Article Text
Abstract
Objectives Although studies including TCGA have selected for pure tumors to enhance detection of cancer-related biomarkers, many impure tumors are associated with poor prognosis raising concerns over historical selection bias. Enrichment techniques to prep tumor micro-compartments was aligned with comprehensive proteogenomic analysis in an advanced stage high grade serous ovarian cancer (HGSOC) patient cohort to detect novel, clinically-relevant alterations.
Methods 87 fresh-frozen tumor specimens were selected from a cohort of over 630 patients to reflect a continuum of tumor purity balanced by progression and disease distribution. A whole tumor (WT) specimen and one enriched for tumor epithelium was prepared for each case using laser microdissection (LMD). Specimens were analyzed by whole genome sequencing (WGS), mRNA-seq, quantitative global/phosphoproteomics, and reverse phase protein array.
Results LMD enrichment increased median tumor purity estimated by WGS from 56% in WT to 79% (P<4e-11, MWW U test) and significantly enhanced identification of somatic single nucleotide variants (SNVs) (27%, P<3e-7) and short indels (16%, P<4e-4). Following LMD, 83% of cases characterized as mesenchymal expression subtype (C4) in WT samples were reclassified to other molecular subtypes (P<0.001). LMD tumors with an immune expression subtype (C1) had improved progression-free survival (PFS) compared with other molecular subtypes (p=0.009). Differential proteomic analyses focused on signaling alterations correlating with altered PFS, homologous recombination deficiency and immune signaling.
Conclusions These data demonstrate feasibility of cohort-level proteogenomic characterization of the tumor microenvironment and establishes a non-restrictive paradigm for patient inclusion and specimen prep in support of the prospective mission-scale analyses associated with APOLLO.