Best practices for targeting colon cancer was the topic of discussion during a Targeted Oncology live case-based peer perspectives presentation led by John H. Strickler, MD.
John H. Strickler, MD
Best practices for targeting colon cancer was the topic of discussion during aTargeted Oncologylive case-based peer perspectives presentation led by John H. Strickler, MD, medical oncologist, Duke Health. Along with a group of physicians, Strickler discussed treatment options and the research supporting his chosen therapies for left-sided colon cancer and right-sided colon cancer based on the real-life case of a patient with metastatic adenocarcinoma of the right colon.
A 53-year-old Caucasian man who had never had a screening colonoscopy presented to his primary care doctor with fatigue, melena, and abdominal tenderness, and he was found to have anemia. His medical history was fairly unremarkable, showing well-controlled high blood pressure. His mother died from breast cancer but was diagnosed after the age of 60.
The patient underwent a colonoscopy, and this revealed an ulcerated non-obstructing mass on the right side of the colon, the ascending colon. Pathology confirmed an invasive, poorly differentiated adenocarcinoma. There was limited panel testing done, which included extendedRASandBRAFtesting. ABRAFV600E mutation was picked up, and he had microsatellite stable disease. The CT of the chest, abdomen, and pelvis showed widely metastatic disease with multi-lobar liver lesions as well as bilateral pulmonary nodules, and his disease was staged as cT4N0M1.
TARGETED ONCOLOGY:How are you ordering molecular profiling for this patient?
Strickler:Each institution is different. There’s not a standard approach nationally, and it’s also chaotic with insurance reimbursement. I know in my state, no commercial payer will pay for comprehensive genomic profiling, period. Even if it’s covered by the Centers for Medicare & Medicaid Services [CMS], it will not be covered by Blue Cross. Medicare will cover certain diseases, like lung cancer, for Foundation Medicine, but no other payer will reimburse for it.
Initially, [we would call this] next-generation sequencing [NGS]. But the problem with calling it next-generation sequencing is that it doesn’t capture all the ways you can profile a tumor. You’ve got immunohistochemistry [IHC], and there’s RNA sequencing that’s coming in now. There are other types of technologies, and NGS is probably too limited. Comprehensive genomic profiling refers to all the potential biomarkers you can test for beyond just a DNA-based NGS panel. [When I say limited], that’s [about] 50 genes or less, so it usually excludes things like amplification and fusions. When you talk about an expanded panel, I think about that being beyond 50 genes, along with amplifications and fusions and other things that are potentially “druggable.” But it can be subjective.
TARGETED ONCOLOGY:What do you think is the best approach for treating this patient?
Strickler:It’s a right-sided cancer, so chemotherapy plus an anti-EGFR is probably not the right choice for this patient, based on the 80405 and FIRE-3 [trials].1,2
When you look at the median survival of metastatic colon cancer, typically, [patients treated in line with the] 80405 and FIRE-3 [trials] see survival times in the 30-month range. These patients in these studies with aBRAF-positive tumor typically have survival of around 14 months. Their tumor, in terms of natural history, has more in common with pancreatic cancer than it does with colon cancer. It’s a very aggressive and diffi­cult-to-treat cancer type, and that has an influence on how we approach these cancers.
Based on the TRIBE trial,3which was an Italian trial, they looked at FOLFOXIRI [irinotecan, 5-fluorouracil (5-FU), folinic acid, and oxaliplatin] versus FOLFIRI [irinotecan, 5-FU, and folinic acid] and bevacizumab [Avastin] and found much higher response rates for FOLFOXIRI, with a strong signal of benefit in those patients who have the worse prognos­tic features,BRAF.
The patient was treated with FOLFOXIRI and bevacizumab. Therapy was well tolerated after management of grade 2 neutropenia.
A follow-up scan 2 months later showed a 35% decrease in his liver lesions and the lung lesion. He continued on FOLFOXIRI and bevacizumab for 4 months and then developed grade 1 neuropathy. Eventually, he was switched to FOLFIRI and bevacizumab.
