Multimodal profiling of STIC lesions to identify precursor states with genomic features of high grade serous ovarian cancer.

Abstract

10581 Background: High-grade serous ovarian cancer (HGSC) is a lethal gynecologic malignancy characterized by near-universal TP53 inactivation, chromosomal instability (CIN), and frequent whole-genome doubling. Because population screening is ineffective, risk-reducing salpingo-oophorectomy (RRSO) remains the only proven risk-reduction strategy. Nonetheless, some individuals develop primary peritoneal carcinoma (PPC), and identification of serous tubal intraepithelial carcinoma (STIC) at RRSO is associated with > 30-fold higher PPC risk. We sought to define the molecular steps required for HGSC initiation and identify features of high-risk STIC. Methods: We assembled a 201-patient cohort spanning stages of HGSC development, including normal fallopian tube epithelium, p53 signatures, STIC, and early- and advanced-stage HGSC. All specimens underwent multiplex immunofluorescence (mpIF). Forty-five samples from 36 patients were profiled using Visium HD spatial transcriptomics ( > 20 million 8-µm bins). We analyzed epithelial state transitions using unsupervised clustering and trajectory inference. Copy-number variation (CNV) was inferred using spatial inferCNV; CNV calls were validated with matched formalin-fixed paraffin-embedded (FFPE) single-cell whole-genome sequencing. Results: mpIF demonstrated lesion-to-lesion variability in cyclic GMP–AMP synthase (cGAS) staining within STIC, consistent with evolving tolerance to CIN before invasion. Spatial transcriptomic clustering identified distinct, reproducible STIC epithelial programs that mapped to histologic features. STIC programs were enriched for NFκB and TGF-β signaling and lacked evidence of the JAK/STAT activation observed in advanced HGSC. CNV inference showed that normal fallopian tube and p53 signature lesions were chromosomally stable, whereas STIC displayed broad, genome-wide alterations resembling advanced HGSC, including losses encompassing TP53 , BRCA1 , and BRCA2 and amplifications of MYC and CCNE1 . In a germline BRCA2 carrier, we identified polyclonal STIC with distinct CNV profiles; trajectory analysis suggested early CNV changes within an expanded secretory population, followed by loss of chromosome 17 ( TP53 ), loss of chromosome 13 ( BRCA2 ), and subsequent oncogene amplifications (e.g., MYC ). As lesions became invasive, spatial CNV profiling revealed neighborhoods of clonally related invasive cells with distinct microenvironments, including differences in fibroblast states and vascularization. Conclusions: A multimodal atlas of fallopian tube transformation supports a model in which TP53 loss precedes BRCA1/2 inactivation and indicates that STIC is genomically similar to advanced HGSC. These findings provide a framework to study early HGSC evolution, stratify STIC risk after RRSO, and inform early-interception strategies.