Research in gastrointestinal cancers has evolved rapidly. In 2015, there were a few landmark studies that either dramatically transformed our clinical practice or significantly advanced our understanding of tumor biology.
Immunotherapy has emerged as the mainstay treatment in solid tumors, especially with the recent media coverage on the dramatic disappearance of melanoma lesions in former president Jimmy Carter. Similarly, immunotherapy has made a striking appearance in gastrointestinal (GI) cancers. The study led by Dung Le et al was based on the concept that tumors with “neoantigens” encoded by somatic mutations may be responsive to immunotherapy, so tumors with a high mutation burden due to mismatch-repaired (MMR) deficiency would generate more “neoantigens” and potentially respond better to immunotherapy (N Engl J Med2015;372:2509-20).
Indeed, this phase II study reported an immune-related objective response rate of 40 percent (4 out of 10) in patients with MMR deficient colorectal cancer while on pembrolizumab, an anti-programmed death protein 1 (PD-1) inhibitor, compared with 0 percent (0 out of 18 patients) in the MMR proficient group.
Additionally, the study confirmed a high somatic mutation load of 1,782 (mean) in the MMR deficient group, compared with 73 in the MMR proficient group, and a high mutation burden was associated with significantly longer progression-free survival. The drastic difference in treatment response between two groups illuminated the fact that a large sample size may not be necessary for a well-designed, hypothesis-driven trial. The similar impressive treatment response toward immune checkpoint inhibitors is anticipated across different tumor types with MMR deficiency. Meanwhile, anti-PD-1/PD-L1 inhibitors have continued to flourish in other GI tumors, including hepatocellular carcinoma and gastroesphogeal cancer, among others.
Comprehensive genomic analysis has not only identified novel oncogenic drivers, but also shed the light on potential therapeutic targets. A study led by Jeffery Ross et al has shown that two-thirds of patients with bile duct cancer have potential clinically actionable genetic alterations (2015 Gastrointestinal Cancers Symposium, abstract 231). Next-generation sequencing for 315 cancer-relevant genes was applied to 554 tumors, including intrahepatic cholangiocarcinoma, extrahepatic cholangiocarcinoma and gallbladder cancer. Interestingly, actionable alterations such as ERBB2 amplifications and PIK3CA/mTOR alterations were observed in gallbladder cancer in nearly the same proportion of breast cancers. Other therapeutic targets—FGFR fusion, IDH1/2 substitution and BRAF substitution—were seen in intrahepatic cholangiocarcinoma.
In an uncommon disease with few active agents such as bile duct cancer, individualized treatment matched to its genomic alteration carries great translational promise. The current challenge is to test this approach in the prospective setting, preferably by enrolling patients with actionable mutations in the NCI-MATCH trial.
Along the same line of genomic research, instead of looking for individual actionable mutations, a mutation signature approach has been developed as a new strategy tailoring therapeutic intervention. In particular, signature 3 is a mutation signature of a defective double-strand DNA break repair, which has great clinical implications, as it may predict response to DNA damaging agents. The study, led by Ludmil Alexandorv and his colleagues, examined the signature 3 (mutation signature) by analyzing 607 whole-genome sequences across 36 different human tumor types (Nat Commun 2015;6:8683). Signature 3 was not only observed in breast, ovarian and pancreatic cancer, but also 7-12 percent gastric cancers. This signature approach is highly translational as it expands the potential therapeutic benefit of platinum compounds and PAPR inhibitors beyond the known deleterious germline BRCA 1 and BRCA 2 mutations.
A similar observation was made in pancreatic cancer by Nicola Waddle et al through deep whole-genome sequencing and copy number variation analysis of 100 pancreatic ductal adenocarcinomas (Nature2015;518:495-501).
The patterns of structure variation were used to classify pancreatic cancer into four subtypes with potential clinical utility. The subtype “genomic unstable” (14% of all samples) showed a large number of structure variations, suggesting the genomic instability associated with defective DNA repair. Therefore, patients with the genomic unstable subtype may be more sensitive to DNA damaging agents.
Hopefully, the concept that tumors harboring a signature of defective DNA repair respond better to the DNA damaging agents is being tested in a prospective trial.
Early Detection and Screening Tools
Pancreatic cancer has a dismal prognosis with the only cure being surgical resection. Early detection remains the key to improving survival in this deadly disease. Exosomes are lipid-covered extracellular vesicles containing cellular material; they can also be derived from cancer cells.
A study led by Sonia Melo et al demonstrated the feasibility of using glypican-1 enriched cancer exosomes in serum as a biomarker for early stage pancreatic cancer detection (Nature2015;523:177-82). The presence of glypican-1 (GPC1) exosomes can be used to distinguish healthy subjects and those with benign pancreatic disease from individuals with pancreatic cancer. Although the use of GPC1 exosomes as a screening biomarker for early stage pancreatic cancer needs to be largely validated, this novel approach has reignited the interest in the field of cancer early detection.
While exosomes appear to be novel, panel 3-protein biomarkers (LYVE1, REG1 and TFF1) in urine for early pancreatic cancer detection have also been recently explored. A study led by Tomasz Radon and his colleagues has validated a biomarker panel in urine samples from 192 pancreatic cancers, 87 healthy volunteers and 92 chronic pancreatitis. The urinary biomarker panel performed well in the study by ROC curve analysis in the training and validation (Clin Cancer Res2015;21:3512-21). This urine test has an advantage of its non-invasive nature, although how to move this test forward as a screening test in a cancer with low incidence remains to be a challenge.