Molecular testing for colorectal cancer continues to evolve at a rapid pace. Biomarkers already have the potential to predict outcomes and better target therapies, with the eventual goal to completely personalize treatments based on the patient’s individual biomarker profile and other tumor markers.
John L. Marshall, MD
Molecular testing for colorectal cancer continues to evolve at a rapid pace. Biomarkers already have the potential to predict outcomes and better target therapies, with the eventual goal to completely personalize treatments based on the patient’s individual biomarker profile and other tumor markers. However, knowing which molecular testing to perform or how to interpret the results can be challenging, particularly as new discoveries continue to be made.
Defining the Essential Molecular Tests
KRAS and RAS results
During a recent OncLive Peer Exchange panel titled “Optimizing Systemic Therapy in Advanced Colorectal Cancer,” the panelists provided an overview of how they are using molecular markers in their practice to help with risk stratification and guide treatment. They also reviewed new advances in molecular testing and discussed the impact of tumor location (left- vs right-sided) on prognosis and treatment, the latter of which was a major finding unveiled at the 2016 American Society of Clinical Oncology (ASCO) Annual Meeting.With regard to molecular tests, “it seems every year the story changes,” said moderator John L. Marshall, MD, as he first addressed the panel. Based on currently available data, the panelists agreed that all patients with newly diagnosed metastatic colon cancer should receive KRAS, RAS (extended analysis), and BRAF testing and also be assessed for microsatellite instability (MSI). The panel suggested that upfront testing for HER2 could also be considered, but is not regarded as mandatory yet.impact treatment decisions, and an extended RAS analysis should be performed even in the setting of KRAS wild-type tumors, the panel noted. “[With this testing], you’ll identify at least an additional 13% to 16% of extended RAS mutations that could impact your treatment despite being KRAS wild-type,” said Cathy Eng, MD. She recommended clinicians send out for extended RAS testing if their centers are not routinely performing it.
Cathy Eng, MD
Although BRAF testing does not direct therapy, it should be assessed because it is a poor prognostic indicator. “The patient with a BRAF mutation probably has a survival of 1 year, or a little more,” said Daniel G. Haller, MD. However, many good clinical trials are ongoing for patients with BRAF-mutant metastatic colorectal cancer, offering hope to these patients, noted Eng.
MSI testing assesses tumors for mutations in DNA mismatch repair genes, such as MLH1, MSH2, or MSH6, which cause repair errors in repetitive sequences and, subsequently, MSI of tumors. In the past, MSI testing was only done for stage II cancers to determine if patients would benefit from 5-fluorouracil (5-FU). “Now, in stage IV, we realize if patients’ MSI is high, they’ll benefit from a checkpoint inhibitor,” said Tara E. Seery, MD.
A major challenge with MSI testing is that its results are difficult to interpret, particularly because not all results are actionable. In January 2015, ASCO started tumor oncology boards (http://university.asco.org/motb) to help clinicians get a better handle on tumor molecular profiling tests and studies, noted Haller, who is involved with the initiative.
HER2 testing for colon cancer is still largely experimental. Although it can be considered because a clinical trial will be investigating trastuzumab (Herceptin) for colorectal cancer, data on HER2 testing in this tumor type are not sufficient to give it high-priority status or to support repeat HER2 testing, as is currently the standard of care for breast cancer, noted Tanios Bekaii-Saab, MD. In contrast, Eng was more supportive of HER2 testing. “Less than 10% of patients are going to be HER2/neu positive, but I still think we need to educate people [since we’re now going to have trials for them],” she said.
Tara E. Seery, MD
Liquid Versus Tissue Biopsies and Actionable Mutations
The panel felt the aforementioned molecular tests are sufficient for patients with newly diagnosed metastatic colorectal cancer. “I don’t think that there is a need to expand [further], at least when we’re thinking about our first line or two of treatment,” said Bekaii-Saab. “Beyond that, I think [additional testing] may become relevant, especially when we have more targeted trials.”Blood-based colorectal cancer tests are under development and showing promise, but the panelists note that they are not ready to replace tissue-based testing yet. Currently, none of the panelists are using liquid biopsies as their primary testing modality. “I think ultimately we’re going to move away from tissue-based testing to liquid-based testing, which will essentially give you a snapshot of the changing genetic profile of some of the tumors across time,” said Bekaii-Saab. “That probably would be more beneficial than just a glimpse: one biopsy of a single site at a given time.”
Although blood-based testing has the potential to detect a wider variety of genetic mutations at any given time, the panel noted that a major challenge is finding actionable mutations for which a drug is available.
