Plasma and urine free glycosaminoglycans as monitoring and predictive biomarkers in metastatic renal cell carcinoma: A prospective observational study.

Abstract

e16540 Background: No liquid biomarkers are approved in renal cell carcinoma (RCC). In metastatic RCC (mRCC), there is a need to predict and monitor response noninvasively to guide the choice of treatment. Urine and plasma glycosaminoglycan (GAGs) profiles – or GAGomes - are promising biomarkers reflective of RCC metabolism. Here, we explored if GAGomes could predict and monitor response noninvasively. Methods: A single-center prospective consecutive series of mRCC patients elected for first-line therapy was enrolled between June 2016 - April 2019 at Sahlgrenska University Hospital, Gothenburg, Sweden. Response was assessed by the investigator as progressive disease (PD) versus non-PD. Plasma and urine GAGomes were measured at treatment start, after 6 weeks, and every 3rd month in a blinded central laboratory. We conducted Bayesian estimation to correlate GAGomes to response and to design GAG scores to classify PD. So-trained GAG scores were validated to predict PD vs. non-PD at treatment start or after 6 weeks. Results: Fifty patients with treatment-naïve mRCC were enrolled (median age: 68.5 years, 33% female). All received tyrosine-kinase inhibitors (37 sunitinib, 4 pazopanib, 9 cabozantinib). The median follow-up time was 3.5 months, totaling 65 response evaluation visits - 23 PD and 42 non-PD. PD was compatible with alterations in 40% of the detectable GAGome features. These were used to design a plasma, urine, and combined GAG progression score to classify PD vs. non-PD at response evaluation. The area-under-the-curve (AUC) was 0.91 in plasma, 0.98 in urine, and 0.99 when combined. In validation, the AUC to predict PD at treatment start was 0.64 in plasma, 0.63 in urine, and 0.71 when combined (N = 50); and, after 6 weeks, 0.75 in plasma, 0.66 in urine, and 0.80 when combined (N = 47). The combined GAG progression score had 62% sensitivity and 86% specificity to PD at treatment start and 67% and 87% after 6 weeks. Conclusions: GAGomes correlated with treatment response in mRCC. GAG scores were validated for the early prediction of response. Their clinical utility remains to be ascertained. Clinical trial information: NCT02732665.