Integrating submandibular gland transfer with proton therapy to preserve salivary function Authors: Sreenija Yarlagadda, Geoffrey D Young, Robert M Rohe, Valeriane McNeill, Joanna R Nemec, Chelsea P Tohtz, Lorein Gonzalez, Kristen A McConnell, Noah S Kalman Published: 2026 ABSTRACT AbstractObjectiveSubmandibular gland (SMG) transfer to the submental space in patients undergoing head and neck radiation therapy was historically performed to decrease radiation-associated xerostomia in the era of 2D/3D radiotherapy. In this study, we evaluated the dosimetric significance of SMG transfer with intensity modulated proton therapy (IMPT).MethodsWe reviewed patients with oropharyngeal carcinoma who received curative intent definitive radiation and had contralateral (C/L) SMG transfer. We registered the baseline diagnostic computed tomography (CT) scan of these patients with the CT simulation images and delineated pre-transfer and post-transfer SMG locations. IMPT pre-transfer plans were generated, taking into consideration the initial reference SMG position, and then re-optimized based on the post-transfer SMG position. Each patients' dosimetric parameters from both plans were compared for mean SMG doses, target coverage and other OAR doses.ResultsSeventeen patients were included in the analysis. All patients received definitive radiation (66–70 Gy/30–35 fractions). Pre-transfer plans had a C/L SMG mean dose of 32.92 Gy (range [R]: 17.82–47.6 Gy), which decreased to 20.22 Gy (R: 9.75–38.87 Gy), p < 0.001 on re-optimization. Post-transfer plans met the mean dose constraints more often than pre-transfer plans. The difference in C/L SMG mean dose with gland transfer was 12.02 Gy (95% CI: 9.46–14.59 Gy; p < 0.001). There was no significant difference in target coverage, ipsilateral SMG, and any other OARs between the plans.ConclusionThis pilot study demonstrates that SMG transfer enables better C/L SMG dose sparing during IMPT. Future prospective studies with larger patient cohorts are warranted to support clinical translation.Keywords: SMG transfer, IMPT, HNC, OPSCC, xerostomia Introduction Xerostomia is a common and potentially debilitating toxicity of head and neck radiotherapy, leading to difficulty with swallowing, talking, and maintaining oral hygiene. [1], [2] The parotid and submandibular glands contribute distinct yet complementary components of salivary secretion: the parotid glands account for the majority of stimulated salivary output, whereas the submandibular glands generate approximately two-thirds of unstimulated saliva, [3] which is critical for maintaining basal oral moisture. Mean dose to the submandibular gland (SMG) has demonstrated a significant association with xerostomia, based both objectively through salivary flow rates [4] and subjectively through patient-reported questionnaires. [5] Murdoch-Kinch et al. established a dose-effect relationship between mean SMG dose and salivary flow rates, reporting a decline of 1.2% per Gy in stimulated flow rates, and 3% per Gy in unstimulated flow rates, specifically below the threshold of 39 Gy. [4]. Various strategies have been explored to reduce radiation dose to the SMG. Seikaly et al. reported one such approach that involves surgical transposition of the SMG to the submental space, positioning the gland outside of planned radiation fields. [6], [7], [8] This technique was commonly used in the 2D and 3D conformal radiation therapy (CRT) era. However, with the advent of high precision radiotherapy techniques such as intensity modulated photon radiation therapy (IMRT) and intensity modulated proton therapy (IMPT), the dosimetric benefit of SMG transfer was unclear. In a prior analysis, we observed no statistically significant differences in SMG doses with/without SMG transfer in IMRT/VMAT plans, raising doubts about the procedure's clinical relevance in the high precision radiotherapy era. [9] While IMPT offers dosimetric advantages over IMRT in head and neck radiotherapy [10], [11], practical considerations may limit the benefit of IMPT for SMG sparing. Beam angles are frequently constrained to avoid dental amalgam, and the use of range shifters increase spot size and lateral scatter, which could lead to increased dose penumbra. In this study, we evaluated the dosimetric significance of SMG transfer in patients receiving IMPT. Materials/methods A retrospective review of patients with clinically contralateral node negative oropharyngeal squamous cell carcinoma (OPSCC) was conducted after approval from the institutional review board. Patients who received curative intent definitive radiation therapy after contralateral (C/L) SMG transfer were included. All patients received ≥54 Gy to the ipsilateral and contralateral neck levels II-IV. Patient and disease characteristics were collected from the electronic medical record. Baseline diagnostic computed tomography (CT) scan of these patients were registered with the CT simulation images in Raystation. Baseline SMG locations were contoured using the diagnostic CT, while post-transfer SMG locations were contoured as visualized on the planning CT. The contours were delineated by a trained radiation oncologist and were reviewed by another head and neck expert radiation