Mechanosensitive TRPV4 immunohistochemistry improves deep learning-based classification of ductal carcinoma…

AI / informaticsThe journal of pathology. Clinical research· Jul 2026

Mechanosensitive TRPV4 immunohistochemistry improves deep learning-based classification of ductal carcinoma in situ beyond H&E morphology.

Janghyun Yoo, Raghav Karthikeyan, Kashi Kamat … Inhee Chung

Breast & Gynecologic PathologyAI / informaticsDiagnosisQ1IF 4.2

TRPV4 staining substantially boosts deep learning accuracy classifying ductal carcinoma in situ versus H&E alone.

Mechanosensitive TRPV4 immunohistochemistry improves deep learning-based classification of ductal carcinoma in situ beyond H&E morphology.

Janghyun Yoo … Inhee Chung

The journal of pathology · Jul 2026 · Q1 in Pathology

Background

Ductal carcinoma in situ (ductal carcinoma in situ) spans a biologic continuum from atypical ductal hyperplasia (atypical ductal hyperplasia) to high-grade lesions with variable risk of progression to invasive ductal carcinoma (invasive ductal carcinoma), yet morphologic assessment by hematoxylin and eosin (H&E) remains diagnostically limited, particularly at the benign versus atypical ductal hyperplasia/low-grade ductal carcinoma in situ boundary.

Methods

We assembled a multi-institutional cohort of H&E and TRPV4 immunohistochemical whole-slide images from 108 patients, comprising an internal development cohort (n = 69), an external test cohort (n = 39), yielding 24,248 annotated tiles.

n = 108 patients

Results

Percent relative gain in classification accuracy (macro-F1) using TRPV4 staining instead of H&E.

H&E

0.44macro-

TRPV4 IHC

0.72macro-

More results

We tested whether deep learning models trained on TRPV4 immunohistochemical outperform H&E-based models across the ductal carcinoma in situ progression spectrum.

More results

Xception and EfficientNet-B0 convolutional neural networks were trained with patient-level three-fold cross-validation on the development cohort and evaluated as ensembles on the external test cohort.

“

Conclusion

These findings support TRPV4 immunohistochemical as a mechanistically grounded complement to H&E that improves patient-level discrimination across the ductal carcinoma in situ progression spectrum, with the strongest gains for atypical ductal hyperplasia/low-grade ductal carcinoma in situ and invasive ductal carcinoma, in a pilot multi-institutional setting.

Mechanosensitive TRPV4 immunohistochemistry improves deep learning-based classification of ductal carcinoma in situ beyond H&E morphology.

Mechanosensitive TRPV4 immunohistochemistry improves deep learning-based classification of ductal carcinoma in situ beyond H&E morphology.

Related papers