Minor salivary gland fibrosis in Sjögren's syndrome is elevated, associated with focus score and not solely a consequence of aging

Abstract

Objectives: Evaluate the presence of minor salivary gland (SG) fibrosis in primary Sjögren's syndrome (pSS) as a function of disease pathology or a consequence of ageing.

Methods: Subjects with sicca symptoms attending a Sjögren's research clinic were classified by American European Consensus Group (AECG) criteria as either pSS or non-SS (nSS). Discovery (n=34 pSS, n=28 nSS) and replication (n=35 pSS, n=31 nSS) datasets were evaluated. Minor SG cross-sections from haematoxylin and eosin stained slides were imaged, digitally reconstructed and analysed for percent area fibrosis. Relationships between SG fibrosis, age, and clinical measures were evaluated using Spearman correlations. Association with SS was assessed by: ROC curve, Variable Selection Using Random Forests (VSURF) and uni- and bi-variate regression analyses.

Results: SS subjects had significantly more fibrotic tissue in their minor labial salivary glands (median 24.39%, range 5.12-51.67%) than nSS participants (median 16.7%, range 5.97-38.65%, p<0.0001); age did not differ between groups (average ± SD pSS 50.2 ±13.9 years, nSS 53.8±12.4 years). In both the discovery and replication data sets, multiple regression models showed that the area of minor salivary gland fibrosis predicted pSS significantly better than age alone. Age-corrected linear regression revealed that the area of minor salivary gland fibrosis positively associated with vanBijsterveld score (p=0.042) and biopsy focus score (p=0.002). ROC curve and VSURF analyses ranked fibrosis as a significantly more important variable for subject discrimination than age.

Conclusions: SG fibrosis is an element of pSS pathology that is related to focus score and is not solely attributable to age.

Figure 1

Scoring of minor labial salivary gland fibrosis area. A Digital reconstruction of a minor labial salivary gland paraffin-embedded tissue section. Boxed region is magnified in B. B Representative scoring of labial SG tissue. “0” indicates squares where tissue is present, but is non-fibrotic. “1” indicates both tissue presence and fibrosis. Magnification = 200× Scale = 1.54 pixels/µm, grid square = 2500µm².

Figure 2

Salivary gland fibrosis but not age is increased in pSS compared to nSS subject groups. A Average percent area fibrosis is significantly greater in pSS subjects (23.6% ± 1.1; mean ± SEM) as compared to nSS subjects (16.6% ± 0.90). Mann-Whitney t-test. B pSS subjects (53.8 ± 1.37 years; mean ± SD) are not significantly different in age as compared to nSS subjects (50.2 ± 1.76 years). Student’s two-tailed unpaired t-test. C, D Distribution of percent area fibrosis and subject age are shown.

Figure 3

Significant agreement between fibrosis severity scores and quantified percent fibrosis by area (pSS = 20, nSS n= 15), 2-tailed Spearman correlation.

Figure 4

Receiver operating characteristic (ROC) curves for age (black), average percent area fibrosis (red), and biopsy focus score (blue) are shown compared to the line of no-discrimination (dashed blue line). The optimal cut-point distinguishing pSS from DNMC is indicated by the intersection of the dashed red lines.