Page 32 - Read Online

Padh. J Cancer Metastasis Treat 2018;4:52 I http://dx.doi.org/10.20517/2394-4722.2018.51 Page 3 of 5

While we welcome the trend, we should be wary of the limitations of such testing: (1) the science of
association between a given gene variant and cancer is tenuous and still evolving; (2) the clinical studies
indicating gene-cancer association may have been done in a given population and there are strong
[8]
reasons that the studies need to be validated in other major populations ; (3) the human genome data is
not really that representative of the world’s population. It has been largely collected from the Caucasian
population, and may not have adequate representation from African, Asians including Indian and Chinese
[6]
populations ; (4) the statistical increased or reduced risk assessment may vary from one study to another,
and based on it, it is tricky to counsel the general population about the risk; and (5) the population at
large may not be prepared to understand the associated risk, and may not be prepared to deal with such
predictive risk assessment.

Benefits of pharmacogenetics are two-fold: (1) with a certain probability, we can predict the risk from
cancer for a given individual. This type of pre-symptomatic diagnosis can to a great extent prevent cancer
mortality by adopting frequent screening for the suspected cancer risk and catching it at the very early
stage. BRCA1 and BRCA2 are excellent examples where women with positive cancer biomarker can be
vigilant and catch cancer before it had a chance to spread; and (2) in many cases, pharmacogenetic tests
can help the oncologist to a better treatment regimen with minimum toxicity. Both of these benefits have
become part of the cancer management. In American Society of Clinical Oncology (ASCO) post of May
2016, Dr. Stephen T. Sonis has summarized the role pharmacogenetics can play in cancer patient care
(https://am.asco.org/daily-news/personalizing-supportive-care-pharmacogenomics-and-risk-prediction).

Pharmacogenetics can help us reduce the toxicity of chemotherapy by selecting the right drug and its
dose for a given patient based on his drug response profile [9,10] . Good examples of such an approach are
Herceptin and Xeloda, where predicted non-responders are spared from the respective treatment and the
undue toxicity is minimized. I am reminded of my conversation about a decade ago with an oncologist
who considered the optimum dose is the maximum dose a patient can tolerate, who subsequently agreed
that it would be nice if we can know the effective dose and the toxic dose before treatment for each patient.
For cancer where the time is very critical, the cost and the toxicity of chemotherapy are high, and where
the therapeutic window overlaps with the toxicity window, the pharmacogenetics offers a valuable tool
to select an appropriate drug and its dose for a particular patient, and achieve an optimized therapeutic
outcome. I take liberty to quote Dr. Howard L. McLeod in October 2016 ASCO post, “The somatic genome
can assist oncologists in predicting a patient’s tumor behavior if left untreated (prognosis) or treated (efficacy
prediction), and the germ-line genome can influence prognosis as well as help assess the level of drug-
related toxicity the patient will likely experience.” He further added, “As our data become richer, we will
get to the point where we can predict all severe drug toxicities.”

There are four fundamental limitations in our approach in taking the science of pharmacogenetics to
the clinic: (1) the first and the foremost is the over-emphasis on single nucleotide polymorphism (SNP)
in considering it synonym with genetic variants. The fact is that there are other forms of variants, like
insertion, deletion, copy number variation (CNV) which are abundant in the human genome and may
cover a larger part of the genome than that covered by SNP and may have more influence in cancer
development and drug response [11,12] . Since these variants are relatively new and technologically not as
convenient to type, their impact is undervalued. As the science of pharmacogenetics develops further, it
will be hard to ignore them; (2) the second and equally important limitation is that most of our studies
have tried to link a given gene or its SNP(s) to a very complex biology of cancer. We have realized, but yet
not put to test, that ultimately a disease like cancer may not be associated with a single SNP or a gene.
It has to be a complex combination of several genetic loci and their variants which can collectively lead
a normal cell to become transformed. Since there are many possible permutations and combinations of
variants and genes to be studied for their association with cancer, practically it has remained a daunting