This study was designed to assess the diagnostic value of salivary gland ultrasound in patients highly suspected of Sjögren’s Disease, a cohort that mirrors the typical clinical context in which SGUS is applied. While the incremental diagnostic benefit of SGUS may appear constrained in this preselected high-risk population, we contend that the findings carry substantial clinical significance for several key reasons.

First, SGUS stands out as a non-invasive, convenient, and cost-effective diagnostic modality, offering practical utility for patients with a high suspicion of SjD but lacking a definitive diagnosis. Beyond facilitating confirmation of the condition, SGUS can, in certain instances, diminish the reliance on invasive procedures such as LSGB. This attribute is particularly valuable in resource-constrained clinical settings, where it also serves to mitigate patient discomfort associated with invasive interventions. Although LSGB remains a cornerstone of SjD diagnosis, its procedural complexity and challenges related to patient compliance underscore the appeal of SGUS as a complementary tool. Second, the prospective design of this study specifically targets patients with suspected but unconfirmed SjD, aligning closely with real-world scenarios where SGUS is most relevant. For example, our analysis revealed a strong correlation between grade 3 SGUS findings and positive LSGB results. This suggests that SGUS could function as an effective preliminary test, informing the need for subsequent invasive evaluations and thereby streamlining the diagnostic process. By addressing this specific clinical context, our study provides actionable insights for clinicians managing analogous patient groups.

Our research is based on a semi-quantitative scoring system proposed by the OMERACT Salivary Gland Ultrasound Task Force in 2019 [10]. The reliability of the OMERACT scoring system has been validated in several studies. In 2021, Stephanie Finzel et al. reported that the weighted κ for intrareader reliability ranged from 0.44 to 1 for grading in the PG and from 0.59 to 1 in the SMG. The interreader reliability κ was 0.62 (95% CI: 0.47–0.74) for PG and 0.62 (95% CI: 0.47–0.72) for SMG [15]. In 2022, Rossana Izzetti et al. assessed the reliability of the OMERACT scoring system in minor salivary gland ultra-high frequency ultrasonography using the intraclass correlation coefficient (ICC). The ICC values for grades 0 and 1 exceeded 0.9, while the values for grades 2 and 3 were 0.873 and 0.785, respectively, demonstrating strong inter-observer consistency [16]. In 2023, Nanna S. Schmidt et al. confirmed the reliability of the OMERACT scoring system and suggested that the use of an atlas could further enhance both inter-reader and intra-reader reliability. With the atlas, inter-reader reliability improved to a moderate level (mean κ = 0.52; range 0.31–0.77), while intra-reader reliability reached a good level (mean κ = 0.69; range 0.46–0.86) [17].

The applicability of the OMERACT scoring system has been evaluated in several studies. In 2021, Xia Zhang et al. conducted a multicenter study involving 246 patients with SjD, 140 control subjects with non-SjD conditions, and 27 healthy controls. The study determined that the optimal cut-off value for the left (right) PG and the left (right) SMG was 4, achieving maximal sensitivity (75.6% and 77.2%, respectively) and specificity (91.6% and 92.2%, respectively) [18]. A more recent study investigating OMERACT scores for PG and SMG ultrasound in 242 patients reported a sensitivity of 76% and specificity of 90%, with an AUC of 0.83 at a threshold score of 8 [12].

Comparing our study with previous studies, an undeniable fact is that the diagnostic performance of SGUS may vary across different application scenarios. Variations in patient selection can significantly affect the diagnostic accuracy of a test. Unlike previous studies, which often included broader populations, this study focused on patients highly suspected of having SjD—a group that more closely reflects the clinical scenarios where SGUS is applied for differential diagnosis. These patients had typically remained under suspicion of SjD despite initial evaluations at secondary hospitals or community health centers and were predominantly referred to the rheumatology department of a tertiary care hospital. Of the 171 individuals enrolled, only 41 were ultimately not diagnosed with SjD, and 50 had negative LSGB results. The prevalence of SjD in this cohort was considerably higher than in earlier studies, leading to the underrepresentation of typical normal cases, which may have contributed to the reduced specificity observed for grade 2 ultrasound findings.

