Better MRD Testing May Predict Relapse in Leukemia Subtypes

Jerald P. Radich, MD

Minimal residual disease (MRD) has helped to predict relapse in numerous leukemia subtypes, with novel testing methods helping to identify the biomarker at a higher sensitivity than ever before, according to Jerald P. Radich, MD.¹

“For every cancer, patients can take 1 of 3 paths,” Radich, a professor in the Clinical Research Division and Kurt Enslein Endowed Chair at Fred Hutchinson Cancer Research Center, said in a presentation during the 25th Annual International Congress on Hematologic Malignancies. “One is that [they are refractory] and don’t respond to therapy at all [or] they go into remission. But many patients with acute leukemias initially respond and then relapse…The question is, ‘What can we do to find [those patients] and what can we do about that when they get there?’”

During his presentation, Radich further expanded on how MRD may be able to predict relapse in different subtypes of the disease and how advances such as digital droplet polymerase chain reaction (ddPCR) and duplex sequencing may help to treat patients with improved precision.

The Evolution of Molecular Markers in CML

Peripheral blood BCR-ABL measurements can be used to monitor disease in patients with chronic myeloid leukemia (CML), according to Radich.

In a trial that included 1106 patients with CML in chronic phase, investigators randomized patients to receive either imatinib (Gleevec) or interferon-alpha plus cytarabine as initial therapy and levels of BCR-ABL transcripts were measured in the blood of all who achieved a complete cytogenetic remission.² Results showed that in those who experienced remission, levels of BCR-ABL transcripts following 12 months of treatment had fallen by at least 3 log in 57% of those who received imatinib vs 24% of those given interferon/cytarabine (P = .003).

“This was the first time where we really thought that this was important,” Radich said. “[Looking at] the 12-month level of BCR-ABL based on the international scale, you can see an amazing dose response curve where at 12 months, you can do considerably poorly for patients who have less disease…When we saw this type of measured response based on level, we were blown away.”

The questions that still need to be answered in CML are related to molecular status, including when it’s time to change therapy. Even at 3 months to 6 months, the level of BCR-ABL dictates long-term outcomes, according to Radich; that might be used as a measure of when to change agents in patients. Another question is when to stop treatment.

One source indicates that patients with CML who have a BCR-ABL/ABL of less than 10% have an overall survival (OS) of 93.3%, while those with a BCR-ABL/ABL that is higher than 9.8% have an OS of 54%.³ Additionally, another study indicated that patients who have no measurable disease and discontinued treatment with a TKI had a 12-month probability of having a complete molecular response of 41% (95% CI, 29%-52%).⁴

“The actual measurement of MRD has completely driven decision making,” Radich said.

Additionally, Radich spoke to the quick progress MRD has made in CML.

“In 2000, the IRIS trial was conducted, and in 2003, the first paper on mismatch repair was published in the New England Journal of Medicine,” he explained. “PCR was next mentioned in the National Comprehensive Cancer Network guidelines the next year. Then, in 2007, the nilotinib [Tasigna] and dasatinib [Sprycel] randomized trials actually used 12-month BCR-ABL as 1 of the study end points. Now, of the other TKIs have followed suit and really use that molecular end point at 12 months as a marker of response in FDA-approval trials.”

This indicates that molecular monitoring has successfully been used to reduce trial times from 5 years to 10 years down to 12 months. This is a great increase in efficiency, according to Radich.

Detecting MRD in Acute Leukemias

MRD is typically measured through flow cytometry in acute leukemias, a method that has been fast and fairly sensitive; however, it has also proven difficult to standardize across labs in both acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML), Radich said. PCR can be utilized to detect translocations or IGH and TCR rearrangements in ALL with favorable sensitivity although the process is markedly labor intensive and slow. Lastly, next-generation sequencing (NGS) of immunoglobulin and TCR stands represents the most sensitive and standardizable method but it is one of the more expensive options available and results can take 1 week or longer.

Currently, there are 2 methods for utilizing flow cytometry in patients with ALL, Radich said.

“One is to use the original immunophenotype of the leukemia to track the samples. There’s a disadvantage there; if you don’t have the original material, it’s hard to call,” explained Radich.

“[Another method is to look] for any deviation from the usual phenotypic changes that happen in myeloid development. In that case, you don’t need the initial sample at any point in time; you just look for aberrant cells and call that.”

Sensitivity in ALL goes from 0.1% to 0.01%, while in AML, 0.1% is the most sensitivity that can be achieve, according to Radich.

Previously, in patients with B-cell malignancies, investigators would use PCR based on VDJ rearrangements, a method that is still used in Europe. However, this method is labor intensive, Radich said. Adaptive technology helps to serve a similar purpose but utilizes multiple primers to genotype across thousands of cells and identify clones. This method is thought to be more efficient with better sensitivity compared with flow cytometry, according to Radich.

In one meta-analysis that examined the association between MRD and outcomes in both pediatric (n = 11,249) and adult (n = 2065) patients with ALL,5 the hazard ratio (HRs) proved to be very close in both pediatric (HR, 0.23; 95% BCI, 0.18-0.28) and adult (HR, 0.28; BCI, 0.24-0.33) populations.

“MRD was [shown to be] the most important [factor] to drive overall outcome,” Radich explained. “It doesn’t really matter what therapy [a patient] gets; it doesn’t matter if you measure by flow cytometry or PCR; and it doesn’t matter whether [they’re an] adult or [a child]. MRD is bad in all settings.”

