*4.1. Sialography*

Sialography is a radiographic technique that visualizes the architecture of the ductal system by using X-ray projections after injection of contrast medium. In pSS patients, sialography shows sialectasis, which are collections of contrast material. The degree of sialectasis can be classified according to the scoring system developed by Rubin and Holt [29] (Figure 2). Sialectasis may be found at the location of cystic ductal dilatations in pSS patients. Another explanation could be that sialectasis represents extravasation of contrast material into the glandular parenchyma. A possible explanation for the leakage of contrast medium in pSS patients is dysfunction of tight junctions between striated ductal cells, due to the presence of proinflammatory cytokines [30]. In addition to sialectasis, sparsity of the ductal branching pattern can be found during sialography [3,31,32]. This could be due to obstruction of the ductal system, as a result of lymphocytic infiltration and proliferation of the ductal epithelium. However, direct associations with histopathological findings, such as the area of lymphocytic infiltrate or presence of LELs, have thus far not been reported.

**Figure 2.** Findings on sialography. Sialographies of the parotid gland showing (**A**) no abnormalities in a healthy subject, (**B**) punctate/globular sialectasis in a pSS patient, and (**C**) globular/cavitary sialectasis in a pSS patient [29]. (**D**) Two-dimensional sialo-CBCT image and (**E**) three-dimensional sialo-CBCT image of the parotid gland of a pSS patient, showing normal width of the primary duct, moderate scarcity of ductal branches, and numerous diverse sialectasis. Thanks to Prof. D.J. Aframian and Dr. C. Nadler and colleagues who provided the sialo-CBCT images.

Sialography has been used for diagnosing pSS for decades and shows moderate to high sensitivity and specificity [3]. This technique was excluded from the 2016 American College of Rheumatology/European League Against Rheumatism (ACR-EULAR) classification criteria (Table 1) because of multiple drawbacks. It is an invasive technique with risk of complications and radiation exposure. Furthermore, there are multiple contraindications like acute infection, acute inflammation, and contrast allergy [3,31].

Alternative sialographic techniques have been developed, such as sialo-cone-beam computerized tomography (sialo-CBCT) and magnetic resonance (MR) sialography. These techniques have an increased spatial resolution and provide three-dimensional, instead of two-dimensional, images of the ductal system. Keshet et al. [33] described correlations between sialo-CBCT findings and clinical data, such as xerostomia and serological parameters. However, since only 6 out of 67 sicca patients fulfilled the American European Consensus Group (AECG) classification criteria for pSS in this cohort, the usefulness of sialo-CBCT in pSS should be further investigated. MR sialography can identify changes within the salivary glands without the injection of contrast medium. The typical finding in pSS is the presence of multiple high-signal-intensity spots, which are thought to arise after leakage of saliva from peripheral ducts. Kojima et al. [34] did not find correlations between MR sialography findings and salivary flow rate, which can be explained by the fact that MR sialography visualizes the ductal system instead of saliva-producing acinar cells. Although MR sialography seems to be more sensitive to detect early disease, magnetic resonance imaging (MRI) provides more information on pathological changes in the glandular parenchyma, as we describe below [35].

In conclusion, sialography is not commonly used in the diagnostic work-up and follow-up of pSS anymore. Although alternative sialographic techniques such as sialo-CBCT and MR sialography have been evaluated, their current role in the diagnostic work-up of pSS is limited.

#### *4.2. Magnetic Resonance Imaging*

The role of MRI of the salivary glands in pSS has been investigated during the past decades. The characteristic finding in salivary glands of pSS patients is a heterogeneous signal-intensity distribution on T1- and T2-weighted images. The multiple hypointense and hyperintense areas cause a so-called salt and pepper appearance [34]. In the advanced stages of pSS, cystic changes can be found with MRI, which are thought to arise from destruction of the salivary gland parenchyma and the presence of fibrosis and fatty infiltration [3,31,36]. Although fat fractions in salivary glands seem to increase with higher age and body mass index [37] and can account for 60% of the histological section of the parotid gland in healthy individuals [38], imaging studies found that premature fat deposition found on MRI images is associated with SS [39,40]. Histopathological studies, however, did not make clear whether fatty infiltration is a specific feature of pSS or age-associated. Since biopsies do not represent the entire gland, it remains difficult to correlate MRI findings to histopathological findings. Although correlations between the focus score of the labial gland and MRI findings of the parotid gland were found [41,42], further associations between MRI findings and histopathological findings, such as the area of lymphocytic infiltration, fibrosis, and fatty infiltration, have not been investigated yet.

