Can CLL be cured?

Question for you -

What is better, 10+ years free of CLL in exchange for 6 months of chemotherapy or 10 years of a pill taken daily?  Go one step further: what if the chemo not only got rid of the disease for 10 years but actually cured some patients?

We are incredibly fortunate that there are new therapies approved by the FDA that are non-chemotherapy based - but we should be careful before abandoning something that can be extremely effective for some patients just because it is called "chemo."

Everyone desperately wants to find a cure for CLL so we need to be vigilant and avoid excessive optimism.  The idea of "curable CLL" is debatable among the researchers who study the disease.  For the purposes of this post, I need to be very clear that the difference between very long term remission and cure can become a little blurry.  At what point is a patient with no signs of leukemia considered "cured?" 

I am always careful to define "remission" when I am in the clinic.  To me, it means, "we can't see the cancer but we know that it is there."  That is very different from a cure, where one would assume it isn't.  The only thing that really distinguishes between the two is the test of time.  How much time needs to pass before you say it is no longer a remission but indeed a cure.  In CLL we have never really talked about cure before, so I guess the answer is "a long time."

When carefully identified patients are treated with FCR chemo-immunotherapy, a decent fraction of them may not have any evidence of their CLL for over a decade.  Two studies have now shown statistical outcomes to suggest that some of these patients may be cured (link here and here).  

Frequent readers of my blog likely know that I have periodically taken a skeptical view of the FCR regimen.  The multitude of new drugs such as ibrutinib and idelalisib have forced us to fundamentally re-think the best ways to treat the disease.  After the widespread introduction of bendamustine and rituximab followed by the newer agents, the enthusiasm for FCR has been steadily diminishing.  Database analysis would indicate that it is only utilized in the front line management of patients with CLL in about 20-35% of patients,

I think it is human nature to embrace things that are new and exciting - especially when that means moving away from chemotherapy.  Yet as pendulums swing away from FCR 14 years after it was initially introduced, we may be ignoring some of the most impressive arguments in favor of the regimen that are only just now becoming evident. 

The "chemoimmunotherapy regimens" such as FCR and Bendamustine-rituxan have constituted our treatment backbone for a number of years.  Please see my prior blog post about choosing between the two.  With the new non chemotherapeutic targeted drugs that are coming, there is likely to be quite a "turf war" over what regimens are right in which circumstances.  While the new drugs are primarily approved in patients with relapsed disease, there will be considerable interest in moving them to the front line setting.  Indeed, I've already had quite a few patients ask me if starting with one of the new drugs up front makes more sense than chemotherapy.  I think that in some cases the answer may be yes, in other cases no.  In many cases it is too early to tell.

So where is all this going?

CLL is biologically heterogeneous.  Two patients who look very similar can have very different outcomes with treatment.  Understanding that biologic heterogeneity is essential if you want to make the best choices on behalf of the patient.  At the extremes, I think there are some patients where we need to make every effort to give effective chemoimmunotherapy and others where starting with a targeted agent makes more sense.  Between those two extremes there is a lot of uncertainty.  Over the next several blog posts, I hope to make that spectrum clear.

Let me start by coming back to the question that I began this post with:  If you could get six months of chemotherapy and have 10 years free of disease and not require any treatment, would that be better than taking pills every day for ten years?  What if I upped the ante and asked if that same six months of therapy cured a decent proportion of molecularly defined patients?  Would chemotherapy be preferable to pills in that circumstance?   What fraction of patients would need to be cured?  Would 20% be enough?  What if it was 60%?  If 80% could be cured, would that make chemotherapy better than pills that are not thought to result in cure for anyone (yet)?

Let's start by defining "cure."

When we evaluate the performance of a new drug or a regimen, we plot the efficacy on a "Kaplan Meier" curve.  On the "Y axis (up and down)" is a variable such as overall survival, or progression free survival.  On the "x-axis (left to right)" is time.  At time point zero the curve should be at 100% but then it keeps going down every time someone has an "event" such as disease progression or death.  If a disease is really bad or a treatment really ineffective, the curve goes down very quickly toward the  x-axis.  If a disease is mild, or the treatment very effective, the curve stays very "flat" and doesn't drop from 100 much. 

