This application proposes a genomic prognostic study in osteosarcoma, a relatively understudied cancer with high incidence in younger patients, the molecular underpinnings of which are incompletely understood. Outcome is highly variable and there is a lack of well-studied prognostic factors that would allow stratified management decisions. A major challenge is the relative disease rarity and the lack of widely available well annotated frozen tissue resources. We recently published preliminary data analyzing miRNA expression and sequencing profiles from formalin fixed paraffin embedded (FFPE) osteosarcoma tissues generating testable hypotheses with strong relevance to osteosarcoma clinical outcome. In particular, a strong prognostic miRNA profile has been identified which, interestingly, appears to be clustered mainly in one chromosomal location.
In Specific aim 1, we propose to validate the candidate miRNA prognostic profile in an adult cohort, and a pediatric cohort utilizing data from a large Children's Oncology Group based cohort In addition, we will test for independence from, and potential synergy with, relevant clinicopathologic data such as chemotherapy induced necrosis.
In Aim 2 a novel approach will be tested aiming to develop dynamic miRNA expression models capturing the effect of chemotherapy on the tumor, by utilizing paired pre and post chemotherapy specimens from biopsy and resection, respectively. The hypothesis is that they will offer additional powerful prognostic insights specifically relevan to chemoresistant patients.
In Aim 3, deep miRNA sequence patterns will be studied in order to explore their relation to expression profiles and potential prognostic significance.
In Aim 4 miRNA sequencing will be extended to serial specimens from biopsy, chemoresistant resection and metastasis in order to identify evolving sequence patterns signifying the metastatic phenotype which may not be easily detectable via a snapshot test at diagnosis, due to low abundance. These unique patterns may ultimately be relevant to overall survival outcome and to the observation that many patients with recurrent tumors can still survive a long time with further treatment. Our proposal takes advantage of the unique treatment patterns in osteosarcoma, which include serial biopsies and resections following chemotherapy administration in both primary and metastatic disease, resulting in availability of serial patient specimens during routine care. Additionally, patterns revealed in these Aims will also be studied in FFPE specimens in order to provide the basis for definitive large scale clinical application in the futue. This multi institutional and interdisciplinary proposal utilizes cutting edge genomic technologies and brings together a team of experts in osteosarcoma, genomics, and computational biology and biostatistics from all institutions within the Dana/Farber Harvard Cancer Center.
Osteosarcoma is a cancer with a high degree of clinical and biologic complexity that disproportionately affects young adults or children with a very variable clinical outcome. Despite their complexity, therapeutic approaches, with rare exceptions, have not advanced in almost 20 years, to a large degree because there are no reliable tools to predict their outcome. This grant proposal will utilize cutting-edge tools of tumor genomic analysis to allow improved stratification according to prognosis, so that novel targeted and rational therapies can be developed.
|Chen, Hua; Garbutt, Cassandra C; Spentzos, Dimitrios et al. (2017) Expression and Therapeutic Potential of SOX9 in Chordoma. Clin Cancer Res 23:5176-5186|
|Hill, Katherine E; Kelly, Andrew D; Kuijjer, Marieke L et al. (2017) An imprinted non-coding genomic cluster at 14q32 defines clinically relevant molecular subtypes in osteosarcoma across multiple independent datasets. J Hematol Oncol 10:107|
|Osaka, Eiji; Kelly, Andrew D; Spentzos, Dimitrios et al. (2015) MicroRNA-155 expression is independently predictive of outcome in chordoma. Oncotarget 6:9125-39|
|Glass, Kimberly; Quackenbush, John; Spentzos, Dimitrios et al. (2015) A network model for angiogenesis in ovarian cancer. BMC Bioinformatics 16:115|