oncologist for consistency. Target volumes and organs-at-risk (OARs) were delineated using a standard clinical institutional protocol. IMPT standard of care plans were generated with post-transfer SMG location to limit mean doses ≤39 Gy (prioritizing target coverage); subsequently the plans were re-optimized with the baseline pre-transfer SMG position. Planning parameters such as beam number and beam angles were identical between patient plans. Dosimetric parameters including target coverage and OAR doses were recorded for both the plans using Raystation 12 A (RaySearch Laboratories, Stockholm, Sweden). Descriptive statistics were computed, and median values with range were presented for continuous variables. The number of plans that met SMG mean dose constraints ≤26 Gy, ≤35 Gy, and ≤39 Gy were computed. Wilcoxon signed rank test was performed to compare SMG mean doses, target coverage, and other OAR dosimetric parameters between pre- and post-transfer plans. McNemar test was used to compare the frequencies and proportion of cases that met the dose constraints with SMG transfer. p ≤0.05 was set as the level of significance for all comparisons. Statistical analysis was performed using SPSS, version 27 (SPSS Inc., Chicago). Results Seventeen OPSCC patients were included in the analysis. The majority had tonsil primaries (n = 10, 59%) while the remaining patients had base of tongue tumors (n = 7, 41%). Notably, ten primary tumors (58.8%) crossed the midline. All patients received definitive radiation to a total dose of 66–70 Gy in 30–35 fractions with bilateral neck treatment. A representative case was depicted in Fig. 1A with pre-transfer location of SMG and its proximity to the dose distribution on a sagittal CT slice; and Fig. 1B highlights the physical separation created with SMG transfer in the same case. Pre-transfer plans had an average C/L SMG mean dose of 32.92 Gy (range [R]: 17.82–47.6 Gy), which decreased to 20.22 Gy (R: 9.75–38.87 Gy) with SMG transfer. The median difference in C/L SMG mean dose with gland transfer was 12.02 Gy (95% CI: 9.46–14.59 Gy; p <0.001). Post-transfer plans met the mean dose constraints more often than pre-transfer plans for all the three dose thresholds, but the difference was statistically significant only for ≤26 Gy (p = 0.004). A comparison of pre-transfer and post-transfer plans by overall distribution, absolute dose difference between the plans for each patient, and percentage of plans that met various C/L SMG mean dose constraints was provided in Fig. 2. There was no clinically or statistically significant difference in target coverage, ipsilateral SMG, and any other OARs between the plans (Table 1). A representative case depicting dose distribution: 1A with pre-transfer submandibular gland in orange & 1B with post-transfer submandibular gland in yellow. A comparison of pre-transfer and post-transfer plans. 2A: Paired spaghetti plot combined with split violin plot of contralateral SMG mean dose. Each line represents an individual patient (n = 17), connecting mean dose between pre-transfer and post-transfer plans. Horizontal dashed lines (blue) indicate commonly used clinical dose constraints (26 Gy, 35 Gy, 39 Gy). Overlaid split violin plots depicts the distribution and density of doses at each timepoint. 2B: Waterfall plot of the absolute dose difference for each patient. The dashed horizontal line at 0 Gy indicates no change and bars extending downward indicate a reduction in dose, while bars extending upward indicate an increase. 2C: Bar charts comparing the percentage of cases that met various contralateral submandibular gland mean dose constraints (≤ 26 Gy, ≤ 35 Gy, ≤ 39 Gy). Table 1Comparison of target coverage and other OAR doses between the plans.Pre-transfer plansPost-transfer plansp-valueTarget coverage (D99%)High-risk CTV99.09%99.02%0.660Intermediate risk CTV99.56%99.55%0.190Low risk CTV99.20%99.20%0.977 OAR mean dosesIpsilateral sub-mandibular gland64.49 Gy64.72 Gy0.868Ipsilateral parotid33.86 Gy34.36 Gy0.856Contralateral parotid25.82 Gy25.34 Gy0.981Oral cavity22.41 Gy22.60 Gy0.740Larynx24.62 Gy24.54 Gy0.246 Comparison of target coverage and other OAR doses between the plans. Among the 10 cases with tumors crossing the midline, pre-transfer plans had slightly higher C/L SMG mean doses, with a median of 34.20 Gy (R: 17.82–47.6 Gy), which decreased to 22.04 Gy (R: 12.07–38.87 Gy) with SMG transfer (p = 0.007). The proportion of cases that met the SMG mean dose constraints increased from 70% to 100% (≤39 Gy), 60% to 90% (≤35 Gy) and 20% to 70% (≤26 Gy). For the 7 patients with lateralized lesions, pre-transfer plans had a median C/L SMG dose of 30.38 Gy (R: 20.25–43.39 Gy), which decreased to 20.10 Gy (R: 9.75–24.57 Gy) with SMG transfer (p = 0.018). While 86% (≤39 Gy), 71% (≤ 35Gy), 43% (≤26 Gy) of plans met the C/L SMG mean dose constraints pre-transfer, 100% of plans met the different constraints post-transfer. Discussion In this study, we evaluated the utility of SMG transfer in a uniform cohort of patients with OPSCC receiving head and neck proton therapy. The results demonstrated that SMG transfer was associated with a statistically significant reduction in C/L SMG mean doses. This dosimetric improvement