparation. Hyperparameters for each algorithm were tuned using grid search or Bayesian optimisation to achieve optimal predictive performance.\u003c/p\u003e\u003cp\u003eProcessing Blood Smear Images\u003c/p\u003e\u003cp\u003e For model development, blood smears were prepared for 152 out of 163 participants, comprising 98 negative and 54 positive cases. All images underwent the preprocessing steps described previously, including ROI extraction using Hough Circle Transform, resolution standardisation (224 \u0026times; 224 pixels), and data augmentation to increase variability and reduce class imbalance.\u003c/p\u003e\u003cp\u003eThe image component employed transfer learning to leverage pretrained convolutional neural networks (CNNs) for feature extraction. Multiple architectures were evaluated: VGG16, ResNet50, DenseNet121, MobileNetV2 and ConvNeXtTiny. Each architecture was fine-tuned and hyperparameters were systematically varied using grid search or random search to identify the optimal configuration for each model.\u003c/p\u003e\u003cp\u003eModels were trained using training and validation subsets, with patient-level stratification to prevent data leakage. Performance metrics, including accuracy, sensitivity, specificity, F1-score, and area under the ROC curve (AUC), were calculated to identify the model with the best sensitivity, ensuring that borderline and atypical positive cases were reliably detected.\u003c/p\u003e\u003cp\u003eThe selected image model forms the second-stage screening component.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eModel Training and Optimisation\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eTraining and optimization of each model is described below.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eRBC Indices Model\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe processed RBC dataset was used to train a machine learning model for β-thalassemia carrier classification. The model was implemented in Python 3.9, using scikit-learn v1.2, along with NumPy, Pandas, Matplotlib, and Seaborn for evaluation and visualisation. The MLP architecture comprised hidden layers with ReLU activation, fine-tuned to optimise performance for binary classification. To ensure balanced representation of carriers and non-carriers and to prevent data leakage, stratified 5-fold cross-validation was employed throughout the optimisation process. Performance of the model was evaluated using a comprehensive set of custom metrics, including sensitivity, specificity, accuracy, F1-score, ROC\u0026ndash;AUC, and the Matthews correlation coefficient.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eBlood Smear Image Model\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe second component of the dual-modal model processes blood smear images and was implemented in Python 3.9 using TensorFlow/Keras 2.x alongside standard scientific libraries such as NumPy, Pandas, and Matplotlib. This component was designed to ensure patient-level stratified training and evaluation, preventing data leakage across folds or between images belonging to the same individual. Pretrained convolutional neural networks were used as feature extractors. Their base layers were initially frozen, and a custom classifier head was added, consisting of a global average pooling layer, a dense layer with 256 ReLU-activated units, a dropout layer with a rate of 0.5, and a final softmax output layer with two units for binary classification. After initial training, the top four layers of each CNN backbone were unfrozen for fine-tuning. Training followed a two-stage strategy: 50 epochs with a frozen backbone using the Adam optimiser (learning rate\u0026thinsp;=\u0026thinsp;0.001), followed by 50 fine-tuning epochs with a reduced learning rate of 1e-5. Images were processed at 224 \u0026times; 224 resolution with a batch size of 32, and extensive data augmentation\u0026mdash;such as random flips, brightness and contrast adjustments, saturation shifts, and rotations\u0026mdash;was applied to improve generalisation.\u003c/p\u003e\u003cp\u003eTo maintain rigorous reproducibility, patient-level splitting was enforced using train_test_split to ensure that no images from the same patient appeared in both training and validation sets. A custom PatientFolderDataGenerator, implemented as a subclass of tf.keras.utils.Sequence, handled batch generation, preprocessing, augmentation, and GPU-accelerated loading. Training incorporated callbacks including EarlyStopping (patience\u0026thinsp;=\u0026thinsp;10, monitoring validation accuracy), ReduceLROnPlateau (factor\u0026thinsp;=\u0026thinsp;0.5, patience\u0026thinsp;=\u0026thinsp;5), and ModelCheckpoint for saving the best-performing model.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eImplementation and Deployment\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe proposed dual-modal pipeline was designed with practical deployment in low-resource and primary care settings in mind. A key feature of the system is that it does not require specialised imaging hardware; smartphone-based microscopy was performed using standard laboratory microscopes without custom adapters, modified optics, or controlled imaging environments. This approach reflects the conditions under which peripheral smear evaluation is routinely carried out in many regional hospitals and rural clinics in Sri Lanka, enabling potential adoption without additional capital cost.\u003c/p\u003e\u003cp\u003eIn a real-world workflow, peripheral smear images captured using a smartphone or existing microscope camera could be uploaded to a local workstation or mobile application interface, where automated image analysis and FBC feature processing would be performed. The computational requirements of the final model allow inference to be executed either on-device for newer smartphones or via a low-resource edge or cloud-based server, making the system adaptable to settings with variable technical infrastructure. This supports potential deployment in rural primary care clinics that lack on-site haematologists, providing decision support for early carrier screening and referral.\u003c/p\u003e\u003cp\u003eThe intended operational use-case involves assisting healthcare workers by flagging individuals with a high likelihood of β-thalassaemia trait, thereby prioritising confirmatory testing and improving screening throughput. However, the current implementation remains a research prototype, and several constraints must be addressed prior to clinical integration. These include regulatory approval, standardisation of smear preparation and staining quality, operator training for consistent image capture, and integration with electronic laboratory information systems. Connectivity limitations in remote settings may also affect workflows reliant on cloud-based processing.