People who look at Kaplan Meier curves a lot get really excited when they see a "plateau."  A plateau happens when you do some sort of treatment that is likely to cure a subset of patients.  As time goes on, all those patients who are not cured either relapse or die until you are left with those patients who no longer have the disease and the events stop happening.  When this happens, the curve may start at 100, slowly drop down to the percentage of cured patients (20%?-40%?) and then stays flat - or reaches a "plateau."  If a plateau persists with updates of the data, researchers start to ask if those patients who are no longer relapsing are cured of their disease - particularly if you test them with MRD testing and they remain negative for CLL.  Of course if you follow all patients long enough, it will always go down to zero as patients die of other causes, but few studies follow patients that long and a prolonged plateau is suggestive of something very important when considering a treatment.

At ASH 2012, the MDAnderson group put out an abstract entitled, "Is CLL still incurable?"  This provocative question was asked in response to an apparent plateau in their long term follow up of the original 300 patient sample treated with FCR.  After following their original group of patients treated back in 2000-2003 for thirteen years, a group of them still appear to have no evidence of any active CLL.  That is a very impressive result for a therapy that only lasts six months.

Many researchers however regard data from MDAnderson with a degree of skepticism.  It says a lot about a patient if they get on a plane and fly down to Houston for an opinion.  It says even more if they do that every month for six months to get treated.  Such patients necessarily have a degree of affluence, fitness, and education that makes them different from the average CLL patient.  Multiple different studies have shown that such variables strongly influence outcome.  It ends up being a biased sample set.  Indeed, the average age of patients in the study was 57 years old while the average age of a patient requiring treatment for CLL in the United States is typically 74 years.  That is a massive difference.

I have to admit, I was somewhat dismissive of the 2012 abstract on that basis - until the Germans gave an update of the CLL8 study which compared FC (fludarabine / Cytoxan without rituximab) versus FCR and evaluated outcome on the basis of molecular risk factors.  They show that after six years of average follow up, several groups of patients start to achieve a plateau.  Over the ensuing two years of follow up, if a patient has not already experienced a progression, very few such patients appear likely to do so.

Is this a cure?  It is still probably too early to tell for sure.  All Kaplan meyer curves are prone to becoming "unstable" the farther out in time that you go.  Since fewer patients of the original cohort have been followed that long, single patient changes in status can have disproportionately larger effects on the curve than happens earlier in the follow up.  Furthermore, bias influences become larger if there is a subset of patients with "better" follow up data.  I am very interested though to see if this curve remains flat with subsequent follow up.  It appears to mirror the single center MDA experience but in a multicenter population where the data is more reliable.

So who are these patients? It is interesting to know after the fact that some patients do very well, but it is far more helpful if we know before selecting a therapy if a cure is within reach.  It would likely influence how you think about treating such a patient. 

I previously wrote a post about the mutation status of the B-cell receptor, so called IgVH mutation analysis.  The new update from the German CLL8 study did an impressive job looking at the multitude of new prognostic markers in CLL.  They showed that patients with unmutated IgVH (bad), had a substantially higher rate of other negative prognostic markers (such as NOTCH, SF3B1, TP53 mutations) compared to those patients with mutated IgVH (better) to the tune of about 43% vs 24%.  We also know from recent publications that newer technologies (next generation sequencing) can find a much higher frequency of adverse markers as it is a lot more sensitive to lower levels.  These lower levels appear to be very important because they appear to confer similar overall prognosis.  I wouldn't be surprised if "next-gen" could identify an even larger split between the IgVH mutated and unmutated groups.

Turns out that those patients with the mutated IgVH did MUCH better long term than those with unmutated IgVH.  Indeed if the plateau in their data holds, it may occur in as many as 60% of patients with mutated IgVH whereas no clear plateau is seen in patients with unmutated IgVH. 

Is 60% chance of long term disease control (maybe cure) good enough to take FCR?  It is abundantly clear that not all patients are sufficiently "fit" to receive FCR and NCCN guidelines draw the line for full dose FCR at age 70.  Are there other variables that you can look at to remove "bad actors" within this subset of better risk patients?  If you focus on the "good risk" patients with IgVH mutated BCR and then exclude the patients with 17P or 11Q or bad molecular markers such as TP53, NOTCH1, SF3B1 mutations what is the long term disease control rate in that group - certainly a lot higher than 60% - probably closer to 80-90% chance of long term disease control - possible cure in this subset of patients.

These findings are very similar to those seen by the MD Anderson group.  In their study just under 40% of patients had not experienced any progression at the 10 year point. If you look at the associated table however, it was strongly skewed in favor of those patients IgVH mutated BCR 49% vs 11%.  They did not have access to FISH or molecular markers so that information is not available.  I find it compelling though that the numbers were very similar between the two studies.  It is also interesting to note that the change in the shape of the curve occurred right around the 6-7 year mark in both data sets.  This implies that if you fit this highly favorable profile and make it that far out, your chance or progression over the next few years seems very unlikely.

