Data science and cancer

Data science and cancer


According to research on the American Cancer
Society in 2017, nearly 1.7 million people in the US were expected to be diagnosed with
cancer and over 600,000 would die from cancer. That’s 1,650 per day. Importantly, a huge number of those cases
can be easily prevented if people would change their behavior, for example by stopping smoking
and eating better or even using sunscreen. This lets you know that there is space for
research in terms of things that can predict cancer, also better understanding how cancer
operates and how data scientists can help cancer physicians develop better methods of
treatment. Let’s start by looking at detecting cancer
and how data science can contribute to that. The early detection of cancer is critical. Some common forms of cancer such as breast
cancer and prostate cancer have a survival rate of nearly 100% if they are caught and
treated early, that is at or before stage one, but the survival rate dropped to below
30% if they aren’t caught until they’re at stage four, the most serious stage. As a result, data scientists are working closely
with medical researchers to find methods to identify cancer as soon as possible with the
hopes of improving survival rates for everyone. For instance Dr. Shana Kelley, a researcher
at the University of Toronto, has worked with her colleagues to develop a microchip that
can detect the traces of many forms of cancer such as brain cancer or ovarian cancer which
are difficult to detect until they become too advanced. And by developing this chip, they might be
able to find ways of sorting through the data within that chemical sample from the blood
to give people a better clue on how they can intervene sooner. Also, researchers at MIT’s Computer Science
and Artificial Intelligence Laboratory are developing machine learning models to more
accurately differentiate between different forms of cancer to ensure that patients receive
the right treatment. They cite research that somewhere between
five and 15% of people being treated for a lymphoma, are misclassified into one of the
disease’s 50 different subtypes. So, not only are they improving accuracy with
their model, but they’re making special efforts to ensure that their model is transparent
and can be interpreted by practicing physicians. That’s not something that always happens
with data science. A lot of the models are opaque and difficult
to say exactly what’s happening, but they’re trying to make it as easy as possible to apply
the results which can of course have a direct impact on the efficacy of treatment. And of course, data science has also contributed
to important advances in the science of cancer. That is, the basic research that creates an
understanding of how the disease operates, how it spreads, and really how it can be defeated. For example, Josh Stuart, an associate professor
of biomolecular engineering, and his team at UC Santa Cruz Genomics Institute have analyzed
massive collections of molecular data for 12 different tumor types and found among other
things that approximately one in 10 tumors have been misclassified. This is important of course because if a tumor
is misdiagnosed, then it’s much less likely to respond to the initial treatment and time
is lost for helping that patient. Similarly, the National Institutes of Health
have reported that advanced analytics have also made it possible to understand lifestyle
factors that contribute to the incidence of cancer and they can help develop personalized
treatment plans that rely in part on genomics sequencing for individual patients and that
allows them to adapt the treatment to best suit their particular needs. It’s one of the huge advantages of data
science. We know this from marketing. It allows individualization. And the same thing is true in terms of finding
treatments for cancer. The ability to combine data for medical trials
to the Human Genome Project and individualized genetic analysis for patients to develop treatments
is one of the most important contributions that data science can make. This can be combined with data mining of research
for existing drugs to find novel applications for cancer treatment. And finally, data science has even proven
to be useful in the less glamorous side of cancer treatment by helping with some of the
logistical tasks like scheduling treatment where you take into consideration the doctors’
schedules, the changes in the hospital staff, the typical patterns and the number of patients
in the facility, the time spent on equipment maintenance and room availability, the time
it takes to get lab results, and the presence of clinical trials in the hospital. All those together actually make simply getting
the person in the right place at the right time very challenging and it is in fact one
of the important things that data science can do by simply making the existing treatments
more available and to deliver them at the right time to the patient. In summary, data science could contribute
in terms of understanding how cancer operates and where it comes from, in terms of diagnosing
cases early, and even in terms of helping people get the treatment they need as soon as possible.

One thought on “Data science and cancer

  1. why dont you find the root cause of cancer. How about prevetion of cancer rather than finding better methods of treatment ?

Leave a Reply

Your email address will not be published. Required fields are marked *