151751-Najiba-Chargi

249 Systemic therapy: skeletal muscle mass and anti-cancer drug toxicity SMM and toxicity ≥ grade 3 should be seen as more accurate and trustworthy compared to the meta-analysis for DLT. A limitation of this study is the differences in measurement of SMM and diagnosis of low SMM. All included studies used CT which is the most commonly used and validated technique for SMM measurement 3,8,12 , one study also included MRI measurements 50 . However, studies did measure SMM on different vertebrae levels. Most commonly L3 was used, which is also the most conventionally applied method in literature 12,13 . Several studies in this review used al - ternative vertebrae levels C3 or T4. The methods using these other vertebrae levels have been researched in recent publications but are less frequently used as L3 and some lack validation. The forest plot in Figure 4B shows studies that all used the same measurement technique al - though there was still a difference in the software used, as well as the time between scan and treatment start. This could influence the results, but it is difficult to estimate this influence as there is no previous research on these topics. Especially, the time between scan and treatment start is difficult to interpret as many studies do not report the used time frame. Future research should take this into account for their study design and the results they report. Furthermore, the definition of lowSMM varies between studies. Although some studies use SMI as a continuous variable, most determine a cut-off value to define the presence of low SMM. Most studies use cut-off values from previous publications in similar populations with larger sample sizes. Within this review, the most frequently used cut-offs were those determined by previous studies performed by Prado et al. 46 and Martin et al. 47 . Additional confusion in the already complex field of cut-off values is caused by the incorrect citation of these cut-off values. In this review, five studies cited the cut-off values of Prado et al. 46 but used cut-off values that deviate from those published in the original study. This variation in cut-off values could explain the large range in the occurrence of low SMM, which can be observed in literature as well as in this review (12.2–89.0%). For the optimal diagnosis of low SMM, a universal cut- off value would be preferable. This could be done in a large population of healthy individuals where two standard deviations below average SMI could be seen as a cut-off for low SMM. The leading theory behind the association between low SMM and increased risk of toxicity re- lates to the influence of low SMM on drug distribution. Patients with low SMM have a decreased LBM, as musclemass is a large contributor to LBM. This could cause increased drug levels in the plasma of patients with low SMM and thereby increase the risk of toxicity 6,8,11,17 . Many studies in this review consisted of populations treated with a combination of anti-cancer drugs using different dosing regimens, whichmakes it challenging to compare the drug distribution. There - fore, we specifically focused on studies focused on monotherapy. There was a trend showing increased OR with an increased volume of distribution. This can be seen in the forest plots, as sorafenib has a higher OR for toxicity occurrence when compared to cisplatin and carboplatin, and this correlates with the higher volume of distribution of sorafenib (Figure 5A,B). However, no definitive conclusions can be drawn yet since the sample size in these studies was too low. 13