Clinical Validation

DecisionDx-EC Validated to Predict Extreme
CTRT Resistance

The DecisionDx-EC test has completed extensive technical and clinical validation in localized esophageal adenocarcinoma with initial phases done in the research laboratories of The University of Texas MD Anderson Cancer Center (MD Anderson) and final clinical validation completed in a CLIA certified clinical laboratory under the direction of Castle Biosciences.

Development study results and data from the first clinical validation study were presented at the American Society of Clinical Oncology in 2011 (Ajani, 2011). This study was a single-center, prospectively planned study using archival tissue samples.

A multi-center, prospectively planned study using archival tissue samples was recently completed, which constitutes a second, independent clinical validation study. Importantly, the response to CTRT was re-confirmed under blinded conditions. Data from this second study, and additional ongoing studies, are expected to be presented at an upcoming scientific meeting.

The development program includes Discovery, Development and Clinical Validation studies.
Validation chart

Validation Study #1 Results Reported At
ASCO 2011

Data from the MD Anderson development and validation studies were presented at the annual meeting of the American Society of Clinical Oncology in 2011 (Ajani, JCO 2011). Patients received CTRT followed by surgery. For these studies, researchers defined extreme resistance to chemoradiation therapy as patients having greater than 50% residual tumor upon surgical resection per the Rohatgi research tool. In clinical practice physicians often use a grading tool developed and approved by the College of American Pathologists (CAP). The CAP tool uses Treatment Response Grades (TRG) of 0, 1, 2, and 3. TRG grade 0 is defined as complete response with no residual tumor cells, grade 1 is defined as “moderate” response with minimal residual tumor cells, grade 2 is minimal response, and grade 3 is defined as no or poor response.

In the first DecisionDx-EC clinical validation study, biostatistical analysis of the logistic regression model used to classify extreme resistance samples yielded an uncorrected area under the receiver operating curve, or AUC, of 0.96 and corrected AUC of 0.95. This statistical measure is used to assess new tests on a scale where 0.5 means the result is similar to chance, and 1.0 means that the test perfectly predicts the known outcome.

CLIA Analytical Validation Studies

Castle Biosciences completed technical requirements for DecisionDx-EC analytical validation by processing more than 250 sample runs of archival esophageal adenocarcinoma tumor tissue. Two clinical experts (a pathologist and a doctorate-level clinical scientist) who were blinded to patient outcome scored each sample during development and optimization studies. A 94% concordance was achieved between these two assessors indicating high proficiency and consistency.

Validation Study #2 Results

The second clinical validation study was a multi-center, prospectively planned and blinded study that was initiated and completed following finalization of the CLIA standard operating procedures. This study achieved several important objectives:

-   Confirmed the concordance between the research grading tool used by
    MD Anderson and the College of American Pathologists (CAP)
    Treatment Response Grading (TRG) tool used clinically for esophageal
    cancer. This study employed centralized blinded pathology scoring of
    paired post-treatment resection tissue. Each specimen was graded
    according to both tools. Blinded scoring showed the following:

   TRG  Rohatgi Method  Concordance
 Pathologic complete response    Grade 0  0% residual tumor  100%
 Partial response  Grade 1 or 2  >0% and ≤50%  100%
 exCTRT  Grade 3  >50%  100%

-   Maintained the high level of test scoring concordance between the two
    clinical scientists established in the first clinical validation study.
    The concordance between the two clinical scientists in the second study
    was 98%;

-   Established the positive predictive value (PPV) of the DecisionDx-EC
    test under CLIA conditions. The PPV was found to be 92%, meaning
    that when the assay predicts likelihood of ex-CTRT, it is highly accurate.

Select References
Izzo, et al. Pretherapy nuclear factor-kappaB status, chemoradiation resistance, and metastatic progression in esophageal carcinoma. Molecular Cancer Therapeutics, 2006; 5: 2844-2850

Sims-Mourtada, et al. Hedgehog: an attribute to tumor regrowth after chemoradiotherapy and a target to improve radiation response. Clinical Cancer Research 2006; 12: 6565-6572.

Ajani, et al. Validated biomarker signatures that predict pathologic response to preoperative chemoradiation therapy (CTRT) with high specificity and desirable sensitivity levels in patients with esophageal cancer (EC). J Clin Oncol 29, 2011 (suppl; abstr 4027).

Rohatgi, et al. Failure patterns correlate with the proportion of residual carcinoma after preoperative chemoradiotherapy for carcinoma of the esophagus. Cancer 2005; 104 (7): 1349-1355.

Rohatgi, et al. Characterization of pathologic complete response after preoperative chemoradiotherapy in carcinoma of the esophagus and outcome after pathologic complete response. Cancer 2005; 104 (11): 2365-2372.

Lassen, et al. Clinical development of a proprietary immunohistochemistry (IHC) assay to detect pathologic complete response (pathCR) in esophageal adenocarcinoma (EC). Gastrointest Endosc 2013; 77 (suppl; abstr Su1526).