The Predictive Value of Low Skeletal Muscle Mass Assessed on Cross-Sectional Imaging for Anti-Cancer Drug Toxicity: A Systematic Review and Meta-Analysis

Abstract

Low skeletal muscle mass (LSMM) is increasingly recognized for its predictive value for adverse events in cancer patients. In specific, the predictive value of LSMM has been demonstrated for anti-cancer drug toxicity in a variety of cancer types and anti-cancer drugs. However, due to the limited sample size and study populations focused on a single cancer type, an overall predictive value of LSMM for anti-cancer drug toxicity remains unknown. Therefore, this review aims to provide a comprehensive overview of the predictive value of LSMM and perform a meta-analysis to analyse the overall effect. A systematic search was conducted of MEDLINE, Scopus, EMBASE, and Cochrane. Inclusion criteria were skeletal muscle mass (SMM) evaluated with computed tomography (CT) or magnetic resonance imaging (MRI), articles published in English, SMM studied in humans, SMM measurement normalized for height, and patients did not receive an intervention to treat or prevent LSMM. A meta-analysis was performed using a random-effects model and expressed in odds ratio (OR) with 95% confidence interval (CI). Heterogeneity was assessed using χ² and I² statistics. The search yielded 907 studies. 31 studies were included in the systematic review. Sample sizes ranged from 21 to 414 patients. The occurrence of LSMM ranged from 12.2% to 89.0%. The most frequently studied cancer types were oesophageal, renal, colorectal, breast, and head and neck cancer. Patients with LSMM had a higher risk of severe toxicity (OR 4.08; 95% CI 2.48-6.70; p < 0.001) and dose-limiting toxicity (OR 2.24; 95% CI 1.28-3.92; p < 0.001) compared to patients without LSMM. To conclude, the predictive value of LSMM for anti-cancer drug toxicity can be observed across cancer types. This information increases the need for further research into interventions that could treat LSMM as well as the possibility to adapt treatment regimens based on the presence of LSMM.

Keywords: anti-cancer drugs; cancer; low skeletal muscle mass; meta-analysis; toxicity.

Figure 1

Preferred Reporting Items for Systematic reviews and Meta-analyses (PRISMA) flowchart detailing the study selection process.

Figure 2

Quality in Prognostic Studies (QUIPS) for the included studies.

Figure 3

Forest plots for the association between low skeletal muscle mass (LSMM) and the odds to develop anti-cancer drug toxicity, specifically dose-limiting toxicity (DLT). (A) shows the odds to develop toxicity for all included studies with DLT as the toxicity endpoint. (B) shows the odds to develop DLT for a selected group of studies that besides the same toxicity endpoint also share the same cut-off value established by Martin et al., 2013 [46], as well as the same measurement technique using CT at the L3 vertebrae. (C) shows the odds to develop DLT for a second selected group of studies that share the same cut-off value established by Prado et al., 2008 [45], as well as the same measurement technique using CT at the L3 vertebrae. For each forest plot, the combined effect of the studies is plotted with a black diamond. * The patient population in the study by Kurk et al., 2019, received sequential treatments. The odds ratio was determined for each treatment separately and therefore entered separately into the forest plot.

Figure 4

Forest plots for the association between low skeletal muscle mass (LSMM) and the odds to develop anti-cancer drug toxicity, specifically (A) toxicity ≥ grade 3 which was used as the toxicity endpoint in 6 studies. (B) shows a selection of studies that besides the same toxicity endpoint also used the same cut-off values established by Martin et al., 2013 [46], as well as the same measurement techniques using CT at the L3 vertebrae. For each forest plot, the combined effect of the studies is plotted with a black diamond.

Figure 5

Forest plots for association between low skeletal muscle mass (LSMM) and toxicity specifically for monotherapies used in multiple studies. (A) includes patients treated with cisplatin or carboplatin as a monotherapy; (B) includes patients treated with sorafenib as monotherapy; (C) includes patients treated with sunitinib as monotherapy. For each forest plot, the combined effect of the studies is plotted with a black diamond.