SAN FRANCISCO — A 3-protein biomarker assay might be able to predict which patients with esophageal cancer are likely to be highly resistant to the standard regimen of presurgery chemotherapy and radiation. This test could help individualize treatment regimens by identifying patients with extreme resistance to chemotherapy, allowing physicians to direct them to alternative options.
Details about the test were presented here at the 2014 Gastrointestinal Cancers Symposium.
The biomarkers were originally discovered by researchers at the University of Texas M.D. Anderson Cancer Center in Houston, and were developed into a commercial product, the DecisionDx-EC test, by Castle Biosciences.
An initial single-center validation study of 167 patients demonstrated an accuracy of 92% and a specificity of 97%.
A second validation study, conducted at several different centers, found the accuracy to be 79% and the specificity to be 95%.
Choosing Other Options
The standard of care in esophageal adenocarcinoma is fairly consistent for localized disease, explained Derek Maetzold, CEO and founder of Castle Biosciences. "Patients are treated with one of several chemotherapy and radiation regimens and then go on to surgical resection of the tumor."
However, according to the literature, anywhere from 20% to 30% of patients exhibit extreme resistance to standard chemoradiation. "We have a quarter of the patients with esophageal adenocarcinoma that is resistant to treatment," he told Medscape Medical News. "We have some data from M.D. Anderson and some from the Moffitt Cancer Center in Tampa, Florida, that show that treatment delays in resistant patients can be deleterious to their health."
If patients with extreme resistance can be identified at the time of their diagnosis, they might benefit from an alternative neoadjuvant regimen or go directly to surgery. "Resistant patients can then avoid a highly toxic chemotherapy and radiation regimen that isn't going to benefit them," he said.
Second Validation Shows Consistency
The NF-κB and Hedgehog cellular pathways have been shown to mediate resistance to chemoradiation therapies. The DecisionDx-EC test uses compartmental localization of biomarkers NF-κB, Gli1, and SHH in pretreated tumor biopsies to determine a localization index score for each biomarker. This information is then analyzed using a proprietary algorithm to predict extreme or nonextreme resistance.
The initial validation of the test was conducted with 167 patients at M.D. Anderson, and showed a "clinically useful accuracy of the 3-biomarker set of 92%," the investigators report.
Weiwei Shan, PhD, from Castle Biosciences, and colleagues conducted the second validation test at an independent CLIA laboratory with samples obtained from 71 patients from several centers. NF-κB, Gli1, and SHH proteins were detected using immunohistochemistry, and a blinded pathologist and expert clinical scientist evaluated resection specimens and assigned a labeling index score.
The labeling index for each sample was compared with the established training set, and the predicted response was determined using a logistic regression algorithm. The College of American Pathologist Tumor Response Grade (TRG) and the Rohatgi research scale were used to distinguish resistant from nonresistant samples.
Of the 71 samples tested, 24 (33.8%) showed extreme resistance to chemoradiation (tumor reduction of <50 47="" after="" and="" disease="" nbsp="" neoadjuvant="" no="" of="" reduction="" resistance="" showed="" treatment="" trg="" tumor="">50%; TRG 0 to 2).
The receiver operating characteristic (ROC) curve for the validation cohort was 0.85, and the positive predictive value for the validation study was 83%. This "reflects an assay that can accurately predict those likely to have some degree of response to chemotherapy/radiation," note Dr. Shan and colleagues.
In contrast, the negative predictive value of 83% and the specificity of 94% indicates that this test could accurately identify patients not likely to benefit from therapy.
The test is now available at Castle's laboratory in Arizona, and the ordering process is identical to the process for any other test that needs to be sent out, explained Maetzold.
"It is easy for any physician working anywhere in the 50 states," he said. Castle had considered putting it into a device kit and seeking approval from the US Food and Drug Administration so that it could be sold to individual facilities.
However, "considering the relative rarity of the disease, with only about 17,00 to 18,000 cases a year, most institutions aren't going to want to spend the money on something that will be used very infrequently," he added.
Several of the authors are employed by Castle Biosciences.
2014 Gastrointestinal Cancers Symposium (GICS): Abstract 51. Presented January 16, 2014.
Δεν υπάρχουν σχόλια:
Δημοσίευση σχολίου