No one molecular biomarker is likely to guide treatment of RCC in the foreseeable future. Multipredictor models might be a way forward, an expert concluded.
Despite important treatment advances over the past decade, researchers are unlikely to soon identify a single silver-bullet biomarker for predicting treatment responses in kidney cancer, according to an Education Session review of available evidence at the 2018 American Society of Clinical Oncology (ASCO) Annual Meeting, held June 1–5 in Chicago.
“We’re making very small steps toward precision,” said Daniel Y. Heng, MD, MPH, of the Tom Baker Cancer Centre, University of Calgary, in Alberta, Canada. “PD-L1 [programmed death ligand 1] is not good enough. All of the available biomarkers are not good enough. But all are ‘enriching’ of patient experience, so maybe we should combine them into a [multivariate] model. I think that’s the future.”
Prognostic biomarkers are associated with patient outcomes, whereas predictive markers correlate with tumor responses to specific treatments. To fully realize the promise of precision oncology in renal cell carcinoma (RCC) management, predictive biomarkers are needed to identify which patients are most likely to benefit from which treatments, and which are likely to suffer immune-related adverse events or other toxicities related to immunotherapy.
“There are so many new drugs out there now, it’s hard know how to sequence these treatments,” Heng explained. “There’s a need for biomarkers.”
But to date, few validated prognostic or predictive biomarkers for RCC exist, Heng cautioned. Perhaps the best available tool is a clinical prognostic International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) model, which is based not on genetic mutation signatures but on hemoglobin and calcium values, platelet and neutrophil counts, and time between diagnosis and treatment.
The IMDC model classifies patients with metastatic RCC as having favorable, intermediate, and poor status, with median overall survival times for patients undergoing first-line targeted therapy of 43.2 months, 22.5 months, and 7.8 months, respectively.
IMDC risk group predicts overall survival in patients with non–clear-cell and papillary RCC, and IMDC classification can be used for clinical trial risk stratification and for statistical adjustment of retrospective data, he noted. It can also be used in patient counseling. Temsirolimus can be used with poor-risk patients and cabozantinib can be used in poor- and intermediate-risk patients, Heng said. Prognostication can also inform decision-making about whether or not to attempt cytoreductive nephrectomy, and aid in weighing immunotherapy vs tyrosine kinase inhibition as treatment options.
IMDC has also shown promise in predicting responses to anti–PD-1 immune checkpoint inhibition, Heng noted-even though PD-L1 expression considered by itself offers poor negative-predictive value for this immunotherapy.
But with the growing number of effective therapies available for first-, second-, and third-line treatment of RCC, it is becoming “increasingly important” to develop and validate tools that will allow treatments to be tailored to each patient’s disease, Heng said.
Such tools will likely combine molecular biomarkers and clinical tools like the IMDC model, he said.
Heng and colleagues propose a personalized algorithm for first-line RCC treatment based on IMDC risk stratification and informed by PD-L1 expression. High-dose interleukin-2 should be considered for favorable-risk and intermediate-risk patients, and PD-L1 testing should be considered for favorable-risk patients, for example, with those testing below 1% PD-L1 expression being treated with sunitinib, pazopanib, or ipilimumab plus nivolumab, and those with 1% or greater PD-L1 expression being considered for pazopanib or for ipilimumab plus nivolumab.
Heng disclosed consulting and advisory roles at, and institutional research support from, Astellas Pharma, Bristol-Myers Squibb, Janssen, Novartis, and Pfizer.