\u003c/p\u003e\u003cp\u003eFuture development will focus on a full mobile application interface, streamlined clinician-facing reporting, and incorporation into existing national thalassaemia screening programmes. Prospective field evaluation and usability testing will be required to ensure that the system is safe, scalable, and operationally feasible in real-world clinical environments.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eRBC Indices Data Classification\u003c/p\u003e\u003cp\u003eThe PCA results revealed that MCV and MCH contributed most strongly to the first two principal components, indicating that these parameters alone capture the majority of the variance distinguishing β-thalassemia trait from normal individuals. This observation aligns with clinical knowledge, as microcytosis (low MCV) and hypochromia (low MCH) are hallmark features of thalassemia carriers.\u003c/p\u003e\u003cp\u003eTo further illustrate this, all nine algorithms were initially trained using only MCV and MCH as input features. The resulting classification performance is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, which displays the separation of positive and negative cases by each algorithm in the test set. The graph demonstrates that even with just these two parameters, the models are able to distinguish BTT-positive from normal cases with reasonable accuracy, highlighting the predictive power of MCV and MCH in preliminary screening. This also informed subsequent model development, where additional parameters were incorporated to further improve classification metrics such as sensitivity, specificity, and overall accuracy.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the performance metrics of various machine learning models trained on RBC parameters for differentiating BTT-positive (BTT) and normal cases. Overall, the models demonstrated high performance, with Classification Accuracies exceeding 92% across most configurations and AUC values consistently above 95%, reflecting strong discrimination between BTT-positive and normal cases. Sensitivity and specificity were well-balanced, with specificity ranging from approximately 93\u0026ndash;94% and sensitivity from 88\u0026ndash;94%, which is critical for a screening approach where minimising false negatives is essential. True positive counts (BTT \u0026rarr; BTT) ranged from 48 to 50, and true negative counts (Normal \u0026rarr; Normal) ranged from 84 to 85, indicating a low rate of misclassification. Models trained on slightly different subsets of RBC parameters, including age, sex, or history of iron supplementation, showed minor variations in metrics, suggesting robustness to feature selection. Among the algorithms, the MLP (100) model achieved the highest AUC (~\u0026thinsp;96.88%) with well-balanced sensitivity and specificity, whereas Naive Bayes models exhibited the highest sensitivity (~\u0026thinsp;92.45%) while maintaining competitive overall accuracy. Logistic Regression and SVM models also performed strongly and offer the additional advantage of interpretability.\u003c/p\u003e\u003cp\u003eTable 2 Performance metrics of various machine learning models trained on RBC parameters for differentiating BTT-positive (BTT) and normal cases.\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" width=\"724\" height=\"238\"\u003e\u003c/p\u003e\n\u003cp\u003eBlood Smear Image Analysis\u003c/p\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarises the performance of each CNN model on the independent test set, with all metrics reported along with their 95% confidence intervals.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance of each CNN model on the independent test set, with all metrics reported along with their 95% confidence intervals.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCA (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eF1 (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003ePrecision (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eRecall (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eSpecificity (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eMCC (%)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMobileNetV2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e37.2% (23.0\u0026ndash;53.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e52.6% (35.3\u0026ndash;67.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e35.7% (21.6\u0026ndash;52.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e100.0% (78.2\u0026ndash;100.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3.6%\u003c/p\u003e\u003cp\u003e(0.1\u0026ndash;18.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.113\u003c/p\u003e\u003cp\u003e(0.000\u0026ndash;0.216)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eConvNeXtTiny\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e37.8% (23.8\u0026ndash;53.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e51.7% (34.6\u0026ndash;65.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e34.9% (21.0\u0026ndash;50.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e100.0% (78.2\u0026ndash;100.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e6.7% (0.8\u0026ndash;22.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.152\u003c/p\u003e\u003cp\u003e(0.000\u0026ndash;0.264)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInceptionV3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e60.5% (44.4\u0026ndash;75.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e63.8% (45.5\u0026ndash;77.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e46.9% (29.1\u0026ndash;65.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e100.0% (78.2\u0026ndash;100.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e39.3% (21.5\u0026ndash;59.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.429\u003c/p\u003e\u003cp\u003e(0.283\u0026ndash;0.586)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVGG16\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e72.1% (56.3\u0026ndash;84.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e68.4% (48.5\u0026ndash;83.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e56.5% (34.5\u0026ndash;76.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e86.7% (59.5\u0026ndash;98.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e64.3% (44.1\u0026ndash;81.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.487\u003c/p\u003e\u003cp\u003e(0.229\u0026ndash;0.715)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eResNet50\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e65.1% (49.1\u0026ndash;79.