Parotid gland excretion recovery at 1 year and mean parotid gland dose were strongly correlated, based on Spearman’s correlation, and mean parotid gland dose was the only significant predictive factor for xerostomia. This finding differs from the report by Beetz et al. [30, 31], which proposed that multiple factors are likely to have separate impacts on xerostomia, although there may be no racial differences in the parotid gland response to irradiation.
Whole-mouth salivary function has been shown to be related to QoL determined by questionnaires . With LKB NTCP modeling in the present study, the TD50 for xerostomia 1-year after RT was 43.6 Gy in the SEF analysis and 44.1 Gy in the QoL analysis. Although these values are higher than that reported by Moiseenko et al. (32.4 Gy) , they are similar to the TD50 reported by Dijkema et al. (39.9 Gy), who analyzed the combined and updated results from two institutions . Deasy et al. suggested that the wide range of reported TD50 values (28.4 to 52 Gy) may result from differences in dose distribution, salivary measurement methods, segmentation, intragland sensitivity, and/or patient geographical location .
Xerostomia-specific questionnaires are reliable and valid for measuring patient-reported xerostomia . In the present study, QoL data were shown to be as valid as SEF values for NTCP parameter modeling. Based on Pearson’s chi-squared test, SEF and QoL data gave similar results. Furthermore, both the Nagelkerke’s R2, which describes overall performance, and the AUC demonstrated that both datasets produced similar results, and the Hosmer–Lemeshow test showed no significant disagreement between the results determined from the SEF and QoL data. No significant difference was noted regarding dose distributions to the parotid glands.
For the IMRT planning goal, the mean dose to each parotid gland should be as low as possible while providing the desired clinical target volume coverage . In our analysis, the incidence of grade 3+ xerostomia at 1 year was only ~1% or ~2% for the QUANTEC-recommended cut-off points of 20 Gy or 25 Gy, respectively. Hence, the severe xerostomia would usually be avoided when at least one parotid gland is spared to a mean dose ≤20 Gy or when both glands have been spared to a mean dose ≤25 Gy . A lower parotid mean dose also results in better QoL for patients .
Potential limitations of the present study include the low number of patients with xerostomia toxicity. Although the SEF values before RT were normally distributed, confirmation in a larger sample is needed to validate the NTCP model. Grade selection for the endpoint is another potential limitation in the present study. Choosing a lower grade of xerostomia may provide more valuable dose constraints for preventing complications, as even grade 2 xerostomia significantly diminishes QoL for patients . We used a previously definition for moderate to severe xerostomia based on the QLQ-HN35 questionnaire [2, 30, 31]. However, to our knowledge, no direct evidence exists to clarify this definition or to determine whether it is similar to the grade 3+ xerostomia definition by the subscales of LENT-SOMA criteria. Here, we showed the similar NTCP mapping results between moderate-to-severe xerostomia and LENT-SOMA subscales grade 3+ xerostomia. The practical implications of our results are validation of the use of a QoL form (EORTC QLQ-H&N35) as a surrogate for whole mouth salivary function, and also an important validation of previously proposed QAUNTEC treatment planning constraints to avoid xerostomia. Therefore, follow the QUANTEC guidelines have benefits to result in generic QoL improvement. Preserving more gland function should be pursued as a planning goal when consistent with adequate target dose coverage. Further researches will investigate as to whether new radiation techniques or different study cohorts (combining multiple institutional or cooperative group data sets) could be further validated this finding.