Anne Paxton
March 2020—Three decades ago, when Jerald Radich, MD, started doing BCR-ABL testing, the process was highly hands-on. In those days, “we used water baths to do PCR,” said Dr. Radich, an oncologist and director of the molecular oncology lab and member of the Clinical Research Division, Fred Hutchinson Cancer Research Center, in a recent CAP TODAY webinar.
BCR-ABL testing is much simplified and faster today. But more important, it is central to the strikingly successful diagnosis, treatment, and monitoring of chronic myeloid leukemia. Dr. Radich views the use of BCR-ABL for CML treatment and monitoring as “the poster child of what molecular medicine should really be about.”
CML patients treated in North America or Europe now have an average lifespan roughly the same as that of the normal population, he said, and thus “there has been a continued rise in the numbers of cases in the U.S.”
“We may only have roughly 5,000 cases a year in the U.S., but they’re now going to be living long lives.” That’s why understanding monitoring is going to become more and more important, he said. “As we get more patients, if we aren’t monitoring carefully, we are going to have failures go on to accelerated phase/blast crisis. And that would basically eliminate all of the progress we’ve made to date in getting people to a point of good response and potential discontinuation of the drug.”
Dr. Radich
With an incidence of roughly two per 100,000, CML is no public health menace, Dr. Radich said. But it has played a crucial role in translational research. “It really has taught us more about the bench to bedside than any other disease. It was the first disease where a specific genetic abnormality was found, the Philadelphia chromosome. It was the first molecular lesion where the genes involved in the translocation were discovered. It was the first disease where we could use that breakpoint as a way to monitor disease burden in patients.” And it was the first disease, he said, for which a targeted therapy to a chimeric tyrosine kinase was established, first with imatinib and then with other tyrosine kinase inhibitors.
The introduction of tyrosine kinases and targeted therapy marked a turning point, he noted. Before that, when patients would typically present in a chronic phase, without definitive therapy they would invariably progress to accelerated phase/blast crisis. Their median survival was roughly six years.
Targeted therapy with imatinib brought dramatic improvement. “There was an unbelievable change in outcomes in just a matter of a few years,” Dr. Radich said in the webinar, made possible by a special educational grant from Cepheid. Before imatinib, “we saw 70 percent of patients go into blast crisis in 10 years with various therapies.” Now, evolution of blast crisis is rare. “It’s really a heroic story of progression.”
Four FDA-approved front-line drugs and several second-line therapies are now available to treat CML. Clinical trials of the second-generation drugs nilotinib, dasatinib, and bosutinib show they all have a similar major molecular response (defined as a 3-log reduction in the BCR-ABL with therapy from time of diagnosis) and do it a little better than the first-generation drug, with better early response and fewer progressions. But all the drugs produce excellent survival rates, Dr. Radich noted.
Only about 60 percent of patients on imatinib therapy are still on it at five years, owing to a combination of poor early response in 10 to 20 percent of patients, toxicity, and resistance. The field has known since the 2003 IRIS trial (O’Brien SG, et al. N Engl J Med. 2003;348[11]:994–1004), which compared imatinib with interferon treatment, that monitoring molecular response with BCR-ABL from peripheral blood, not bone marrow, “pretty much predicts your long-term results,” Dr. Radich said.
Patients who have a BCR-ABL on the International Scale of less than or equal to one percent have excellent results, he said, while those with higher BCR-ABL are more likely to have an adverse event. As far as treatment of those with a BCR-ABL IS percent greater than 10 percent at the end of 12 months, “you have to think about plan B. This was our first indication that BCR-ABL monitoring with peripheral blood was going to make a difference in defining disease burden and predicting outcomes.”
Later studies have found that early response can indicate how patients will fare, he said. Patients who do worse at three months (>10 percent) have a dramatically lower probability of survival than those who have a better response at three months. So “even earlier monitoring can give us interesting and actionable information on how to treat patients.” A 2010 study found that patients who were above 10 percent at three months were likely to progress to accelerated phase/blast crisis (Hughes TP, et al. Blood. 2010;116[19]: 3758–3765).
However, an early suboptimal BCR-ABL level may occur for two reasons: bad biology or poor adherence to therapy. If bad biology is the culprit, “we have to consider changing therapy because of the risks involved with poor early molecular response,” including a greater than 10 percent risk of transformation, a greater than 10 percent risk of death due to transformation, and a greater than 90 percent risk of failing to achieve MR4.5 (a deep molecular response).
