Logo image
Explainable artificial intelligence for label-free red blood cell morphometrics using brightfield microscopy towards clinical applications
Journal article   Open access   Peer reviewed   PUBLISHED

Explainable artificial intelligence for label-free red blood cell morphometrics using brightfield microscopy towards clinical applications

Subalakshmi Ramprasad, Sanu Susan Jacob, Prathap M Baby, Sreehari E and Ramesh Babu M. G.
Discover Artificial Intelligence, Vol.6(1), 1655
12/2026

Abstract

Artificial Intelligence General Computer Science Engineering
Background For diagnosing certain hematological disorders, variations in the red blood cell (RBC) morphology are considered as a marker. Advances in artificial intelligence (AI) offer the potential for automated, reproducible, label-free analysis of RBC morphology; however, most AI classifiers lack explainable, clinician-friendly outputs. Methods Single-cell RBC images of 17,781 across 10 morphological classes were acquired from three brightfield microscopes as a multicentric dataset. Mask R-CNN-based segmentation and 74 morphometric feature extractions were performed using the RedTell tool. Seven different machine learning classifiers were compared, and the best classifier was evaluated using SHapley Additive exPlanations (SHAP) analysis for interpretability and confidence calibration analysis for decision reliability. This led to the development of a comprehensive framework for evaluating the AI workflow through explainable AI in label-free RBC morphometric studies. Results The random forest classifier achieved an F1 score of 0.89. SHAP analysis indicated that intensity features strongly influenced the model predictions, followed by shape and texture features. These intensity-specific features were tested for PCA and were preserved across all three microscopes (0.49–0.88) with no meaningful device-driven confounding (silhouette ~ − 0.09). Expected calibration error was less (0.050), and the prediction confidence was good for clinically meaningful morphologies, supporting a practical reject-option strategy for flagging low-confidence predictions. Conclusion This fully explainable workflow, with feature-level interpretability and decision reliability, connects the AI and label-free RBC imaging domains, enabling trust and cost-effective adoption of label-free RBC morphology analysis, even in resource-limited clinical settings. Highlights Novel Framework: To our knowledge, this is the first study to integrate label-free RBC image acquisition, using a simple brightfield microscope for imaging, AI-based classification, and explainable AI within a single evaluated framework for label-free RBC morphometrics. Classifier performance: The random forest classifier achieved an F1 score of 0.89 across 10 RBC morphology classes, including rare morphologies and technical artifacts. Cross-Device Generalization: Leave-one-microscope-out validation demonstrated substantial generalization to unseen acquisition devices (mean accuracy 85%), with intensity-based feature discrimination found to be relevant with a biological basis (preserved independently across all three microscopes) rather than confounded by the acquisition device. Explainable AI Interpretability and Reliability: SHAP analysis revealed that the random forest classifier model relies primarily on intensity-based features, supported by shape descriptors, to capture global structural organization and texture features to encode finer-scale spatial variability. A complementary confidence-calibration analysis (ECE = 0.050) revealed that prediction confidence reliably distinguished trustworthy predictions from those requiring manual review, supporting a practical reject-option workflow. Toward Clinical Translation: Transparent feature-level explanations and calibrated decision reliability together link AI predictions to biophysical mechanisms, supporting trust, quality control, and cost-effective adoption in hematology diagnostics.
pdf
Explainable artificial intelligence for label-free red blood cell morphometrics using brightfield microscopy towards clinical applications3.06 MBDownloadView
Published (Version of record) Open Access CC BY V4.0  — This license enables reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator.
url
Link to publisher versionView
Published (Version of record) Open Access Open CC BY V4.0  — This license enables reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator.

Metrics

1 Record Views

Details

Logo image

Usage Policy