151751-Najiba-Chargi

423 Summarizing discussion and future perspectives the study by Swartz et al. The results of this study show a good correlation (r=0.75, p<0.01) between segmented cross-sectional skeletal muscle area at the level of C3 and L3. When using the multivariable prediction formula proposed by Swartz et al. to calculate the skel - etal muscle area at the level of L3 using skeletal muscle area at C3, gender, age and weight, the correlation between skeletal muscle area at C3 and L3 improved (r = 0.82, p<0.01). There was some difference in the identification of patients with low skeletal muscle mass based on the calculated lumbar skeletal muscle index and the actual lumbar skeletal muscle index (Cohen’s κ: 0.57; 95% CI 0.45-0.69), although the sensitivity of identifying patients with low skeletal muscle mass using the estimated lumbar skeletal muscle index was high (84.4%). This study shows that a measurement of skeletal muscle mass at the level of C3 provides an easy and robust alternative for estimation of the skeletal muscle mass of a patient. Chapter 3 presents an association study for skeletal muscle mass measurements at the level of C3 on CT and magnetic resonance imaging (MRI). MRI does not allow for radiodensity-based seg - mentation of muscle tissue and is therefore subject to observer interpretation (i.e., the total muscle surface). A large proportion of head and neck cancer patients undergo diagnostic MRI instead of CT imaging in the diagnostic process. Therefore, patients’ skeletal muscle mass was segmented on both CT and MRI and the association between these measurements was analyzed. An excellent intraclass coefficient was found (0.97; 95% CI 0.94-0.98, p<0.01). The mean difference of skeletal muscle area measurements between CT and MRI was less than 1cm 2 . Occasionally, skeletal mass measurement at the level of C3 may be impaired by exten - sion of primary tumor, lymph nodes or previous treatment. Therefore, Chapter 4 presents a correlation study to investigate whether skeletal muscle mass measurements of a single muscle, the masseter muscle, correlates with measurements of skeletal muscle mass on a single slice at the level of C3 and L3. Themasseter muscle was chosen because it is consistently present on routine head and neck imaging, is rarely impacted by disease or treatment and is quick and easy to characterize. Several masseter muscle parameters (muscle mass volume, musclemass area, muscle thickness) were significantly correlated with skeletal muscle area at the level of C3 and L3. However, these correlations varied frommoderately to strong, with the strongest correlation found between skeletal muscle area at C3 and masseter mass volume (r=0.67). Skeletal muscle area of the masseter muscle had only moderate correlation with skeletal muscle area at the level of L3 (r=0.47) and C3 (r=0.57). The prognostic impact of low skeletal musclemass, obtained bymeasurement of themasseter muscle index, for survival was additionally investigated and this showed that low masseter muscle index was a significant prognostic for decreased survival (HR 3.0, p<0.05). In patients without cross-sectional imaging at level L3 or C3 or with impaired C3 measurements, masseter muscle parameters could serve as an alternative for skeletal muscle mass assessed by skeletal muscle area measurements at these vertebral levels. Because of the lack of reference values of skeletal muscle mass in the general population and the heterogeneity in cut-off values for low skeletal muscle mass in literature, Chapter 5 presents a study performed in a large cohort of head and neck cancer patients (n=1415) to develop cut-off values for low skeletal muscle mass measured at the level of C3. Because of the significant correlation between skeletal muscle mass index with gender (r 2 =0.4, p<0.01) and body-mass index (BMI) (r 2 =0.4, p<0.01), gender and BMI-specific 21