But resistance happens not infrequently, Dr. Radich said, often because of mutations in the ABL tyrosine kinase domain. People with such mutations have a lower probability of progression-free survival compared with those who don’t have the mutations. Before the second-generation drugs, patients who developed resistance progressed quickly to accelerated phase or blast phase CML. Now there is the option of salvage therapy, though it may not be curative. If a patient on imatinib had a mutation, for example, “you would want to change to dasatinib or bosutinib or nilotinib, based on the specific mutation.” After switching therapy, careful monitoring is warranted. “These patients especially need to be monitored carefully because if they don’t respond molecularly, they again have a higher risk of progression.” If they haven’t shown a response by PCR or cytogenetics at the end of three months, “at that time we move to a transplant.”
The National Comprehensive Cancer Network and the European LeukemiaNet have developed guidelines on monitoring and changes in therapies based on the response. “These are fairly easy guidelines, based now on PCR testing with peripheral blood.” They call for BCR-ABL testing every three months, so the average patient would have three or four tests per year. However, studies have shown that the medical community is falling short in following guidelines.
In a survey of 5,446 patients reported at ASCO in 2014, the annual testing rates were one test or fewer per year for more than 60 percent of those patients, which Dr. Radich called “spectacular underperformance.” A second survey of 40,000 patients looked at how often BCR-ABL mutations were tested for based on the NCCN guidelines, which say if you have two BCR-ABL tests three months apart and have a 1-log increase in level, you should get a BCR-ABL mutation test. Less than 10 percent with a rising BCR-ABL level were tested for mutations. In a third survey of 100 community oncologists, all said they believe the guidelines and that they follow them faithfully.
“So there seems to be a disconnect between what is out there as far as monitoring guidelines and what is actually done,” Dr. Radich said, suggesting that better communication with the medical community about the importance of these tests may be needed.
A study of people in two administrative health facilities, which looked at how adding BCR-ABL tests changed inpatient admissions, inpatient days, and ER visits, found that for every one extra test after a one-test year (so one to two tests a year), there was a large reduction in inpatient admissions and days and ER visits, he said (Latremouille-Viau D, et al. J Manag Care Spec Pharm. 2017;23[2]:214–224). “The testing does make a difference in patient care.”
Patient adherence to medication regimens can also be improved with BCR-ABL testing. “We tell people to take their drugs, but if we also emphasize that that’s why we’re monitoring them, it may have an effect on the overall outcome.” He pointed to the ADAGIO study in the U.K. (Mazzeo F, et al. Anticancer Res. 2011;31[4]:1407–1409), in which 70 percent of patients said they were taking their medications as prescribed but did not know that a radiofrequency device in the cap of their pill bottle showed how often they took their drug. The device revealed that only 14 percent of them took the drug as prescribed.
“If we look at levels of [CML] patients who achieve major molecular remission [MMR], the absence of BCR-ABL detection, you can see that those patients who took greater than 90 percent of their drug had a spectacular response” compared with those who took less than 90 percent. With frequent monitoring, “you can basically keep an eye on both how the disease burden is behaving and whether patients are taking their drug or may actually need secondary therapy.”
Discontinuation of therapy can be another option. When Dr. Radich started to work in CML, “we thought people would have to be on drugs for life,” he said. “We figured there’s always going to be a reservoir of disease and if you took people off they would relapse.” When treated with tyrosine kinases, patients’ BCR-ABL level at first has a very rapid drop down, “thought to be probably from committed cells in the periphery. They are killed off very quickly, and then you get this slower slope of decrease in the BCR-ABL level as stem cells begin to cycle and then get killed by the tyrosine kinase.” Mathematical modeling might predict then that it would require at least 30 years of tyrosine kinase to get rid of the stem cells.
Nevertheless, François X. Mahon, MD, PhD, of France believed that TKI discontinuation might be possible and took patients who had been on the drug for 10 years and had undetectable BCR-ABL and discontinued TKI therapy (Mahon FX, et al. Lancet Oncol. 2010;11[11]:1029–1035). “About 40 percent of his patients got off tyrosine kinase and did not relapse,” Dr. Radich said.
Now, discontinuation trials have exploded across the world, he said. “All of the treatment-free remission rates are stacking up to be between 40 and 60 percent. And this is quite a surprise.” Other trials have found that most people will lose their response within three to six months of therapy. “We basically say they need frequent molecular testing, once a month, once they discontinue. Once they’ve lost their response, half of them have regained their response by about eight weeks. About 90 percent or more retain their original response. So discontinuation seems to be fairly safe, but you have to be monitored carefully and frequently because you don’t want someone who’s relapsing to go a long period of time without any tyrosine kinase therapy.”
