Development of a biomarker panel for screening and early detection of gynecological cancers

PROJECT SUMMARY It is estimated that more than 80,000 women will be diagnosed with ovarian (OvCA) and endometrial (EndoCA) cancers this year in the U.S. and this will result in the death of 26,000 women.

There are no screening tests for either of these two female-specific cancers which also disproportionally affect ethnically distinct populations.

For both cancers, detection of early stage, localized disease is associated with 5-year survival rates in excess of 90%, while diagnosis with late stage, metastatic disease results in dramatically reduced 5-year survival rates of ~25%.

MDDx, Inc. has been leveraging access to >12-years of longitudinally collected and deeply annotated biobanked plasma samples from the Gynecologic Cancer Translational Research Program (Icahn School of Medicine at Mount Sinai; New York, NY and Nuvance Health; Danbury, CT) to identify tumor-associated autoantibodies (AAb) that could serve as diagnostic biomarkers.

By performing AAb profiling against the entire human proteome and applying our novel machine-learning based method for classification of molecular profiles we have determined that classification signatures of <20 biomarkers can differentiate between women: 1. with and without cancer with accuracies of ~90% or higher (area under receiver operating curve, AUROC=0.92), 2. with OvCA vs EndoCA (AUROC=0.97), and 3. type I vs II EndoCA subtypes.

This current approach requires that each patient is screened using the 21,000 protein array, which while powerful, would be prohibitively expensive and inefficient, and complicates the process of assigning a risk score; however, our data indicates that there is a minimum and common set of ~100 biomarkers that could be used for screening all women.

This STTR program aims to refine our current platform into an affordable, easy-to-use, high confidence biomarker panel that can be used to screen all perimenopausal and older women as well as those with known risk factors for these two cancers.

The goal of this Phase I STTR proposal is to select the optimized set of AAbs required to assign a high specificity risk score and perform preliminary validation of this diagnostic panel.

In Aim 1 we will expand our proprietary database of plasma autoantibodies to ensure that we have a sample size (~650) that will enable us to confidently apply our machine learning approaches to identifying the minimal panel of AAbs for the diagnostic.

We will use this enhanced database to identify a prototype panel of ~200 AAbs for construction of classification scoring functions to distinguish between cancer and no cancer, as well as OvCA vs. EndoCA, and type I vs II EndoCA.

In Aim 2 we will perform a blinded validation and performance study using an independent set of 200 biobanked blood samples of women with and without these cancers to identify the minimal panel of ~100 biomarkers for large scale prospective validation in Phase II.

Successful completion of this Phase I program will identify the optimized panel of AAbs for an affordable, laboratory-based screening test to enable identification and provide actionable management recommendations for asymptomatic women with high specificity risk scores for OvCA and EndoCA when they are still curable.