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e63.4% (43.9\u0026ndash;79.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e50.0% (29.9\u0026ndash;70.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e86.7% (59.5\u0026ndash;98.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e53.6% (33.9\u0026ndash;72.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.392\u003c/p\u003e\u003cp\u003e(0.131\u0026ndash;0.624)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDenseNet121\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e62.8% (46.7\u0026ndash;77.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e61.9% (42.4\u0026ndash;77.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e48.1% (28.7\u0026ndash;68.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e86.7% (59.5\u0026ndash;98.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e50.0% (30.6\u0026ndash;69.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.362\u003c/p\u003e\u003cp\u003e(0.106\u0026ndash;0.590)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eVGG16 achieved the highest classification accuracy (72.1%) and the highest specificity (64.3%) among the models, while maintaining a strong recall (86.67%), indicating a balanced ability to identify both positive and negative samples. Models such as MobileNetV2, ConvNeXtTiny, and InceptionV3 reached perfect recall (100%), but their specificity was extremely low, leading to a high number of false positives and a risk of unnecessary follow-up testing. ResNet50 and DenseNet121 showed moderate performance across all metrics, offering more balanced but overall lower accuracy compared to VGG16.\u003c/p\u003e\u003cp\u003eOverall, the results demonstrate a clear trade-off between recall and specificity among the CNN models. Since the primary objective of this step is screening, avoiding missed positive cases is critical. Therefore, VGG16 was selected for blood smear analysis because it provides high sensitivity (recall 86.67%), ensuring that most potential β-thalassemia carriers are flagged for confirmation testing. Although this increases false positives, the pipeline tolerates this, as all flagged cases will undergo HPLC confirmation. This approach maximises patient safety and supports effective population screening within the dual-modal workflow.\u003c/p\u003e\u003cp\u003eOverall, the results demonstrate a clear trade-off between recall and specificity among the CNN models. Models such as ConvNeXtTiny and MobileNetV2 achieved very high recall, correctly identifying nearly all positive cases, but their extremely low specificity led to many false positives and unnecessary follow-up testing. In contrast, VGG16 provided the best balance, with the highest classification accuracy and specificity while maintaining high sensitivity (86.67%), reducing the risk of missed positive cases.\u003c/p\u003e\u003cp\u003eSince the primary objective of this second step is screening, avoiding false negatives is critical. Therefore, VGG16 was selected as the preferred model for blood smear analysis, as its high sensitivity ensures that most potential β-thalassemia carriers are flagged for confirmation testing. Although this increases false positives, the pipeline is designed to tolerate this, as all flagged cases will undergo HPLC confirmation. This approach maximises patient safety, aligns with clinical screening practice, and supports the overall efficiency of the dual-modal pipeline.\u003c/p\u003e\u003cp\u003eError Analysis\u003c/p\u003e\u003cp\u003eTo further interpret the performance of the proposed thalassemia screening pipeline, an error analysis was conducted to examine the characteristics and potential causes of model misclassifications. This analysis aimed to identify patterns among false positives and false negatives across both the RBC-based and image-based models, evaluate the complementary effects of the two-step approach, and reveal factors contributing to residual errors. Understanding these limitations provides valuable insight into model behaviour and guides future improvements in data preprocessing, feature selection, and threshold optimisation.\u003c/p\u003e\u003cp\u003eError Analysis of Red Blood Cell Indices-based MLP model\u003c/p\u003e\u003cp\u003eTo further investigate the misclassifications of the Red Blood Cell Indices-based MLP model, scatter plots were generated for multiple pairs of haematological parameters. Each plot highlights true positives (TP), true negatives (TN), false positives (FP), and false negatives (FN), allowing visual assessment of the model\u0026rsquo;s decision boundaries and error patterns. The plots also indicate patient sex and adjust point transparency to emphasise misclassified cases. This multi-pair approach provides an intuitive way to explore the distribution of misclassified samples across different parameter combinations, identify regions where the model struggles, and guide potential improvements in feature selection or thresholding. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e shows Multi-pair scatter plots showing MLP model predictions on the RBC dataset. Visualisations illustrate the distribution of errors, reveal parameter regions contributing to misclassifications, and provide insights for refining the screening pipeline.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAnalysis of the MLP misclassifications revealed that false negatives (FNs) cluster in regions of the MCV vs MCH scatter plot that largely overlap with true negatives (TNs). These samples exhibited relatively higher MCV and MCH values compared to true positives, while their RBC counts were lower than those of typical positive cases but similar to those of negative samples. Biologically, this pattern may reflect borderline or atypical β-thalassemia carriers whose red blood cell indices do not fall within classical microcytic and hypochromic ranges. Such cases are known to occur in carriers with mild phenotypic expression, coexisting iron deficiency, or recent blood transfusions, leading to partial normalisation of haematological parameters. In the present dataset, the observed false negatives did not have recent blood transfusions or iron treatment, suggesting that their borderline haematological profiles may be due to mild genetic variants, heterozygosity with less severe mutation expression, or individual physiological compensation mechanisms that partially normalise RBC indices. Consequently, the model may classify these borderline carriers as negative, illustrating the inherent limitation of using Red Blood Cell indices alone for precise detection in all patient subpopulations.