RPS 1408 - Hot Topic: hybrid solutions for salivary gland imaging

Author Block: A. A. Basheer Ahmed¹, F. Abubacker Sulaiman², R. Praveenkumar², D. Velan², J. Lydia², P. K. Anbazhagan²; ¹Melmaruvathur/IN, ²Chennai/IN
Purpose: To evaluate the diagnostic potential of dynamic MR sialography following lemon juice stimulation as a non-invasive, functional biomarker for detecting early salivary gland involvement in Sjögren’s syndrome.

The objective was to assess both morphological and secretory changes of major salivary glands and to correlate these imaging findings with serological and clinical parameters.
Methods or Background: A prospective observational study was conducted on 40 participants—20 with clinically confirmed Sjögren’s syndrome (based on ACR/EULAR criteria) and 20 healthy controls. Imaging was performed using 3T MRI with high-resolution, heavily T2-weighted 3D MR sialography sequences (SPACE/DRIVE technique). Each subject received 5 mL of freshly squeezed lemon juice orally to stimulate salivary flow. Sequential scans were acquired at baseline, 3 minutes, and 10 minutes post-stimulation.

Quantitative evaluation included ductal visualization grading, glandular parenchymal signal intensity, and dynamic excretory flow assessment. All results were correlated with Schirmer’s test, unstimulated salivary flow rate, and anti-Ro/La antibody positivity.
Results or Findings: Healthy subjects showed a prompt increase in ductal visualization and salivary excretion following stimulation, while Sjögren’s patients exhibited reduced or delayed flow, ductal beading, and glandular atrophy. Combined dynamic and structural parameters achieved superior diagnostic accuracy (AUC = 0.93) compared to conventional morphological imaging alone.

The technique demonstrated high interobserver reproducibility and excellent gland-to-duct contrast resolution.
Conclusion: Lemon juice–stimulated MR sialography represents a promising, radiation-free, and functionally dynamic imaging modality for early detection and grading of Sjögren’s syndrome, offering a potential alternative to invasive diagnostic techniques such as parotid sialography or labial gland biopsy.
Limitations: The study was limited by a relatively small cohort size, absence of histopathological validation in all cases, and interindividual variability in lemon-induced stimulation response. Larger multicenter studies with standardized stimulation protocols are warranted to establish reproducibility and optimize diagnostic thresholds.
Funding for this study: Not Applicable
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: L. Levante¹, C. Fichera², G. Fontana¹, V. Dolcetti¹, J. Fragomene¹, B. Vischioni¹, E. Orlandi¹, L. Preda¹, S. Imparato¹; ¹Pavia/IT, ²Lodi/IT
Purpose: To identify early radiological features of recurrent adenoid cystic carcinoma (ACC) of the head and neck after carbon ion radiotherapy (CIRT).
Methods or Background: Radiological features (shape, margins, mass effect, contrast enhancement, DWI, mean ADC on ROI, necrosis, side relative to treated area, location relative to initial tumor volume, growth trend, T2 and T1 signal) were retrospectively analyzed in 51 patients with recurrent ACC treated with CIRT between 2013 and 2022 at CNAO. The same variables were assessed in 51 patients without recurrence (control group). In recurrent cases, imaging was evaluated at three time points (t.p.): 6 months after treatment (t.p.0), 6 months before recurrence (t.p.1), and at recurrence (t.p.2). For controls, t.p.2 corresponded to the mean recurrence time of the other group (34 months). ΔADC values across time points were also calculated in both groups.
Results or Findings: Compared with controls, recurrent ACC exhibited significantly more frequent features (p<0.001), including rounded or lobulated shape, mass effect, homogeneous enhancement, mean ADC of 1098 mm²/s, ΔADC of –460 mm²/s, progressive growth, peripheral localization, and intermediate T2 signal. At t.p.1, these characteristics were already comparable to those observed at t.p.2, except for mean ADC, which showed a progressive reduction, and ΔADC, which had already increased compared with t.p.0 (1306 and –250 mm²/s, respectively). A reproducible radiological pattern of recurrent ACC after CIRT was identified, with some features being objectively quantifiable, particularly ΔADC, a parameter scarcely investigated to date.
Conclusion: This preliminary study outlines the main radiological features of ACC recurrence after CIRT and provides a basis for developing response criteria related to salivary gland tumors of the head and neck treated with particle therapy.
Limitations: This is a single-center study
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Approved by local ethical committee of Pavia (PV)