translated into a significantly higher proportion of cases meeting established SMG mean dose constraints. Interestingly, our current results of improved SMG doses post-transfer with proton therapy stand in contrast to the lack of improved SMG doses post-transfer with IMRT plans [9]. This likely stems from unique limitations in current proton therapy planning techniques, including beam number and beam angle constraints, as well as increased dose penumbra from range shifter use. This can uniquely increase SMG doses given their abutment of the level II neck targets and proximity to the mandible and possible dental amalgam. Dose-response relationships for SMGs are varied. Murdoch-Kinch et al. demonstrated an exponential decline in selective SMG salivary output as mean SMG dose increased, identifying the critical threshold around 39 Gy. [4] The QUANTEC guidelines advocate for sparing of at least one parotid and one SMG to preserve salivary function, with suggested SMG mean doses ≤35 Gy. [12] Overall, variability in clinical practice persists, with some centers adopting stricter constraints down to ≤26 Gy with advanced techniques such as IMPT. Overall, the statistically significant reduction in mean SMG dose provided by SMG transfer would be expected to improve gland function and patient quality of life. Our findings highlight that SMG transfer retains utility across a range of presentations. The procedure benefited not only patients with tumors extending beyond the midline, but also for well lateralized tumors (Fig. 2B). This broader applicability addresses a notable challenge in IMPT planning for the head and neck region, where despite superior dose conformity, the C/L SMG still receives a low-to-intermediate dose spill due to target geometry or beam arrangement. In this study, target coverage was maintained across plans. The improved dose sparing with the SMG transfer thus represents a complementary approach to the inherent advantages of proton therapy. Given the favorable survival outcomes of OPSCC, preservation of salivary function has become an increasingly important consideration in contemporary head and neck radiotherapy. High precision radiotherapy starting with IMRT has demonstrated benefit in reducing doses to the parotid glands [13] as well as the SMGs. [14], [15] Importantly, such dose reductions have translated into improved patient-reported outcomes [5]. The safety and effectiveness of SMG transfer in prevention of RT-induced xerostomia has previous reported. [7], [8], [16], [17], [18] The dose reductions demonstrated in this study would similarly be expected to result in improved SMG function. Study limitations include its retrospective nature and relatively small sample size, which may limit the generalizability of our results. Additionally, patients undergoing SMG transfer in this cohort tended to receive somewhat higher elective volume doses compared to contemporary HPV-positive oropharynx protocols in which nodal de-escalation strategies are being investigated, which could potentially offset some of the observed relative dosimetric gains. Clinical salivary function outcomes were also not available for this cohort, which could further limit the potential clinical impact of the study. Future prospective trials incorporating patient-reported outcomes and objective salivary flow assessments will be essential to validate the clinical significance of the observed dosimetric benefits. Overall, in patients receiving head and neck proton therapy, SMG transfer has a role in mitigating salivary toxicity. As the use of proton therapy for head and neck cancer continues to expand [10], [11], multidisciplinary approaches may offer the greatest potential for reducing xerostomia and preserving long-term quality of life in patients receiving head and neck radiotherapy. Conclusion In this pilot study, SMG transfer provided a statistically significant benefit in reducing C/L SMG doses in OPSCC patients treated with IMPT. By facilitating SMG sparing, this approach complements parotid-sparing techniques in optimizing overall salivary function preservation. The observed benefits are relevant for both well-lateralized tumors and those crossing midline. Future prospective studies with larger patient cohorts are warranted to support clinical translation. CRediT authorship contribution statement Sreenija Yarlagadda: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Geoffrey D. Young: Methodology, Investigation. Robert M. Rohe: Writing – review & editing, Methodology, Investigation. Valeriane McNeill: Writing – review & editing, Methodology, Investigation. Joanna R. Nemec: Writing – review & editing, Methodology, Investigation. Chelsea P. Tohtz: Writing – review & editing, Methodology, Investigation. Lorein Gonzalez: Methodology, Investigation. Kristen A. McConnell: Writing – review & editing, Methodology. Noah S. Kalman: Writing – review & editing, Writing – original draft, Supervision, Methodology, Conceptualization. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used (Grammarly) to improve the manuscript's language and punctuation. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. Funding None. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments None. References