It's Not Quite Prime Time for Up-Front Genomic Sequencing in Early-Stage CRC

Wafik S. El-Deiry, MD, PhD, FACP

Deputy Cancer Center Director

Coleader, Molecular Therapeutics Program

William Wikoff Smith Chair in Cancer Research

Fox Chase Cancer Center

Mismatch Repair Deficiency

Genomic aberrations are becoming increasingly significant in the diagnosis and treatment of patients colorectal cancer (CRC), but the optimal timing and nature of the testing needed for these biomarkers is the subject of robust debate. Although testing for some genetic defects has been incorporated into the treatment paradigm, particularly for advanced disease, the clinical utility of up-front tumor sequencing that would yield wide-ranging information has not yet been established.In the past several years, universal testing for mismatch repair deficiency (dMMR) has been adopted as a standard for all patients diagnosed with CRC. Tumor tissue is stained by immunohistochemistry (IHC) for mismatch repair (MMR) proteins. If these proteins are present, it is consistent with microsatellite stability (MSS) or a lack of dMMR. If there is loss of expression of MMR proteins, tumor DNA can be subjected to classical microsatellite instability (MSI) testing, which involves polymerase chain reaction amplification across a number of repetitive sequences found in the human genome.

When the DNA from an MSI CRC sample is run on a gel, one observes a smear rather than discrete bands (1 from the female parent and 1 from the male parent in “normal individuals”). This is due to slippage during DNA replication and the highly polymorphic nature of these genomic regions. In addition to universal testing for dMMR, patients with stage IV CRC require tumor testing for KRAS, NRAS, and BRAF mutations. Patients with right-sided versus left-sided colon tumors have different prognoses and are now treated differently, most notably when the KRAS and NRAS genes are not mutated.

Growing Genomic Evidence

With immunotherapy for MSI-high tumors approved since 2017, finding dMMR is not just relevant to altered surveillance procedures but also actionable and could significantly affect patient survival. Extensive molecular analysis of tumors by including larger sequencing gene panels in stage IV CRC generally occurs later in the disease course after multiple lines of therapy and may identify actionable targets such as HER2, ALK, NTRK, DNA damage and repair pathways, and other DNA repair defects. When potentially actionable targets are identified for which therapeutics exist, patients can consider enrolling in open clinical trials (such as the NCI MATCH and ASCO TAPUR).In 2017, our research team at Fox Chase Cancer Center in Philadelphia, Pennsylvania, reported that CRC tumors with dMMR have increased rates of BRCA2 mutations, as well as a higher frequency of activating EGFR and NTRK gene mutations.¹ Although it remains unclear whether PARP inhibitors, anti-EGFR, or anti-NTRK therapeutics would be effective in dMMR tumors or add value if combined with immunotherapy in MSI-high tumors (those with biallelic BRCA2 mutations or activating mutations in EGFR or the NTRK genes), there is certainly compelling evidence that targeting activated NTRK genes is effective.²

Today, tumor sequencing with limited panels or a few genes is not performed in early-stage (I-III) CRC; such testing is not reimbursable because the results are not immediately actionable. Additionally, the current standard of therapy for stage II to III CRC is not tailored toward genetic abnormalities. For example, National Comprehensive Cancer Network guidelines for adjuvant therapy recommend that patients with dMMR stage III tumors be treated with 5-fluorouracil (5-FU) chemotherapy—based therapy. However, it is well known that MSI-high CRC tumors tend to be more resistant to 5-FU than other subtypes. (There is perhaps less concern for reduced 5-FU efficacy in MSI-high CRC tumors because oxaliplatin, which has been established as effective in this population, is frequently part of combination therapy, such as in the FOLFOX or CAPOX regimens used in the adjuvant setting.)

Testing for dihydropyrimidine dehydrogenase (DPD) levels in patients with CRC also is not standard of care, given that it is very rare to have a severe deficiency of this enzyme leading to excessive toxicity from 5-FU therapy. Nevertheless, we previously provided evidence that pharmacokinetic (PK)-guided dosing of 5-FU influences the toxicity and quality of life for patients with CRC in the adjuvant setting.³ However, there is currently no available approved or reimbursable clinical test for evaluating patients for PK-guided 5-FU therapy.

