NcXHF

INVITED ARTICLES
HEAD AND NECK CANCER
Towards a Personalized Treatment of Head and Neck Cancer
Andrew M. Gross, CPhil, and Ezra E.W. Cohen, MD
There is increasing excitement surrounding the surge of
data that has emerged throughout various sectors of oncology
research, which often seems like hyperbole. Although the
use of data to inform decision making is far from novel, the
emergence of inexpensive and ubiquitous computing power
and storage has bolstered our ability to produce data-driven insights
and has ushered in the era of “big data.” The essence of the
big data paradigm is simple: given enough data, machines can
generate insights equivalent to or sometimes even exceeding experts
in the fıeld. One example of the application of this paradigm
lies in the fıeld of natural language processing where the
use of giant collections of text provide better results with much
simpler models and less expert input. 1 In contrast, notable failures
exist including the buildup and subsequent implosion of
Google Flu Trends to predict the onset of the flu season. 2,3
Alongside these developments, the fıeld of cancer has seen a
bolus of data emerging from a combination of old and new
sources. The digitization of hospitals and the spread of electronic
medical records have created large archives of traditional
data, making meta-analyses of large patient cohorts almost facile.
Furthermore, the adoption of high-throughput molecular
profıling and the plummeting costs of genetic sequencing are
creating new opportunities to probe the inner-workings of
healthy and diseased cells. Although the idea of personalized
medicine is clearly not new, our ability to accurately quantify the
defıning characteristics of an individual patient’s tumor lends
hope that a targeted attack against its specifıc biology is possible.
In head and neck squamous cell carcinoma (HNSCC), such
molecular biology has proven pivotal to our understanding
of the disease. Starting with targeted studies of mRNA, protein
expression, and analysis of chromosomal gains and
losses, the fıeld slowly pieced together a model in which epidermal
growth factor receptor (EGFR) expression is activated
and chromosomal abnormalities accumulate, often
alongside mutations to the TP53 gene. 4-6 As cohorts became
larger and more high-throughput analysis such as microarrays
and sequencing technologies became available, it became
clear that the picture was far more complex and that
there exist different subtypes of the disease.
As data acquisition becomes readily available with the rise
of high-throughput sequencing, our understanding of tumor
biology is actually beginning to blur. Although more and
more driver genes are being discovered, it is exceedingly clear
that the molecular characteristics of no two tumors are
identical and that the biology driving tumor initiation and
progression may be far more complex than originally envisioned.
As genomic profıling in cancer becomes ubiquitous,
the promise of the big data paradigm may affect patients with
HNSCC by enabling personalized monitoring, diagnosis,
and treatment of this disease. Although the dystopian vision
of machines treating patients is not likely to become a reality
anytime soon, the ability to use systematic data to inform better
patient treatment is all but inevitable.
MOLECULAR STRATIFICATION OF HEAD AND NECK
CANCER
In the late 1990s, the convergence of epidemiological and
molecular studies on the influence of HPV in cancers of the
oropharynx led to the fırst molecular stratifıcation of head
and neck cancers. The ability within the research community
to detect HPV DNA by sequencing-based methods allowed
for comparison of HPV-positive and HPV-negative groups
with other molecular measurements and clinically obtained
variables. In parallel, analysis of increasingly large molecularly
characterized cohorts has enabled a more refıned understanding
of the heterogeneity of this disease. The rise of such
cohorts alongside the continued use of model systems has
allowed for exquisite understanding of the key events that
drive HNSCC and even how these events interact with each
other. Whereas the molecular details of HPV-positive and
HPV-negative HNSCC are fascinating, we refer the reader to
a number of recent papers and reviews that comprehensively
cover these facets of the biology of HNSCC tumors. 7-10
The picture that eventually emerged was one of very different
cancers, with HPV-positive patients being far less likely
to have the traditional risk factors of smoking and drinking
while having substantially better cure rates than their HPV-
From the Bioinformatics and Systems Biology Program, University of California San Diego, La Jolla, CA; Moores Cancer Center, University of California San Diego, La Jolla, CA.
Disclosures of potential conflicts of interest are found at the end of this article.
Corresponding author: Ezra E.W. Cohen, MD, Moores Cancer Center, University of California San Diego, 3855 Health Sciences Drive, #0658, La Jolla, CA 92093; email: ecohen@ucsd.edu.
© 2015 by American Society of Clinical Oncology.
28 2015 ASCO EDUCATIONAL BOOK | asco.org/edbook