EURURO Vol. 72 No. 4

642 656

1. Introduction
Upper tract urothelial carcinoma (UTUC) is histologically
similar to urothelial bladder cancer yet several clinical,
biological and molecular features are unique to UTUC,
prompting the term disparate twins when considering the
similarities and differences between bladder urothelial
cancer and UTUC . Major knowledge gaps remain in our
understanding of the biology and genomic landscape of
UTUC, a rare disease in Western countries but of potentially
epidemic proportions in the Far East . Biologically
interesting features along the environmental-genetic spec-
trum include the strong association between known
exposure to agents such as tobacco and aristolochic
acid and genetic predisposition in patients with Lynch
syndrome .
The Cancer Genome Atlas (TGCA) reported the mutation
landscape in muscle-invasive bladder cancer, which has a
high somatic mutation frequency among adult solid tumors,
similar to melanoma and lung adenomas and squamous
carcinomas . The most common mutation was TP53 , with
frequent alterations in chromatin modifier genes (MLL2,
ARID1A, KDM6A) . Four expression-based subtypes were
described, and were shown to be associated with overall
survival and response to immune checkpoint inhibition and
potentially cisplatin-based chemotherapy
The largest targeted genomic study of UTUC to date
evaluated 300 cancer-associated genes in 83 patients using
next-generation sequencing . Mutations of FGFR3,
CREBBP, and STAG2 were commonly found in low-grade
tumors, while TP53 mutations were more common in high-
grade tumors. FGFR3 mutations were observed at a similar
rate in high- and low-grade tumors
In this study we report the first integrated comprehen-
sive genomic analysis of UTUC using whole exome
sequencing (WES), gene expression profiling, and protein
expression analysis to further characterize the genomic
landscape of UTUC and provide deeper insights into the
biology of this rare cancer.
2. Materials and methods
UTUC samples were obtained from 31 patients under protocols approved
by institutional review boards using endoscopic biopsy or surgical
resection, and were stored frozen at 80 8C. Ten samples were primary
ureter and 21 were renal pelvis in origin. Histology slides were reviewed
by genitourinary pathology experts at each respective institution (M.I.,
C.G.). All tumors were composed of conventional urothelial carcinoma,
and no variant histology was present. Microdissection was not
performed, as all specimens were enriched with tumor cells. Patients
were excluded if they had inadequate clinical data or prior treatment, or
if the histological tumor purity was <30% tumor cells. White blood cells
from peripheral blood were used as a normal control for somatic
mutation discovery.
Of the 31 samples, DNA was purified from 27 tumor and matched
normal tissues and used for WES. RNA was purified from 28 tumors, and
the polyA+ mRNA fraction was used to generate stranded cDNA libraries.
Protein was extracted from 20 tumors and used in analysis via reverse-
phase protein array (RPPA). Supplementary Table 1 summarizes the
molecular data available for each sample.
Somatic mutations were called via a standard cancer analysis
pipeline at the Baylor College of Medicine Human Genome Sequencing
Center and by using VARSCAN2 (Supplementary methods).
Copy number alterations were assessed using VARSCAN2. Microsatellite
instability was evaluated in all WES samples using our previously
published method, involving evaluation of insertions and deletions in
sequencing reads coving regions of homopolymer, for a length of 6–
10 bp . Somatic mutation data were also used to evaluate
mutation signatures in all patients .
Expression levels were computed for all genes from RNA sequencing
(RNAseq) data, and consensus clustering was used to classify patients
into groups according to expression patterns. Gene fusions were detected
in the RNAseq data using deFuse and SOAPfuse . Identified
fusions were validated by reverse transcriptase–polymerase chain
reaction (RT-PCR).
RPPAwas performed by the Functional Proteomics RPPA core facility at
MD Anderson Cancer Center using standardized protocols as previously
described The Supplementary material provides further details.
3. Results
3.1. Patient demographic and clinical data
Patient demographic and clinical data are shown in
The male/female ratio of 2:1 is similar to that in
previous reports for UTUC . The majority of patients were
white and former or current smokers. The majority had high-
grade tumors and 32.3% had muscle-invasive or higher-stage
disease (pT2+). Recurrences that were local, distant, or in the
bladder were detected in approximately half of patients
during median follow-up of 20 mo (range 3–66) for living
patients. Median overall survival was 18 mo (range 1–66).
There was no significant difference in overall patient survival
between the two institutions (log-rank test p > 0.9).
3.2. Genomic alterations in UTUC
WES for samples from 27 patients identified 2784 somatic
mutations. Three patients exhibited a high mutation
frequency. Among these, one patient had more than
750 mutations, including an MSH2 frame-shift deletion,
and mild microsatellite instability (MSI) was identified by
mutational analysis. Two other patients each had more than
300 mutations, including mutations in the helicase ATP-
binding domain of ERCC2 (Supplementary Fig. 1).
Patient summary: We conducted a comprehensive study of the genetics of upper urinary
tract urothelial cancer by evaluating DNA, RNA and protein expression in 31 tumors. We
identified four molecular subtypes with distinct behaviors. Future studies will determine if
these subtypes appear to have different responses to treatments.
# 2017 European Association of Urology. Published by Elsevier B.V. All rights reserved.
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 6 4 1 – 6 4 9 642