Abstract BACKGROUND Primary central nervous system lymphoma (PCNSL) is a rare, aggressive diffuse large B-cell lymphoma, confined to the central nervous system. Age and performance status are validated prognostic variables in PCNSL. Frailty may also be a factor but is not easily quantified. Temporal muscle thickness (TMT), as parameter of sarcopenia, is believed to reflect frailty, is easily measurable and may be a valuable additional prognostic factor: mean TMT has been described as an independent prognostic factor in PCNSL patients. This study aims to validate these results in a prospective cohort. MATERIAL AND METHODS All patients from a large phase III trial (HOVON 105/ALLG NHL 24; EudraCT 2009-014722-42) of whom baseline axial T1 weighted magnetic resonance images (MRI) were available were included. Sarcopenia is defined as a mean TMT <6.3mm for men and <5.2mm for women, measured on axial T1-weighted MRI. The TMT was measured by a radiologist (RG) with and a neurologist (MvdM) without correction for angulation of images. Sarcopenia, together with known other prognostic factors were analyzed in a Cox-regression analysis for progression-free (PFS) and overall survival (OS) for both investigators. RESULTS TMT measurements were possible in 164/199 trial patients. The median mean TMT was 7.62mm (IQR: 6.58-9.00; measured by RG) and 5.94mm (IQR: 5.09-7.01; measured by MvdM). The mean TMT was not significantly different between those who used dexamethasone, and those who did not. Sarcopenia was present in 75 patients (46%), measured by MvdM without correction for angulation of the images and in 21 patients (13%), measured by RG with correction for angulation of the images. In univariate analyses, age (>60 years-old) and baseline Mini-Mental State Examination-score <27 were significantly associated with adverse PFS and OS. In multivariate analyses, only age remained as independent prognostic factor for adverse PFS and OS. Sarcopenia, as measured by either of the investigators, was not associated with survival. For measurements by RG, presumed to be most accurate, the Hazard Ratio (95% confidence interval) was 0.68 (0.33-1.40), p=0.30 for PFS, and HR (95% CI): 0.97 (0.42-2.20), p=0.93 for OS in the multivariate analysis. CONCLUSION There was a low interobserver agreement between a neurologist and a radiologist regarding TMT measurements, which questions the use of this parameter in daily practice by non-radiologists. Sarcopenia, based on the mean TMT, was not a prognostic factor in PCNSL patients for PFS and OS. The small number of patients with sarcopenia in this series may have influenced results.
Read full abstract