He did well until 4 months later, when he developed more fatigue and started losing weight. Imaging showed a new 2-cm left-sided lung lesion and associated pleural effusion. HisECOG performance status was 1.
TARGETED ONCOLOGY:What data do we have on BRAF-targeted strategies for a patient like this?
Strickler:We are talking about tumor-agnostic strategies when we talk about inhibiting BRAF, NTRK, or HER2. We started to think about all tumors being the same, but in truth, each tumor has its own wiring, and what we found is, the single-agent response rate for vemurafenib [Zelboraf], a BRAF inhibitor for colon cancer, is about 5% versus 45% for melanoma. So, something is different about a colon cancer that makes it poorly responsive to single-agent BRAF inhibition, which is what inspired this latest wave of anti-BRAF strategies.
The [SWOG S1406] trial was presented at ASCO [American Society of Clinical Oncology Annual Meeting] a couple of years ago by Scott Kopetz, MD, PhD. [It was] a large, phase II, coopera­tive group study looking at an anti-BRAF strategy for patients withBRAF-mutated tumors.4Now, these patients, as I’ve said, have poor survival. The primary endpoint of this [study] was progres­sion-free survival [PFS] because they allowed crossover. So, you can’t have survival as a primary endpoint if you’re allowing cross­over, [because] it will confuse the data.
The control arm got cetuximab [Erbitux] and irinotecan, and the treatment arm included the vemurafenib, irinotecan, and cetuximab [VIC] triplet combination. [It showed] that PFS was more than doubled in patients who got that anti-BRAF combined with antiepidermal growth factor receptor [EGFR] with systemic chemotherapy, and that was statically significant. Based on [this study], we now have this in our National Comprehensive Cancer Network [NCCN] guidelines as an option. There was a trend toward survival benefit, as well, but the study was underpow­ered to look at survival specifically.
[These are] very modest results. A median PFS of 4.3 months is not that exciting. But, at the same time, [we have to] remember where we started with these patients. They do poorly in general.
That was our first-generation attempt at drugging BRAF, and there are new BRAF inhibitors on the market, [and] there are new anti-BRAF strategies now. Studies have found, at least preclini­cally, that if you drug multiple points in the MAP kinase pathway, you can increase the benefit of your anti-BRAF strategy.
The BEACON trial combines 3 drugs: encorafenib [Braftovi], a BRAF inhibitor; binimetinib [Mektovi], a MEK inhibitor; and then cetuximab, an EGFR inhibitor. The safety lead-in, nonrandomized data, was in 30 patients.5The phase III study just finished enroll­ing and reported positive survival data, so this is going to be our new standard of care forBRAF-positive colon cancer within the next 12 months.
This is impressive data. [For the] survival data for the safety lead-in, the median overall survival [OS] is 15 months. All of these patients were in the second and third line, with a median survival of 15 months when, as a population, they live, on average, 12 to 14 months. It’s impressive survival data, with a 1-year OS rate of 62%. The response rate was 48% compared with around 20% for the VIC regimen. So, once again, for patients who are in the chemotherapy-refractory settingjust as a point of reference, first-line FOLFOXIRI has a response rate of [only] 40%.
The primary endpoint of this trial was survival. The problem that we have in the United States is that we have such easy access to off-label therapies that they had to expand the trial to outside of the United States to finish enrollment, and it turned out to be a positive study. We’re getting access to these drugs based on the safety lead-in, so it becomes difficult to randomize patients in the United States.
The patient received irinotecan, cetuximab, and vemurafenib.Following a good response for 6 months, progressive disease was detected on imaging with multiple new lesions.
TARGETED ONCOLOGY:What are the choices for therapy at this point: regorafenib (Stivarga), TAS-102 (Lonsurf), or clinical trial?