There is now a push, Dr. Radich said, to use possibly more sensitive BCR-ABL methods for those patients who have ongoing discontinuation trials, to see if the test can pick up more of the patients who will relapse. “Because if someone is PCR negative by a conventional assay, maybe the ones who are going to relapse are the ones who have slightly more disease that you would pick up by a more sensitive test.” There are basically two ways digital PCR can be done, he said: by partitioning samples via channels into thousands of wells on a plate format (Fluidigm), or by partitioning into thousands of droplets (Bio-Rad). Preliminary studies by several groups suggest that these more sensitive detection methods may help predict those patients who will relapse (or not) after discontinuation.
The NCCN and ELN guidelines strongly recommend that discontinuation be attempted only on a clinical trial. Both recognize, however, that attempts are being made in the community, Dr. Radich said. “If you can’t be on a clinical trial, then we want some minimum requirements.” One test is not enough, he said; multiple tests are needed to prove a certain level of response. “You have to have at least a molecular response to 4-log reduction for at least two years, documented by at least four tests performed at least three months apart.”
“You can only do this if you have patients adhering to getting their BCR-ABL tested by their peripheral blood. No history of accelerated phase, no history of resistance, and most important—because these people again relapse within three to six months—you want access to reliable BCR-ABL testing with a sensitivity down to 4 or 4.5 and results within two weeks. You need to get results back rapidly so that if they are relapsing, you can jump on it while they have a low burden of disease.”
“So we have come a long, long way in testing BCR-ABL, and this is a model of how we are starting to do business in all hematologic malignancies—and eventually I think solid tumors as well, with cell-free DNA and the like. But CML was the first one,” Dr. Radich said.
BCR-ABL testing has been pivotal in other ways too. By 2007 and up to 2010, “we started to use BCR-ABL testing at the end of 12 months as an endpoint for phase three trials, the registration trial for the second-generation drugs. We essentially turned five- to 10-year studies into one-year studies just based on the power of being able to do accurate peripheral blood BCR-ABL testing.”
In 2007, Dr. Radich worked with Cepheid to develop the assay that could be used in monitoring these CML studies. “The early molecular response patients were in 2012, and now the early marker response has made it into the NCCN and ELN guidelines. This is a remarkable amount of progress in a relatively short time.”
For many years, laboratories did their own BCR-ABL tests. But in the last year three different FDA-cleared BCR-ABL tests have come on the scene. Asuragen was the first with its Quantidex qPCR BCR-ABL IS kit. “It’s a kit that can be done on standard thermocyclers, but the initial FDA approval was on limited PCR instruments,” he said. “You can do digital droplet PCR with the Bio-Rad system [QXDx BCR-ABL %IS kit]. It’s especially good for large batching since it has a 96-cell plate format. It’s very sensitive and very accurate. And then we have the Cepheid system [Xpert BCR-ABL Ultra] which is cartridge based, fast, very good for point of care, and easy to learn and perform because it has the least technical hands-on need. It’s very good for a lower-volume point-of-care system.” The Bio-Rad and Cepheid systems each have pluses and minuses, he said. “We actually have both of them because we find they’re quite complementary.”
Among the challenges ahead are the need to predict early response, Dr. Radich said. “That means using BCR-ABL at a very early timepoint, maybe earlier than three months. Some biological markers we’re working on in diagnosis may be able to tell who’s going to be a good early responder. Hopefully we can both predict who is going to have early versus bad response based on biomarkers and early BCR-ABL, and maybe then drug patients so we can turn a poor responder into an early response.”
“Similarly, we will want to use peripheral blood BCR-ABL and some biomarkers to predict who we should be pushing on to get to a deep molecular response so they can get off therapy.” Monitoring beyond MR4.5, he added, will be important, especially in the setting of discontinuation.
One of the biggest challenges will be to get some kind of testing into low-resource areas, where it can be done without electricity, because most of the world’s CML occurs in places that have no monitoring and no drug availability, Dr. Radich noted. Innovation will also be needed to bring more sensitive and more reproducible assays, to handle the movement of more and more testing to the point of care, and to find ways to be able to do DNA-based assays in BCR-ABL. “Right now, that’s difficult because the breakpoint is so huge. But in the future we may be able to use some of the new sequencing technologies to define breakpoints quickly, and then we would have the advantage of dealing with a DNA assay rather than an RNA assay.”
This is important, he said, because with CML patients having a normal lifespan, “in all of our clinics, CML will become more and more frequent. So understanding the nuts and bolts of monitoring and discontinuation will become more and more important.”
Anne Paxton is a writer and attorney in Seattle. The full webinar is at captodayonline.com.