Transcript:John L. Marshall, MD: We need a better biomarker to see who’s responding, because even with all-RAS and BRAF wild-type, not everybody responds. It’s not trastuzumab and it’s not imatinib, yet, for colorectal cancer, but we have new evidence. And Dan, you did your homework, so you get to present it to us today.

Daniel G. Haller, MD: We did the homework. It turns out we may have a superb biomarker, in fact, but it’s not a biomarker. It’s an anatomic marker of tumor biology. The back story is the right colon comes from the midgut and the left colon comes from the hindgut. We’ve seen data from Tejpar and others that they vary dramatically in the types of tumor markers that we typically see. What’s going to be presented on Sunday is a look at right- versus left-sided tumors from CALGB 80405. If you look at overall survival—left side at 33.3 months, right side at 19.4 months—for all patients independent of the biologic, and these are all wild-type, obviously, if you look at biologic choice plus the side, it’s 16.7 months for cetuximab with right-sided tumors and it’s 31.4 months for left-sided tumors with bevacizumab. It’s all over the map, frankly, but the bottom line is that cetuximab may not be a good drug for a right-sided tumor.

John L. Marshall, MD: Even RAS wild-type, right?

Daniel G. Haller, MD: For cetuximab right side, it’s 16.7 months, which is 8 months’ difference. Remember, bevacizumab, ramucirumab, aflibercept, TAS-102, and regorafenib were all approved for overall survival benefits of 1.4 to 1.8 months. And just by looking where the tumor is, it’s 8 months.

John L. Marshall, MD: It’s not improving by 8 months, it’s dragging down by 8 months.

Daniel G. Haller, MD: Yes, exactly.

John L. Marshall, MD: It’s identifying a bad-risk patient.

Daniel G. Haller, MD: So, are we actually harming people on the right side if you give them up-front cetuximab?

Cathy Eng, MD: Dan, do we know about the primary tumor, if it was removed or not in those cases?

John L. Marshall, MD: In most of 80405, the tumor was not removed.

Cathy Eng, MD: Yes. So, does it make you feel inclined then to maybe remove the tumor?

John L. Marshall, MD: Well, let’s talk removing the tumor since you brought it up. So, should we? I’ve got a patient right now who’s got a nice response to the drugs. Should I take it out?

Cathy Eng, MD: With metastatic disease, and we’re treating the patient for palliative purposes?

John L. Marshall, MD: Yes, yes. Should I take it out? Is it kidney cancer? Should I take it out?

Cathy Eng, MD: The tendency currently is to not remove the tumor unless the patient is symptomatic with a near obstruction, bleeding, etc.

John L. Marshall, MD: But, there are increasing data and meta-analyses that say maybe they do better if you take their primary out.

Daniel G. Haller, MD: Didn’t the CAIRO-3 trial show that?

John L. Marshall, MD: Yes.

Cathy Eng, MD: It’s a pooled analysis from Europe, but I would say the majority of the time, we don’t take the primary tumor out.

John L. Marshall, MD: No, I know we don’t. I’m saying, we’re in charge here. Should we tell these people to start thinking about it?

Daniel G. Haller, MD: Well, it could be like renal cell carcinoma, where people do better with nephrectomy.

John L. Marshall, MD: So, it’s back to, is it the source of bad cells? I don’t know.

Tanios Bekaii-Saab, MD: The clear answer to this is we don’t know. We do not know.

Cathy Eng, MD: That’s why I brought it up.

Daniel G. Haller, MD: It’s clear that it’s unclear.

Cathy Eng, MD: Now it makes it even more confusing.

John L. Marshall, MD: But, it’s an interesting question.

Tanios Bekaii-Saab, MD: There are data that suggest that maybe there’s an improvement, and there are data that suggest otherwise. The example of renal cell is very different; it’s a cytokine-driven tumor. There are other reasons. You actually remove the primary, and you may have a chance to actually have the metastatic disease disappear, which is not going to happen with colorectal cancer. I think most of the meta-analyses and the original data, that suggest that primary tumor removal improves outcome, come from the days of 5-FU, where response rate was 10% or 15%. With more modern chemotherapy plus a biologic where the response rate is 60% or 70%, the primary tumor actually melts away as quickly.

John L. Marshall, MD: Dan, does the control arm change for colon cancer? Are these two different diseases? Do we need to do something different for right versus left tumors? We now have two trials for colon cancer, right-sided and left-sided.

Daniel G. Haller, MD: I’m waiting for the ASCO statement on this one.

John L. Marshall, MD: People are going to be talking about it this weekend, right?

Daniel G. Haller, MD: As more and more data come in, just like with RAS testing 7 to 9 years ago, yes, people will have to address this up front. There certainly will have to be stratification, at the very least, in controlled trials of any new agent.

John L. Marshall, MD: So, when you grumpily have to enter in your ICD-10 code for which side…

Cathy Eng, MD: You’re supposed to do it already.

John L. Marshall, MD: You’re going to have to do it; you’re supposed to. But those people who said unspecified, we need them to put in right versus left. I think we do now.

Cathy Eng, MD: We do.

Tanios Bekaii-Saab, MD: It’s important to keep in mind that right versus left, if you look at some of the genetics of the two, there’s definitely a preponderance of BRAF mutations on the right. That may actually be one of the reasons why a lot of these tumors do a little bit worse. Unfortunately, 80405 did not have the BRAF mutational status.

John L. Marshall, MD: But, we have those specimens. I know it’s a basement they sit in, but they’re going to get analyzed.

Tanios Bekaii-Saab, MD: It’s going to happen at some point, but there are additional studies actually that do show that they’re genetically very different cancers, which makes sense. They behave differently because, biologically, they’re driven differently. The question of whether side matters, in terms of the choice of biologic, is difficult to answer from these questions because you have to randomize in the left and randomize in the right, which hasn’t been done yet. You can say that it’s at least prognostic. If it’s predictive, it would be great. It would be a cheap way to predict for biologic or no biologic, but we don’t have that data yet.

Transcript Edited for Clarity