\u003c/p\u003e\u003cp\u003eConversely, false positives (FPs) were primarily located within regions characteristic of true positives, exhibiting lower MCV and MCH values alongside higher RBC counts. This pattern suggests that while these individuals do not carry the β-thalassemia trait, their haematological parameters resemble those of carriers, possibly due to benign microcytosis, variations in haemoglobin concentration, or other haematological conditions. Such overlap underscores the trade-off between sensitivity and specificity in screening approaches and highlights the necessity of confirmatory testing to accurately distinguish true carriers from individuals with overlapping but non-pathological profiles.\u003c/p\u003e\u003cp\u003eError Analysis of Image-Based Model (VGG-16)\u003c/p\u003e\u003cp\u003eTo further interpret the performance of the image-based screening step, misclassifications from the VGG-16 model were examined. False negatives (FNs) and false positives (FPs) were analysed in relation to morphological characteristics observed in the blood smear images, providing insight into patterns that contribute to residual errors. This analysis aimed to identify visual features associated with borderline cases, evaluate limitations in feature extraction by the CNN, and highlight scenarios where morphological ambiguity affects predictive accuracy.\u003c/p\u003e\u003cp\u003eGrad-CAM visualisations were generated for representative true negative (TN), true positive (TP), false positive (FP), and false negative (FN) samples to investigate the image-based model\u0026rsquo;s attention patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). For the TP sample (ID: 63542, Positive Probability: 99.29%), attention was strongly focused on the central regions of RBCs, particularly highlighting morphological features consistent with β-thalassemia. Similarly, the TN sample (ID: 63769, Positive Probability: 1.71%) showed minimal attention on RBCs, with the model correctly ignoring normal cell morphology. In contrast, the FP sample (ID: 63627, Positive Probability: 91.32%) exhibited concentrated attention on RBC clusters, likely misinterpreting normal variations or overlapping cells as thalassemia-like features, resulting in high predicted probability despite the sample being negative. The FN sample (ID: 4303, Positive Probability: 11.72%) displayed diffuse and less distinct attention across RBCs, indicating that subtle or borderline morphological abnormalities were insufficient for confident detection. Collectively, these patterns suggest that misclassifications arise from either over-interpretation of normal variations (FPs) or under-recognition of mild, atypical morphological features (FNs), highlighting inherent limitations of relying solely on image morphology for screening in borderline cases.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eEvaluation of Alternative Screening Pipelines\u003c/p\u003e\u003cp\u003eTo further evaluate the robustness and optimal configuration of the dual-modal screening system, two alternative pipeline architectures were tested by altering the order of Red Blood Cell Indices and image-based analysis. This served as an ablation-style assessment to determine how sequencing the analytical stages influences diagnostic accuracy, sensitivity, and specificity.\u003c/p\u003e\n\u003ch3\u003ePipeline 1 (Red Blood Cell Indices Analysis → Image Analysis):\u003c/h3\u003e\n\u003cp\u003eIn this configuration, the Multi-Layer Perceptron (MLP) model based on full blood count parameters was applied first. Samples classified as negative were subsequently analysed using the VGG-16 CNN model. This approach aimed to minimise the number of smear analyses required while ensuring that no β-thalassemia carriers were missed. The combined pipeline achieved 100% sensitivity and 60.7% specificity, indicating that all positive cases were successfully detected but at the cost of increased false positives. Importantly, this configuration reduced the overall need for smear preparation and image processing by 37.7%, representing a practical advantage in large-scale screening workflows.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance metrics of the Red Blood Cell Indices \u0026rarr; Image Analysis pipeline for β-thalassemia screening, with all metrics reported along with their 95% confidence intervals. The MLP model was applied to Red Blood Cell parameters first, followed by VGG-16\u0026ndash;based image analysis for Red Blood Cell indices analysis-negative cases.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStep\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSensitivity (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eSpecificity (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eClassification Accuracy (%)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFirst step\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRed Blood Cell Indices Analysis (MLP)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e88.2%\u003c/p\u003e\u003cp\u003e(63.6\u0026ndash;98.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e92.9%\u003c/p\u003e\u003cp\u003e(76.5\u0026ndash;99.1%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e91.1%\u003c/p\u003e\u003cp\u003e(78.8\u0026ndash;97.5%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSecond step\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eImage Analysis (VGG-16)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100.0%\u003c/p\u003e\u003cp\u003e(15.8\u0026ndash;100%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e57.7%\u003c/p\u003e\u003cp\u003e(36.9\u0026ndash;76.6%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e60.7%\u003c/p\u003e\u003cp\u003e(40.6\u0026ndash;78.5%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOverall Pipeline\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100%\u003c/p\u003e\u003cp\u003e(80.5\u0026ndash;100%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e60.7%\u003c/p\u003e\u003cp\u003e(40.6\u0026ndash;78.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e75.6%\u003c/p\u003e\u003cp\u003e(60.5\u0026ndash;87.1%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\n\u003ch3\u003ePipeline 2 (Image Analysis → Red Blood Cell Indices Analysis):\u003c/h3\u003e\n\u003cp\u003eThe second configuration reversed the order, applying VGG-16\u0026ndash;based image analysis first and subsequently re-evaluating negative cases with the MLP model. This sequence also achieved \u003cb\u003e100% sensitivity\u003c/b\u003e, confirming that no positive samples were missed. Although specificity increased slightly to \u003cb\u003e64.3%\u003c/b\u003e, all samples required blood smear preparation and image processing. This introduced \u003cb\u003esignificant manual effort, time consumption, and computational workload\u003c/b\u003e. Consequently, despite maintaining perfect sensitivity, the operational overhead makes this architecture less practical for large-scale or resource-limited screening programs.