Treating Colorectal Cancer in the COVID-19 Era – Medscape

This interview was originally published as part of MDedge's Blood & Cancer Podcast series. In this episode, podcast host David H. Henry, MD, of Pennsylvania Hospital, Philadelphia, spoke with David J. Kerr, CBE, MD, DSc, professor of cancer medicine at the University of Oxford in England and past president of the European Society for Medical Oncology. Kerr recently cowrote an article outlining recommendations for treating colorectal cancer during the COVID-19 pandemic. In their discussion, Henry and Kerr review those recommendations and compare notes on the practice of cancer medicine in the United States and the United Kingdom. This transcript has been edited for length and clarity.

David H. Henry, MD: As a general hematologist/oncologist, I usually think of stage II colorectal cancer (CRC) as something not to treat unless it meets several conditions. What typically pushes you to treat a patient with stage II CRC with adjuvant therapy?

David J. Kerr, CBE, MD, DSc: I was a part of the QUASAR (Quick And Simple And Reliable) trial group that published a study in 2007 showing that there was a survival benefit for offering adjuvant chemotherapy to stage II patients. The benefit was modest, with an approximately 4% survival improvement, but I feel that's enough to at least discuss adjuvant chemotherapy with patients who are under 70 and fit. We find that maybe 60% of the younger patients are keen to go ahead with adjuvant treatment, even with that very modest survival improvement. Older patients tend to look at the pros and cons and more often than not say no. But I think it's legitimate and intellectually reasonable to have the discussion with the patient.

Henry: That's something that also occurs in adjuvant chemotherapy in breast cancer, where the benefit is around 2%.

Kerr: Exactly. The QUASAR trial had about 3000 patients in it, so its reliability comes from large numbers and tight confidence intervals. Around 75% of these patients were not considered high risk.

I've got a wee bit of an ax to grind about using pathologic markers of high risk in stage II. When its capacity to identify patients with, for example, vascular and lymphatic invasion has been studied, the degree of correlation between different pathology centers can be as low as 50%. We think that these risk factors are carved in stone, yet they're often more subjective than objective. Therefore, I tend not to be too swayed by them. That's why we've got a big research group looking at more objective molecular and other markers to define those truly at high risk.

Henry: In breast cancer, we do the Oncotype DX testing all the time. Do you think that's ready for us to use yet in CRC?

Kerr: Our QUASAR group validated the Oncotype DX for colon cancer. The problem is that the risk factors associated with this assay for colon cancer are not as strong as for breast cancer. Although it is used and is a validated discriminant, the hazard ratios are relatively small, around 1.5. Whereas some of the more recent work, particularly looking at digital pathology and artificial intelligence, appears a bit more powerful, with hazard ratios up to 3 and 4. There are much stronger discriminants out there.

Henry: Discussing another patient scenario: If you saw less than 12 lymph nodes or perforation, would that push you to offer this patient adjuvant therapy?

Kerr: That would be true for the latter, but these days, less so for the first one. In the past decade and possibly longer, I've seen a remarkable global improvement in the quality of surgery and the uniformity of pathology. It's unusual now for anybody in a significant cancer hospital to see a pathology report that's got less than 12 nodes. And although I agree that it's a recognized prognostic factor, it just doesn't count as much now.

Henry: Shifting to the COVID-19 era, what are you giving and for how long?

Kerr: We use a lot of oral capecitabine. In the original placebo-controlled QUASAR studies, we were using bolus 5-fluorouracil (5-FU)/leucovorin, the so-called Mayo Clinic regimen. When the X-ACT trial came out comparing capecitabine with that regimen, the capecitabine came out very nicely. It's a much more cost-effective way for us to deliver the drug. It keeps patients out of the hospital facility. It's not a trivial drug; we still need to watch over patients. But we wanted to try to reduce the comorbidity of chemotherapy in the time of COVID, keep patients out of hospital, and reduce the burden on our chemotherapy front-line staff that is, fiddling around with infusional pumps and so on. We wanted to take the hassle factor down as low as we could. We use 6 months of capecitabine in a conventional dose (2 weeks on, 1 week off), as was reported in the X-ACT study. We were pretty happy with that.

Henry: You've also recommended measuring for enzymes to see if there's fluoropyrimidine toxicity. Do you check that on your capecitabine patients?

Kerr: Increasingly in the United Kingdom, we do that in all of our patients receiving fluoropyrimidine treatment. 5-FU, which is one of the most widely used anticancer drugs in the world, is broken down and degraded by an enzyme called dihydropyrimidine dehydrogenase. If there are genetic variants that reduce the activity of the enzyme, then consequently you get higher-than-average levels of 5-FU in the bloodstream. With some of these genetic variants, it's associated with an almost 100% risk for death.

