Where are we with prognostic markers today?

Who should have their prognostic markers done?

When I first started working in this area, I was motivated by the fact that many people were labelled as having leukemia who would never need any treatment for it. I spent a good deal of my career trying to figure out a means of distinguishing between those who would be in this bucket and those who definitely would need treatment. Many of these problems have been cleared up by the change in definition of CLL to require a minimum of 5000/cu mm B lymphocytes for the diagnosis. Many of those who had fewer than this would have been diagnosed as stage 0 CLL under the old definition. Now they are called monoclonal B cell lymphocytosis (MBL) and no-one thinks such people should be called leukemia patients or even followed up. They comprise about 4% of the population over-40 and no-one is going to contemplate making 4% of the mature population into invalids.

There is no reason to think that such people need their prognostic markers done.

What about people who really do have CLL. Experts are still recommending that outside clinical trials that the only prognostic indicator that makes any sense is the clinical stage. In other words, if you have Rai stage 0 or Binet stage A then that should be enough; you have a long period of life ahead of you.

I disagree. We know very clearly that Binet stage A patients who have unmutated IgVH genes will live on average 8 years (range 4-12 years). With better treatment this may be prolonged some, but it contrasts with an average survival of 25 years for those with mutated IgVH genes.

If any prognostication is to be done then the experts rely on the lymphocyte doubling time. Now, there is no doubt that unlike clinical stage, this is a dynamic indicator; it tells you something about the rate of progression. But it only tells you that when the disease progresses and it is unreliable. Because the blood compartment is only one of those where CLL cells accumulate, it does not tell you that your marrow cavity is filling up, nor that the lymph nodes that you cannot feel (like retroperitoneal nodes) are enlarging. In fact even the experts won't prognosticate on lymphocyte doubling times if the absolute lymphocyte count is less than 30.

The mutational status of the IgVH genes, on the other hand is a given characteristic of the disease, available at diagnosis.

Now this doesn't mean that I advocate that everyone with CLL should go out and get their IgVH genes done. If you are 90-years old, it hardly makes sense and people have to make their own decisions on how much information they want. Remember that these survival curves only predict averages and there is no such thing as an average person. Stephen J Gould, the Biology Professor at Harvard wrote a brilliant piece entitled 'The Median is not the Message' when he was diagnosed with mesothelioma and given less than a year to live. He survived for 17 years and died of something else.

It does make sense for the doctor looking after the patient to have this information since it could influence management, but I see no compelling reason that the patient should be burdened with it unless it is going to be acted on or unless he demands it.

In my previous article "You can't handle the Truth" I explained how difficult I found it to handle information about my CT scan. Patients want to be given hope and not despair. Too much openness is sometimes harmful. It used to be the case that the patient would leave the worrying to the doctor and in Japan they still do. Perhaps that's why people live longer in Japan.

If prognostic tests are to be done, which ones are useful?

For most tests, the results are not set in stone. In fact, the only unchanging test is the mutational status of the IgVH genes. The type of CLL you start with is the one you finish with. As far as I can determine this is not necessarily true for all other markers. But the effect of IgVH mutational status can be influenced by the other markers. It seems that the ability to signal through the B-cell receptor is what sets the pace of the disease, and in mutated IgVH cases this is impaired.

Proliferation in CLL occurs in proliferation centers in the marrow and lymph nodes and it is only when the cell is dividing that further genetic damage can occur. Every time a cell visits a proliferation center it up-regulates CD38 and possibly ZAP-70 as well. Every time it divides it shortens its telomeres and lays itself open to deletions and translocations on its chromosomes. There are simple serum tests like beta-2 M and thymidine kinase that give some indication of the rapidity of cell division, but they also measure tumor bulk, so that a high level might mean a slowly growing large tumor or a quickly growing small one.

This means that the baseline test that everyone who has full blown CLL should have done - whether or not they want to know the result - is the mutational status of their IgVH genes. Mutated and unmutated CLLs behave like two separate diseases and for the doctor not to know which one the patient has is like him saying the patient has lymphoma but time will tell us which type it is.

Why is it not done for everybody?