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance metrics of the Image Analysis \u0026rarr; Red Blood Cell Indices analysis pipeline for β-thalassemia screening. VGG-16\u0026ndash;based image analysis was applied first, followed by MLP analysis for image-negative cases.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStep\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSensitivity (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eSpecificity (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eClassification Accuracy (%)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFirst step\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eImage Analysis (VGG-16)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e88.2%\u003c/p\u003e\u003cp\u003e(63.6\u0026ndash;98.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e64.3%\u003c/p\u003e\u003cp\u003e(44.1\u0026ndash;81.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e73.3%\u003c/p\u003e\u003cp\u003e(58.1\u0026ndash;85.4%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSecond step\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRed Blood Cell Indices Analysis (MLP)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100.0%\u003c/p\u003e\u003cp\u003e(15.8\u0026ndash;100.0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e94.4%\u003c/p\u003e\u003cp\u003e(72.7\u0026ndash;99.9%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e95%\u003c/p\u003e\u003cp\u003e(75.1\u0026ndash;99.9%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOverall Pipeline\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e100.0%\u003c/p\u003e\u003cp\u003e(80.5\u0026ndash;100.0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e64.3%\u003c/p\u003e\u003cp\u003e(44.1\u0026ndash;81.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e77.8%\u003c/p\u003e\u003cp\u003e(62.9\u0026ndash;88.8%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\n\u003ch3\u003eComparative Interpretation:\u003c/h3\u003e\n\u003cp\u003eBoth pipelines maintained perfect sensitivity, confirming the value of combining haematological and morphological features for comprehensive β-thalassemia screening. However, starting with Red Blood Cell indices analysis followed by image evaluation of negatives provided greater operational efficiency by reducing smear preparation and processing requirements. This indicates that the Red Blood Cell Indices \u0026rarr; Image pipeline is more suitable for scalable, resource-conscious screening in clinical and public health settings.\u003c/p\u003e\u003cp\u003eOverall Dual-Modal Pipeline Performance\u003c/p\u003e\u003cp\u003eThe proposed dual-modal screening pipeline combines Red Blood Cell indices analysis and blood smear image classification to maximise sensitivity while reducing unnecessary laboratory workload. In this study, all positive cases (BTT carriers) were successfully identified, with zero missed cases, demonstrating that the pipeline is highly reliable for screening purposes. By first using the Red Blood Cell indices analysis model to classify cases, only the negative cases from the first step were forwarded to blood smear analysis. Consequently, the number of blood smear preparations was reduced from 152 total samples to 97, effectively saving 37% of laboratory time, reagents, and personnel effort.\u003c/p\u003e\u003cp\u003eThis two-step approach ensures that almost all positive cases are flagged for confirmation while minimising unnecessary laboratory procedures, making the pipeline both practical and cost-effective, particularly in settings with limited access to haematologists or specialised laboratory infrastructure. Additionally, by using a high-sensitivity VGG16 model for the second step, the pipeline preserves patient safety without compromising screening efficiency.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study demonstrates the feasibility and potential utility of a dual-modal machine learning pipeline for β-thalassemia screening in Sri Lanka, combining Red Blood Cell (RBC) indices and peripheral blood smear images. By integrating these two complementary modalities, the system leverages both quantitative haematological features and morphological patterns, providing a sensitive, cost-effective, and scalable approach suitable for deployment in low-resource settings.\u003c/p\u003e\u003cp\u003eDiagnostic Performance and Key Findings\u003c/p\u003e\u003cp\u003eAnalysis of the RBC dataset revealed clear microcytic and hypochromic patterns among β-thalassemia trait (BTT) carriers, characterised by lower MCV (59.89\u0026thinsp;\u0026plusmn;\u0026thinsp;4.20 fL) and MCH (19.05\u0026thinsp;\u0026plusmn;\u0026thinsp;1.37 pg) and elevated RBC counts (5.72\u0026thinsp;\u0026plusmn;\u0026thinsp;0.65 \u0026times;10\u0026sup1;\u0026sup2;/L), compared to non-carriers. RDW was also higher among carriers, reflecting increased variability in red cell size. These differences provided a strong discriminatory signal for the initial RBC-based screening step. The MLP machine learning classifier, optimised using these parameters, achieved high specificity, allowing positive cases to be confidently referred for confirmatory HPLC testing while minimising unnecessary smear preparation for individuals screened as negative.\u003c/p\u003e\u003cp\u003eThe blood smear image analysis component further enhanced screening sensitivity, effectively detecting subtle morphological changes, including target cells, pencil cells, and other characteristic features of thalassemia. Using the VGG16 CNN architecture with transfer learning, the system reliably identified borderline and atypical carriers that could be missed by RBC indices alone. The integration of structured and image-based data in a sequential pipeline ensured that no true positive cases were overlooked, while maintaining operational efficiency by limiting smear analysis to individuals not flagged in the first step.