We run genetic screens in all of our patients. That allows us to identify rarely but importantly patients who would die if we gave them 5-FU, patients who are at a high risk for grade 4 toxicity, particularly neutropenia, septic neutropenic fever, grade 4 hematology drops, and so on. In the 6 months since we instituted this, we found a significant reduction in the number of patients being admitted to our wards because of fluoropyrimidine toxicity, so that's a nice thing. It's a test that we helped to invent. We've got a fantastic genetics outfit in Oxford, and we've done quite a bit of work in improving the sensitivity and specificity of the test.

Henry: Patients with Gilbert syndrome don't metabolize irinotecan very well. Is this is a different enzyme system than with Gilbert syndrome?

Kerr: It is. But you're exactly right to mention it. It's encompassed within a whole field of pharmacogenetics, in that if we can understand what the genetic variants are that may cause somebody to either overmetabolize or undermetabolize a drug, it would allow us to individualize those.

Precision medicine has been taken over by the molecular biologists, which we like. But you and I both know, being sort of graybeards, that the thing that we control most about the delivery of any drug to any patient is its dose and schedule. And that will be determined by a host of other factors. We overdose patients, but we also underdose them. If we could individualize that better, that's got to be an important component of precision medicine.

Henry: It's certainly a step in the right direction. Sadly, I've given capecitabine to patients and one or two times had someone admitted for weeks with diarrhea that wouldn't stop tremendous toxicity by proper meter-squared dosing.

Here where I practice, there's been a bid to keep people out of the hospital and clinic by using alternate-week capecitabine. Would you be a fan of that?

Kerr: It's a clever idea for reducing patterns of toxicity. We've gone with the enzyme because we're pretty comfortable with it. But we're also prepared to dose-reduce in those patients over the age of 70. In terms of clinical pharmacology, we old folk tend to be a wee bit gentler with the chemotherapy anyway, maybe offering 80% of the conventional dose up front, for sensible reasons.

Henry: In the COVID-19 era, what are you giving your patients with stage III T3N1 disease?

Kerr: I think the T3N1 patients do as well as the stage II patients. We wouldn't be inclined to offer them anything different. I don't think there's any role for the use of oxaliplatin in the treatment of stage II disease. If we think that the clinical outcomes, natural history, and survival patterns are similar for T3N1, as has been shown repeatedly in large community studies, we treat them in the same way.

Now, wouldn't it be lovely just to give 3 months of capecitabine? But that feels a step too far now. We try to base everything on published evidence, and 6 months of capecitabine feels fine to me. My wife Rachel [Kerr, an oncologist specializing in gastrointestinal cancers at the University of Oxford] was part of the SCOT trial, which contributed to that lovely New England Journal of Medicine publication in 2018, in which there was good evidence that 3 months was as effective as 6 months. If we were seeing patients with T4N2 disease in normal circumstances, we would often offer 6 months of treatment rather than 3, splitting hairs a little. But in the time of COVID, we just wouldn't do it. We'd do 3 months.

Henry: Are you persuaded by the International Duration Evaluation of Adjuvant Therapy (IDEA) trial showing that FOLFOX (fluorouracil, leucovorin, and oxaliplatin) was beat out by CAPOX (capecitabine/oxaliplatin)?

Kerr: We cannot work in Oxford and not be terrified of Sir Richard Peto, who is the enemy of all subgroup analyses. Richard has beat us over the head with that. So I think there's a huge statistical instability in subgroup analysis.

It would be surprising if CAPOX is better than FOLFOX, because if you look at the data overall, they're indistinguishable. If anything, the needle slightly favors FOLFOX. In this particular case, the needle went the other way. I didn't pay too much attention to it. But during COVID-19, we recommend 3 months of CAPOX for reasons of reducing hassle of infusion and making life that bit safer and easier, hopefully for our patients as well as our frontline chemotherapy staff.

Henry: You offered recommendations for advanced cancer as well. I just had a patient with sigmoid colon cancer with a couple of very large metastases in the liver. We began her on CAPOX. Is that the regimen you're suggesting?

Kerr: It is. That's sort of our default go-to chemotherapy at the moment.

Henry: What are your thoughts about using bevacizumab? Not just during the COVID-19 era, but before it as well.

Kerr: We don't use bevacizumab. This is a stark difference between our two healthcare systems. I used to be quite involved in governmental health policy around cancer planning. Our National Institute for Health and Care Excellence (NICE) decided early on that the clinical benefits accrued by bevacizumab were too small to warrant its at that stage enormous cost. So we don't use bevacizumab at all.

Henry: That's very interesting, because we're looking at the same data across the Atlantic and practicing differently. I can't disagree with you. I've been underwhelmed with the difference, yet our patients sometimes say, "What's the most you could possibly give to me?" Of course, then we'd start clicking off the bevacizumab and sink the national healthcare budget.

