Last month, the Oncologic Drugs Advisory Committee (ODAC) for the US Food and Drug Administration (FDA) voted in favour of supporting the use of minimal residual disease (MRD) as a surrogate endpoint to enable accelerated approvals of new therapies in multiple myeloma (MM). Given the ODAC’s unanimous stance, the US regulatory agency is expected to use the surrogate endpoint in practice, although the advisory committee’s recommendations are not binding.
To date, the FDA has granted regular approvals for myeloma drugs and combination therapies based on progression-free survival (PFS) and overall survival (OS) data. Whereas, the overall response rate, supported by the duration of response data, has been used to support accelerated approvals.
The ODAC’s recommendation was based on the data from two meta-analyses validating MRD negativity as a surrogate endpoint. One of these was conducted by the i2TEAMM (International Independent Team for Endpoint Approval of Myeloma MRD) in collaboration with the International Myeloma Foundation (IMF), who also presented the data at the ODAC meeting.
Need for using MRD
The use of MRD as a surrogate endpoint can lead to faster trial readout and accelerated approval. All of this can improve patient access to newer MM therapies. Furthermore, MRD provides oncologists with tools to not just escalate but deescalate treatment, says Dr. Rahul Banerjee, oncologist, and assistant professor at Fred Hutchinson Cancer Center.
“We currently use overall response rate and complete remission to evaluate if one treatment is better than another, but this is not 100% complete,” says Dr. Brian G.M. Durie, one of i2TEAMM members. “It just means that the myeloma protein has not been detected using immunofixation.” Dr Durie explains that approximately half of patients with complete remission still have significant MRD, meaning that the traditional endpoint of complete remission is not ideal.
Dr. Durie added that new immune therapies like chimeric antigen receptor (CAR)-T cell therapies and bispecific antibodies can induce a very deep response, and so virtually all patients achieve complete remission. But only some of these achieve MRD negative, which is defined as no cancer cells being detected in the background of at least 100,000 (10-5) normal cells. So, to assess these deeper responses, a sensitive MRD test is needed.
Dr. Durie explained that an MRD of 10-5 or better has an impact on the length of remission, i.e., PFS and OS, as per the i2TEAMM’s analysis. He added: “MRD is a better sort of evaluation tool where we are kind of raising the bar, we are not just looking for treatments that might be slightly better. We are looking for treatments that will have a major impact at the patient level.”
Dr. Banerjee agrees that MRD is a better evaluation tool as it can take years to accrue OS if the therapy is very effective and the numeric nature of MRD removes the reliance on arbitrary measures. However, some patients can be MRD-positive but still be fine as the cancerous cells that are left do not necessarily behave cancerously.” I have a patient who is eight years out, in remission but MRD positive,” says Dr Banerjee.” At the of the day, not everyone who fails to achieve MRD negativity is destined to do poorly.”
Impact of using MRD on trial design
Dr. Banerjee notes that having MRD as an endpoint does not impact trial operations as most trials include MRD as a secondary endpoint. However, it will impact both the sample size and the number of events needed to get to the primary endpoint.
Although the MRD endpoint could be used to obtain accelerated approval, the trials still need to capture PFS and OS data, says Dr. Durie.
Using MRD as a primary endpoint may compress the trial timeline as the MRD difference can be evaluated much more quickly. For example, a patient can achieve MRD negativity within 28 days after a CAR-T therapy, says Dr. Banerjee.
Currently, the only way to reliably measure MRD is through a bone marrow biopsy. There is a need to set guidelines on the timing of testing to avoid unnecessary repeat testing, as obtaining a bone marrow sample is a painful procedure. Dr Durie says that as of now it is indicated that MRD testing should be done at nine or 12 months. Furthermore, a patient should be in complete remission at the time of testing.
According to Dr. Banerjee the need for complete remission may delay the MRD testing. MM is the only cancer where to establish remission, clinicians look not at the cancer cells but at the protein produced by the cancer cells. The complete remission diagnosis is made by the absence of M-spike protein in the blood, which can take about six months to clear.
The M-spike protein can be detected for about 12-18 months using more sensitive technologies like mass spectrometry, explains Dr. Banerjee. So, the patient may still have M-spike present in the blood whilst being MRD negative. A case can be made to test for MRD negativity if a patient shows a very good partial response or better, owing to the lag in the clearance of M-spike protein.
Dr. Durie notes that although the ODAC recommended MRD as a surrogate endpoint in MM irrespective of the type of therapy there is limited MRD testing data for CAR-T therapies and bispecific antibodies. Adding that the FDA will likely evaluate the MRD timing and trial protocols for these therapies on a case-by-case basis to balance both MRD testing time and safety monitoring.
A need for advancements in MRD testing
Although only MRD negativity can be established at a sensitivity level of 10−5, the available test allows for more sensitive testing up to 10−7. There is data which suggests that a deeper limit of 10-6 or 10-7 could predict a better PFS. However, there are practical limitations that make getting a 10-6 response harder, says Dr. Banerjee.
For a test to reliably say that someone is MRD negative at 10−5,you would need to go one level deeper and have 10 cells in 10-6, explains Dr. Banerjee. The difficulty in trying to reach the lower limit of these tests has to do with how many cells you can draw. MRD testing is done using the sample collected on the first pull from the bone marrow, which he says usually never contains more than five or six million cells.
“You cannot reliably test MRD negative at 10−6, as the test does not have enough cells,” adds Dr. Banerjee. “So, you either draw a larger sample which leads to haemodilution and distorts the sample quality, or you take two bone marrow samples and possibly cause unnecessary harm to the patients. Furthermore, even if a limited number of labs are able to reliably test MRD to 10-7 in the future, the limited availability of this technical expertise will absolutely add to the cost of the test.
The International Myeloma Foundation is working on developing a blood test equivalent to bone marrow testing, says Dr. Durie. “We already have some pretty good preliminary data on blood testing, that can substitute for the bone marrow testing”.
Dr. Durie also mentions that Thermo Fisher Scientific’s Exent test is a more sensitive M-protein test, which could also serve as a surrogate of MRD. It uses mass spectrometry to provide a serum/plasma-based quantitative assessment of M-protein in the blood. Dr Banerjee suggests using this sensitive blood test as a screening test for bone marrow MRD, thereby, possibly alleviating the need for unnecessary bone marrow biopsies.
Dr. Durie says that the i2TEAMM is actively engaging with the European Medicines Agency (EMA) regarding the adoption of MRD as a surrogate endpoint to enable accelerated approvals of new therapies in MM. The European regulatory body has set a deadline of 1 July for the evaluation. In addition to the efficacy data, the EMA also require cost analysis of the MRD test to support this clinical value of the MRD testing.
Dr. Banerjee believes that MRD as an endpoint could potentially be adopted for use in other cancers. Multiple people have reached out to me and have expressed interest at looking at the use of MRD in other blood cancers including leukaemia and lymphomas, says Dr. Durie. “There is precedent for it, with chronic myeloid leukaemia (CML) using BCR-ABL1 transcript tests to measure MRD,” adds Dr. Banerjee.