\u003c/p\u003e\u003cp\u003ePractical Implications\u003c/p\u003e\u003cp\u003eThis dual-modal approach addresses several challenges faced by national screening programmes in low- and middle-income countries. Firstly, it reduces reliance on specialised laboratory equipment, as smartphone-assisted microscopy provided sufficiently high-quality images. Secondly, the sequential workflow optimises resource utilisation by minimising the number of smears requiring manual examination, thereby saving time, reagents, and labour. Thirdly, given the limited number of haematologists in Sri Lankan hospitals, conducting mass screening is extremely tiring and exhausting for available staff; the model\u0026rsquo;s portability and compatibility with widely available smartphones enable decentralised screening in rural and primary care settings. This aligns with ongoing efforts to expand thalassaemia prevention programmes in Sri Lanka, including premarital screening initiatives.\u003c/p\u003e\u003cp\u003eBy combining RBC indices and smear morphology, the system also improves diagnostic confidence, reducing false positives and negatives compared to single-modality approaches. This is particularly relevant in LMICs, where conventional confirmatory tests like HPLC or genetic analysis may be limited by cost, infrastructure, or accessibility. The proposed pipeline could serve as a triage tool, prioritising individuals for confirmatory testing while reducing the burden on central laboratories.\u003c/p\u003e\u003cp\u003eLimitations\u003c/p\u003e\u003cp\u003eDespite its promising performance, the study has several limitations. The dataset was derived from a single tertiary referral centre in Kurunegala, limiting geographical and ethnic diversity. While age and sex distributions were balanced, other demographic factors, such as socioeconomic background, ethnicity, and regional prevalence variations, were not captured. Consequently, external validity remains to be established, and multi-centre studies are necessary to confirm generalisability across diverse populations.\u003c/p\u003e\u003cp\u003eThe sample size is modest, comprising 152 participants, which may affect the robustness of the findings and the ability to capture less common presentations of β-thalassaemia.\u003c/p\u003e\u003cp\u003eAlthough smartphone-assisted image acquisition reflects real-world conditions, image quality may vary depending on operator experience, lighting, and device specifications. Data augmentation and CNN training mitigated some of these effects; however, further field validation is necessary to assess robustness under variable community-level conditions.\u003c/p\u003e\u003cp\u003eThe current pipeline remains a research prototype and has not yet undergone regulatory review, integration with laboratory information systems, or prospective clinical evaluation.\u003c/p\u003e\u003cp\u003eFuture Directions\u003c/p\u003e\u003cp\u003eFuture work should focus on expanding the dataset to include multi-centre, nationwide samples and additional demographic variables, enhancing model generalisability and robustness. Prospective studies are also required to evaluate real-world usability, workflow integration, and the impact on screening coverage, particularly in the context of mass thalassaemia prevention programs.\u003c/p\u003e\u003cp\u003eEngagement with policymakers, healthcare administrators, and frontline health workers will be essential to understand operational constraints, gather feedback, and secure support for implementation. The development of a clinician-friendly mobile application with streamlined reporting could facilitate adoption, improve operational efficiency, and enable decentralised screening in rural and primary care settings.\u003c/p\u003e\u003cp\u003eThe dual-modal framework could be adapted for other haemoglobinopathies, supporting broader public health screening initiatives and contributing to scalable, sustainable, and equitable approaches to genetic disorder prevention.\u003c/p\u003e\u003cp\u003eImplementation and Ethical Considerations\u003c/p\u003e\u003cp\u003eWhile the dual-modal machine learning pipeline demonstrates high sensitivity and practical feasibility, several considerations must be addressed prior to deployment.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eData Security and Privacy\u003c/strong\u003e\u003cp\u003eThe system relies on patient data and blood smear images, which are sensitive health information. Secure storage, anonymisation, and controlled access are essential to prevent unauthorised use or breaches. Any mobile or cloud-based deployment must comply with national and international data protection regulations, including secure data transmission and storage protocols.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eExplainability and AI Reliability\u003c/strong\u003e\u003cp\u003eSupervised machine learning models, including deep learning, can produce incorrect or unexpected predictions, particularly when presented with inputs that differ from the training data. Ensuring explainability of these predictions is essential when they inform clinical decisions. Providing interpretable outputs, such as feature importance for RBC indices or saliency maps for blood smear images, can enhance clinician trust, facilitate oversight, and support safe integration into screening workflows.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eOperator and Workflow Variability\u003c/strong\u003e\u003cp\u003eThe quality of smartphone-assisted images may vary with operator skill, lighting, and device type. Standardised training, clear imaging protocols, and continuous quality control are necessary to maintain model reliability across diverse settings.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eEthical and Regulatory Oversight\u003c/strong\u003e\u003cp\u003e Clinical deployment must align with ethical guidelines, obtain regulatory approvals, and integrate seamlessly with existing laboratory and screening workflows. Continuous monitoring for potential biases, misclassifications, or inequitable outcomes is essential, particularly when expanding to new populations.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eStakeholder Engagement\u003c/strong\u003e\u003cp\u003eSuccessful implementation requires input from healthcare workers, administrators, and policymakers to ensure the system is feasible, acceptable, and sustainable within national screening programmes. Addressing practical concerns, such as workload, cost, and training, will be critical for uptake and long-term adoption.\u003c/p\u003e\u003c/p\u003e\u003cp\u003eProactively addressing these considerations will enable safe, effective integration of the system into national screening pathways while maintaining patient trust and clinical reliability.