Expert Outlines Clinical Implications of Precision Medicine in GI Cancers

Shirley Michelle Shiller, DO

Regarding precision medicine in gastrointestinal (GI) cancers, pathologists are working to determine which patient populations should undergo which molecular tests.

Colorectal Cancer

In a presentation during the 2018 OncLive^® State of the Science Summit™ on Gastrointestinal Cancers, Shirley Michelle Shiller, DO, a pathologist at Baylor University Medical Center, discussed the clinical application of precision medicine in GI cancer.For patients with colorectal cancer (CRC), Shiller said that testing is recommended for KRAS, NRAS, and BRAF mutations. These 3 mutations constitute over 50% of the genetic changes in CRC, she added. Identifying these markers can guide therapy, provide prognostic data, and can be relevant in Lynch syndrome work-up. BRAF V600E is the relevant target in the colon that does not necessarily apply in other sites of origin, and does allow a patient to get targeted therapy, Shiller explained.

"It is important to remember when ordering tests on solid tumors that cancer is genetically heterogeneous," said Shiller. "Sometimes you may not get a result that you expect or that is consistent with a previous test in the same patient with the same disease. Cancer and tumor heterogeneity may be dictating that."

Additionally, when clinicians are ordering molecular-based assays, Shiller said it is important to know whether the lab is College of American Pathologists or Clinical Laboratory Improvement Amendments certified.

Gastric Cancer

Other recommended testing from collaborative groups includes screening for Lynch syndrome on all colorectal adenocarcinomas. This is not only in an effort to guide treatment, but also as a method to guide the patient for surveillance of extraintestinal manifestations and overall global management.In gastric and gastroesophageal junction (GEJ) cancers, the guideline-recommended testing for inoperable locally advanced, recurrent, or metastatic disease is HER2, mismatch repair deficiency (dMMR), microsatellite instability (MSI) status, and PD-L1 expression.

Esophageal Cancer

The same interpretive guidelines for breast cancer are recommended for HER2 testing in gastric/GEJ cancers; however, Shiller said she is unsure this is the best approach. The National Comprehensive Cancer Network (NCCN) guidelines recommend using immunohistochemistry (IHC) first for HER2. Moreover, screening for dMMR and MSI can potentially guide therapy in the pan-cancer setting. Shiller said that it is important to keep in mind that gastric cancer is one of the extraintestinal manifestations for Lynch syndrome.In esophageal cancer, NCCN guidelines recommend hereditary predisposition testing with tylosis, nonepidermolytic palmoplantar keratosis (PPK), Howel-Evans syndrome, Familial Barrett’s Esophagus, Bloom’s syndrome, and Fanconi anemia. Additionally, the guidelines recommend hereditary predisposition testing to be overseen by a genetics specialist.

"De novo mutations are known to occur in each of these syndromes, meaning that they may have a relatively innocuous and uninformative pedigree or family history. However, it could be originating with that individual, and the rate of that varies upon entities," Shiller explained.

Pancreatic Cancer

Approximately 5% to 10% of gastric cancer is familial, while 3% to 5% is hereditary. Shiller said that these percentages are likely underestimated because testing is underutilized for patients in this setting. Hereditary cancer predisposition syndromes that include gastric cancer are hereditary diffuse gastric cancer, Lynch syndrome, juvenile polyposis syndrome, and Peutz-Jeghers syndrome.For patients with pancreatic cancer, the NCCN guidelines recommend dMMR/MSI testing. Shiller said that the overwhelming majority of pancreatic cancer is KRAS mutation—positive, which potentially suggests that these patients may respond to MEK inhibition.

With respect to hereditary predisposition to pancreatic cancer, at least 10% of cases are familial. The entities that have pancreatic cancer as part of their syndromic presentation are BRCA1/BRCA2-associated hereditary breast and ovarian cancer; Lynch syndrome with mutations of MLH1, PMS2, EPCAM, MSH2, and MSH6; Fanconi Anemia and BRCA pathway genes ATM, PALB2, FANCC, and FANCG; familial malignant melanoma tracking with CDKN2A mutations; familial pancreatitis which includes PRSS1, SPINK1, and CFTR; Peutz-Jeghers syndrome; familial adenomatous polyposis-adenomatous polyposis coli; and Li-Fraumeni syndrome with a mutation in the TP53 gene.

