DeepFusion-HSK: a multimodal deep learning for subtype classification of herpes simplex keratitis with multicenter validation
- 1National Clinical Research Center for Ocular Diseases, Eye Hospital, Cixi Biomedical Research Institute, Wenzhou Medical University , ,
- 2School of Electronic Engineering, Xi’an University of Posts and Telecommunications , ,
Abstract
Introduction: To develop and evaluate a multimodal deep learning (DL) model for subtype classification of viral keratitis.
Methods: This retrospective multicenter study included 3,501 slit-lamp images (807 eyes) as the internal development set and two external validation cohorts (2,957 images total). DeepFusion-HSK integrates three complementary illumination modalities—diffuse light imaging (DLI), slit-beam imaging (SBI), and fluorescein staining imaging (FSI)—through a novel fusion architecture. The framework uses a tri-way higher-order interaction layer to capture cross-modal pathological signatures, dual-task learning with auxiliary heads to prevent modality-specific feature degradation, and a dynamic gating mechanism that uses high-confidence coarse priors to calibrate fine-grained predictions and mitigate softmax probability spillover. To simulate real-world data deficiency, we evaluated model resilience under complete FSI dropout.
Results: On the internal test set, DeepFusion-HSK achieved an accuracy of 85.1% (95% CI: 78.5%–90.0%) and AUCs of 0.947–0.949 for discriminating epithelial herpes simplex keratitis (E-HSK) and non-epithelial herpes simplex keratitis (NE-HSK), significantly outperforming Swin-Transformer, EfficientNetV2, and ResNet50. External validation confirmed robust generalization, with accuracies of 81.7% on Dataset I and 78.2% on Dataset II, demonstrating clinical resilience. Ablation studies confirmed the hierarchical synergy of multimodal inputs: the best single modality (FSI) yielded 78.5% accuracy, the optimal dual-modality combination (DLI + FSI) reached 81.7%, and full tri-modal fusion improved performance to 85.1%, validating that integrated global morphology, depth contour, and superficial ulcer signatures are essential for robust HSK subtyping.
Discussion: DeepFusion-HSK offers a robust, generalizable, and clinically deployable solution for therapeutic HSK subtyping. Its resilience to missing imaging modalities makes it well suited for heterogeneous real-world ophthalmic practice, providing a decision-support tool to reduce vision-threatening misdiagnoses and guide therapy.
Related
- CSwinFK: a bidirectional cross-attention CNN-Swin Transformer fusion model for fungal keratitis classification in in vivo confocal microscopy images
- Comparative Evaluation of Multiple Deep Learning Models for Six-Class Keratitis Classification
- Multi-Scale MRI, Radiology Reports, and Blood Biomarker–Guided Multimodal Deep Learning for Predicting Postoperative Recurrence in Cervical Cancer: A Multicenter Study
- Multimodal prediction of neoadjuvant chemotherapy response in locally advanced gastric cancer: a multicenter cohort study
- Multimodal patch-based deep learning for early Alzheimer’s disease classification using MRI and PET
BibTeX
@article{zhang2026deepfusionhsk,
title = {{DeepFusion-HSK: a multimodal deep learning for subtype classification of herpes simplex keratitis with multicenter validation}},
author = {Zuhui Zhang and Tao Lv and Jiahao Tang and Kunhui Xu and Fen-Fen Li and Xinxin Yu and Chenchen Wang and Shuangqing Wu and Ya-na Fu and Jie-Wei Jiang and Fan Lu and Qi Dai},
journal = {Frontiers in cell and developmental biology},
year = {2026},
month = oct,
doi = {10.3389/fcell.2026.1978214},
url = {https://europepmc.org/article/PMC/PMC13724611},
note = {CC BY},
}
RIS
TY - JOUR
TI - DeepFusion-HSK: a multimodal deep learning for subtype classification of herpes simplex keratitis with multicenter validation
AU - Zuhui Zhang
AU - Tao Lv
AU - Jiahao Tang
AU - Kunhui Xu
AU - Fen-Fen Li
AU - Xinxin Yu
AU - Chenchen Wang
AU - Shuangqing Wu
AU - Ya-na Fu
AU - Jie-Wei Jiang
AU - Fan Lu
AU - Qi Dai
PY - 2026
DA - 2026/10/19
JO - Frontiers in cell and developmental biology
DO - 10.3389/fcell.2026.1978214
UR - https://europepmc.org/article/PMC/PMC13724611
AB - Introduction: To develop and evaluate a multimodal deep learning (DL) model for subtype classification of viral keratitis. Methods: This retrospective multicenter study included 3,501 slit-lamp images (807 eyes) as the internal development set and two external validation cohorts (2,957 images total). DeepFusion-HSK integrates three complementary illumination modalities—diffuse light imaging (DLI), slit-beam imaging (SBI), and fluorescein staining imaging (FSI)—through a novel fusion architecture. The framework uses a tri-way higher-order interaction layer to capture cross-modal pathological signatures, dual-task learning with auxiliary heads to prevent modality-specific feature degradation, and a dynamic gating mechanism that uses high-confidence coarse priors to calibrate fine-grained predictions and mitigate softmax probability spillover. To simulate real-world data deficiency, we evaluated model resilience under complete FSI dropout. Results: On the internal test set, DeepFusion-HSK achieved an accuracy of 85.1% (95% CI: 78.5%–90.0%) and AUCs of 0.947–0.949 for discriminating epithelial herpes simplex keratitis (E-HSK) and non-epithelial herpes simplex keratitis (NE-HSK), significantly outperforming Swin-Transformer, EfficientNetV2, and ResNet50. External validation confirmed robust generalization, with accuracies of 81.7% on Dataset I and 78.2% on Dataset II, demonstrating clinical resilience. Ablation studies confirmed the hierarchical synergy of multimodal inputs: the best single modality (FSI) yielded 78.5% accuracy, the optimal dual-modality combination (DLI + FSI) reached 81.7%, and full tri-modal fusion improved performance to 85.1%, validating that integrated global morphology, depth contour, and superficial ulcer signatures are essential for robust HSK subtyping. Discussion: DeepFusion-HSK offers a robust, generalizable, and clinically deployable solution for therapeutic HSK subtyping. Its resilience to missing imaging modalities makes it well suited for heterogeneous real-world ophthalmic practice, providing a decision-support tool to reduce vision-threatening misdiagnoses and guide therapy.
SN - 2296-634X
ER -