Profile: Results from Multiplexed Mass Spectrometric Genotyping of 2178 Cancer Patients
Neal I Lindeman, Laura E MacConaill, Elizabeth Garcia, Frank C Kuo, Janina A Longtine, William C Hahn, Philip W Kantoff, Barrett J Rollins. Harvard Medical School, Boston, MA; Brigham and Women's Hospital, Boston, MA; Mt. Sinai Hospital, New York City, NY; Dana-Farber Cancer Institute, Boston, MA
Background: The Profile initiative is a joint venture between the Brigham and Women's Hospital (BWH) and the Dana-Farber Cancer Institute (DFCI) to perform high throughput molecular analysis of samples from all new cancer patients. The initial assay was designed to detect sequence alterations in known oncogenes and tumor suppressor genes, to identify patients suitable for clinical trials and to foment new biological investigations.
Design: Test requisitions were submitted for all new patients who provided informed consent. Tested samples included fresh and formalin-fixed paraffin-embedded tissues, marrow, and blood. Analysis was not performed on samples without invasive cancer, <50% malignant cells, <5 mm in greatest dimension, or samples that were decalcified. 471 mutations in 41 genes were assessed by a PCR-single nucleotide extension-mass spectrometry assay (Sequenom). Mutations were classified into three interpretive categories: Tiers 1 (standard-of-care), 2 (useful for clinical trials), and 3 (unknown significance).
Results: Over one year, requisitions were received from 8624 patients who consented to testing. Of these, 3309 (38%) had adequate samples within the BWH pathology department; 2258 (27%) of samples were archived at the patients' primary institutions and unavailable for analysis, and 3057 (35%) were insufficient. Tests were completed on 2178 samples, of which 879 (40%) had at least one mutation and 215 (10%) had multiple mutations. 25% of mutations were in either Tier 1 or 2. The highest mutation rates were seen in GI (64%), GYN (59%), and Endocrine (59%) cancers, whie the lowest (∼30%) were seen in GU, Head/Neck and sarcomas. The most frequent genes altered were KRAS (18%), PIK3CA (16%), and TP53 (10%). Of the targetable oncogenes, PIK3CA was seen in the widest range of cancer types. 10% of samples with PIK3CA mutations had another mutation.
Conclusions: Multiplexed genotyping of all new cancer patients' samples in a referral cancer center is feasible. By genotyping 471 mutations in 41 cancer-related genes, at least one mutation was found in ∼40% of samples; of these ∼25% were relevant for either standard therapy or clinical trial agents. Transition to next generation sequencing will expand testing to a greater number of samples and gene targets.
Category: Special Category - Pan-genomic/Pan-proteomic approaches to Cancer
Tuesday, March 5, 2013 1:00 PM
Proffered Papers: Section H1, Tuesday Afternoon