The point I wish to make is this.  The new drugs are very "sexy."  It is very appealing to think of  taking a non-chemo pill rather than chemotherapy, but if I am ever a patient with CLL and IgVH mutated BCR lacking 17P/11Q/TP53/NOTCH1/SF3B1 abnormality, I will absolutely take chemoimmunotherapy because there is a VERY GOOD chance I will not have to think about my CLL for many years, and based upon these two studies, I think he plateau in the survival curve is very provocative.   

This blog post was originally intended to also talk about what I would do if I had 17P deletion, IgVH unumtated BCR, or other high risk markers but the post became too long and unwieldy.  I will tackle those possibilities in upcoming posts. 

Thanks for reading.

Stereotypes

One aspect of B cell receptor (BCR - aka antibody) physiology that I think is really remarkable plays out in several types of lymphoma (mantle cell and marginal zone lymphoma) but is most striking in CLL.  It is called BCR “stereotyping.”

You make antibodies to fight off flu, e. coli, salmonella, etc.  In fact you can make an enormous number of different antibodies to fight of just about any sort of invading micro-organism imaginable.  People who study this have determined you can actually make about 1,000,000,000,000 different antibodies (one trillion).  Recall that DNA is the master plans for RNA and RNA is the template for making proteins (such as an antibody).  Even though we have a TON of DNA in each and every cell (4 billion base pairs), if you had to have a different DNA segment for every antibody you could make, you wouldn’t have enough DNA (almost sounds like the federal budget type of numbers)!

So how do we possibly get one trillion different antibodies out of four billion base pairs and still have enough DNA left over for the rest of the things our cells need to do?  The answer is best illustrated in the childhood game of Mr. Potato Head.  Recall that brown kidney bean shaped doll where you could put different arms, different hats, and different legs on it to make an infinite number of different Potato Heads?  B cells do the same thing – except instead of a trademarked game from Hasbro, it involves a process called VDJ recombination.

On chromosome 14, you have a relatively small segment of DNA where most of the BCR gets made.  In order to make an antibody, the cell has to pick a “V=variable” a “D=diversity” and “J=joining” segment sort of like picking a set of arms, legs, and bow ties for Mr. Potato Head.  I forget the exact numbers but there is something like 70 V’s to choose from, 40 D’s to choose from, and something like 10 different J segments.  The smart math guy in the back of the room would quickly point out that can only give you 28,000 different possible combinations.  The B cells have some special tricks up their sleeves though and can add in additional base pairs between the segments that they  produce out of thin air (non-templated base pair addition) and in the process called “somatic hyper-mutation” can swap out base pairs (this same process may actually cause lymphoma in the first place).  It takes three base pairs to select an amino acid (the building blocks of proteins) but if you add one to the chain it can cause a “frame shift” so that everything moves to the right and now you have one base pair from one amino acid partnering up with two base pairs from another amino acid to give you a totally different amino acid.  Everything downstream is also messed up too so you get a chain reaction.

With those sorts of tricks – you get to one trillion different antibodies.  No two people should EVER have a case of CLL where their BCR looks anything like another patients BCR.  Heck, there are something like 6 billion people on planet earth.  Even if every single one of them had CLL you would only have a 1/166 chance of having the same BCR as another CLL patient.

BUT about 1/3 patients with CLL have a BCR that looks functionally identical to somebody else – what gives?

Turns out CLL is really strongly driven by signaling through the BCR.  In the routine processing of cellular debris, some of the "cellular junk" can stimulate the B cells into action when they should probably just look the other way.  When cells die, they may expose proteins such as vimentin, myosin, etc.  Sometimes those proteins can stimulate a B cell when it shouldn’t.  Since lots of our cells are always dying there is a persistent stimulus for those B cells.  Leave that stimulus on long enough and those B cells keep growing and growing and growing trying to fight off what it thinks is an invading micro-organism when in fact it is just some cellular debris that your body will never actually get rid of.

Since we all have the same myosin, vimentin, etc. we can stimulate BCR’s that look pretty much the same.  Therefore we can make, “stereotyped” B cell receptors.   

So what you ask?

The “why it all happens” is far less interesting than what it means in the clinic.  If you accept that 1/3 patients with CLL have similar receptors, they break up into very distinct subgroups and those subgroups can behave in very characteristic ways.  Knowing about stereotypes can have a big impact on prognosis. It may help to review my prior post about new prognostic markers.