Although MRI showed added value in the diagnostic work-up of pSS by detecting pSS-specific abnormalities of the salivary glands, this technique is not routinely applied in pSS. Findings on MRI showed good agreement with salivary gland ultrasonography (SGUS). Since SGUS has several advantages over MRI, such as its high spatial resolution in superficial organs and the fact that SGUS is more easily accessible, SGUS is a better alternative for the diagnostic work-up of pSS [3,43].

Kojima et al. [34] demonstrated, in a group of pSS patients, that a higher degree of glandular heterogeneity and a smaller volume of the parotid and submandibular glands on MRI images were associated with lower stimulated and unstimulated salivary flow rates. These associations were even more pronounced for the submandibular glands, compared to the parotid glands, indicating that MRI findings of the submandibular glands can reflect hyposalivation. A possible explanation for the differences in associations between both glands is that the function of the submandibular gland is impaired earlier in the disease process than the function of the parotid gland [34,44,45]. Collection of saliva from the individual glands would be a more direct approach to relate MRI findings with salivary gland functioning in pSS. Similar to the MRI findings of Kojima et al. [34], lacrimal flow rates were associated with lacrimal gland volumes of pSS patients. Lacrimal flow rates were lower in pSS patients with atrophic lacrimal glands compared to patients with hypertrophic and normal-sized glands [46]. No studies have been performed to evaluate associations between MRI findings of the salivary glands and systemic disease activity, but MRI is the most appropriate imaging technique to evaluate central or peripheral nervous system involvement in pSS [47,48].

MRI is also used for the evaluation of pSS-associated lymphomas in the head and neck region (Figure 3). MRI findings of salivary and lacrimal gland MALT lymphomas vary. Findings that have been described are glandular enlargement, (micro)cystic changes, and calcifications [49–51]. Zhu et al. [51] found that solid cystic appearances of MALT lymphomas can help to differentiate MALT from non-MALT lymphomas. However, benign and malignant lesions of salivary and lacrimal gland show overlap, which makes MRI a less reliable technique to differentiate between benign and malignant disorders of the exocrine glands [3,52,53]. Despite the indolent nature of pSS-associated lymphomas, these malignancies are able to disseminate to other mucosal sites or organs. MRI is used in local staging of the disease, by assessing the ingrowth in adjacent structures and spread to lymph nodes or other organs [50,54] (Table 2).

**Figure 3.** 18F-fluorodeoxyglucose (FDG) positron emmison tomography/computed tomography (PET/CT) and magnetic resonance imaging (MRI) findings in a pSS patient with salivary gland mucosa associated lymphoid tissue (MALT) lymphoma. (**A**) Whole-body FDG-PET showing high heterogeneous FDG uptake in both parotid and submandibular glands. No other pathological lesions were found (axillary and clavicular regions with increased uptake represent brown fat). (**B**) FDG-PET/CT image showing pathological uptake in the parotid glands and physiological uptake in the tonsils. (**C**) MRI stir sequence showing a pathological, heterogeneous aspect of both parotid glands. (**D**) Manually fused FDG-PET/MRI image, showing pathological uptake in the parotid glands and physiological uptake in the tonsils. (**E**) Whole-body FDG-PET and (**F**) FDG-PET/CT image after treatment, showing no pathological uptake in the parotid glands, indicating complete remission.

Together, MRI is not often used in the standard diagnostic work-up of pSS. However, due to its high spatial resolution, MRI is the most useful imaging technique for local staging of pSS-associated salivary and lacrimal gland lymphomas.

#### *4.3. Salivary Gland Ultrasonography*

Within the past decade, salivary gland ultrasonography (SGUS) has gained more and more attention, and was proven to be effective for the detection of typical structural abnormalities in pSS [55, 56]. Furthermore, various studies demonstrated that addition of SGUS improves the performance and feasibility of the 2016 ACR-EULAR classification criteria [57–60]. However, many different SGUS-based scoring systems are available, and international consensus on which scoring system should be used is lacking. This hampers addition of SGUS to the classification criteria [55,56]. Therefore, the Outcome Measures in Rheumatology Clinical Trials (OMERACT) SGUS task force group has recently developed ultrasound definitions and a novel SGUS scoring system with good and excellent inter and intraobserver reliabilities, respectively [61]. Further studies should validate this scoring system before SGUS can be added to the 2016 ACR-EULAR classification criteria.

