NEW YORK—Can chronic myeloid leukemia (CML) patients safely stop taking a tyrosine kinase inhibitor (TKI) after achieving a complete molecular response (CMR) for at least two years? That was the topic of a pro-con session discussed here at the Great Debates and Updates in Hematologic Malignancies meeting.
Mauro: Yes, TKIs Can Be Safely Stopped, and Then Restarted If Relapse Occurs
“Can we safely stop TKI therapy after achieving a complete molecular response (CMR) for more than two years? Sure, you can always restart TKI therapy if a relapse occurs,” said Michael Mauro, MD, Leader of the Myeloproliferative Neoplasms Program in the Leukemia Service at Memorial Sloan-Kettering Cancer Center.
CML is increasing in prevalence, he said: “We expect 10 times the number of cases in 2050. We now face a pharmaco-economic discussion on how to treat CML. Lifelong therapy with TKIs may not be feasible. Patients with copays are more likely to be non-adherent and to stop drug therapy due to the high costs.”
He noted that in Mexico, the cost of a transplant is equal to three months of imatinib treatment. “Cost is a burden—We need to try discontinuation.”
A growing number of CML patients are achieving CMR, Mauro said. About one-third of CML patients can expect to achieve treatment-free remission off-treatment, which is “quite reasonable.”
He pointed to several studies of imatinib discontinuation. For example, French researchers had CML patients with undetectable BCR-ABL stop therapy and observed them monthly. At 60 months, 61 percent of the patients relapsed, and the treatment-free remission was about 40 percent, he said, noting that the researchers estimated that the cost of imatinib would have been more than four billion Euros if the patients had stayed on therapy for the length of the trial.
In addition, in the TWISTER study, Australian researchers found a 43 percent success rate among patients who discontinued imatinib. “No patient progressed, and there was no development of mutations,” Mauro said. “There is consistent data. We can predict how many patients will be successful by discontinuing imatinib therapy.”
One predictor is to use the patient's Sokal score (the system developed in 1984 by Joseph E. Sokal, MD). “At 60 months, 73 percent of patients with a high or intermediate score are at risk of relapse,” Mauro said. “This is a universal finding across studies,” he said. “Another predictor is imatinib duration. Those on imatinib therapy for longer are less likely to relapse and may be able to discontinue.”
Another study, A-STIM, he continued, found that half of CML patients with undetectable, intermediate transcripts not above threshold levels needed treatment. “Patients who lose major molecular response [MMR] should be considered in need of treatment for remission.” This study found that 60 percent of patient lost their CMR; stability did not have an impact on long-term success, he noted.
Discontinuation studies have also looked at second-generation TKIs. Another French group found a failure rate of 60 percent due to loss of major molecular response. “Patients in CMR may be defined as non-proliferative, rather than disease-free,” Mauro said. “They may not be different from patients with stable levels of undetectable disease. MMR loss may be a more pragmatic, accepted endpoint for failure and the need for treatment.
“I'm nervous about letting a patient who may have consistently undetectable disease rise to the level of MMR loss. We can't predict the biology of these patients.”
In addition, slower-growing leukemia-initiating cells may remain even after imatinib treatment. A threshold of molecular response (MR) 4.5 is most feasible, and treatment decrease after two years is now the standard, he said.
Loss of MMR is a true sign of failure to get the patient back on treatment, but this should be done only in the context of a clinical trial, not in clinical practice.
The Euro-SKI discontinuation trial is going even further in using an MR 4 threshold, Mauro noted. The researchers will watch these patients for relapse and non-proliferative disease. So far, “it's quite rare that a patient needs retreatment,” he said.
In conclusion, Mauro said, “current approaches, even imatinib, offer the potential for treatment-free remission. Second-generation TKIs offer a greater opportunity. Treatment-free remission success can be predicted with old tools, such as Sokal score, and newer tools will give greater clarity.”
In addition, the need for retreatment of a TKI is nearly always evident within six months, he said. “A non-proliferative state is likely a reality of treatment-free remission. Retreatment, by all reports in discontinuation studies, has been successful. Retreatment opportunities have increased with alternate TKIs, and will continue with experimental approaches, such as PD-1 and other immunotherapies.”
Ritchie: No, We Need More Clinical Trials Before Stopping Can Be Recommended
Ellen K. Ritchie, MD, Associate Professor of Clinical Medicine at Cornell Weill Medical College, took the other side of the debate, saying: “No, we need more clinical trial data before we recommend stopping TKIs. It's a gamble that is not ready for prime time. Be careful of doing this with an average CML patient.”
