Minimal Residual Disease

MRD testing after treatment.

This is a topic that needs to be aired.

Many years ago, treatment for CLL was designed to relieve symptoms. It was generally accepted that CLL was an incurable disease, but for many patients they were able to live a very long time with it, and might well die of something else without the CLL causing problems. If the CLL did cause problems then they could be relieved, at least for a while, with simple oral treatment. Even now there are no clinical trials that show that any treatment prolongs life over what would have happened naturally, with simple oral treatment to relieve symptoms.

But now we know very much more about CLL. We can pick most of the patients who are going to run into problems from those who are unlikely to need much in the way of treatment. We also have treatments that are more effective in killing more CLL cells.

The question then is "Can we cure CLL?"

In order to attempt a cure we would need to get rid of all detectable CLL cells. This state is called a complete remission. In acute leukemia, a complete remission is defined as fewer than 5% blasts in the bone marrow with a return of the blood count to normal. The 5% level is chosen because normal people have up to 5% normal myeloblasts in their bone marrow. But we know that patients with acute leukemia in complete remission still have up to one thousand million (in America that is called a billion) leukemic cells in their body. So after obtaining a complete remission patients are given 2 further courses of chemotherapy to get rid of their residual disease. Even then there may be 100,000 leukemic cells left behind.

At this level perhaps 50% of patients will still relapse.

Get this in perspective - a teaspoonful of blood contains 25 billion blood cells.

When such small numbers of leukemic cells are involved more sensitive methods of detecting residual disease are needed than just counting blasts down a microscope. For AML the test used is PCR or polymerase chain reaction. It depends on their being a piece of DNA that is unique to the leukemic cells. Usually this is provided by a conjoined chromosome. For example in chronic myeloid leukemia (CML) the ends of chromosomes #9 and #22 swap places so a new gene is formed at the junction known as bcr/abl. bcr from one chromosome joins to abl from the other to produce a gene that is not present in normal cells. The are lots of other translocations that are characteristic of different types of leukemia. The PCR works by matching to this characteristic piece of DNA.

This is how PCR works. The basic apparatus is called a thermal cycler. This is just a machine that changes temperature at regular intervals. It switches between two temperatures - one where an enzyme can cause DNA to make a copy of itself and one where the two strands of DNA fall apart. Apart from the thermal cycler you need the enzyme (called a polymerase - it is special because it is not destroyed by the high temperatures that make the DNA fall apart), a soup of the component parts of the DNA, and DNA probes (known as primers) that start the copying process off at the right part of the DNA. The first cycle makes 2 copies, the next 4, then 16, then 256 and so on. Normally 30-40 cycles are used. As long as the particular DNA sequence was there at the beginning it will eventually be detected when enough copies are made.

Using this technique one abnormal cell in a million can be detected.

In CLL detecting small amounts of disease is more difficult. First, because there may be up to 30% small lymphocytes in a normal bone marrow and a Complete remission is defined as having fewer than 30% small lymphocytes in the marrow with a normal blood count. Second, because CLL in the marrow is often present in nodules rather than evenly distributed so there might be a sampling error and even if nodules are found they might be normal B cells or T cells rather than the CLL cells. Third, CLL cells don't have to be in the bone marrow, they could be in any lymph node, the spleen or liver or just about anywhere else in the body. Fourth, CLL doesn't have chromosomal translocations to use in PCR.

The unique bit of DNA in CLL is in the sequence of the immunoglobulin genes, but unless the patient has had his VH genes sequenced, exactly what this unique sequence is, is unknown. Therefore, scientists have used what they call consensus primers. They will amplify a monoclonal population, but they are not a specific as the unique primers that are possible if the sequence of the VH genes is known. Since most patients don't have their VH genes sequenced, consensus primers are generally what have been used when people with CLL have their MRD (minimal residual disease) estimated by PCR. Because of the risk of false positives, the sensitivity of this assay cannot be set so high. If specific VH probes are used then PCR can detect 1 cell in a million, but with consensus primers, it can only detect 1 cell in 10,000.

Another way of detecting MRD is by flow cytometry. This relies on the leukemia having a unique set of cell markers in its immunophenotype. Again to be really sensitive you need to detect at least 3 and better, 4 unique markers. In CLL many labs have just detected the CD5/CD19 population. This test is very insensitive, being able to detect 1 in 100 cells. Some quite famous centers have used this test and told their patients that they are MRD negative. On the other hand the Leeds group have produced a set of markers that is much more sensitive - they estimate that they can detect 1 cell in 500,000. Andy Rawstron is leading an international group that is exploring the detection of MRD in CLL. They have identified several groups of antigens that might be useful and which group should be used is still in the melting pot. It may be that a different group should be used in different circumstances. For example, the original Leeds group included CD20, but after treatment with rituximab, the CD20 antigen may be temporarily removed from CLL cells even though the cells a re still alive, so this would give a false negative.

So who needs their MRD estimated? It is a general maxim in clinical pathology that you only do a test if you are going to act on the result. So if you are on a type of treatment that may stop or continue depending on the result, or if it might change depending on the result then MRD is for you, as long as you are in a clinical complete remission. Obviously there is no point if you still have enlarged lymph nodes or a big spleen. The only other patients who need it done are patient in clinical trials where it is being used to evaluate treatments or indeed to evaluate tests for MRD.