Identifying low muscle mass and monitoring body composition changes in newly diagnosed cancer patients: Agreement between multifrequency bioelectrical impedance analysis and computed tomography

Objectives: Low muscle mass (MM) is significant in cancer patients, and computed tomography (CT) is considered the reference standard for MM assessment. We investigated the consistency of CT and multifrequency bioelectrical impedance analysis (mBIA) in detecting body composition at baseline and during anticancer treatment and the relationship between MM and malnutrition as well as complications in lung and cervical cancer patients.

Methods: Abdominal CT and mBIA were conducted to assess body composition at baseline for all patients and repeated for patients with cervical cancer after 4 wk of chemoradiotherapy. Concordance was compared by intraclass correlation coefficient and Bland-Altman plots. Receiver operating characteristic analysis was used to assess the diagnostic efficacy of mBIA for low MM. Correlation analysis was conducted to assess the relationship between MM and Nutritional Risk Screening 2002 and Global Leadership Initiative on Malnutrition. Furthermore, we assessed whether there was a difference in the incidence of chemoradiotherapy side effects in the low MM group derived by CT or mBIA.

Results: A total of 77 cervical and 73 lung cancer patients were enrolled. mBIA showed fair discriminative capacity (area under the curve = 0.651) for detecting low MM, the concordance of skeletal MM and visceral fat area between CT and mBIA was good (intraclass correlation coefficient = 0.712 and 0.698, respectively), and mBIA and CT had consistent observations of muscle and fat changes (P = 0.051 and 0.124, respectively). There was no difference in the incidence of chemoradiotherapy side effects in the low MM group compared with controls regardless of whether derived by CT or mBIA (P > 0.05). MM was correlated with Nutritional Risk Screening 2002 and Global Leadership Initiative on Malnutrition but showed unsatisfactory prediction of malnutrition (area under the curve <0.7).

Conclusions: mBIA- and CT-derived body composition was highly correlated, and agreement was reached on body composition changes during treatment.