Stratifying diagnostic cut-off values of tumor markers based on genetic variants may improve detection of pancreatic cancer, according to investigators.
Stratification had the greatest positive impact on accuracy of carbohydrate antigen 19-9 (CA19-9), reported lead author Toshiya Abe, MD, PhD, of Johns Hopkins Hospital, Baltimore, and colleagues.
“Despite the evidence that genetic factors influence tumor marker levels, the potential utility of using a genetic test to improve the interpretation of tumor markers has drawn limited attention,” the investigators wrote in Clinical Gastroenterology and Hepatology.
And improvements are needed, the investigators noted, particularly for early cancer detection in high-risk individuals.
“[T]he toughest hurdle for a pancreatic cancer detection blood test is the detection of stage I disease,” the investigators wrote. “Cancers generally shed biomarkers in proportion to their size, and small stage I pancreatic cancers shed fewer diagnostic biomarkers into the circulation, making diagnosis more difficult.”
Although a 2016 study by Dr. Guopei Luo and colleagues demonstrated that diagnostic accuracy of CA19-9 could be improved via genotyping, tumor marker performance was not characterized by high-specificity cut-off values, which the present study aimed to do.
The control group included 504 high-risk individuals who were prospectively enrolled in the Cancer of the Pancreas Screening (CAPS) studies from 2002 to 2018, while the case group included 245 patients with pancreatic ductal adenocarcinoma (PDAC) who underwent resection at Johns Hopkins from 2010 to 2017.
The control group was randomly divided into discovery and validation sets in order to achieve 99% specificity cut-off values, which were used to measure sensitivity in the case group. According to the investigators, high-specificity cut-off values are necessary for surveillance of asymptomatic high-risk individuals in order to minimize false-positive results.
In all patients, tumor markers and genotype were analyzed. Tumor markers included carcinoembryonic antigen (CEA), CA19-9, and cancer antigen 125 (CA-125). Genotyping included 16 single-nucleotide polymorphisms (SNPs) in 9 genes, including FUT2 and FUT3, which are known to influence levels of CA19-9.
In contrast with previous findings, which identified three relevant subgroups of FUT2/FUT3, the present study found that four distinct subgroups were significantly associated with CA19-9 levels: FUT3-null, FUT3+/-, FUT3+/+, and FUT2-null.
When CA19-9 cut-off levels were stratified by these four subgroups and applied to the 245 patients with pancreatic cancer, the investigators achieved a sensitivity of 60.8%, compared with 52.7% without stratification. The new cut-off values led to reclassification of 28 (11.4%) patients with pancreatic cancer, including 24 who switched from negative to positive, and 4 who switched from positive to negative.
Sensitivity of the SNP-adjusted CA19-9 test was improved to 66.4% when used exclusively in patients with functional FUT3 genes. Conversely, sensitivity was markedly lower, at 36.7%, when the test was used for patients with stage I disease.
While CA19-9 testing was notably improved by SNP-based stratification, results from CEA and CA-125 testing were more modest. Standard CEA testing had a sensitivity of 13.8%, compared with 15.9% when cut-off values were stratified by FUT2 status and ABO blood group. Similarly, modifying CA-125 values based on SNPs in GAL3ST2 raised sensitivity from 15.5% to 17.6%.
Although combining SNP-modified tumor marker results did increase overall sensitivity to as high as 66.1%, this also reduced specificity to as low as 95.4%
Still, Dr. Abe and colleagues suggested that the findings demonstrate proof of concept.
“Our results show that a tumor marker SNP test can improve the diagnostic accuracy of CA19-9 and, to a lesser extent, CEA and CA-125, but further work is needed to improve the diagnostic accuracy of our panel for the detection of early-stage pancreatic cancer,” they concluded.
The investigators also suggested that the technique could have value for surveillance of ovarian cancer; however, again, they emphasized the need for more research.The study was funded by the National Institutes of Health, Susan Wojcicki and Dennis Troper, the Pancreatic Cancer Action Network, and others. The investigators reported no conflicts of interest.
SOURCE: Abe T et al. Clin Gastro Hepatol. 2019 Oct 29. doi: 10.1016/j.cgh.2019.10.036.