Abstract 2812: Detection of mesothelioma in asbestos exposed individuals with aptamer proteomic technology

Abstract

Abstract Malignant pleural mesothelioma is an aggressive, asbestos-related pulmonary cancer which is increasing in incidence. This disease causes an estimated 15,000 to 20,000 deaths per year worldwide. Between 1940 and 1979, approximately 27.5 million people were occupationally exposed to asbestos in the United States. The incidence of pleural mesothelioma in the US is 3,000 new cases/year and will not peak for another 20 years. Mesothelioma has a latency period of 20-40 years from asbestos exposure, but once diagnosed this aggressive disease is often fatal within 14 months. Because diagnosis is difficult, most patients present at a clinically advanced stage where possibility of cure is minimal. Therefore, we have conducted a broad search for new serum biomarkers with our aptamer-based proteomic platform and defined a classifier for the detection of mesothelioma in asbestos exposed individuals. Secreted proteins and those released during apoptosis from tumor cells and surrounding tissues contain important biologic information that may enable early diagnosis and prognostic and therapeutic decisions in oncology. However, there is great difficulty in finding and quantifying such signals for large numbers of low abundance proteins. We therefore created a highly multiplexed proteomic assay that currently measures ∼850 proteins simultaneously from 15ul blood, with throughput of 300 samples/day. The average dynamic range of each protein in the assay is >3 logs – with nearly seven logs of dynamic range achieved through multiple dilutions – and the median lower limit of quantification is below 1 pM. The median coefficient of variation for each protein is <5%. This assay performance arises from the selection of high affinity aptamers that bind selectively to their target proteins with slow off-rates. The objective of this study was to discover proteins which are involved in malignant mesothelioma and to develop algorithms and classifiers for detection of the disease. To this end, blood samples from three study centers were analyzed with the aptamer proteomics platform in a prospectively designed case:control study. We compared 170 serum samples from 90 patients diagnosed with malignant mesothelioma to 80 asbestos exposed controls. These samples were divided into 75% for training and 25% set aside as a blinded test set for classifier development and verification. Nineteen significant biomarkers were discovered by applying a backwards selection strategy. Classifiers were built with subsets of these biomarkers resulting in an AUC of 0.95 or better with an overall accuracy of 93%. Applying a 13-plex Random Forest classifier to the blinded test set resulted in a specificity of 100% and sensitivity of 80% for distinction of asbestos exposed controls from mesothelioma, including detection of 15/19 Stage I/II cases. Refinement and confirmation of classifier performance will be established through ongoing validation studies. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 2812. doi:10.1158/1538-7445.AM2011-2812