Subgroup 1?  These patients typically need treatment very soon after they walk into clinic the first time.  If you compare the typical BCR mutated/unmutated status, these guys blow them both out of the water.  They may also have a higher frequency of NOTCH mutations and trisomy 12 or other high risk abnormalities like p53 mutations.

Subgroup 2?  They are bad actors.   It is clear that they have a high frequency of SF3B1 mutations which indicate that fludarabine won’t work very well.  They tend to progress early and may also have a higher frequency of del 11q.

Subgroup 4?  The average age of a patient with subgroup 4 is actually 43 years old (whoa nelly – I thought CLL was average 71).  Furthermore, these patients have exceptionally slow growing CLL.  I recently met a young woman with CLL in her early 30’s who was pregnant at the time.  In my mind I bet she is subgroup 4 and I would love to know that because it might put her at ease that she is likely to see her baby grow up.  Their BCR is typically IgG subtype instead of the customary IgM variety.

Subgroup 8? They have an extremely high rate trisomy 12 and Notch mutations.  We know Notch is a bad thing to have – but it is really bad in this subgroup because 75% of these patients will experience Richter’s transformation within 5 years - and those patients destined to transform all have the Tri12/Notch combo.

All told, about 1/3 of patients will have a particular stereotype but there are a ton of different stereotypes someone can have.  About 1/10 CLL patients will have one of the frequent ones  (subgroups 1-8).  As we get to know these subgroups more, I think it will add quite a bit of value to knowing if you are stereotyped or not.

Some of the relationships get really interesting.  We know that certain lymphomas arise as a response to an infection.  In fact, you can sometimes cure gastric marginal zone lymphoma by treating the bacteria that causes stomach ulcers.  Subgroup 4 often has evidence of a viral infection with either CMV or EBV.  Another subgroup looks like it is trying to fight off certain types of yeast.  It is really not that farfetched (THOUGH NEVER FORMALLY TESTED – don’t try this at home) that you might be able to treat some of these cases of CLL with either anti-viral drugs or anti-fungals. 

The problem is that we don’t currently test for stereotypes. Darn!  Sort of like Notch, SF3B1 and other new markers – testing for stereotypes is not part of current CLL management.

I think we will be testing for these soon enough though and it will be interesting to see how these markers get utilized into clinical practice.  Ask most general oncologists and they are totally unaware of this topic.  Even a lot of hematology oriented docs are aware of the topic but may not have much familiarity with the different subgroups.

Frankly, the Europeans have been kicking the Americans backside on this topic.  It is the Greeks and the Italians who have been the champions of the topic. 

Anyhow – hope you find it as interesting as I do.  My flight is coming in for a landing – have to turn this off now….

CLL Prognosis Markers Defined

One of the "legnedary" papers in CLL literature is the Dohner paper in New England Journal of Medicine from 2000.  It is the landmark paper that taught us about 13q, 11q, 17p, trisomy 12, and normal cytogenetic CLL.  The FISH technology it employed was developed in the early 1980's.  For the last 13 years knowing the "FISH" status helped with prognosis and treatment selection.  We are now on the eve of a major change of how we think about molecular markers in CLL.  These markers will help us pick treatments that are best for a patient, monitor dangerous subclones, and give a much more clear picture of prognosis when the disease is diagnosed and at each relapse. 

The human genome project took 13 years and 6 billion dollars to “sequence” the genomes of four individuals.  DNA is the “building plans” for just about every important task a cell has to do.  Even though it is given this amazing task it does so with only four different building blocks called “bases.”  There are two purines: guanine (G), adenine (A) and two pyrimidines: thymidine (T) and cytosine (C).  The complexity comes by putting these together in very long sequences that make them unique.  Add some extra bells and whistles and you have a “gene.”  Actually determining the sequence (ie. g-a-a-t-c-c-a-a-c-a-t-g-c and so forth) or order of a particular segment of DNA is called sequencing.

What is remarkable is that the same amount of work that went into the human genome project can now be done in a matter of days to weeks with considerably higher resolution for several thousand dollars.  The cost and efficiency of sequencing is dropping faster than microchips are getting faster.  We are getting very close to being about to sequence an entire genome in only a day for a thousand dollars.

With that diagnostic power comes an incredible ability to probe the very fundamental causes of a particular cancer.  CLL has been a beneficiary of this effort and we now have a very nice short list of the most common mutations found in CLL and several groups have done a great job figuring out the clinical significance of each of them.  Since most of these are likely to be new terms, I thought a brief write up on what these mutations do and what they mean would be great.  I think we are very close to incorporating these markers into our routine work up of a new CLL patient.