The primary factor that makes a difference in relapse and survival in ALL is whether a patient experiences a complete response (CR) without MRD, according to Radich.

Optimizing MRD Testing in AML

Similar studies examining the association between measurable MRD and survival outcomes have been conducted in patients with AML. Another meta-analysis that examined 61 studies with over 9000 patients with AML demonstrated similar findings to what had been observed in ALL.^5,6

“It doesn’t matter if you measure by PCR or flow cytometry; it doesn’t matter if you measure induction therapy after consolidation or prior to transplant; and it doesn’t matter which treatment [you used]. MRD is bad in all contexts with a similar HR of about 0.35 compared with ALL,” Radich said.

As interest in MRD has expanded on a national level, the National Cancer Institute has launched an AML Precision Medicine initiative where all trials under their prevue will feature MRD as a trial end point, and MRD eraser trials are also underway. Additionally, the Foundation for the National Institutes of Health is in the process of funding a resource for MRD testing in AML, wherein novel methods will be compared head-to-head. This effort will help determine which methods should be further tested in precision medicine clinical trials.

Emerging Strategies Under Exploration

Several developments have been made with regard to new MRD testing methods, including ddPCR, ultra-sensitive DNA sequencing, single cell multi-omics, and point-of-care devices.

The issue with standard PCR, is that providers are typically looking at real-time amplification based on standards and it is dependent on the kinetics of that exponential amplification, explained Radich.

“Sometimes that’s difficult to do because of sample quality or [factors] that are interfering with the chemistry. It’s a really interesting task because you’re looking for a needle in a haystack,” Radich explained. “It’s all an issue of signal and noise. One way you can potentially reduce signal and noise is instead of doing a PCR in 1 tube, you divide up each target into a separate well. Therefore, you’ll be either positive or negative, and in each of those wells there is no background. There’s either target or non-target…This allows for absolute quantification.”

A more recent trial examining outcomes after discontinuing treatment with TKIs in patients with CML found that if patients had MRD negativity detected by ddPCR, the odds of relapse were low.7 By comparison, patients who had MRDpositivity by this method had a significantly high risk of relapse. “We think this is a better mousetrap for looking at PCR,” said Radich.

Additionally, sequencing has the potential to identify MRD status, although Radich noted that conventional sequencing is not incredibly sensitive. Methods such as Sanger sequencing and NGS have sensitivities of roughly 10% and 1%, respectively, while duplex sequencing has a sensitivity of approximately 0.001%, according to Radich.

“Duplex sequencing makes libraries out of both strands,” Radich explained. “The only thing you call is a complimentary mutation in the exact same basepair location. That brings your sensitivity down to roughly about 1 in 107.”

Lastly, there is a lot of work being done with point-of-care devices, which may have particular importance amidst the COVID-19 pandemic. Investigators at the Fred Hutchinson Cancer Research Center are examining the use of a mechanical leech to collect blood from patients for MRD testing. In 1 clinical trial, investigators are examining whether this device can play a role for monitoring patients with CML.

Radich explained that it may be possible to distribute these devices to patients via a drone, which will, in turn, deliver devices back to a lab for analysis. His institute is also working to develop a CML “pregnancy test” of sorts that can identify amplifications in patients and operate without electricity.

“[Imagine that] you can do a test for any of these diseases from the comfort of your own home in 20 minutes,” Radich concluded.

References

1. Radich JP. Role of MRD assessment in hematologic malignancies. Presented at: 25th Annual International Congress on Hematologic Malignancies; February 25-28, 2021; Virtual. Accessed March 1, 2021.
2. Hughes TP, Kaeda J, Branford S, et al. Frequency of major molecular responses to imatinib or interferon alfa plus cytarabine in newly diagnosed chronic myeloid leukemia. N Engl J Med. 2003;349:1423-1432. doi:10.1056/NEJMoa030513
3. Marin D, Ibrahin AR, Lucas C, et al. Assessment of BCR-ABL1 transcript levels at 3 months is the only requirement for predicting outcome for patients with chronic myeloid leukemia treated with tyrosine kinase inhibitors. J Clin Oncol. 2011;30(3):232-238. doi:10.1200/JCO.2011.38.6565
4. Mahon FX, Réa D, Guilhot J, et al. Discontinuation of imatinib in patients with chronic myeloid leukaemia who have maintained complete molecular remission for at least 2 years: the prospective, multicentre Stop Imatinib (STIM) trial. Lancet Oncol. 2010;11(11):1029-1035. doi:10.1016/S1470-2045(10)70233-3
5. Berry DA, Zhou S, Higley H, et al. Association of minimal residual disease with clinical outcome in pediatric and adult acute lymphoblastic leukemia. JAMA Oncol. 2017;3(7):e170580. doi:10.1001/jamaoncol.2017.0580
6. Short NJ, Zhou S, Fu Chenqi. Association of measurable residual disease with survival outcomes in patients with acute myeloid leukemia. JAMA Oncol. 2020;6(12):1890-1899. doi:10.1001/jamaoncol.2020.4600
7. Atallah E, Schiffer CA, Radich JP, et al. Assessment of outcomes after stopping Tyrosine kinase inhibitors among patients with chronic myeloid leukemia. JAMA Oncol. 2021;7(1):42-50. doi:10.1001/jamaoncol.2020.5774