Cholangiocarcinoma

"This is my warning to you to not let the syndromes catch you off guard. The tricky thing about syndromes is that they are defined as a constellation of symptoms, and so it requires us to make sure that we are thinking of multiple things all at one time," Shiller said. "The time is now, in my opinion, to start shifting our thinking in having a low threshold in considering these hereditary aspects.”Another GI malignancy that is a part of the Lynch syndrome spectrum is cholangiocarcinoma. Within the disease are genetic differences, particularly between intrahepatic cholangiocarcinoma (ICC) and extrahepatic cholangiocarcinoma (ECC).

Liquid Biopsies, PD-L1, and Tumor Mutational Burden

The relevant markers in ICC include IDH1/2 (10%-23%), FGFR2 fusions (8%-14%), NRAS (17%), and BAP1 (14%). Negative prognostic indicators are ARID1A, PIK3C2G, STK11, TGFBR2, and TP53. In ECC, relevant markers are KRAS (47%), HER2 amplification (18%), TP53 (24%), and ARID1A (16%). Mutations in ALK, TP53, and IDH1/2 are all negative prognostic indicators.There has been a lot of conversation about liquid biopsies in the oncology space, said Shiller. The role of liquid biopsies in some of the GI malignancies would be to detect recurrence, identify treatment resistance, or early CRC detection. Liquid biopsies could also be useful in pancreatic cancer, as the pancreas is difficult to reach and get diagnostic tissue from.

Strengths of liquid biopsies are that it is less expensive than a standard biopsy, can provide a more rapid turnaround time, is less invasive, and has a very high sensitivity. Factors that limit the results are tumor heterogeneity, tumor shedding, and tumor burden.

As for PD-L1 testing, the clinical implications are underdeveloped. Pembrolizumab (Keytruda) is approved for patients with PD-L1—positive recurrent or advanced gastric/GEJ adenocarcinoma, and in dMMR/MSI-high CRC. The indication is for patients who failed previous therapy, as this inhibitor is not a frontline option.

Moreover, the PD-L1 IHC 22C3 pharmDx assay, which is approved by the FDA, can be an aid in identifying patients with non—small cell lung cancer and gastrointestinal cancers who are eligible for pembrolizumab monotherapy.

One of the molecular testing topics on the horizon is tumor mutational burden (TMB), said Shiller, though she added it is not yet ready for primetime. TMB is defined by number of mutations per megabase, and patients can be categorized having low, intermediate, or high TMB. One of reasons that TMB testing is not in the mainstream is because it has not yet been standardized. There is an effort to standardize the technique to avoid the challenges that have been experienced with the integration of PD-L1 testing.

Variables that impact the result of TMB testing are whether targeted or whole-exome sequencing is being done, the depth of coverage, and the number of genes being sequenced.

Something that is underestimated in molecular reports are ploidy and copy number variations, Shiller said. Copy number changes should be considered, as they can impact response to therapy. There is also a nearly ubiquitous and critical role for c-MYC in CRC. Other potential areas of growth are cell-free DNA and stool studies.

Other new targets include proteins in the RTK/RAS and PI3K pathways, including IGF-2, IGFR, ERBB2, ERBB3, MEK, AKT, and mTOR.

"Do I think it is time for us to be screening every single cancer patient for a hereditary predisposition? No. The technology has to catch up—our testing has to catch up. We are detecting new mutations daily and we don't know what they mean. Though, we do want to start thinking about these considerations more and more," concluded Shiller. “Trust your genetics experts, counselors, and molecular-trained pathologists to help guide you if you ever have a question—especially if the patient has a unique cancer or is young with a family history that is otherwise uninformative."

Shiller SM. Precision medicine in gastrointestinal cancers. In: Proceedings from the 2018 State of the Science Summit on Gastrointestinal Cancers; August 16, 2018; Dallas, Texas.