Currently, several trials are collecting data on gene mutations via tissue biopsy samples to determine additional uses of molecularly targeted cancer drugs outside of the indications approved by the FDA. These include ASCO’s TAPUR (Targeted Agent and Profiling Utilization Registry; http://www.tapur.org) and the National Institutes of Health’s MATCH (Molecular Analysis for Therapy Choice; https://clinicaltrials.gov/ct2/show/ NCT02465060) trials.
While TAPUR and MATCH are important initiatives, they have faced several hurdles. The MATCH trial had to be put on hold because an overwhelming number of tissue samples were submitted, noted Haller.
Daniel G. Haller, MD
The problem was that several biopsy specimens were required for each patient and it took up to 6 weeks for clinicians to find out whether their submitted samples were sufficient for molecular testing. If not, the biopsy had to be repeated, which was not always possible. The trial has since been redesigned and restarted, improving testing time to closer to 14 days and adding additional arms, noted Eng.
Furthermore, even if samples are sufficient and actionable mutations are found, the information is not always easy or straightforward to act upon.
“It’s going to get very complicated because we used to simplify things by saying you treat the organ from which the tumor came from,” said Haller. “And then people said, ‘we’re going to totally change that, it’s only based on the genomic pattern,’ but although melanoma and colon cancer have BRAF mutations, in one inhibitors work and in the other they don’t work by themselves, although they may with EGFR combination therapy.”
Anatomic Tumor Marker
The general consensus among panelists was that finding optimal treatments for actionable mutations will continue to come down to trial and error. Nevertheless, these molecular profiling trials are opening the door to new treatments for patients who have exhausted standard options or cannot afford to pay for novel medications out of pocket. In MATCH, for example, the pharmaceutical companies will provide the medications in exchange for reporting whether the drug worked, noted Marshall.Mounting evidence suggests that tumor location might be a good prognostic marker and may have important treatment implications. The panelists discussed data from a retrospective analysis of CALGB/SWOG 80405, a large, federally funded phase III trial that compared bevacizumab and cetuximab in combination with chemotherapy as initial therapy for metastatic colorectal cancer.1
The analysis revealed that patients with tumors originating on the left side of the colon (ie, descending colon, sigmoid colon, rectum) had significantly longer median overall survival (OS) compared with those with tumors originating on the right side (ie, cecum, ascending colon), with survival rates of 33.3 months versus 19.4 months, respectively.
This trend remained when patients received cetuximab or bevacizumab. With cetuximab, patients with left- and right-sided tumors had an OS of 36 months and 16.7 months, respectively, and with bevacizumab, it was 31.4 months and 24.2 months, respectively.
The reason for the discrepancy in OS between left- and right-sided tumors is not completely clear. “The right colon comes from the midgut and the left colon comes from the hindgut…data show that they vary dramatically in the types of tumor markers that we typically see,” explained Haller.
“There’s definitely a preponderance of BRAF mutations on the right, which may actually be one of the reasons why a lot of these tumors do a little worse,” said Bekaii-Saab.
CALGB/SWOG 80405 did not consider BRAF mutational status and was initiated before KRAS mutation status was known to be an important determinant of cetuximab use. The panel agreed that more studies are needed to better determine the biologic profile of left- versus right-sided tumors, especially as studies are increasingly showing these cancers to be very different from each other biologically.
They also agreed that clinicians should make sure to note the side of the tumor when entering their ICD-10 codes, particularly because location is likely to play an increasing role in risk stratification and become an important consideration in clinical trials of new treatments.
Tanios Bekaii-Saab, MD
“The question of whether side matters in terms of the choice of biologic is kind of difficult to answer because you have to randomize in the left and randomize in the right, which, of course, hasn’t been done,” said Bekaii-Saab. Nevertheless, analysis of CALGB/SWOG 80405 indicates that cetuximab may not be a good drug for right-sided tumors, as OS decreased in patients receiving this treatment, whereas those receiving bevacizumab had a modest improvement in OS.The panelists agreed that the biggest colorectal cancer story at the ASCO annual meeting was the finding regarding left- versus right-sided tumors, and they urge clinicians to pay close attention to the side their patients’ tumors are on. They also urged clinicians to consider clinical trials for their patients whenever they are available.
Finally, although it wasn’t discussed much during the panel, Haller suggested the next exciting step to look for in colorectal cancer research will be the development of checkpoint inhibitors, which are already showing promise for other tumors, with response rates in the double digits. “Checkpoint inhibitors work when you have a lot of mutations,” he said. This might become especially important as clinical trials and better biopsy modalities, such as liquid biopsies, identify new mutations and treatment targets.