Typical ultrasonographic abnormalities in pSS are hypoechogenic areas, hyperechogenic reflections, and poorly defined salivary gland borders [55] (Figure 4). Mossel et al. [62] demonstrated that the presence of hypoechogenic areas is the most important SGUS feature. However, it is still unknown what these hypoechogenic areas reflect at a histological level. It has been suggested that hypoechogenic areas consist of foci containing inflammatory cells. Histopathological foci, however, are smaller in size compared to the hypoechogenic areas. Preliminary results of Mossel et al. [63] show a good correlation between hypoechogenic areas and percentages of CD45+ leukocytic infiltrate. These results indicate that, despite the differences in size, hypoechogenic areas are somehow associated with foci of inflammatory cells. One explanation for this association could be that hypoechogenic areas originate from leakage of saliva that is transported through the ductal system into the periductal infiltrate, and eventually into the salivary gland parenchyma. Leakage of saliva from the ductal system can be a comparable phenomenon to leakage of contrast medium during sialography, due to dysfunction of tight junctions between striated ductal cells [30]. However, collections of saliva in the periductal infiltrates or parenchyma are not commonly seen in salivary gland biopsies of pSS patients. Another hypothesis is that the hypoechogenic areas represent fatty infiltration. However, fat tissue is most often visible as a hyperechogenic instead of hypoechogenic area [64]. Furthermore, preliminary results of Mossel et al. [63] show poor associations between hypoechogenic areas and the percentage of fat cells in the total salivary gland parenchyma, which contradicts the latter hypothesis.

As described before, the area from which the parotid biopsies are taken may not be representative for the ultrasonographic images. The biopsy is taken from the periphery of the gland, which does not contain larger excretory ducts. Therefore, correlating histopathology to SGUS findings remains difficult. A possibility to get a better understanding of what SGUS features in salivary glands of pSS patients represent is taking ultrasound-guided core needle biopsies. A recent study by Baer et al. [65] showed that taking core needle biopsies in pSS patients suspected of salivary gland lymphoma is a safe and useful procedure. Although the morphology of core needle biopsies is inferior compared to that of open biopsies, the core needle method allows biopsies to be taken from the exact location of hypoechogenic areas or hyperechogenic reflections.

Another unresolved question is whether current SGUS scoring systems are sensitive enough to assess (treatment-induced) differences that are seen histopathologically. Current SGUS scoring systems use subjective categorical scales. Hypoechogenic areas, for instance, are scored on a 0–3 scale. Therefore, major changes need to occur in order to change from one category to another. This could be an important drawback when using SGUS as an objective tool to assess disease progression as well as to assess changes in salivary gland involvement in clinical trials [56,66]. Scoring SGUS findings on a continuous scale and in a more objective way could increase the sensitivity to change. A possible way to do this is by using image segmentation and artificial intelligence. HarmonicSS, a multicenter and EU-supported project, is applying artificial intelligence to SGUS images in pSS. Preliminary results show that, among the tested algorithms, the multilayer perceptron classifier is the best performing algorithm. Since the HarmonicSS cohort will increase in size over time, further validation will follow [56,67].

**Figure 4.** Salivary gland ultrasonography findings in pSS. Presence of hypoechogenic areas and hyperechogenic reflections in the (**A**) submandibular gland and (**B**) parotid gland of a pSS patient.

Despite the potential usefulness of SGUS in diagnosis and classification of pSS, the value of SGUS to assess disease activity and disease progression and to detect salivary gland MALT lymphoma needs to be established. SGUS scores seem to correlate with objective salivary gland function, as unstimulated salivary flow rates were found to be lower in SGUS-positive patients, compared to SGUS-negative patients [68–70]. Zabotti et al. [71] described that of all SGUS findings, the presence of hyperechogenic bands was independently associated with salivary flow rates. Although they suggested that damage of the glands is reflected by hyperechogenic bands, it is still unclear what these hyperechogenic bands reflect at a histological level. Several studies showed associations between SGUS scores and clinical parameters of disease activity, such as ESSDAI scores, IgG levels, and rheumatoid factor (RF) levels [68,72–75]. In contrast, other studies did not find correlations between SGUS scores and ESSDAI [69,76]. These discrepancies can be explained by differences in patient characteristics between cohorts, and by the fact that severe salivary gland involvement might not reflect systemic disease activity in all pSS patients.

