Explainable artificial intelligence for predicting inflammatory signatures and spatial molecular heterogeneity from nasal polyp histology.

Background: Chronic Rhinosinusitis with Nasal Polyps (CRSwNP) is a heterogeneous disorder characterized by diverse inflammatory signatures and endotypes.

Objective: To develop a histology-based deep learning network for predicting inflammatory gene signatures and spatial patterns in CRSwNP.

Methods: We developed HE2Signature, a deep learning model, using 70 H&E-stained whole-slide images (WSIs) of nasal polyps paired with corresponding endotypic signature gene expression profiles derived from transcriptomic data. The model was validated in an internal cohort (n = 30) and tested in an independent external cohort (n = 224) from four medical centers. The performance was evaluated using correlation and confusion matrix analyses and receiver operating characteristic (ROC) curves. Spatial predictions were validated by immunohistochemistry.

Results: The HE2Signature demonstrated strong positive correlations between the predicted and actual expression levels of 28 of the 33 signature genes. The internal validation cohort was stratified into distinct endotypes based on predicted signature gene expression, achieving AUC values of 0.833, 0.903, and 0.935 for the T1, T2, and T3 endotypes, respectively. In the external cohort, the combination of HE2Signature-predicted FCER2 and CST1 effectively identified the T2 endotype, with an ROC value of 0.716. Histopathological examination of high-prediction patches revealed characteristic features associated with signature gene expression. Heatmaps and immunohistochemistry confirmed the accuracy of the model in mapping spatial gene expression patterns. HE2Signature-predicted spatial expression of signature genes correlated with blood and tissue eosinophil counts, Lund-Kennedy, and SNOT-22 scores, particularly within subepithelial regions.

Conclusion: Our study introduces the first histology-based, explainable deep learning model capable of predicting inflammatory gene signatures and spatial molecular heterogeneity. This proof-of-concept highlights that artificial intelligence-powered histopathological analysis can generate digital biomarkers by linking tissue patterns with molecular profiles, providing a clinically applicable framework for endotype-guided precision medicine in CRSwNP.

Clinical Implications: Explainable deep learning applied to standard histology enables rapid, cost-effective prediction and spatial mapping of CRSwNP endotypes, supporting precision, endotype-guided management in routine clinical practice.