Quick note about biology: genes are found in DNA and DNA pretty much hangs out in the nucleus of a cell.  They serve as a template for making RNA.  Once a gene gets “transcribed” from a region of DNA into a much shorter strand of RNA (often times one RNA molecule per gene), it goes out into the main part of the cell called the “cytoplasm” where the RNA gets “translated” into a protein.  Proteins are the tools that do most of the tasks in the cell.   When there is a mutation in DNA, it gets copied into the RNA (which is a lot like DNA but gets out of the nucleus), and leads to the synthesis of an altered / mutated protein.  Sometimes we speak of mutations as though they occur in a protein but really it is in the DNA.  Just in case I am sloppy in my descriptions, I wanted to clarify the biology.

NOTCH1

NOTCH1 is the most interesting of the new markers to me.  It is highly associated with CLL cases that have trisomy 12 as the chromosome change and especially those cases that have an “unmutated” B cell receptor.  NOTCH1 has been a well-known protein because it is extremely important in childhood acute lymphoblastic leukemia where it is present in almost half of all cases.  Over the past few years, there have been a number of efforts to find drugs for mutated NOTCH.  So far I wouldn’t consider those efforts successful, but I am really hopeful about a new class of drugs just entering the clinic now.

NOTCH hangs out in the plasma membrane which keeps the inside of cells in and the outside of the cells out.   NOTCH is like a light switch stuck in the off position waiting to be turned on by another cell.   When that other cell comes by and “flips the switch” a piece of NOTCH gets cut free from its membrane anchor so that it can float away from the membrane.  NOTCH then travels to the nucleus where it interacts with the DNA and makes a bunch of other genes get turned on.  Those genes get copied (transcribed) into RNA and then proteins are synthesized (translated) to do their tasks.  For this reason NOTCH is called a “transcription factor.”  Once it has the right cue, it turns on the transcription of a bunch of genes and therefore determines a whole bunch of important functions.

The genes turned on by NOTCH are really important.  One critically important NOTCH regulated gene that helps cause Richter’s syndrome is MYC.  That is a protein that is a really bad actor in a bunch of different types of lymphoma and leukemia. 

Once NOTCH has done its job and turned on / off a bunch of other genes it gets marked for its own destruction.  The cell wouldn’t want to leave that signal on forever so it needs to turn it off.  Sure enough there is an entire system in place to make sure NOTCH gets shut down after it has done its task.  The particular mutation in this case makes it harder for the cell to turn off NOTCH so it ends up being a signal that won’t stop – sort of like a car where the brake pedal isn’t actually attached to the brakes.  Press all you want and the car won’t stop.

Clinically, the most important thing about NOTCH mutations is that they pretty much split the trisomy 12 patients into two groups, the good ones and the bad ones.  The good ones who lack a NOTCH mutation end up behaving as though they have normal cytogenetics (chromosomes).  The bad ones with a NOTCH mutation are now considered high risk.  They undergo transformation to Richter’s syndrome a lot more frequently and survival is shortened.  See my other post on “new risk groups.”

FBXW7

If NOTCH is important you were probably all expecting that this protein should be on the list too (well ok, maybe just some of you).  Remember all that business about turning off NOTCH?  FBXW7 is the protein that does it.  Take the same car analogy – now just throw out the brake pedal altogether. 

FBXW7 has not been evaluated as closely in terms of clinical significance so I can’t really tell you what it means to have a mutation here – but safe money would bet that will be a lot like a NOTCH mutation.

P53

I have written about P53 before.  It is the protein encoded by the TP53 gene which lives on the short arm of chromosome 17 (yes – that would be 17p).  I would encourage the interested reader to read my post about 17p deletion as well as my post about new risk groups in CLL because it really goes into deep detail about this protein.

Turns out that P53 can be mutated even when 17p is normal and they are just as bad.  The problem is that right now we don’t test for P53 mutations.  Fortunately most of the time you have a mutation in P53 you also have a deletion of 17p on the other chromosome (remember – we have a pair of each chromosomes) but that relationship isn’t air tight.  You can have mutation without deletion, deletion without mutation, deletion with mutation, or normal/normal.  The more 17p dysfunction the worse off you are.  In other words, having one good copy is better than none.  It has been a while since I have seen the number so I might be off a little bit, but something like 20% of cases with P53 mutation do not have 17P deletion so you might have a high risk marker and have no idea based on current testing.