It's true that several prospective studies, as well as retrospective studies from Japan and Korea, have demonstrated that TKI discontinuation is possible, she said. For example, the STIM trial enrolled 100 CML patients at an average age of 69, and followed them for a median of 24 months, assessing their CMR levels.
“A total of 61 percent of patients had a molecular relapse and were put back on imatinib. Molecular relapse-free survival was 45 percent at six months, 41 percent at 12 months, and 38 percent at 24 months. The majority of patients relapsed. Those who did well did well for a long time.”
The duration of imatinib therapy appears to be important, she said. After a median follow-up of 30 months, 61 percent of patients relapsed, most of them within seven months. Three patients relapsed at 19, 20, and 22 months.
The complete molecular response rate at both 24 and 36 months was 39 percent. “Patients who had an early restart regained CMR on imatinib,” she said, noting that patients who did not do well had high Sokal scores, which is “a significant predictor of prognosis.”
The TWISTER discontinuation trial included 40 chronic-phase CML patients. Of the 40 patients, 22 relapsed, 15 of them within six months. Five of 18 patients had detectable BCR-ABL transcripts without relapse. Treatment-free remission was 43 percent. All 22 patients who relapsed responded to retreatment with imatinib. Another 16 patients regained stable, undetectable minimal residual disease (MRD).
The predictors of success in this study, Ritchie said, are prior interferon therapy, the time to switch to imatinib, and Sokal score.
Still, even with these study results, discontinuation remains an ongoing gamble, she said. “From a patient perspective, is the patient willing to gamble that the CML will come back? A patient will consider this only if he thinks he can beat the odds.”
Patients may have many questions about discontinuing TKI therapy. These include: What is the chance of a sustained remission for five years or more? Will I need bone marrow biopsies? Will I need to see my doctor more? Will I need to have more blood taken? How likely am I to go into relapse again?
The probability of maintaining CMR off imatinib is about 39 percent, Ritchie said. “There are no data for more than 36 months. If a patient has a low Sokal score, the probability of CMR is 55 percent. With a high Sokal score, it's unlikely a patient will be able to be maintained off drug.” She added that she would feel more comfortable if the patient had a longer duration of imatinib before discontinuing.
There is no need for more bone marrow biopsies, Ritchie said. “The TWISTER trial showed no benefit of ongoing bone marrow biopsies over peripheral blood monitoring. But patients will definitely have to see the doctor more and have more blood taken. I would tell patients: ‘Yes, you will see me more the first three years, and for the first three months you will see me a lot.’”
If a patient has molecular data that shows two positive samples in a row, then there is a concern for relapse. “To measure molecular relapse, we need to use real-time PCR from a high-quality laboratory,” she said. “We don't have a great understanding of the mechanisms of relapse. Before stopping becomes widespread, we need to know how and why patients relapse.”
There are still unresolved issues with regard to measurement of major and complete molecular responses, she said. “There is no single definition of CMR. Response is defined by an upper boundary, and the disease still may be detectable by real-time PCR or nested PCR at a lower level. There is substantial variation caused by technical differences across different laboratories, and it's very difficult to standardize definitions of undetectable BCR-ABL.”
In addition, there is no evidence supporting the ability of TKIs to eliminate precursor stem cells. “We have no reliable method to measure CML stem cells,” Ritchie said, noting that the TWISTER trial measured BCR-ABL junctions from genomic DNA that characterized individual patient breakpoints to detect minimal residual disease. Factors other than the number of residual CML cells are important for a sustained CMR.
“Stopping should be done only in a clinical trial, and the trial should include peripheral blood monitoring for CMR off drug,” she said. “The patients need to be willing to come into the doctor's office monthly for the first year for close monitoring to look for early molecular relapse.”
CML patients who are candidates for drug discontinuation are those who are not able to tolerate TKIs, are pregnant, or are otherwise unable to take the drugs.
In conclusion, Ritchie posed as a CML patient and asked, “Why should I stop taking my TKI? My insurance company pays for the drug. It took me three years to get in CMR. I might just barely be in CMR. Chances are I will relapse if I go off. I have to see the doctor more often and get more blood drawn. I am going to be really anxious the whole time, like I was when I was first diagnosed.”
Voting and Q&A
According to the vote by the audience, the debate was a draw. An equal number of “Undecided” voters switched their votes to each position presented.
In the question-and-answer period, an audience member asked how much the decision as to stop TKIs is due to cost concerns? Ritchie answered: “Cost is the driver if we look at developing countries and the U.S. About 12 percent of the U.S. population is over age 80. We have a revolutionary expansion of the population who are using more drugs.
“We need to learn how to use new therapies to manage costs in the interests of every government and health care paying entity.”