151751-Najiba-Chargi

38 CHAPTER 2 average, suggesting a reasonably good agreement. The ICC between estimated SMA at L3 and actual SMA at L3 was good: 0.78 (95% CI: 0.61 – 0.86, p < 0.01). Figure 3. Correlation between estimated cross-sectional SMA (CSMA) at the level of L3 and actual cross-sectional SMA (CSMA) at the level of L3 AGREEMENT AND ACCURACY IN IDENTIFICATION OF PATIENTS WITH LOW SKEL- ETAL MUSCLE MASS Using Formula 2, the estimated LSMI and actual LSMI were calculated. The previously pub - lished cut-off value of ≤43.2cm 2 /m 2 was used to determine low SMM. Using this cut-off value, 96 patients were determined to have low SMM using the estimated LSMI, and 77 patients had low SMM using the actual LSMI; see Table 2. The sensitivity of identifying patients with low SMM using the estimated LSMI and a cut-off of ≤43.2cm 2 /m 2 was 84.4% and the specificity was 74.8%. The positive predictive value (PPV) of the estimated LSMI was 67.7% and the negative predictive value (NPV) was 88.5%. The false positive value, indicating the number of patients that incorrectly were identified as having low SMM, was 25.2%. Cohen’s kappa for agreement between low SMM using the estimated and the actual LSMI was 0.57, indicating moderate agreement.