Another potential factor is the significant discrepancy between the two levels in the OMERACT semi-quantitative scoring system. In this study, when grade 2 was used as the diagnostic threshold, the sensitivity was 0.85 for the PG and 0.95 for the SMG, while the specificity was only 0.46 for PG and 0.24 for SMG. However, when grade 3 was selected as the threshold, specificity markedly increased to 1.00 in PG and 0.90 in SMG, but sensitivity dropped sharply to 0.27 in PG and 0.52 in SMG. This suggested that a midpoint between grade 2 and grade 3 may yield a more optimal diagnostic performance. Currently, grade 2 is defined as moderate inhomogeneity with focal anechoic/hypoechoic areas, while grade 3 is characterized by diffuse inhomogeneity with anechoic/hypoechoic areas occupying the entire gland surface [10]. Incorporating more specific features and refining the grading criteria could enhance the diagnostic accuracy of ultrasound in complex cases.

The application of ultrasound in the diagnosis and management of SjD merits careful consideration. Despite the strong correlation between SGUS and LSGB, the prevailing consensus is that SGUS cannot serve as a complete substitute for biopsy. This limitation stems from the fact that ultrasound does not yield immunological evidence, and the focus score derived from biopsy is critically associated with patient prognosis.

Nevertheless, the inclusion of ultrasound as an optional criterion in the classification of SjD may hold critical value. Several studies have explored this possibility with encouraging results. For example, a prospective cohort study by Esther Mossel et al. demonstrated that a positive salivary gland ultrasound, in conjunction with anti-SSA/Ro antibodies, was a strong predictor for SjD classification [19]. In 2020, S. Jousse-Joulin et al. found that incorporating a salivary gland ultrasound score (non-OMERACT) of ≥ 2 increased sensitivity from 90.2 to 95.6%, with only a slight decrease in specificity (84.1% versus 82.6%) [20]. More recently, in 2022, François Robin et al. reported that adding the OMERACT ultrasound score to the 2016 ACR/EULAR classification criteria enhanced sensitivity (91.5% vs. 89.4%) while minimally reducing specificity (96.0% vs. 100%), with an AUC of 0.975 (95% CI: 0.945–1.00) [21]. Importantly, the study by Esther Mossel et al. underscored that gland biopsy, salivary flow rate, and ultrasonography are complementary diagnostic tools and should not be considered interchangeable [22].

Research has established a significant correlation between various SGUS scores and salivary flow rates. Yasemin Yalcinkaya et al. reported that elevated Milic and Hocevar ultrasound scores were associated with decreased salivary flow rate, while the homogeneity score correlated with low USFR [23]. S. Garcia-Cirera et al. further demonstrated that both USFR and stimulated salivary flow rates were significantly correlated with ultrasound scores from the De Vita (0–3), Salaffi (0–4), and OMERACT (0–3) scoring systems [24].

Our findings corroborated that ultrasound grading is closely associated with USFR, suggesting that abnormalities identified through ultrasound reflect compromised salivary gland function. This supports the integration of ultrasound grade into the classification criteria for Sjögren’s Disease.

In addition to being a possible item in the classification criteria, ultrasound grade can also provide some predictors for LSGB in clinical practice. Omar Al Tabaa et al. proposed that negative anti-SSA antibodies together with SGUS score < 2 could avoid LSGB [25]. Our findings indicated that a grade 3 for either the PG or SMG is a strong predictor of a positive biopsy result. This evidence suggests that ultrasound, as a non-invasive and convenient diagnostic modality, is suitable as a pretest for LSGB.

It must be acknowledged that the diagnosis of SjD is a multidimensional process. While this study primarily investigates the application of SGUS in diagnosing SjD, the critical role of ophthalmological involvement in disease assessment cannot be overlooked. SjD is characterized not only by xerostomia but also by keratoconjunctivitis sicca, with the latter serving as a pivotal diagnostic feature. Routine ophthalmological evaluations, including the Schirmer’s test, tear breakup time, and ocular surface staining, are widely employed to identify tear deficiency and ocular surface damage [26, 27]. These assessments have been formally recognized and integrated into existing classification criteria [1, 2]. By combining ophthalmological evaluations with serological markers and salivary gland assessments such as SGUS, a comprehensive diagnostic framework can be established, offering a more complete representation of the clinical characteristics of SjD.

A notable limitation of this study is its design as a single-center investigation, which may have resulted in a regionally biased sample. Additionally, patients often had previously undergone evaluations at secondary or community healthcare facilities but continued to be suspected of SjD. This scenario contributed to an unusually high prevalence of SjD among the study participants, potentially explaining the discrepancies between our findings and those reported in prior studies. Nevertheless, this emphasizes the significance of our research, as the characteristics of the population on which the diagnostic test is applied are paramount. This study is prospective and closely mirrors real-world applications of ultrasound, thereby minimizing the selection bias associated with specialized populations.