Since previous studies found associations between SGUS findings and risk markers of lymphoma, such as cryoglobulinemia, lymphopenia, and persistent salivary gland swelling, Theander et al. [72] and Coiffier et al. [77] stated that SGUS can identify patients at risk of developing lymphoma. However, these results were found in retrospective cohorts, and longitudinal studies should be performed to assess whether SGUS findings are predictive of lymphoma [78]. Furthermore, the capability of SGUS to detect lymphoma compared to histopathology and MRI should be clarified.

The studies presented thus far provide evidence that SGUS has added value in the diagnostic work-up of pSS (Table 2). Further research should be performed on the development of a consensus scoring system. Furthermore, to assess the usefulness of SGUS in follow-up and lymphoma detection in pSS, longitudinal studies are needed. Several initiatives have started already, such as OMERACT and HarmonicSS projects, which will give us more insight into the potential of this imaging tool in pSS.

#### **5. Sialendoscopy**

Sialendoscopy is a minimally invasive technique used for both diagnosis and management of obstructive salivary gland disorders, such as sialolithiasis, anatomic ductal abnormalities and mucus plugs. With this gland-sparing technique, a sialendoscope is entered through the ductal orifice of major salivary glands for inspection and irrigation of the ductal system, after local or general anesthesia.

In pSS patients, sialendoscopic examination mainly shows strictures, but mucous plugs and a pale, minimally vascularized ductal wall of the larger excretory ducts can also be observed [79–81]. Ductal strictures can cause ductal obstruction and could therewith account for glandular swelling and pain in pSS. It is still unclear what these strictures reflect at a histological level. One hypothesis is that strictures are caused by large LELs that obstruct the ductal system, but no histological studies have been performed yet to prove this hypothesis. Since the sialendoscopic findings in pSS are not specific for the disease, the diagnostic value of this technique in pSS is limited. However, various studies show that dilatation of strictures in combination with irrigation of the ductal system with saline and/or corticosteroids by using sialendoscopy is a safe and effective treatment option for salivary gland dysfunction in pSS patients. Studies showed that both subjective and objective oral dryness improved after sialendoscopy [80,82,83]. Furthermore, visual analogue scale pain scores decreased after the procedure [79,84], and the number of episodes of glandular swelling declined after treatment [85]. However, the above mentioned (pilot) studies included relatively small numbers of patients. The clinical benefits for pSS patients after sialendoscopy should be further explored in larger cohorts.

Although complications, such as infections and postoperative pain, seem to be limited [86], multiple studies reported that the sialendoscopic procedure was not successful in all pSS patients because of technical issues [79,80]. The most common difficulties reported were problems with identification or dilatation of the papilla before introducing the sialendoscope, which occurred more often during sialendoscopy of the submandibular gland compared to the parotid gland. These problems might be associated with characteristic features of severe stages of pSS, such as the presence of extreme hyposalivation and atrophic changes. Using salivary gland ultrasonography to assess the stage of the disease was suggested to predict whether patients would benefit from sialendoscopic treatment [80]. It would be of value to study this suggestion.

In summary, although sialendoscopy has no added value in the diagnostic work-up of pSS, recent studies have drawn attention to the fact that rinsing the ductal system, a procedure that accompanies sialendoscopy, might be useful in the management of oral symptoms.

#### **6. Nuclear Medicine Techniques**

#### *6.1. Conventional Nuclear Medicine*

Nuclear medicine imaging uses specific radiopharmaceuticals to visualize (patho)physiological processes in the body. Imaging with a a gamma camera system that provides two-dimensional planar images forms the basis of conventional nuclear medicine. However, it is often difficult to determine the exact location of increased tracer uptake by using these two-dimensional images. Three-dimensional images can be created by collecting images from different angles around the patient. This technique, called single photon emission computed tomography (SPECT), leads to a higher contrast and improves sensitivity, compared to two-dimensional nuclear medicine. Combining SPECT with a low dose or contrast-enhanced CT scan enables determination of the exact location of the area with increased uptake.
