Open Access Highly Accessed Research article

Normal tissue complication probability model parameter estimation for xerostomia in head and neck cancer patients based on scintigraphy and quality of life assessments

Tsair-Fwu Lee1*, Pei-Ju Chao12, Hung-Yu Wang1, Hsuan-Chih Hsu2, PaoShu Chang34 and Wen-Cheng Chen5

Author Affiliations

1 Medical Physics and Informatics Laboratory, Department of Electronics Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung, Taiwan, ROC

2 Department of Radiation Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan, ROC

3 Department of Radiation Oncology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan, ROC

4 Department of Medical Imaging and Radiological Sciences, Kaohsiung Medical University, Kaohsiung, Taiwan, ROC

5 Departments of Radiation Oncology, Chang Gung Memorial Hospital, Chiayi, Taiwan, ROC

For all author emails, please log on.

BMC Cancer 2012, 12:567  doi:10.1186/1471-2407-12-567

Published: 4 December 2012



With advances in modern radiotherapy (RT), many patients with head and neck (HN) cancer can be effectively cured. However, xerostomia is a common complication in patients after RT for HN cancer. The purpose of this study was to use the Lyman–Kutcher–Burman (LKB) model to derive parameters for the normal tissue complication probability (NTCP) for xerostomia based on scintigraphy assessments and quality of life (QoL) questionnaires. We performed validation tests of the Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) guidelines against prospectively collected QoL and salivary scintigraphic data.


Thirty-one patients with HN cancer were enrolled. Salivary excretion factors (SEFs) measured by scintigraphy and QoL data from self-reported questionnaires were used for NTCP modeling to describe the incidence of grade 3+ xerostomia. The NTCP parameters estimated from the QoL and SEF datasets were compared. Model performance was assessed using Pearson’s chi-squared test, Nagelkerke’s R2, the area under the receiver operating characteristic curve, and the Hosmer–Lemeshow test. The negative predictive value (NPV) was checked for the rate of correctly predicting the lack of incidence. Pearson’s chi-squared test was used to test the goodness of fit and association.


Using the LKB NTCP model and assuming n=1, the dose for uniform irradiation of the whole or partial volume of the parotid gland that results in 50% probability of a complication (TD50) and the slope of the dose–response curve (m) were determined from the QoL and SEF datasets, respectively. The NTCP-fitted parameters for local disease were TD50=43.6 Gy and m=0.18 with the SEF data, and TD50=44.1 Gy and m=0.11 with the QoL data. The rate of grade 3+ xerostomia for treatment plans meeting the QUANTEC guidelines was specifically predicted, with a NPV of 100%, using either the QoL or SEF dataset.


Our study shows the agreement between the NTCP parameter modeling based on SEF and QoL data, which gave a NPV of 100% with each dataset, and the QUANTEC guidelines, thus validating the cut-off values of 20 and 25 Gy. Based on these results, we believe that the QUANTEC 25/20-Gy spared-gland mean-dose guidelines are clinically useful for avoiding xerostomia in the HN cohort.

NTCP; Xerostomia; Scintigraphy; Quality of Life (QoL); Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC)