151751-Najiba-Chargi

298 CHAPTER 16 the level of L3 for squared height, as shown in formula 2. Sarcopenic obesity was defined as the combination of low SMM in combination with a BMI ≥ 27 kg/m 2 . Formula 1: CSA at L3 (cm 2 ) = 27.304 + 1.363 * CSA at C3 (cm 2 ) – 0.671*Age (years) + 0.640 * Weight (kg) + 26.442*Sex (Sex=1 for female and 2 for male) Formula 2: Lumbar SMI (cm 2 /m 2 ) = CSA at L3/length (m 2 ) STATISTICAL METHODS The optimal stratification method was used to determine cohort specific cutoff values of the SMM. This method is the preferred method in literature and is based on log rank statistics to find the most significant cutoff value for SMM with respect to overall and disease-free sur - vival. 29 Endpoints (OS and DFS) specific cutoff values were determined for the lumbar SMM index and these were used to categorize patients into patients with low SMM and without low SMM for each endpoint. Data analysis was performed using IBM SPSS statistics 25. Descriptive statistics for continuous variables with a normal distribution were presented as mean with standard deviation (SD). Normality was investigated using the Kolmogorov-Smirnov test. The variables age at diagnosis and units of alcohol intake per day were not normally distributed. Variables with a skewed distribution were presented as median with interquartile range (IQR). Categorical variables were presented as frequencies and percentages. Chi-square statistics were used for analyzing differences between the frequencies of each categorical variable with the presence or absence of low SMM. Independent sample student’s t-tests were used for comparing the means of the normally distributed continuous variables with the presence or absence of low SMM. Statistical significance was evaluated at the 0.05 level using 2-tailed tests. Survival was visualized using Kaplan Meier survival curves and number at risk tables. Survival analysis was performed for the subset of patients with known HPV-status, patients with missing HPV-status (n=42) were excluded. A Cox proportional hazard regressionmodel was used for univariate andmultivariate analysis of overall and disease-free survival. Covariates used in the univariate analysis were selected based on clinical relevance based on literature. Covariates used in the multivariate analysis were selected based on statistical significance (p<0.05) in univariate cox regression analysis. In multivariate Cox regression analysis, two models were constructed, each examining the role of low SMM and sarcopenic obesity separately. Statistical significance was evaluated at the 0.05 level using 2-tailed tests.