Strickler:Regorafenib never went to a phase II trialit went straight to a registrational phase III trial, and I think they paid the price for that. The ReDOS trial came out of a debate Axel Grothey, MD, and Tanios Bekaii-Saab, MD, had about how to give this. Grothey said, “I never give the full dose out of the gate, because my patients can’t tolerate it,” and Saab said, “You can’t do that—there [are] no data indicating that you can do that. The FDA approval is for 160 mg; you have to start at 160 mg.” So they had a bet and they ran a trial to see whose approach would win.
Arm A [of the ReDOS trial] was a dose escalation, where you start at the half dose and escalate by 1 pill each week up to the full dose.6Arm B was the standard dose at 160 mg, or 4 pills. Then they had a further randomization for hand-foot skin reaction. The primary endpoint was unique: They were looking at the proportion of patients who completed 2 cycles, or 2 months, of regorafenib and initiated the third, so this was encompassing not just PFS but also tolerability. If a patient completed 2 months, didn’t progress, but hated itcouldn’t stand it anymore—that would be factored in here.
This study did meet its primary endpoint. Axel Grothey won the debate. The dose-escalation approach was successful, and it was statistically significant, where 43% initiated the third cycle compared with 25%. Grade 3/4 toxicities were improved for the dose-escalation approach; quality of life was better on some parameters, particularly in that week 2 of the first cycle, where you see the toxicity get bad at the 160-mg dose level. This is now part of our NCCN guidelines.
This study was not meant to look at OS, but it looked at it as a secondary endpoint, and they found that the OS with dose esca­lation was clinically significant but not statistically significant in improvement9 months versus 5.9 months,Pvalue with a slight edge, and PFS of 2.5 months versus 2 months. We’re certainly not giving up any toxicity by doing a dose-escalation approach, and you could argue that there’s a signal here that would suggest that maybe in certain cases your patient would do better, but this study is not powered to say that.
The data would suggest that 25% of patients truly experience benefit from regorafenib. That’s hard to pick up in a Kaplan-Meier curve, where you’re looking at a median. There are similari­ties among the patients who benefit. There is not a molecular biomarker that predicts benefit, but there is a suggestion that patients do better when you give the regorafenib earlier and when the performance status is better. It has all those multiki­nase toxicitiesfatigue, diarrhea, hand-foot skin reaction. When patients are fifth line, they are just not able to hold up under that toxicity. So, if I do give it, it tends to be in a healthier, more robust patient, and I’ll tend to give it earlier, not later.
The patient was started on regorafenib.
TARGETED ONCOLOGY:What if the patient exhibited anNTRKfusion on next-generation sequencing?
Strickler:NTRKis a gene that is involved in neurologic signaling like pain and proprioception. That’s its normal function. Now, when through a fusion event it gets fused to the active promoter of another gene, it becomes activated and can lead to malignancies.
Larotrectinib [Vitrakvi] is a highly selective TRK [tropomyosin receptor kinase] inhibitor. BecauseNTRKfusions are extremely rare, they had to do a development strategy where they tested all solid tumors together. [Looking at] the collective data set for 122 patients withNTRKfusions, there’s a greatNew England Journal of Medicinearticle that shows the results from the initial cutoff, and the European Society of Medical Oncology [ESMO] data are an update of this, and the trial is still ongoing. Looking at theNew England Journal of Medicinedata,7colorectal cancer can have anNTRKfusion; the rate is about 0.1%, so it’s about 1 in 1000 patients. The response rate out of this was 80% [in the overall patient cohort]. You can see huge and deep responses for these patients across a broad range of tumors [including] pancreatic cancer, breast cancer, or salivary gland cancer.
A couple of patients developed resistance. One had a nonex­pressing fusion, so the gene wasn’t producing a protein. Another patient had a solvent front resistance mutation, so that patient had a prior TRK inhibitor and then developed a gatekeeper muta­tion, which also conferred resistance against larotrectinib. So, almost all patients had some form of benefit except for those few cases where there wasn’t.