Debate Over Testing Approaches

We also reported that the host tumor suppressor p53 protein regulates DPYD gene expression, including some differences among common TP53 gene polymorphisms, and that mutated TP53 in CRC tumors derepresses DPYD, leading to 5-FU resistance of mutated p53-expressing tumors.⁴ Looking forward, it certainly makes sense to try to tailor therapy in a manner that is individualized to patients and their tumors by keeping track of the status of DPD and how it may be influenced in potentially opposing ways in the host versus the tumor.In a recent JAMA Oncology paper, investigators for the Ohio Colorectal Cancer Prevention Initiative Study Group took a bold and important step, comparing up-front universal testing for Lynch syndrome (a subset of patients with MSI CRC due to an inherited defect in an MMR gene such as MSH2, MLH1, PMS2, or MSH6) with tumor sequencing.⁵ The results show that tumor sequencing is simpler, more sensitive, and, with large gene panels, could identify DPYD deficiency, as well. Major issues include whether such a strategy is cost effective and whether it improves clinical outcomes. Neither issue was addressed by the manuscript but could be in future studies.

Tumor and germline genomic testing certainly makes sense for patients strongly suspected of having Lynch syndrome, such as younger patients or those with a strong family history of CRC or other associated tumors. However, although universal testing for tumor dMMR is considered a standard of care for all patients with CRC, in part due to the low cost of the IHC or MSI tests, large gene panel tumor tests are more expensive and not covered by insurance for patients with early-stage CRC.

In fact, routine tumor profiling is the subject of national debate. Vinay K. Prasad, MD, MPH, has argued that until randomized clinical trials demonstrate survival benefit for patients due to tumor genomic testing, it should not be reimbursed and should be considered experimental.⁶ David Hyman, MD, has argued that if response rates from available therapy are very high—for example, approximately 90% in patients whose tumors have NTRK fusions— then it makes sense to do the testing to find the actionable alterations in patients with advanced disease.² The costs are coming down, and some commercial platforms go further with whole exome or RNA sequencing.

Overall, the merits of doing tumor sequencing up front to improve care of patients with CRC go well beyond identification of Lynch syndrome. However, doing so should remain mostly within the realm of clinical trials until it can be shown to be cost effective and/or improve clinical outcomes of patients with CRC with earlystage disease.

References

1. Deihimi S, Lev A, Slifker M, et al. BRCA2, EGFR, and NTRK mutations in mismatch repair-deficient colorectal cancers with MSH2 or MLH1 mutations. Oncotarget. 2017;8(25):39945-39962. doi: 10.18632/oncotarget.18098.
2. Drilon A, Laetsch TW, Kummar S, et al. Efficacy of larotrectinib in TRK fusion—positive adult and pediatric cancers. New Engl J Med. 2018;378(8):731-739. doi: 10.1056/NEJMoa1714448.
3. Kline CL, Schiccitano A, Zhu J, et al. Personalized dosing via pharmacokinetic monitoring of 5-fluorouracil might reduce toxicity in early- or late-stage colorectal cancer patients treated with infusional 5-fluorouracil-based chemotherapy regimens. Clin Colorectal Cancer. 2014;13(2):119-126. doi: 10.1016/j.clcc.2013.11.001.
4. Gokare P, Finnberg NK, Abbosh PH, Dai J, Murphy ME, El-Deiry WS. P53 represses pyrimidine catabolic gene dihydropyrimidine dehydrogenase (DPYD) expression in response to thymidylate synthase (TS) targeting. Sci Rep. 2017;7(1):9711. doi: 10.1038/s41598-017-09859-x.
5. Hampel H, Pearlman R, Beightol M, et al; Ohio Colorectal Cancer Prevention Initiative. Assessment of tumor sequencing as a replacement for Lynch syndrome screening and current molecular tests for patients with colorectal cancer [published online March 29, 2018]. JAMA Oncol. doi: 10.1001/jamaoncol.2018.0104.
6. Kaiser J. A cancer drug tailored to your tumor? Experts trade barbs over ‘precision oncology.’ Science. sciencemag.org/news/2018/04/ cancer-drug-tailored-your-tumor-experts-trade-barbs-over-precisiononcology. Published April 24, 2018. Accessed May 21, 2018.