Study of tumor-specific signals in pancreatic cancer is a particularly challenging due to its low cellularity and extensive desmoplastic stroma. To tackle that problem, we used a novel approach utilizing computational methods to separate normal from cancerous tissue and stroma. We analyzed a cohort of 359 patient samples including primary and metastatic tumors and normal tissues and found 2 tumor-specific and 2 stroma-specific subtypes with differential prognostic value across large independent datasets (https://icgc.org and http://cancergenome.nih.gov/). Tumors displaying a ?classical? subtype were less aggressive and patients survived significantly longer (HR 1.93). Retrospective analysis of a small number of patients with the ?basal- like? subtype suggested that they may derive greater benefit from adjuvant therapy after surgery than patients with classical subtype tumors. Our ?basal-like? subtype was consistent with basal subtypes in both external breast and bladder cancer datasets which also show better response to certain therapies. Together these data suggested to us that our tumor and stroma subtypes may require distinct therapeutic regimens. Given the potential relevance of our tumor-specific subtypes to therapy response, we have developed a single sample classifier that is platform agnostic, using a top scoring paired genes approach, with a >92% accuracy for subtype calling on a single sample and have validated our findings in multiple external RNAseq and microarray datasets. There has been great interest in the field such that our subtypes are now integrated markers in two ongoing prospective clinical trials. Preliminary findings from the initial patients enrolled in the COMPASS trial suggest that RNA subtypes may be important for tailoring treatment decisions. Therefore, in this proposal we propose to complete our clinical validation of the RNAseq classifier assay and propose to perform the validation of a Nanostring platform classifier assay, as Nanostring is more widely accessible, with the goal of using our classifier for medical decision making for patients enrolling in clinical trials for pancreatic cancer.
We have developed a single sample classifier that is platform agnostic, using a top scoring paired genes approach, with a >92% accuracy for subtype calling on a single sample and have validated our findings in multiple external RNAseq and microarray datasets. In this proposal we propose to complete our clinical validation of the RNAseq classifier assay and propose to perform the validation of a Nanostring platform classifier assay for medical decision making for patients enrolling in clinical trials for pancreatic cancer.
| Torphy, Robert J; Wang, Zhen; True-Yasaki, Aisha et al. (2018) Stromal Content Is Correlated With Tissue Site, Contrast Retention, and Survival in Pancreatic Adenocarcinoma. JCO Precis Oncol 2018: |
| Aguirre, Andrew J; Nowak, Jonathan A; Camarda, Nicholas D et al. (2018) Real-time Genomic Characterization of Advanced Pancreatic Cancer to Enable Precision Medicine. Cancer Discov 8:1096-1111 |
| Krulikas, Linas J; McDonald, Ian M; Lee, Benjamin et al. (2018) Application of Integrated Drug Screening/Kinome Analysis to Identify Inhibitors of Gemcitabine-Resistant Pancreatic Cancer Cell Growth. SLAS Discov 23:850-861 |
| Aung, Kyaw L; Fischer, Sandra E; Denroche, Robert E et al. (2018) Genomics-Driven Precision Medicine for Advanced Pancreatic Cancer: Early Results from the COMPASS Trial. Clin Cancer Res 24:1344-1354 |
| Cancer Genome Atlas Research Network. Electronic address: andrew_aguirre@dfci.harvard.edu; Cancer Genome Atlas Research Network (2017) Integrated Genomic Characterization of Pancreatic Ductal Adenocarcinoma. Cancer Cell 32:185-203.e13 |