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study demonstrates that a dual-modal machine learning pipeline combining RBC indices and peripheral smear morphology can provide a sensitive, cost-effective, and practical approach for β-thalassemia screening. By enabling early detection and prioritisation of at-risk individuals, the system has potential to strengthen national prevention programs, reduce healthcare burden, and support decentralised screening in low-resource settings. It\u0026rsquo;s design, optimised for smartphones and minimal infrastructure, illustrates a scalable model for implementing AI-driven diagnostic support in LMICs.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical approval for this study was obtained from the Institutional Ethical Review Committee of the Teaching Hospital, Kurunegala. Written informed consent was obtained from all participants prior to enrolment using a certificate of consent.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Not applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was funded by University of Colombo School of Computing, Sri Lanka.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAR, CR and DL work collaboratively in conceptualizing the study. DL\u003c/strong\u003e collected data, performed data analysis, and was a major contributor in writing the manuscript. \u003cstrong\u003eCR\u003c/strong\u003e led data collection and preparation, model development, data analysis and reviewed and approved the final manuscript. \u003cstrong\u003ePM\u003c/strong\u003e assisted with data collection and reviewed the final manuscript. \u003cstrong\u003eAR\u003c/strong\u003e contributed in model development, \u0026nbsp;data analysis and interpretation of the results.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe research team expresses sincere gratitude to Teaching Hospital Kurunegala for granting permission for data collection. A very special thanks goes to the medical laboratory technologists at the National Thalassemia Centre, Kurunegala, for their invaluable support in data collection. We also extend our appreciation to the University of Colombo School of Computing for providing necessary guidance and funding for this research. Finally, we acknowledge all individuals who contributed their time, expertise, and resources to make this study possible.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAksu T, \u0026Uuml;nal Ş, Thalassemia. Trends Pediatr. 2023;2(1):1\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5222/TP.2021.10820\u003c/span\u003e\u003cspan address=\"10.5222/TP.2021.10820\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePerera O, Ranathunga T. Socio-economic impact of thalassemia: a case study of Kurunegala district, Sri Lanka. Fac J Humanit Social Sci (Univ Ruhuna). 2020;9(1):69\u0026ndash;81. (hss.ruh.ac.lk).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMarengo-Rowe AJ. The thalassemias and related disorders. Proc (Bayl Univ Med Cent). 2007;20(1):27\u0026ndash;30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/08998280.2007.11928230\u003c/span\u003e\u003cspan address=\"10.1080/08998280.2007.11928230\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAmarasinghe N, et al. Redesigning new policy options for thalassemia prevention in Sri Lanka. Thalassemia Rep. 2022;12(4):135\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/thalassrep12040018\u003c/span\u003e\u003cspan address=\"10.3390/thalassrep12040018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ede Silva S, Fisher CA, Premawardhena A, Lamabadusuriya SP, Peto TE, Perera G, Old JM, Clegg JB, Olivieri NF, Weatherall DJ, Sri Lanka Thalassaemia Study Group. Thalassaemia in Sri Lanka: implications for the future health burden of Asian populations. \u003cem\u003eLancet.\u003c/em\u003e 2000;355(9206):786\u0026thinsp;\u0026ndash;\u0026thinsp;91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S0140-6736(99)08246-X\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(99)08246-X\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 10711926.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReed-Embleton H, Wijesiriwardana B, Perera S, Fernando M, Pathiraja R, Padeniya A. A cost-of-illness analysis of β-thalassaemia major in children in Sri Lanka: experience from a tertiary level teaching hospital. BMC Pediatr. 2020;20(1):160. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12887-020-02160-3\u003c/span\u003e\u003cspan address=\"10.1186/s12887-020-02160-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eD\u0026iacute;az G, Manzanera A. Automatic analysis of microscopic images in hematological cytology applications. In: Gonz\u0026aacute;lez FA, Romero E, editors. Biomedical Image Analysis and Machine Learning Technologies: Applications and Techniques. Hershey (PA): IGI Global; 2009. Chapter 8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4018/978-1-60566-956-4.ch008\u003c/span\u003e\u003cspan address=\"10.4018/978-1-60566-956-4.ch008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee SY, Chen CME, Lim EYP, Shen L, Sathe A, Singh A, Sauer J, Taghipour K, Yip CYC. Image analysis using machine learning for automated detection of Hemoglobin H inclusions in blood smears \u0026ndash; a method for morphologic detection of rare cells. J Pathol Inf. 2021;12:18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4103/jpi.jpi_110_20\u003c/span\u003e\u003cspan address=\"10.4103/jpi.jpi_110_20\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRodellar J, Alf\u0026eacute;rez S, Acevedo A, Molina A, Merino A. Image processing and machine learning in the morphological analysis of blood cells. Int J Lab Hematol. 2018;40(Suppl 1):46\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/ijlh.12818\u003c/span\u003e\u003cspan address=\"10.1111/ijlh.12818\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSandanayake TC, Thalewela ATPMN, Thilakesooriya HP, Rathnayake RMAU, Wimalasooriya SAYA. Automated thalassemia identifier using image processing. IJIRT. 2016;3(7):97\u0026ndash;105.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhan MS, Khan M, Khan SA, Imran M, Ullah F, Khan H, Irfan M, Nawaz M, Mehmood I. Deep learning assisted automated assessment of thalassemia from haemoglobin electrophoresis images. Diagnostics (Basel). 2022;12(10):2405. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/diagnostics12102405\u003c/span\u003e\u003cspan address=\"10.3390/diagnostics12102405\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin YH, Liao KY-K, Sung K-B. Automatic detection and characterization of quantitative phase images of thalassemic red blood cells using a mask region-based convolutional neural network. J Biomed Opt. 2020;25(11):116502. