Longitudinal Multiparametric MRI Assessment of Irradiated Salivary Gland in a Rat Model: Correlated With Histological Findings

Application of Diffusion Kurtosis Imaging in Evaluating Acute Xerostomia in Nasopharyngeal Carcinoma Treated With Induction Chemotherapy Plus Concurrent Chemoradiotherapy

Purpose

The aim of this study was to identify the efficacy of diffusion kurtosis imaging (DKI) in tracking and monitoring the dynamic change of parotid glands (PGs), submandibular glands (SMGs), sublingual glands (SLGs), and acute xerostomia in nasopharyngeal carcinoma (NPC) patients treated with induction chemotherapy (IC) plus concurrent chemoradiotherapy (CCRT).

Methods

The prospective study recruited 42 participants treated with IC+CCRT. All patients underwent DKI scanning six times: before IC, before RT, in the middle of the RT course, immediately after RT, and 1 and 3 months post-RT. Mean diffusion coefficient (MD) and mean kurtosis (MK) of PG, SMG, SLG, saliva flow rate measured under resting (uSFR) and stimulated condition (sSFR), and xerostomia questionnaire (XQ) scores were recorded.

Results

At each time point, sSFR was significantly higher than uSFR (p < 0.05 for all). MD of the salivary glands and XQ scores increased over time while MK, uSFR, and sSFR decreased. After IC, the significant differences were detected in MD and MK of bilateral SMG and MK of the left SLG (p < 0.05 for all), but not in MD and MK of PG, uSFR, sSFR, and XQ scores. After RT, sSFR at 1m-RT decreased significantly (p = 0.03) while no significant differences were detected in uSFR and XQ scores. Moderate-strong correlations were detected in ΔMD-PG-R%, ΔMK-PG-R%, ΔMD-PG-L%, ΔMK-PG-L%, ΔMD-SMG-R%, ΔMK-SMG-R%, ΔMD-SMG-L%, ΔMK-SMG-L%, and ΔMD-SLG-R%, with correlation coefficients (p < 0.05 for all) ranging from 0.401 to 0.714. ΔuSFR% was correlated with ΔMD-SMG% (p = 0.01, r = -0.39), ΔMD-SLG% (p < 0.001, r = -0.532), and ΔMK-SMG% (p < 0.001, r = -0.493). ΔsSFR% correlated with ΔMD-PG% (p = 0.001, r = -0.509), ΔMD-SMG% (p = 0.015, r = -0.221), and ΔMK-PG% (p < 0.001, r = 0.524). ΔXQ% was only correlated with ΔMK-PG% (p = 0.004, r = 0.433).

Conclusion

DKI is a promising tool for tracking and monitoring the acute damage of PG, SMG, and SLG induced by IC+CCRT in NPC patients.