The quick explanation of why this marker is SO IMPORTANT is that it is the protein that pretty much tells the cancer cell to die in response to damaged DNA.  Since drugs like fludarabine, bendamustine, chlorambucil, cyclophosphamide and so forth attack DNA – you need a functional P53 for the chemotherapy to work.

Patients with P53 mutations are considered “ultra-high risk” – it would be nice if we routinely tested for this – but we don’t!

ATM

ATM is to 11q as P53 is to 17p (are you following me?)  ATM lives on the long arm of chromosome 11 (long arms are designated “Q”).  When patients have deletion of chromosome 11q it is a pretty big chunk of DNA that goes missing and includes a handful of genes but ATM is one of the ones that almost always goes missing.

Like P53/17P above, you can have mutation of ATM with or without deletion of the other chromosome.  While a high frequency of 17P deleted cases (70-80%) ALSO have P53 mutation, only about 30% of 11q deleted cases of ATM mutation.  On the other hand, mutations of ATM without deletion of 11q can happen too and once again although it isn’t too common.

Like P53, ATM is important for sensing DNA damage.  If you recall DNA is what we call “double stranded.”  It is like a set of train tracks that gently twist around each other.  When DNA gets damaged it can result in a single or a double stranded break.  ATM is one of the sensors of this broken DNA and it sounds the alarm to stop cell division and also activates our friend P53.

In some studies we’ve seen that having both 11q deletion and ATM mutation is worse than just having one or the other.  Once again, current testing does not look for this.  ATM is an ENORMOUS protein.  It is hard to measure all the possible alterations but new technology is making it a lot easier.

I’ve written previously about clonal evolution both here and here.  It might not be immediately obvious if you haven’t thought about it before but I think it is fairly intuitive that when you use DNA damaging chemotherapy, the cells that survive are the ones that have higher frequency of alterations in 11q/ATM or 17p/P53.  It is sort of like taking a short course of antibiotics for a sore throat and finding that those same antibiotics don’t work well the next time around.  We therefore see a lot more alterations in 11q/17p in patients with relapsed disease than we do in newly diagnosed patients.  This is why it is so imperative to repeat molecular testing before each new line of therapy.

Clinically we think it isn’t enough to give fludarabine / rituxan for patients with 11q.  There is some data to suggest that they do better with cyclophosphamide and fludarabine than just fludarabine alone.  Add in the rituxan (ie. FCR described here) and you overcome some of the negative prognosis associated with 11q.

BIRC3

BIRC3 is another new marker of considerable importance and guess where it lives in the genome?  It lives at the far end of the same 11q deletion that knocks out ATM.  Interesting not all 11q deletions are created equal.  Most include BIRC3 but not all do – so it is possible to have an 11q deletion and have either normal or deleted BIRC3 depending on the size of the deletion.  Sadly FISH doesn’t tell us which is which because BIRC3 is a bad thing to go wrong.

Since BIRC3 is one of the newest abnormalities, we know less about how it interacts with all the permutations of 11q / ATM etc.  For now I think we can just summarize that having a mutated BIRC3 puts you in a high risk category even if everything else appears normal or favorable such as 13q deleted.

BIRC3 is a protein from a family known as IAP or “inhibitor of apoptosis.”  BIRC3 therefore helps regulate cell death and influences another very important protein known as NF-kB.  BIRC3 is another way cells can become resistant to fludarabine.

SF3B1

This is a new marker that burst onto the scene just about two years ago.  Right now, we do not routinely test for it (catch a theme here?). 

When you make an RNA copy of DNA it often consists of long segments of RNA called “introns” that need to be cut out of the final RNA strand (I remember it by saying “introns interrupt”).  Once all the introns have been removed you are left with the “exons.”  When all the exons are lined up end to end it can be copied (translated) into a protein.  We used to think this was just a bunch of cellular waste from millennia of evolution, but now we know that these introns have a bunch of important functions.

SF3B1 has the task of cutting out all those introns and creating the uninterrupted sequence of exons.  Right now, I don’t think we totally understand what happens at a cellular level when SF3B1 is mutated but we do understand some of the clinical implications.   Patients with SF3B1 mutations are resistant to fludarabine.  The other thing about SF3B1 mutations is that it makes you “high risk.”  It isn’t as bad as 17P deletion or P53 mutation but you are still worse off with it than you are without it.