What’s also impressive is the durability of these responses84% of these responses were ongoing at the time [the study] was published, and the median duration of response had not been reached at the time of publication. So, this is a home run for a patient, if you ever find [anNTRKfusion]. This led to an FDA approval, [which was] the second tumor-agnostic approval by the FDA, the first being pembrolizumab [Keytruda] for microsat­ellite instabilityhigh cancers. This was a landmark achievement for larotrectinib, and it’s a great drugthe problem is that few patients are eligible for it.
Entrectinib is another TRK inhibitor, [but] it’s more of a nonselec­tive TRK inhibitor. It hits NTRK,ALK, and ROS1, as well. These data were presented at ESMO in 54 patients.8The primary endpoint is response rate, and although it’s ongoing, the response rate is 57%so not quite as impressive. And they were finding the same rate of patients with colorectal cancer, about 7% of their data set. The median duration of response has been reached for that drug at 10 months, with a median PFS of 11 months.
I’ll also point out that these drugs are generally very well toler­ated. They can have grade 1 or 2 adverse events, and typically, you’ll hear patients describe a mind-fogginess sensation, which is hard to grade, but I wonder if that has something to do with the fact that TRK is involved in the central nervous system. They both cross the blood-brain barrier, they’re small molecules, and they’re active against brain metastases.
TARGETED ONCOLOGY:How would you treat the patient if he hadHER2-amplified colon cancer?
Strickler:When you talk to your pathologists, they will tell you that breast cancers tend to have homogenousHER2expression under a microscope, [but] gastric cancers are different. They’re gener­ally heterogeneous, so there will be patches ofHER2positivity and negativity. Colon cancer you would think would be more like a gastric [cancer] and have heterogeneous [expression], but, in fact, it looks more like a breast cancer under the microscope. It tends to have a homogeneous expression ofHER2. It’s rare, but we will see it if we test enough patients. The prevalence is 3%, and it confers resistance to EGFR antibodies in the same way that aKRASmutation does, although those data are still contro­versial, andHER2amplification is a gradient, from low grade to high grade. It’s those high-gradeHER2amplifications that confer the most resistance. [With] a low grade, you might still get some mileage out of an anti-EGFR therapy.
[In the HERACLES] study,9which was conducted in Italy, [inves­tigators] screened about 1000 patients withKRASwild-type disease forHER2amplification, and they found and treated about 30 patients. The response rate was 30%. [The previous data showed] regorafenib has a 1% response rate, and all these patients would have otherwise gotten regorafenib. TAS-102 has over a 2% response rate. This is definitely much better than our best available therapies. The PFS was 21 weeks, the 6-month PFS [was 33%, and] 84% of patients had tumor shrinkage. The IHC 3+ patients cluster towards the best responses. The IHC 3+ patients have the high­est level of HER2 expression, [which] predicts the greatest level of benefit from a dual HER2 therapy, trastuzumab [Herceptin] and lapatinib [Tykerb].
I personally treated a patient like this withHER2-positive colon cancer, and we couldn’t get lapatinib over 750 mg because of the diarrhea. MyPathway was a large basket trial that also tested an anti-HER2 strategy with 2 Genentech drugs, trastuzumab and pertuzumab [Perjeta], in a much larger groups of patients, 57 patients with HER2-positive disease.10They used response rate as the primary endpoint. There’s 1 major difference between HERACLES and MyPathway: MyPathway allowed patients on if they wereKRASmutated andHER2positive…the response rate was 30%, [but] most of the patients who hadKRASmutations had no benefit from trastuzumab plus pertuzumab. What we learned from MyPathway is that if you are going to give this strategy, it should be forKRASwild-type. Listen to the Italiansthey were right. They should beKRASwild-type. The disease median survival is about 11.5 months.
We now have the SWOG [S]1613 study [NCT03365882], which is ongoing nationally in the United States. This is, hopefully, going to lead a registrational approval for trastuzumab/pertu­zumab. It’s a randomized study of patients withHER2-positive colon cancer that areRASwild-type, and they’ll be randomized to trastuzumab/pertuzumab versus cetuximab/irinotecan, and when they progress, they’ll be allowed to cross over.