Author Block: C. Zito, R. Cannella, V. Pampalone, F. A. Cordaro, F. Pandolfo, A. Lo Casto; Palermo/IT
Purpose: This study aims to evaluate the diagnostic accuracy of shear wave elastography (SWE) and multiparametric MRI in differentiating benign from malignant parotid gland lesions, based on quantitative assessment of tissue stiffness and perfusion parameters.
Methods or Background: From October 2022 to September 2025, 47 patients (24 females, 23 males; age range 15–85 years) presenting with 53 parotid lesions (size range 8–60 mm) were prospectively evaluated at our Department of Radiological Sciences. Each patient underwent ultrasound with SWE using a 5–14 MHz linear probe. Regions of interest (ROI) were sampled multiple times per lesion (10 measurements each). Additionally, all patients underwent 3-Tesla MRI with dynamic contrast-enhanced sequences. MRI curve patterns were classified based on parameters including peak signal intensity, time to peak (T-peak), wash-in, and wash-out rates. Surgical excision and histopathological analysis were performed for definitive diagnosis.
Results or Findings: A total of 35 parotid lesions underwent histopathological analysis, of which 26 were benign (including 10 cystadenolymphomas and 9 pleomorphic adenomas) and 9 were malignant. Malignant lesions were significantly larger than benign ones (median size: 31 mm vs. 20 mm; p = 0.013). The median SWE values were 40.6 kPa in malignant lesions and 37.5 kPa in benign lesions, with no significant difference (p=0.437). On dynamic contrast-enhanced MRI, the brevity of enhancement showed a tendency toward statistical significance (median values: 88.3 in malignant vs. 49.2 in benign lesions; p = 0.078).
Conclusion: Parotid tumors demonstrated overlapping stiffness values on SWE. However, dynamic contrast-enhanced MRI revealed a trend toward significant differences, suggesting potential value in differentiating malignant from benign lesions.
Limitations: Single-center study. Small number of included patients.
Funding for this study: No funding
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: Z. N. Doghramachi, T. Cankurtaran, M. ALIYEV, T. Yıldırım, A. Isik, G. Kahraman, C. S. Oygur, D. KALEMCİ, O. Kayık, A. Beheshtırooy; Ankara/TR
Purpose: To investigate structural and diffusion-related heterogeneity in salivary glands of Sjögren patients using quantitative MRI and histogram-derived parameters.
Methods or Background: Fifty-six participants (26 with Sjögren’s syndrome, 30 controls; mean age 47 years) underwent MRI of parotid and submandibular glands. Volumetric measurements, inter-rater agreement for Kojima and Makula grading, and quantitative metrics (T1 mapping, ADC, fat fraction, R2*) were assessed. Histogram analyses were performed with dedicated software and included mean, standard deviation, interquartile range (IQR), robust mean absolute deviation (rMAD), root mean square (RMS), entropy, skewness, and kurtosis. Group comparisons were performed.
Results or Findings: Salivary gland volumes were reduced in Sjögren patients (p<0.05). Inter-rater reliability was high for Kojima grading and moderate for Makula. In parotid glands, mean T1 values tended to be higher without significance (p=0.085), while entropy, IQR, rMAD, and SD were significantly elevated (all p<0.05); skewness and kurtosis showed no differences. Parotid glands ADC metrics (mean, RMS, rMAD, SD) were also higher in Sjögren (all p<0.05), whereas submandibular glands showed no significant differences. Fat fraction and R2* values were comparable between groups.
Conclusion: Sjögren’s syndrome is associated with increased parotid gland heterogeneity in both T1 and ADC histogram metrics, while fat and R2* values remain unchanged. Histogram-derived MRI metrics may provide non-invasive biomarkers for salivary gland involvement.
Limitations: Single-center cohort with a moderate sample size; voxel-based or whole-gland segmentation was not performed.
Funding for this study: Funding was provided by Siemens Healthineers, Erlangen, Germany.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by Baskent University Institutional Review Board and Ethics Committee. (reference number KA22/180).

Author Block: B. Wen¹, Z. Zhang¹, Y. Zhang¹, D. Zheng², H. Cui³; ¹Zhengzhou/CN, ²Beijing/CN, ³Hangzhou/CN
Purpose: To establish and validate MR-based radiomics models for differentiating benign from malignant parotid tumors (BT vs MT), and pleomorphic adenomas (PA) from Warthin tumors (WT).
Methods or Background: This two-center study included 200 patients (BT=133 [PA=87, WT=46], MT=67). Radiomics features were extracted from T1WI, fs-T2WI, and fs-CE-T1WI, reduced with Pearson correlation, and used to build machine learning models. A clinical model was developed from demographics and MR features. Fused models combining radiomics and clinical variables were evaluated by ROC and decision curve analysis.
Results or Findings: Radiomics and fused models showed robust diagnostic performance. For BT vs MT, AUCs reached 88.3% (validation) and 86.5% (external test). For PA vs WT, AUCs reached 98.7% and 91.4%, respectively. Clinical models showed lower accuracy. Fused models provided greater clinical benefit than radiomics or clinical models alone.
Conclusion: Fused models integrating radiomics and clinical features achieved high accuracy in differentiating BT from MT and PA from WT, and may support preoperative decision-making.
Limitations: Limited external validation; larger multicenter studies are warranted.
Funding for this study: None
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: