Analytical Chemist Livia S. Eberlin | 2018 MacArthur Fellow

Analytical Chemist Livia S. Eberlin | 2018 MacArthur Fellow


If you know exactly what diagnosis and
what type of cancer a person has, then you can really tailor and personalize
treatments. My name is Livia Eberlin. I’m a scientist
and I’m a researcher doing work in analytical chemistry in cancer diagnosis.
I like to define analytical chemistry as the technology branch of
chemistry, we’re really developing new methods for measurement, for
characterization of molecules, and in particular I work with mass spectrometry,
which is a tool that can measure the mass of molecules in complex samples,
like tissues, biofluids, or plants. I’m very passionate about research at the
interface of chemistry and medicine and through my research we’re developing new technologies using mass spectrometry to improve accuracy of diagnosis, but also
to make the entire clinical diagnosis process faster for the cancer patients.
So, disease tissue in healthy tissue have very different molecular patterns. The
reason why ambient ionization mass spectrometry methods are incredibly
valuable for cancer diagnosis is that this technology can measure the change
of molecules that are going on in tissue in real time. So based on this patterns
of molecules we can make predictive diagnosis of what’s the disease state of
this tissue. One of the main problems that we’re trying to address is to
improve accuracy of diagnosis during surgery and in a surgical procedure time
is very valuable, both for the healthcare system and the cost, but also for the
patient so you’re really minimizing their risk during that procedure. We
use machine learning algorithms, and what machine learning does is we provide the
machine the computer with data that’s been confirmed to be okay this is data
that’s characteristic of normal tissue or this is the data that’s characteristic of
cancer tissue, and then we train this algorithm to based on the subset of the
molecules to be able to predict okay this is what cancer looks like this is
what normal looks like based on these molecules and
now when we have a case that is unknown the machine learning algorithm will
provide this predictive diagnosis. Cancer surgeries are very complex procedures,
and the major goal in the majority of the cancer surgeries is to maximize the
removal of the cancer tissue, but to do that you really need to know where the
boundary is, right where does the cancer tissue finishes and when where the
normal tissue starts. So, our current version of what we call the MasSpec pen takes about 10 seconds from the touching of the tissue to a predictive
diagnosis of if the tissue is normal or cancer and even what type of cancer it
is its subtype of cancer. That’s the power of molecular information, and if we put
that all together into one tool that’s very easy to use and very fast, I think
that can be huge very transformative and change how surgeons are treating their
patients, and you know that’s just really thrilling and exciting. I was really
fortunate to grow up in a family that valued science and research. My dad is a
chemist, my mom is a biochemist, so early on I was exposed to those concepts it has really motivated me to pursue research in science as a career.

Leave a Reply

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