Castle Biosciences, Inc. has announced the publication of the second clinical validation study of its multi-analyte test for esophageal cancer. The DecisionDx-EC test analyzes the localization of three protein biomarkers within a patient's tumor to classify the cancer as either responsive to or resistant to pre-surgical chemoradiotherapy. The paper, "Prospective prediction of resistance to neoadjuvant therapy in patients with locoregional esophageal adenocarcinoma," was published in the peer-reviewed journal Gastrointestinal Cancer: Targets and Therapy.

According to the study, DecisionDx-EC demonstrated strong accuracy and specificity in identifying patients with tumors that are unlikely to respond to standard pre-surgical (neoadjuvant) chemotherapy and radiation.

"Neoadjuvant chemotherapy and radiation therapy will achieve no clinical benefits in up to a quarter of patients with esophageal cancer. However, this subset of patients is subject to potential toxicity, along with associated costs," commented Kenneth Kesler MD, Professor of Surgery, Thoracic Oncology Program, Indiana University. "The ability to identify resistance may help with selection of alternative chemotherapy agents or treatment strategies."

Study Highlights

The test, a three-protein biomarker assay, was discovered by researchers at the University of Texas MD Anderson Cancer Center and licensed to Castle Biosciences. Pre-treatment tumor biopsies were used to evaluate resistance (exCTRT) or responsiveness to (non-exCTRT) standard pre-surgical CTRT regimens under CAP-accredited, CLIA-certified laboratory protocols. According to validation studies, the test can reliably differentiate patients who are complete or partial responders to chemoradiotherapy from those who are non-responders. An initial, single center clinical validation study of 167 patients, which was used as a training set for the current validation study, achieved an area under the curve (AUC) of 0.96 and an overall accuracy of 90%.

The second validation study, published today, enrolled 64 patients from two independent institutions, and achieved an AUC of 0.96 and an overall accuracy of 84% for classifying which patients are likely to be highly resistant to pre-surgical chemotherapy treatment for esophageal cancer.

The DecisionDx-EC test uses a predictive algorithm to classify patients as either exCTRT (College of American Pathology Treatment Response Grade 3) or non-exCTRT (College of American Pathology Treatment Response Grade 0, 1, or 2). Compartmental localization of the biomarkers NF-kB, Gli1 and SHH is assessed in pre-treated tumor biopsies to determine a localization index score for each biomarker. This information is then analyzed using the proprietary algorithm to predict exCTRT or non-exCTRT. The test is currently available for health care providers to order through Castle Biosciences.

"Publication of these results is a culmination of our extensive program to analytically and clinically validate a new predictive test for esophageal cancer," said Derek Maetzold, President and CEO of Castle Biosciences. "DecisionDx-EC fits well within our strategy of developing and commercializing valuable prognostic tests that help physicians to select the most appropriate care for their patients."