SF3B1 can be sneaky, it can hide in the background of cases with normal chromosomes or even in the 13q deletions where you might otherwise expect a patient to do fairly well. There are now several markers for fludarabine resistance and including P53, BIRC3, and SF3B1. In my mind it would be pretty helpful to know a patient’s markers when they are first diagnosed or when you are picking out a treatment. 

There are several other important new molecules such as XPO1, MYD88, etc.  I have not really seen good data yet that indicates that they influence treatment choice or prognosis.  I wouldn’t be surprised if we learn more about them in the next 1-2 years.

It is an alphabet soup out there but right now these markers are not readily available.  I anticipate we might have a test for them soon and it will be helpful but unfortunately it will add a whole new dimension to the way so many patients worry about their future.  In the future people will now no longer say, “phew, I am a 13q, BCR mutated CLL.”  Instead they may say, “I am 13q deleted, BCR mutated, P53/BIR3 normal, SF3B1 6% subclone mutated.”  It is going to get very complicated very soon!

CLL Prognosis

Many CLL patients identify themselves by their prognostic markers when writing in social media outlets.  "Diagnosis age 63, unmutated, trisomy 12, treated FCR age 67, still in remission 2 years later"  is the sort of "tag line" I've seen people write.  For individuals who visit social sites frequently it is a way to tell your story in a few words.  For individuals who are new to CLL, it can all seem very confusing.  Well it is about to get a whole lot more complicated for everyone very soon.

A lot of what follows is very technical but I wanted to get it all written in one place.  I hope patients actually read and re-read this material several times.  For people who are prone to sleeping every time they read one of my posts, here are two videos I did with Brian Koffman in Sept 2013 that goes over the same material in video format:

Part 1: New Prognostic Markers
Part 2: Another on New Prognostic Markers

I've been wanting to write this post for a while but a recent paper has really brought this to the forefront of management of our CLL patients.  Unfortunately the names are strange and I worry this post may fall toward the technical side - sorry.  I will create a separate post that specifically defines many of these terms.

Integrated mutational and cytogenetic analysis identifies new prognostic subgroups in chronic lymphocytic leukemia

For people who have read all my posts on FISH testing, you are probably aware that it is an antiquated technology that has served us well for 20 years but needs desperately to be replaced. Sequencing technology has advanced incredibly quickly and is now poised to refine our understanding of CLL risk groups with new molecular detail.

While most patients are aware of the incredible advances in CLL therapies (ibrutinib, CAL-101, GA-101, ABT-199), fewer are aware of the really important advances in molecular markers that have been recently discovered.  Once these are rolled out to the general public we will be able to understand with much more precision how a patients disease will behave.  Pretty soon, folks will not only be talking about 13q, 17p without also talking about BIRC3, SF3B1, and NOTCH.

In the last 24 months, genomic sequencing has been applied to cases of CLL with pretty remarkable results (see New England Journal of Medicine article or Journal of Experimental Medicine article)

Several key findings have emerged from these data sets.

1) CLL has a relatively simple genome.  While some "smart cancers" (cancers that quickly gain resistance to our treatments and are far more aggressive) like small cell lung cancer may have 50,000 mutations per tumor, CLL (a comparatively dumb cancer - which is typically slow, responds well to most treatments, does not gain resistance all that fast) may have fewer than 100 mutations per case and only a small fraction of those (around 10-20) affect important proteins (the enzymes that make all things happen inside a cell).

2) Certain mutations seem to be observed fairly commonly in CLL and have some defined prognostic or predictive value.  For instance BIRC3 turns out to be a really bad thing to have - it is the new 17p.  NOTCH probably is one way to get to Richter's and helps sort out the trisomy 12 cases, SF3B1 makes you resistant to fludarabine chemotherapy.

3) Certain mutations are seen early in the disease, while others seem to accumulate with time.  Furthermore, some of the ones present later on are actually present early but only emerge through "clonal selection."

4) Some cases of "familial CLL" (ie those cases that run in families) have an unifying genomic explanation that point toward things we already knew were important.


With all of this new information, it was only a matter of time before someone took on the herculean effort to figure out which of these were most important and what they all mean when you analyze them simultaneously in a large group of patients (1300 of them to make this model).

The old risk groups were:
High risk: 17p changes (home of the p53 protein)
Intermediate risk: 11q changes
Low risk; normal cytogenetics & trisomy 12
Very Low Risk: Isolated 13q changes

Unfortunately, there is a lot of biologic diversity that FISH testing misses since it only looks at large chunks of missing or added DNA.  Using sequencing technology (think microscope compared to telescope) as an adjunct to FISH we can now help sort all of these out.

