Insights From: Axel Grothey, MD, West Cancer Center, University of Tennessee; John Marshall, MD, Ruesch Center for the Cure of Gastrointestinal Cancers
Axel Grothey, MD: In colorectal cancer, MRD [minimal residual disease] testing has 2 implications at this point. One would be to treat where we would normally not treat. A positive ctDNA [circulating tumor DNA] test, MRD test, in a situation where we would normally not deliver, for instance, adjuvant chemotherapy, like low-risk stage II cancer, would allow us to identify patients who are at high risk for recurrence. We do know that if we have a positive ctDNA test, it predicts recurrence with almost 100% probability. This is a high-risk factor. We also know that adjuvant chemotherapy can lower this risk and actually reduce the risk of recurrence by half, according to recent data. So if I have a patient in the situation with stage II disease based on clinical factors for which I would normally not offer adjuvant chemotherapy, if this patient has a ctDNA-positive MRD test, I would of course offer this patient chemotherapy to potentially lower the risk of recurrence. That is the easy part—escalating treatment where we would normally not use treatment.
Now, the more complicated question and ethical discussion is, what if we have a patient for whom we would normally administer chemotherapy in an adjuvant setting, like stage III colon cancer, routinely patients with lymph node-positive disease, and we see the MRD test, ctDNA test is negative? The problem is that we are not 100% sure that this will really lead to a tumor-free state forever in patients without adjuvant therapy. So are we allowed to withhold treatment?
The question is, should we withhold treatment or deescalate treatment, for instance, going from a combination therapy with a fluoropyrimidine plus oxaliplatin to just a fluoropyrimidine, just capecitabine, for instance? Or, which is probably another interesting factor, should we sequentially test these patients with let’s say repeated MRD testing, a repeated ctDNA test, potentially, with this personalized test and then only administer adjuvant therapy when the test turns positive? We know there is a lead time difference between positivity of the test and emergence of metastasis on scans of about 9 to 11 months, which is potentially the time when we could utilize adjuvant treatment for patients. So it would be adjuvant therapy when you need it and not adjuvant therapy for every patient, because we know that a lot of patients really don’t need adjuvant therapy.
In stage II colorectal cancer, at this point we use clinical factors—mainly T [original tumor] stage and other factors like perforation, obstruction, lymphovascular invasion, perineural invasion—to make a decision for patients about whether they should or should not receive adjuvant therapy. Now, I think it’s the consensus that for patients with a high-risk stage, like T4N0, we should administer chemotherapy. However, we still don’t have any clear data whether, for instance, an oxaliplatin-based regimen is needed. A lot of patients will probably only benefit from a fluoropyrimidine alone.
Now, a molecular, let’s say, residual disease test like the Signatera test would allow us to identify a patient population with a very high risk of recurrence in a situation like stage II cancer where we definitely need to administer chemotherapy to give this patient a higher chance to be cured from their disease. This is actually not trivial. For several years, we did not know whether adjuvant therapy in this setting could actually influence patient outcomes. What if the positivity of an MRD test means there’s a point of no return? Patients will definitively, without any question, develop metastases in spite of adjuvant therapy.
Now we’ve seen recent data and that’s not the case. There is a consequence of a positive test in a setting where we would normally not administer chemotherapy because we can improve survival for patients and potentially even cure patients. We need longer follow-up, but we definitely can make a difference from the data that we’ve seen in some patients with MRD-positive disease.
John Marshall, MD: When you think of applying ctDNA to a patient population, one of the first and most obvious places is in stage II colon cancer. Here, the patient starts with about an 80% chance of being cured. But we know that some stage II patients have a much higher risk—T4 disease, poorly differentiated, perforated tumors. Those patients have a much higher risk than say only 20%. They have a much higher risk of relapse. On the other hand, there are some stage II patients who probably have a lower risk, that do better. And so, we’re guessing when we see a patient like this.
And so, this stage II setting is an obvious place to incorporate this kind of test. If it’s positive, then you know that patient still has persistent cancer in a stage II controversial setting. “I’m not controversial any more. I need to give chemotherapy.” If it’s negative and the patient is low risk, then I am very comfortable not giving that patient chemotherapy. So that’s the easiest place, in my opinion, to apply this kind of ctDNA analysis. Where it will get trickier is in stage III disease, because our strategy now is to treat everybody. And how certain are we going to have to be that a negative test is negative and withhold any chemotherapy from a stage III patient? That’s the goal, because we’re overtreating a lot of people with chemotherapy.
And so for right now, I don’t think we’re there yet. Some of the prospective studies that are being done are addressing these groups of patients so that our datasets are bigger and we can answer that question.
Axel Grothey, MD: We’ve seen very consistent data from various studies, whether it’s a tumor-informed or uninformed test. The presence of ctDNA after surgery, after a washout period of several weeks, if ctDNA is still there, if we find mutations that can be attributed to the resected cancer, the risk of recurrence is virtually 100%. So it’s a very strong prognostic factor. Actually, it kind of trumps almost all of these factors that we see clinically. We have actually had discussions that it might need to be integrated into the AJCC [American Joint Committee on Cancer] staging system. We’ll go beyond TNM [tumor, node, metastasis] stage, for instance, to assess the prognosis of our patients, because it clearly has treatment consequences right now.
I can easily foresee that in the very near future, almost the present right now, the ctDNA-MRD assessment is going to become standard of care and would help guide our treatment decisions, at least in the area where we escalate adjuvant therapy, and at some point, potentially also deescalate treatment in ctDNA-negative cases and only use adjuvant therapy when patients turn positive with sequential testing.