A Biomarker for Transformed FL?

Very, very interesting article published a few days ago in the medical journal Blood.

Some background first. As I'm sure you all know, one of the great fears of being a Follicular Lymphoma patient is that our relatively slow-growing cancer will transform, and turn into a fast-growing cancer (usually DLBCL, or Diffuse Large B Cell Lymphoma). Transformed FL has a lower Overall Survival rate, requires more aggressive treatment, and most importantly, is just about impossible to predict. And the statistics for how common it is for patients to transform are all over the place. Some say up to 50% of FL patients will transform, but most studies that actually look at large populations of FL patients put it at around 15%. All of that makes it a scary thing that kind of lurks in the background for many of us.

The lack of a way to predict to predict transformation is frustrating for lymphoma experts. And so far, there has been no biomarker to help predict this -- no feature of an FL cell or a gene that is different from other FL cells or genes. 

Maybe, just maybe, this research has an answer. 

The article is called "Genetic Subdivisions of Follicular Lymphoma Defined by Distinct Coding and non-coding Mutation Patterns." It features some pretty heavy science, and to be honest, I was greatly helped by a Twitter thread by Ryan Morin, whose lab did the work for this. Dr. Morin is not an oncologist or Medical Doctor, but a PhD Molecular Biologist at Simon Fraser University who studies DNA sequencing. His lab looks at full DNA sequences to study tumors by comparing the genetics of tumors with healthy cells. This is not so much about looking at tumor cells as it is about using computers to look at differences.

(See? Lots of heavy science.)

I'll try to make this simple for you (which is my nice way of saying that I have to keep it simple for myself).

The research looked at the whole genome sequences of 423 patients. This means that they looked at all of their genes, roughly 20,000 per person (this is why computers are involved here). Then using the computer, they were able to compare patients with FL, patients with transformed FL, and patients with DLBCL (that developed on its own, not DLBCL that came from transformed FL).

What they found was that there were enough genetic differences to put Follicular Lymphoma patients into two groups.

They call the first DLBCL-like Follicular Lymphoma, or dFL. 

The call the second group Constrained Follicular Lymphoma, or cFL.

Each of the two groups has different patterns for mutating or changing, and has some different biological characteristics, and also may different clinical characteristics (they seem different to a doctor). 

Most importantly, they found that the cFL group has a lower rate of transforming. 

How they did this was pretty interesting (and again, I have a basic sense of how this worked). Research has already shown that FL and DLBCL have a lot of the same features. So the research team looked instead at what features do not overall. They found that many of the differences occurred in places involving somatic hypermutation. Here's what that means -- FL and DLBCL are both cancers of B cell lymphocytes, which are immune cells. As B cells develop, they have to be flexible, so they can go after different types of invaders. During that period where they are flexible, where they can in change in lots of different ways, sometimes the changes go wrong, and cancer develops.

Specifically, they found that the dFL group (the one more likely to transform) had more problems with something called the CREBBP gene. This is a gene that helps control cell growth. So when there are problems, cells grow uncontrollably, which is what cancer is. The cFL (the one less likely to transform) had fewer problems with this gene, so cell growth was kept in check. In other words, it was "constrained" -- cFL = Constrained FL. 

When they tested this idea, they found that the dFL group (more likely to transform) had a median time to transformation of about 5 years. The cFL group (less likely to transform) had a median time to transformation of about 15 years.     

(A reminder about medians -- that means half of the group transformed sooner than that number, and half transformed later, or not at all. A 15 year median for time to transformation means a lot of them probably died of old age before they ever transformed).

The researchers are guessing that the CREBBP gene has a big affect on transformation.

So what does all of this mean for us as patients?

At the moment, it means nothing. This is a bunch of computer nerds, not a bunch of oncologists.

But they are pretty amazing computer nerds (and you all know how much I appreciate a nerd). But ti will take some time to test this out on a larger group of patients, and then begin to develop treatments that perhaps target the CREBBP gene in some way. Whether that happens before transformation or after, I have no idea. It will likely take a while to figure all of that out.

But for now, this seems to me like the most promising attempt at isolating a biomarker that I have read about. I do see articles like this a lot, and they often hedge ("It seems like this this might be important for lots of patients, but not all of them..."). The genetics of cancer is incredibly complex, and there's probably no one simple answer to transformation.  

But every new step toward an answer is exciting.