Dosimetric, Volumetric, and Positional Stability After Submandibular Gland Transfer

PURPOSE/OBJECTIVE(S) Xerostomia and its effects are significant challenges for patients receiving radiation therapy (RT) for head and neck cancers (HNC). Submandibular gland (SMG) transfer has demonstrated promise in reduction of subjective and objective xerostomia. However, SMG transfer investigations have largely employed conventional RT or adjuvant RT, with little evaluation in patients treated with definitive modern RT. Furthermore, there has been no assessment of whether anatomic changes during RT may affect the potential benefits of SMG. We sought to characterize the preliminary dosimetric, volumetric, and positional stability of SMG transfer on an ongoing prospective case-control study. MATERIALS/METHODS Previously untreated patients with oropharyngeal HNC, N0-N2b (unilateral neck disease) were enrolled. SMG transfers were performed by experienced surgeons per the Jha and Seikaly technique, transposing the gland anteriorly into the submental neck. Definitive VMAT plans were optimized according to institutional guidelines (goal mean SMG dose < 26 Gy), patients were treated to 70 Gy in 33-35 fractions. To assess the stability of planned dose and change in SMG volume and position, repeat CT simulation was performed at week 4. These were rigidly registered to the initial simulation centered about the mandible and hyoid, with the original dose superimposed. Glands were meticulously re-contoured on both scans. Statistical comparisons were by Wilcoxon paired rank tests or Mann-Whitney U tests. RESULTS To date, 11 patients with SMG transfer and 5 controls have been treated, all male. In 3 cases, no re-simulation was available and a week 4 cone-beam CT was used. The mean time from gland transfer to RT start was 28.3 days. From simulation to week 4, transposed SMGs shrunk by a mean of 3.49 cc (27.6%), no different than corresponding contralateral SMGs at 2.37 cc (23.5%) or control SMGs (SMGs in control patients on the N0 side) at 1.71 cc (22.8%). There were no associations between SMG volume change and dose by linear regression or Spearman's rank correlation. Transposed SMGs had an average mean dose of 28.2 Gy on initial sim. At week 4, extrapolated to a full course, this increased to 32.8 Gy (P = 0.01). Average mean dose to control glands was 46 Gy, statistically greater than transposed glands (P = 0.03), with a non-significant 1.5 Gy increase at week 4. Mean change in geometric position was 0.45 cm for transposed glands, 0.37 cm for contralateral glands (P = .19), and 0.32 for control glands (P = .37). CONCLUSION SMG transfer achieved significant reduction in dose compared to controls. While SMG shrinkage and movement were not separately different from contralateral or control glands, the higher dose at re-simulation indicates these factors may together have an effect, calling for careful SMG delineation, attention to daily position and image guidance, and planned re-simulation in order to maximize dosimetric benefits. Our trial is ongoing and clinical and QoL endpoints will be examined with these findings in mind.