New Cancer Classifications?

Some exciting news from the well-respected journal Cell a couple of days ago: a large research study of tumor tissues suggests a new way to classify cancers that might help explain why some patients don't do well with certain treatments.

For a very long time, cancers have been classified by their origin. Someone says they have cancer, and they get asked "What kind?" and the answers is "Breast" or "lung" or "colon." The cancer is classified by what part of the body it came from. Lymphomas are the same  -- they are, broadly, "blood" cancers, although most types of cancer also have sub-types (there are differnt types of lung cancer, just as there are different types of blood cancer).

But, for many years, cancers were classified (or sub-typed) mostly by the way they looked under a microscope. Blood cancers could be leukemia, or myeloma, or one of a bunch of different lymphomas, based on what they looked like, at least when they are initially diagnosed.

That has started to change somewhat over the last few years, as the human genome has been mapped, and we are getting a better sense of some of the differences between cancers, and within cancer types: we can see, for example, that not all follicular lymphomas are alike, when one looks at them on a genomic level -- deeper than just a look through a microscope, but looking at how messed up the DNA is, and how that has affected the cells.

The research published in Cell builds on that more recent, deeper way of looking at cancer cells.

The researchers looked at 3527 tumor specimens from 12 different types of cancer. (No lymphomas among the 12 -- the closest we get is a type of leukemia.) What they found was pretty exciting. Certain cancers that had been based on location were actually related to other types of cancer that you wouldn't think of as being related.

For example, bladder cancer could be divided into three types: one that is now recognized; one that is related to lung and head and neck cancers; and one that is related to a different type of lung cancer. Breats cancer is divided into two types, basal and luminal. Other cancers can also be classified by their similarities to other types of cancer.

So the upshot is, unlike in real estate sales, location doesn't necessarily matter. Cell type, based on genomic analysis, is what does.

(There's a joke in there about one of my father's favorite expressions, something about not knowing your ass from your elbow. I'm not going to be the one to make that joke, though.)

Lots of articles online in the last couple of days about what a breakthrough this might be (like this one). I agree that it is absolutely the next step in the evolution of cancer diagnosis. But it's a step. And not an unexpected one, given the type of work that has been done in diagnosis in the last few years. More and more research has suggested this kind of difference between types of cancers. Not all that long ago, you could count the number of types of NHL on two hands; now, because of this kind of genomic analysis, I've heard some people say we have as many as 60 different types. It's not that new types of cancer have developed, just that we've been able to go deep enough to see the differences between them.

The bigger deal is the consideration on how this might change the way we classify them, and that will be important. People get locked into classifications sometimes. "It's DLBCL, so this is what we do." And that can be dangerous. A new or alternate classification system might get doctors to step back a little and think more carefully about how to treat something.

There are dangers, of course. I like a quote from an article about all of this from the LA Times -- a researcher from Dana Farber points out that this new classification would have "grouped two distinct types of breast cancer — one that feeds on estrogen and another on an epidermal growth factor captured by a receptor called HER2 — that require very different treatments into the same category." And that would mean missing out one some very important factor.

My guess is, though, that the danger of that happening -- of this classification system overtaking the current, location-based system -- is slim. More likely is that we see a continuation of the trend, with more and more factors being considered at diagnosis. I could see more of a flow chart approach: 
Did the tissue sample come from the blood? YES.
Does it look like Follicular Lymphoma? YES.
Does it have a 14;18 translocation? NO.
Does it have CD20 cell marker? YES.
CD22? NO.
CD19? YES.
Your treatment should be: IBRUTINIB + RITUXAN, WITH 2 YEARS RITUXAN MAINTENANCE.

Something like that. In other words, we aren't going to rely on a single classification system, but we'll find a whole bunch of factors that distinguish different sub-sub-sub-types of different cancers and determine the best treatment from there.

But that's going to take a lot of work. More research into genomic differences in cells. More development of treatments based on those differences.

And more clinical trials based on that research. And, of course, more volunteers for those trials. Someone has to do it -- right?