AUTHOR=Martini Dylan J. , Olsen T. Anders , Goyal Subir , Liu Yuan , Evans Sean T. , Magod Benjamin , Brown Jacqueline T. , Yantorni Lauren , Russler Greta Anne , Caulfield Sarah , Goldman Jamie M. , Nazha Bassel , Kissick Haydn T. , Harris Wayne B. , Kucuk Omer , Carthon Bradley C. , Master Viraj A. , Bilen Mehmet Asim TITLE=Body Composition Variables as Radiographic Biomarkers of Clinical Outcomes in Metastatic Renal Cell Carcinoma Patients Receiving Immune Checkpoint Inhibitors JOURNAL=Frontiers in Oncology VOLUME=11 YEAR=2021 URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.707050 DOI=10.3389/fonc.2021.707050 ISSN=2234-943X ABSTRACT=Background

Immune checkpoint inhibitors (ICI) have revolutionized the treatment of metastatic renal cell carcinoma (mRCC). Biomarkers for mRCC patients treated with ICI are limited, and body composition is underutilized in mRCC. We investigated the association between body composition and clinical outcomes in ICI-treated mRCC patients.

Methods

We performed a retrospective analysis of 79 ICI-treated mRCC patients at Winship Cancer Institute from 2015-2020. Baseline CT images were collected at mid-L3 and segmented using SliceOMatic v5.0 (TomoVision). Density of skeletal muscle (SM), subcutaneous fat, inter-muscular fat, and visceral fat were measured and converted to indices by dividing by height(m)2 (SMI, SFI, IFI, and VFI, respectively). Total fat index (TFI) was defined as the sum of SFI, IFI, and VFI. Patients were characterized as high versus low for each variable at gender-specific optimal cuts using overall survival (OS) as the primary outcome. A prognostic risk score was created based on the beta coefficient from the multivariable Cox model after best subset variable selection. Body composition risk score was calculated as IFI + 2*SM mean + SFI and patients were classified as poor (0-1), intermediate (2), or favorable risk (3-4). Kaplan-Meier method and Log-rank test were used to estimate OS and PFS and compare the risk groups. Concordance statistics (C-statistics) were used to measure the discriminatory magnitude of the model.

Results

Most patients were male (73%) and most received ICI as first (35%) or second-line (51%) therapy. The body composition poor-risk patients had significantly shorter OS (HR: 6.37, p<0.001), PFS (HR: 4.19, p<0.001), and lower chance of CB (OR: 0.23, p=0.044) compared to favorable risk patients in multivariable analysis. Patients with low TFI had significantly shorter OS (HR: 2.72, p=0.002), PFS (HR: 1.91, p=0.025), and lower chance of CB (OR: 0.25, p=0.008) compared to high TFI patients in multivariable analysis. The C-statistics were higher for body composition risk groups and TFI (all C-statistics ≥ 0.598) compared to IMDC and BMI.

Conclusions

Risk stratification using the body composition variables IFI, SM mean, SFI, and TFI may be prognostic and predictive of clinical outcomes in mRCC patients treated with ICI. Larger, prospective studies are warranted to validate this hypothesis-generating data.