Evaluation of minor labial salivary gland focus score in Sjögren's disease using deep learning: a tool for more efficient diagnosis and future tissue biomarker discovery

This is the authors' abstract. We don't add key points for this paper.

BACKGROUND: Sjögren's Disease (SjD) is histopathologically characterized by focal sialadenitis in minor labial salivary gland biopsies (mLSGB), which is evaluated by utilizing the focus score (FS). Focus score ≥1 identification is a critical step of the diagnostic approach and SjD classification. Nonetheless, during mLSGB analysis, FS reporting is neglected in a staggering 17 %, and a degree of inter-observer variability is introduced, even among specialized university centers. As the unmet need for reliable FS reporting is displayed, leveraging artificial intelligence in mLSGB evaluation shows encouraging potential and mandates to be investigated. METHODS: Minor LSGBs stained only with hematoxylin and eosin (H&E) during evaluation of individuals with a clinical suspicion of SjD, were randomly chosen from our archive. All mLSGBs were scanned digitally as whole slide images (WSI) and the final dataset was partitioned into a training (70 %) and a test set (30 %). An attention-based deep learning binary classification model was employed for evaluation of mLSGBs positivity (FS ≥ 1 or FS < 1). RESULTS: The final dataset consisted of 271 mLSGBs, with 153 (56 %) having FS < 1 and 118 (44 %) FS ≥ 1. In the FS ≥ 1 subset, 74 (63 %) were in the FS = 1-2 range, and the remaining biopsies had FS > 2, following the expected FS distribution among the typical SjD population. Our model resulted in: AUC = 0.932 (0.881-0.984), sensitivity 87 % (0.733-0.944), specificity 84 % (0.71-0.915) and accuracy 85.2 % (0.763-0.912), achieving better performance from previous works. CONCLUSION: Artificial intelligence models may overcome the intra-observer biases and inter-observer variability in FS evaluation, reinforcing the diagnosis and biomarker discovery in SjD.