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1117/1.JBO.25.11.116502\u003c/span\u003e\u003cspan address=\"10.1117/1.JBO.25.11.116502\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSadiq S, Khalid MU, Mui-Zzud-Din, Ullah S, Aslam W, Mehmood A, Choi GS, On B-W. Classification of β-Thalassemia carriers from red blood cell indices using ensemble classifier. IEEE Access. 2021;9:45528\u0026ndash;38. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/ACCESS.2021.3066782\u003c/span\u003e\u003cspan address=\"10.1109/ACCESS.2021.3066782\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePurwar S, Singh P, Singh A. Classification of thalassemia patients using a fusion of deep image and clinical features. \u003cem\u003eMater Today Proc.\u003c/em\u003e 2021;46:10418-22.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Thalassemia screening, β-thalassemia trait detection, Machine learning in healthcare, Clinical decision support, Red blood cell indices, Blood smear image analysis, Convolutional neural networks (CNN), Transfer learning, Low-resource settings, Smartphone-based microscopy","lastPublishedDoi":"10.21203/rs.3.rs-8281069/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8281069/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThalassemia is the most prevalent single-gene disorder in Sri Lanka, imposing a significant socioeconomic and healthcare burden. Early carrier detection is essential for genetic counselling and the prevention of thalassemia major births. Current screening programs rely on red cell indices, manual examination of blood smears by expert haematologists, and expensive confirmatory tests such as High-Performance Liquid Chromatography (HPLC) or genetic analysis, which limit large-scale applicability in low-resource settings. This study aims to develop a machine learning–based automated screening system that integrates Red Blood Cell (RBC) indices and blood smear image analysis to support cost-effective and scalable β-thalassemia trait detection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn a cross-sectional study of 152 individuals (54 confirmed β-thalassemia trait, 98 negative), 30% of the dataset was allocated for independent testing. The remaining 70% was used for model training, with 5-fold cross-validation for the RBC analysis model and 10-fold cross-validation for the image analysis model. RBC features were classified using a Multi-Layer Perceptron (MLP; scikit-learn), while blood smear images, captured through a conventional microscope using a smartphone camera, were analysed using a transfer learning–based VGG-16 CNN (TensorFlow/Keras). Data balancing and image augmentation (rotation, flipping, brightness variation) were applied to address class imbalance and overfitting. A two-step screening pipeline was proposed, applying RBC analysis first, followed by smear image analysis for RBC-negative cases.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe RBC indices analysis model achieved 88.2% sensitivity (95% CI: 63.6–98.5%) and 92.9% specificity (95% CI: 76.5–99.1%), while the image analysis model reached 88.2% sensitivity (95% CI: 63.6–98.5%) and 64.3% specificity (95% CI: 44.1–81.4%). When integrated sequentially, the combined pipeline achieved 100% sensitivity (95% CI: 80.5–100%) and 60.7% specificity (95% CI: 40.6–78.5%), reducing the overall need for smear preparation by 37.7%.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis dual-modal system provides a highly sensitive, cost-effective, and practical solution for β-thalassemia carrier detection, enabling mass screening and supporting sustainable prevention strategies in resource-limited settings. With larger, independent, multi-centre validation, it could be integrated into laboratory workflows to expand screening coverage, while ongoing evaluation will ensure generalisability and inform policy adoption.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Registration \u003c/strong\u003eNot applicable\u003c/p\u003e","manuscriptTitle":"A Dual-Modal Machine Learning Framework Integrating Red Blood Indices and Smartphone-Captured Microscopic Images for β-Thalassemia Screening","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-08 11:34:07","doi":"10.21203/rs.3.rs-8281069/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-12T09:26:04+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-12T05:56:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"20081160847442974919608513179078643320","date":"2026-01-12T05:53:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"201645439096140485236153290733443463387","date":"2026-01-11T18:36:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"288282612315500657456625470140151298241","date":"2026-01-10T17:07:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"25351924115969001772756562518056834161","date":"2026-01-10T16:48:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"269797461482579657388593580422574186120","date":"2026-01-10T16:00:43+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-10T04:19:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"74151255063250697816331778707049426501","date":"2026-01-10T04:09:24+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-12T19:29:59+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-12-09T10:46:12+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-12-09T06:02:42+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-12-09T06:01:06+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Informatics and Decision Making","date":"2025-12-04T15:29:48+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-informatics-and-decision-making","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"midm","sideBox":"Learn more about [BMC Medical Informatics and Decision Making](http://bmcmedinformdecismak.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/midm/default.aspx","title":"BMC Medical Informatics and Decision Making","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9630d04e-02e0-453a-b719-d7b3f979b65d","owner":[],"postedDate":"December 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-03-23T16:04:42+00:00","versionOfRecord":{"articleIdentity":"rs-8281069","link":"https://doi.org/10.1186/s12911-026-03451-y","journal":{"identity":"bmc-medical-informatics-and-decision-making","isVorOnly":false,"title":"BMC Medical Informatics and Decision Making"},"publishedOn":"2026-03-21 15:58:38","publishedOnDateReadable":"March 21st, 2026"},"versionCreatedAt":"2025-12-08 11:34:07","video":"","vorDoi":"10.1186/s12911-026-03451-y","vorDoiUrl":"https://doi.org/10.1186/s12911-026-03451-y","workflowStages":[]},"version":"v1","identity":"rs-8281069","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8281069","identity":"rs-8281069","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}