TKI (tyrosine kinase inhibitor) therapy established a new standard for what medications could achieve in the treatment of CML (and set the bar for all cancer therapies). With TKIs, the goal posts shifted, but what constitutes a success has continued to evolve over time. The most recent benchmark is “deep molecular response”, a term that first appeared in a medical publication only five years ago (Yoshida and colleagues. Int J Hematol 2011;93:618-623).
The first true measure of success was a complete cytogenetic response (CCyR), which meant that when bone marrow cells were examined under a microscope, none carried the defective gene (BCR-ABL) found in most cases of CML. Early medication studies (interferon-alfa at the time) found that people who achieved CCyR survived longer than those who didn’t (Mahon and colleagues. Blood 1998;92:4059-4065). But this “complete” response wasn’t truly complete: testing was only done on a small number of cells and people could still harbour the CML gene.
Meanwhile, molecular biology was developing techniques of its own. PCR (polymerase chain reaction) tests could tunnel down even deeper to detect the molecules (called transcripts) being produced by the CML gene, so it was a much more sensitive method for determining the presence of disease. The new way of testing measured CML gene activity against a reference. If the transcript number was reduced to 10% of what it had been, this was a 1-log reduction; if reduced to 1/100th, this was a 2-log reduction; and if it was 1/1000th, it was a 3-log reduction. CCyR roughly corresponded to a 2-log reduction. But even better was the new benchmark: a major molecular response (MMR), which was a 3-log reduction. MMR was first defined in the IRIS trial of Gleevec, which found that everyone (i.e. 100%) who achieved MMR in the first year of treatment did not progress over the next two years (Hughes and colleagues. N Engl J Med 2003;349:1423-1432).
This was a remarkable finding. But if a 3-log reduction was better, what about a deeper level of response? The best that could be achieved was a complete molecular response (CMR), in which PCR testing could not find any evidence of CML activity. But as in the case of CCyR, “complete” reflected the method used rather than the disease itself. As international standards were developed, and as laboratories improved their techniques, it became possible to go to the level of molecular response of 4-log, 4.5-log and even 5-log (called MR4, MR4.5 and MR5). So a 5-log reduction meant that the amount of CML activity was 1/100,000th of what it had been. So where there might have been 100,000 molecules, there might now be one. This allowed for a possibility of one, so it wasn’t really “complete”. As a result, “complete molecular response” has been largely replaced by the term “molecularly undetectable leukemia” (Baccarani and colleagues. Blood 2013;122:872-884). At the moment, “undetectable” means better than a 5-log reduction, but that may change as testing becomes even more sensitive.
While the term “complete” may have been a casualty, “deep molecular response” has remained as a measure of treatment success. The original definition – a 4-log reduction (MR4) or better – is still used, but includes responses (e.g. MR4.5 or MR5) that are deeper still.
Do these goal posts matter? There has been some debate about whether people who achieve MMR are better off than those who reach the lower benchmark of CCyR (Breccia and colleagues. Crit Rev Oncol Hematol 2011;79:135-143). The most important outcome, of course, is survival, and it has been very difficult to show that people with MMR live longer than those who reach CCyR. In large part this is because CML deaths are (fortunately) uncommon for people who respond well to treatment. However, a deep molecular response is substantially better than CCyR or MMR, so better long-term outcomes are easier to demonstrate. For example, a French study of 206 people taking Gleevec found that treatment was significantly less likely to fail if people achieved MR4.5 rather than CCyR (Etienne and colleagues. Haematologica 2014;99:458-464). In the German CML IV study, no one experienced progression of their CML over an 8-year period if they achieved MR4.5. Eight-year life expectancy was substantially higher for people with MR4 or MR4.5 compared to those with CCyR: 92% vs. 83% (Hehlmann and colleagues. J Clin Oncol 2014;32:415-423). In fact, life expectancy for people with a deep molecular response was the same as it was for people without CML.
Perhaps most importantly, a deep molecular response is the ticket you need to enter a treatment discontinuation trial. (See Treatment-free remission – how well do people do? CML-IQ, November 26, 2015.)
A deep molecular response is achievable with any of the TKIs. However, the likelihood of reaching this goal is better with one of the second-generation medications, such as Tasigna or Sprycel, rather than Gleevec. For example, in the ENEST1st study of Tasigna, 1 in 5 people achieved MR4 within the first six months and about 40% reached this mark within 18 months (Hochhaus and colleagues. Leukemia 2016;30:57-64). In the NordCML006 study of Sprycel, 61% of people achieved MR4.5 within 36 months compared to 21% taking Gleevec (Hjorth-Hansen and colleagues. Eur J Haematol 2015;94:243-250).