Prognostic Utility of Parameters Derived From Pretreatment Dual-Layer Spectral-Detector CT in Patients With Metastatic Renal Cell Carcinoma.

Abstract

<b>BACKGROUND.</b> New therapies have emerged for metastatic renal cell carcinoma (mRCC), though corresponding imaging markers are lacking. Dual-layer spectral-detector CT (DLCT) can quantify iodine concentration (IC) and effective atomic number (Z_effective), providing information beyond attenuation that may indicate mRCC prognosis. <b>OBJECTIVE.</b> The purpose of our study was to assess the utility of the DLCT-derived parameters IC and Z_effective for predicting mRCC treatment response and survival. <b>METHODS.</b> This prospective study enrolled 120 participants with mRCC from January 2018 to January 2020 who underwent DLCT, with reconstruction of IC and Z_effective maps, before treatment initiation. Final analysis included 115 participants (86 men, 29 women; median age, 65.1 years), incorporating 313 target lesions that were clinically selected using RECIST version 1.1 on arterial phase acquisitions of the chest and abdomen. Semiautomatic volumetric segmentation was performed of the target lesions. Voxels from all lesions were combined to a single histogram per patient. The median IC and Z_effective of the combined histograms were recorded. Measurements above and below the cohort median values were considered high and low, respectively. Univariable associations were explored between IC and Z_effective with objective response rate (ORR), progression-free survival (PFS), and overall survival (OS). Multivariable associations were explored between IC and ORR, PFS, and OS, adjusting for treatment (tyrosine kinase inhibitor vs checkpoint immunotherapy) and significant univariable predictors (including tumor histology and International Metastatic Renal Cell Carcinoma Database Consortium [IMDC] risk factors). <b>RESULTS.</b> At baseline, median IC was 2.26 mg/mL, and median Z_effective was 8.49. In univariable analysis, high IC and high Z_effective were associated with better ORR (both, odds ratio [OR] = 4.35; <i>p</i> = .001), better PFS (both, hazard ratio [HR] = 0.51; <i>p</i> = .004), and better OS (both, HR = 0.38; <i>p</i> < .001). In multivariable models, high IC independently predicted better ORR (OR = 4.35, <i>p</i> = .001), better PFS (HR = 0.51, <i>p</i> = .004), and better OS (HR = 0.37, <i>p</i> < .001); neutrophilia independently predicted worse PFS (HR = 2.10, <i>p</i> = .004) and worse OS (HR = 2.28, <i>p</i> = .003). The estimated C-index for predicting OS using IMDC risk factors alone was 0.650 versus 0.687 when incorporating high attenuation and 0.692 when incorporating high IC or high Z_effective. <b>CONCLUSION.</b> High IC and high Z_effective are significant predictors of better treatment response and survival in mRCC. <b>CLINICAL IMPACT.</b> Baseline DLCT parameters may improve current mRCC prognostic models. <b>TRIAL REGISTRATION.</b> ClinicalTrials.gov NCT03616951.