The new risk groups
Very high risk: 17p deletions, p53 mutations, or BIRC3 mutations (10 year survival 29%)
High risk: 11q deletions, SF3B1 mutations, NOTCH mutations (10 year survival 37%)
Low Risk: Normal cytogenetics, trisomy 12 (without NOTCH mutations) (10 year survival 57%)
Very low risk: Isolated 13q deletions (10 year survival same as age matched controls).



There are some really interesting observations contained within this.

1)  It is not a surprise that 17p deletion and p53 mutation are both really bad - we've known that for a long time.  They commonly run together (ie. most 17p deletions also have p53 mutations - but not all cases).

2) BIRC3 is a new kid on the block.  It has only been recognized for about 18 months.  Turns out it is really bad to have.  It confers chemotherapy resistance and is often very discrete from p53 alterations (i.e., if you have one, your probably don't have the other).  We've known for a while that p53 doesn't explain all cases of chemotherapy resistance - BIRC3 explains a lot of them.

3)  We have known for a while that 11q deletions often associate with bulky lymph nodes, unmutated B-cell receptors, faster growth kinetics, requirement for alkylating drugs (cytoxan, bendamustine).  It has often been considered a poor risk feature.  SF3B1 and NOTCH are totally new though and we didn't know where these fit in terms of hierarchy.  Turns out, they are about equal.

4)  Last year the relationship between NOTCH and trisomy 12 was identified.  About half of trisomy 12 cases carry a NOTCH mutation - particularly those with unmutated BCR (ie. cases with unmutated BCR and trisomy 12 have high frequency of NOTCH mutations - sorry if this gets confusing).  We have been aware that trisomy 12 was a bit of a wild card - some did fine, some did poorly.  Turns out that NOTCH mutations can sort the two apart.  Those with mutations do worse, those without mutations are now considered "low risk."  I am very eager to learn if the new NOTCH antibodies turn into personalized medicines for patients with the NOTCH (or even FBXW7 changes).

5)  Our good old friend 13q is still "good risk."  The surprise here is that 25% of 13q cases get put into higher risk categories when you do the mutation analysis.  They might have an SF3B1 mutation or BIRC3 mutation you would have otherwise never known about.  By carving out the bad players, it makes the good group even better.  "Matching age controls" does have some limitations because the model is built upon typical CLL cases.  There are probably not sufficient number of 42 year olds with 13q in the model to say that they necessarily match their peers.

6)  This model holds true no matter when you evaluate a patient.  In other words, if clonal evolution occurs and you go from very low risk to high risk by molecular definition - your clinical outcome changes too.

There are some important questions in all of this.

1) The most obvious is - how do I know what I am?  Right now - you can't easily tell.  There really are not commercial tests to sort this out - I'm trying to make one but seem to running into more walls than doors.  If anyone out there wants to finance this idea, let me know!

2) What defines "positive" for mutation?  For 17p by FISH we do not define a patient as positive until 20% of their cells are positive.  With ultrasensitive testing you may find 0.07% of cells have a BIRC3 mutation.  That patient isn't "positive" but I would be very concerned that clone may evolve in the future.  Do you therefore do anything different when you choose to treat them?

3)  This analysis may miss some of the subtlety of different FISH abnormalities.  We already know there are type I and type II deletions on chromosome 13 with different prognostic value.  We also know that the overall percent of cells with 11q or 13q makes a difference.  This model does not capture that degree of subtlety.

4)  Mutated vs unmutated is not included necessarily in this model - I would like to know if it "sub-stratifies" amongst the various different risk groups (although it is more common to see unmutated with 17p and 11q than the 13q cases so perhaps the model was just not big enough to take it all into account)

5) How do these markers hold up in the face of the new drugs.  ABT-199, ibrutinib, CAL-101, GA-101 are so remarkable.  Will traditional markers hold up in the "new era?"  It is important to note that this model is based upon cases that have already been followed for quite a few years.  Some didn't get rituxan with their first line of therapy.  Presumably none were able to take advantage (since it is an Italian study) of the new drugs.  By definition, this is a backwards looking model and does not capture what I see as a very optimistic future. For example, 29% 10 year survival for poor risk does not reflect the impressive durable control obtained in front line 17p patients treated with ibrutinib.


Though there are questions, the authors of this paper are to be thanked profusely for their remarkable effort to create a single predictive model of this magnitude.  I would imagine that there were thousands of hours put into creating and analyzing the data.  This paper will serve as a landmark for quite a few years and will help guide countless numbers of patients.