Drug response indicator test (DRIT): A predictive diagnostic laboratory service for anticancer treatment of individual patients.

B35

For treatment of breast cancer patients, limited options are available for selecting optimal treatment regimen for a particular patient at a given stage of disease. DRIT is a diagnostic service that predicts treatment outcomes prior to selection of anticancer drug therapy for individual patients. A retrospective study utilizing DRIT was conducted to predict treatment outcomes in advanced breast cancer patients (ABC). DRIT quantitatively measures Drug Response Indicator (DRI) expression levels in formalin fixed, paraffin embedded tumor tissue. DRI chosen for each drug is indicative of tumor response as related to mechanism of drug action. This relationship is verified in vitro with 8-10 cancer cell lines of differing drug sensitivity by correlating drug toxicity (IC50) and DRI expression. For each drug/DRI pairing, a cut off level for dichotomized prediction of sensitive versus resistant tumor is generated based on DRI expression level in tumor. DRI quantitation in over 500 tumor cells is based on fluorescent dye-labeled monoclonal antibody staining, followed by analysis of digital images using computer-assisted microscopy, calibrated to external standard. DRI expression measurement results in tumor classification as sensitive or resistant to a particular drug. A tumor classified as sensitive by DRIT predicts that patient (Pt) will respond to treatment, while a tumor that is classified as resistant predicts that Pt will be non-responsive. Clinically, treatment outcome is based on RECIST guidelines and classified into responsive group (non-progressive disease, CR, PR, SD) and non-responsive group (progressive disease, PD). The drugs and DRI tested are: capcitabine/thymidylate synthase; taxanes /β-tubulin isoform III; trastuzumab/HER-2; endocrine therapy/estrogen receptor; gemcitabine/ribonucleotide reductase. Fifty-one ABC received mono or doublet therapy as first 3 lines of therapy (80 treatments). Retrospective trial results are shown with values representing - Accuracy %, Sensitivity, Specificity, Positive Predictive Value and Negative Predictive Value, respectively:
 Endocrine Therapy - 93 %( 39/42), 1.00, 0.40, 0.93, 1.00; Chemotherapy - 81 %( 22/27), 1.00, 0.58, 0.75, 1.00; Doublet Therapy - 91 %( 10/11), 0.89, 1.00, 1.00, 0.67; Cumulative Results - 89 %( 71/80), 0.98, 0.58, 0.88, 0.92.
 Favorable response rate for standard of care outcome is 75% (60/80), while the potential favorable response rate using DRIT is 88% (60/68). Accuracy for prediction of ineffective treatments is 92% (11/12). In addition, DRIT prediction accuracy for estrogen receptor expression (ER) was compared with that of a FDA approved commercial assay for prediction of ER status. The objective prediction accuracy of DRIT is 93% (39/42), while the objective prediction accuracy of commercial assays is 74 % (31/42). A prospective study using DRIT to predict treatment outcomes in ABC is underway. The above study has been extended to consider Pts with GI cancer. Preliminary results indicate that DRIT may be used to predict patient response to drug therapy in GI cancer as well. These retrospective studies demonstrate that DRIT is a platform technology that can accurately predict treatment outcomes and can potentially improve response rates in treatment of ABC and possibly other cancers.
 Supported in part by Maryland Industrial Partnership Program.