Kerr: The study that provided initial data in metastatic CRC randomized patients to receive bevacizumab in addition to bolus 5-FU, irinotecan, and leucovorin, which is a regimen that's not widely used at all. It showed a very significant benefit for bevacizumab. I suspect that the bevacizumab was making up the gap for a somewhat inferior chemotherapy regimen.

In the big, 2000-patient studies looking at CAPOX or FOLFOX plus or minus bevacizumab, the benefits almost disappeared. There were some improvements in progression-free survival but nothing seen in overall survival at all. Therefore, in our taxation-based healthcare system, it's actually an easy decision to say, sorry, we need to pass on this one.

Henry: If you've gotten 3 months in and a patient's carcinoembryonic antigen is falling and their lesions are smaller, do you press on or give a break?

Kerr: We've done two large trials in the United Kingdom, both of which were published in The Lancet, in which we took patients who were responding to chemotherapy after 3 months and randomized them into stop-and-start or continuous chemotherapy. There was no difference in overall survival in either of those groups. These are old trials with somewhat inferior chemotherapy, but nevertheless it gave us an intellectual basis for offering chemotherapy holidays. It's a nice way to present it to patients.

If patients have got small-volume disease and are relatively well, indolent, and asymptomatic, we would give them a couple of months' holiday break and then scan again. If there's any evidence of progression, we'd restart the chemotherapy.

If patients present with large-volume disease that we know we'll lose control of as soon as we stop the chemotherapy, we'll probably move from CAPOX to single-agent capecitabine, with the same regimen, doing the scan again in a couple of months. If patients were on epidermal growth factor receptor inhibitors, then we would keep that biologic going, probably with single-agent capecitabine, dose reduced.

Henry: My patient a couple of weeks ago had a sigmoid primary, was not obstructed, in pain, or bleeding, and had a couple of liver lesions that I could feel and were a bit tender. What's your threshold for deferring start of chemotherapy in metastatic disease like that?

Kerr: You present interesting cases. With that patient, I think I'd want to do something. That feels intuitively right.

But say we have somebody who has had their primary out, so we're not worried about any obstructive problems, and they come with a couple of small liver or pulmonary metastases. If they're elderly, a bit frail, or comorbid, then I'd consider watching them for a couple of months to see what happened.

There's very little trial evidence to suggest when you and I should start chemotherapy. When I was younger, there were two trials that I knew of, one Italian and one northern European, both of which were very small. Each recruited only about 120 patients. The idea was you observe and give chemotherapy when your patients become symptomatic, or give chemotherapy straightaway and randomize to one or the other management plan. One trial was positive for early intervention and the other was negative. These are very small, difficult trials to run.

Also, I think you and I suffer from two things. First, there's a lack of clear trial information. Second, there's a lack of biological markers that allow us to tell the patient sitting in our consulting room that we think they're going to get indolent disease and we can wait, or we've seen small volume in the scans but it's nasty, so we need to get on with treatment. These are things that we and many other labs around the world are looking at. Can we get a marker for aggressiveness? Wouldn't that be a handy thing to have?

Henry: The last of your recommendations for advanced CRC dealt with chemotherapy after resection and metastases. I believe there are some data pointing to the fact that postoperative chemotherapy may make for better outcomes. But would you hold that or give it in the COVID-19 era?

Kerr: We wouldn't recommend it. I think the whole literature around both neoadjuvant and classically post-resection adjuvant chemotherapy is difficult. It's an area that we've contributed to in the past and which is littered with small trials, some almost single-center. There have been great results with intrahepatic arterial infusion, but those are difficult to apply at other centers, even Penn and Oxford. So it's a rather confused area. And if I have a degree of confusion about benefits in the time of COVID-19, we don't feel it's worthwhile considering.

Henry: I also wanted to ask about rectal cancer. We've been persuaded a lot by data on this side of the world about total neoadjuvant therapy, which I've always been a fan of in my oncology career. In breast cancer, you and the patient can see if it's working or not. And here, you can do the same thing and probably get more drug in, as opposed to waiting. Are you a fan of the total neoadjuvant therapy? And are you somewhat curtailing the radiation piece in the COVID-19 era?

Kerr: I think we are becoming fans of the total neoadjuvant therapy. It's taking a bit longer to be taken up across the United Kingdom. But the data look good; we agree about that.

As things stand now, we are trying to curtail everything. We're going from complex radio-chemotherapeutic regimens, from long-course to short-course treatment. We're looking to omit some of the drugs for example, taking oxaliplatin out rather than giving combination chemo-radiotherapy. Our colleagues in the radiation field again are looking to compress and simplify as best they can and as logically as they can.