Radiation-induced xerostomia is related to stem cell dose-dependent reduction of saliva production

Purpose /Objective: Despite the use of advanced radiation techniques, patients with head and neck cancer (HNC) still experience xerostomia due to radiation-induced salivary gland damage. Previous studies showed that mean dose to the parotid gland Stem Cell Rich regions (DSCR) is the strongest predictor for the risk of patient-reported daytime xerostomia. However, the reason why DSCR results in more xerostomia, is not yet elucidated. Therefore, this study tested the hypothesis that the relation between daytime xerostomia and DSCR is explained by reduced saliva production related to dose to stem cells.
Material/Methods: In 570 HNC patients treated with definitive radiotherapy (RT), parotid glands flow (FLOWPAR), submandibular/sublingual glands flow (FLOWSMSL) and patient-reported daytime (XERDAY) and nighttime xerostomia (XERNIGHT) were prospectively measured before, and at 6 (M6) and 12 months (M12) after RT. Using linear mixed effect models, relations of mean dose to the parotid glands (Dpar), DSCR, non-SCR parotid gland tissue (DnonSCR), submandibular glands (Dsub) and oral cavity (Doral) with salivary flow and xerostomia were analysed while correcting for known confounders.
Results: As expected, Dpar was associated with FLOWPAR (p<0.001), while Dsub affected FLOWSMSL (p<0.001). After dividing the parotid gland into the SCR and non-SCR region, DSCR proved to be responsible for the negative effect of DPAR on FLOWPAR (p≤0.03), while DnonSCR was not associated with FLOWPAR (p≥0.11). Typical trajectories illustrated the enhanced effect of DSCR on FLOWPAR (Figure 1). For example, increasing DSCR by 10 Gy at fixed DnonSCR, reduced FLOWPAR by 0.022 (27%) and 0.024 ml/min (22%) at M6 and M12, respectively. In contrast, increase of DnonSCR while keeping DSCR constant did not significantly decrease FLOWPAR (0.002 [3%] and 0.004 ml/min [4%] at M6 and M12, respectively).
Moreover, besides salivary flow, Dpar also affects patient-rated xerostomia. In line with previous findings, higher DSCR increased XERDAY scores (p=0.09 [M6], p=0.03 [M12]), but not XERNIGHT scores (p≥0.33). Furthermore, Doral affected XERDAY (p=0.05 [M12]) while Dsub affected XERNIGHT (p=0.01 [M6]).
To test whether the negative effect of dose to salivary glands on xerostomia was achieved through reduced saliva production, the observed relations between dose and xerostomia were corrected for salivary flow. Indeed, the significant effect of DSCR on xerostomia disappeared after adding FLOWPAR and FLOWSMSL to the models. Thus, DSCR only increased XERDAY scores via reducing FLOWPAR, whereas the effects of Doral and Dsub on xerostomia were independent of FLOWPAR and FLOWSMSL.
Conclusion: Dose to Stem Cell Rich region (DSCR) reduces parotid gland saliva production, leading to higher patient-reported daytime xerostomia scores. No other influences of DSCR on xerostomia were observed. Consequently, DSCR should be minimized to preserve salivary flow with aiming at reducing the risk of daytime xerostomia.