EURURO Vol. 72 No. 4

Introduction

Over the past decade, a total of six vascular endothelial

growth factor (VEGF)-directed therapies and two mamma-

lian target of rapamycin (mTOR) inhibitors have been

approved by the US Food and Drug Administration for

metastatic renal cell carcinoma (mRCC)

[1]

. VEGF-directed

therapies include both VEGF-tyrosine kinase inhibitors

(sunitinib, pazopanib, axitinib, sorafenib, cabozantinib, and

lenvatinib), andmonoclonal antibodies (bevacizumab)

[2–6]

The two approved mTOR inhibitors include everolimus

and temsirolimus. Beyond these agents, the programmed

death-1 inhibitor nivolumab has also been approved

[7,8] .

Ironically, despite a multitude of targeted therapies

approved for mRCC, there is no biomarker used for

treatment allocation. Rather, treatment sequence is largely

dictated by the regulatory studies used to garner Food and

Drug Administration approval. For instance, claims data

suggest that over 90% patients receive sunitinib, pazopanib,

or bevacizumab in the first-line setting

[9] .

These patterns

reflect the fact that all three agents were assessed in pivotal

phase 3 trials in the front-line setting. In the postfirst-line

setting, a shift towards increasing use of cabozantinib and

nivolumab has been observed, although substantial use of

agents such as everolimus and axitinib still remains.

Many have argued for a more personalized paradigm of

therapy, harnessing the genomic profile of RCC detailed in

efforts such as The Cancer Genome Atlas (TCGA)

[10] .

One of

several caveats to this approach is that TCGA data is derived

from earlier stages of disease, and some degree of evolution

may occur in more advanced stage and through selective

pressures from treatment. In addition, TCGA specimens are

from primary nephrectomies. Repeated tissue biopsies may

not be practical due to issues of cost and risk of physical

harm. Circulating tumor DNA (ctDNA), derived from serum

or plasma, has been applied in other malignancies to

provide a longitudinal assessment of tumor genomics. For

example, in the setting of lung cancer, resistance mutations

arising under treatment pressure from matched therapies

for

EGFR

and

ALK

alterations have been identified and are

used to select specific next-line treatments

[11,12]

. Herein,

we attempt to establish the feasibility of ctDNA assessment

in a large cohort of patients with mRCC, and to define

differences in ctDNA profile across lines of systemic

targeted therapy in RCC.

Material and methods

Patients included in the current series had a diagnosis of advanced RCC

and consented to blood collection for ctDNA assessment as a part of

routine clinical care. ctDNA assessment was performed using a specific

Clinical Laboratory Improvement Amendments-certified, College of

American Pathologists-accredited comprehensive plasma assay (Guar-

dant360; Redwood City, CA, USA). Detailed methods describing this

assay have been previously published

[13]

. In brief, two 10-ml aliquots of

blood are collected, with subsequent extraction of 5–30 ng of cell-free

DNA. Isolated cell-free DNA is subsequently exposed to capture probes

encompassing 73 cancer-related genes, with complete exon coverage of

18 genes and coverage of critical regions of exons in the remainder.

Enriched digital sequence libraries are subsequently analyzed using the

HiSeq2500 Sequencing System (Illumina, San Diego, CA, USA), achieving

an average raw coverage depth of 15 000 (minimum: 2000 , average

Q-score: 20). A full list of genomic alterations (GAs) captured by the

assay are listed in Supplementary Table 1.

Through this commercially available platform, limited patient-related

data is submitted, including patient age, sex, histology, and current/prior

treatment. Deidentified information was supplied to the principal

investigator of this study (SKP). This information was acquired through

an Institutional Review Board-approved protocol allowing for access of

limited personal health information. For purposes of the current study,

treatment was characterized as either first line or later line based on

prevailing practice patterns. Specifically, sunitinib, pazopanib, bevacizu-

mab, and temsirolimus (reflecting

90% of all front-line therapy based on

claims data) were considered first-line treatment

[9] .

All other treatments

were considered later line. The frequency of GAs was assessed in the

overall cohort and compared across line of therapy and histology (when

available). The frequency of individual GAs in the first- and later-line

settings was compared using the chi-square test, with the a priori

hypothesis that the frequency of alterations would increase in later-line

settings relative to the first-line setting.

Results

3.1.

Patient characteristics

ctDNAwas collected from a total of 220 patients withmRCC,

including 145 men and 75 women. The median age of the

cohort was 63 yr (interquartile range [IQR]: 57–70). Sample

collection dates ranged from May 2014 to September

2016. Histologic subtype was specified in 124 cases (56%;

89 clear cell, 14 papillary, 8 sarcomatoid, 6 chromophobe,

4 mixed histology, 2 undifferentiated, 1 collecting duct); the

remainder (96) were undesignated. Notably, for patients

with sarcomatoid disease, the degree of sarcomatoid

Conclusions:

In the largest assessment of ctDNA-detected GAs prevalence in mRCC to date,

the majority of patients demonstrated clinically and biologically relevant GAs. Increasing

p53 and mechanistic target of rapamycin pathway (eg,

NF1

PIK3CA

) alterations in postfirst-

line patients with first-line vascular endothelial growth factor-directed therapy may

underlie mechanisms of resistance. Routine ctDNA assessment during the clinical course

of mRCC patients may have therapeutic implications.

Patient summary:

Collection of circulating tumor DNA is feasible in patients with meta-

static renal cell carcinoma, and analysis of a large cohort demonstrates significant changes

in circulating tumor DNA profile across patients’ clinical course which may have therapeu-

tic implications.

E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 5 5 7 – 5 6 4

558