First, it is a matter of expense. The test costs $200 to do, though commercial labs charge $1000. You can get it done and interpreted in England for $200 plus the cost of the blood draw and a courier. Set next to the cost of one course of fludarabine let alone rituximab, this is peanuts.

What other reason? Sorry can't think of one.

This isn't a foolproof test. Being unmutated does not guarantee early treatment and being mutated doesn't guarantee late treatment or no treatment. I have written before that there are some borderline cases with 97% homology who tend to include some cases that are effectively unmutated and of course, those who use the V3-21 gene behave as if they were unmutated even when they are mutated. There is also the influence of other prognostic markers. Having a high CD38 or ZAP-70 can make the disease behave more badly. However, when patients come to see me I get this test done because it gives me greater insight into the disease than any other.

We know that whatever clinical stage you are in the mutational status of IgVH genes affects your time to first treatment, length of remission and overall survival, and it has a bigger effect than which type of treatment you get. So far this applies whatever type of treatment you receive, though there may come a time when there is a cure-all that acts well whatever your mutational status is - but we're not there yet.

Should other tests be done? You can make a case for CD38, ZAP-70, and Beta-2M. The trouble with CD38 is that the cut-off between positive and negative is disputed, Should it be 7%, 20% or 30%? And what about bimodal cases? In practice I do it because it can be incorporated into the flow cytometry picture necessary for diagnosis, but I find it most useful for cases of MBL. MBL with a CD38 less than 30% will almost never transform to CLL. The other thing about CD38 is that the percentage of positive cells tends to change during the course of the disease - in at least a quarter of patients - frequently when the disease relapses after treatment.

The trouble with ZAP-70 is finding a method that everyone can agree on. The concordance with IgVH mutations was over 90% using the first two published methods, but only 76% when the method developed in Tom Kipps lab was used. That paper suggested that it was a better prognosticator than IgVH mutations, but when it was adopted by commercial labs, very strange results indeed were apparent. Not only that, the Spanish group found that ZAP-70 levels changed in about 10% of patients during the course of the disease. Nowadays most labs don't trust ZAP-70.

Beta-2M is a good test except in renal failure. It is not just a test of disease activity since it also measures bulk. However I would do this test on all my patients.

The next test we should consider is FISH. This stands for Fluorescent In-Situ Hybridization, and it is a way for looking for particular chromosomal aberrations. It is not a perfect test and for 25 years in Bournemouth we preferred to do conventional karyotyping as well, since FISH only picks up the common abnormalities (del 13q, del 11q, del 17p and trisomy 12 - though sometimes del 6q is added). Conventional karyotyping also picks up the rarer abnormalities - trisomy 18, trisomy 19 for example and the unusual translocations - t(14;18) and t(14;19).

Although there is a famous hierarchical model which ranks cases according to prognosis according to FISH findings, this is not always helpful. It is very dependent on mutational status. Del 13q as an isolated finding mainly occurs in mutated cases, and the severity of trisomy 12 depends on whether the patient has mutated or unmutated IgVH genes. In clinical trials del 11q and del 17p almost universally occur in unmutated cases, but in untreated mutated cases we do see both of these FISH patterns in patients who do not have an aggressive disease, and indeed, may never need treatment.

Where FISH is important is in giving some idea of what is going to happen in patients who are treated. Patients who need treatment who are given either fludarabine or chlorambucil or any combination of purine analogue and alkylating agent, are not going to respond well if they have either del 11q or del 17p. Remissions are short with del 11q and almost non-existent with del 17p. If rituximab is added to the mixture then the adverse effect of del 11q seems to be removed, but it does nothing for del 17p cases.

Therefore FISH should be done before treatment since this could be determined by the FISH findings. Very few drugs work well with del 17p. High dose steroids, Alemtuzumab (Campath), flavopiridol and perhaps revilimid. A new drug in clinical trials, acadesine, also looks promising. Some people thing that ofatumumab might also be helpful, but this is not proven.

Del 11q and del 17p are almost always secondary findings and some years ago a |German study showed that almost always they developed in the unmutated cases rather than the mutated cases. All this reinforces my view that for most patients with CLL their doctors would benefit from knowing the mutational status of the disease.