Development of a PANoptosis-Related Pathomics Prognostic Model in Ovarian Cancer: A Multi-Omics Study.

Ovarian cancer (OC) is a high-mortality gynaecological malignancy, and the role of PANoptosis, a comprehensive cell death mechanism, in its prognosis remains unexplored. This study aims to clarify it, potentially guiding OC diagnosis and treatment. We analysed the ovarian data from TCGA and GTEx, and the GSE184880 scRNA-seq dataset from GEO. Spatial data and pathological images were sourced from the 10X Genomics website and GDC Portal. Features were extracted using CellProfiler and ResNet-50, and a PANoptosis-related pathomics prognostic model (PANPM) powered by deep learning was developed. The PANoptosis-related hub gene STAT4 potentially served as a protective factor for patients with OC. A better prognosis in OC was found linked to higher PANoptosis. The PANPM, manifesting distinct advantages for clinical application by accurately extracting pathological features, performed excellently in validation and the high-risk group indicated a poor prognosis. Additionally, STAT4+ T cells may inhibit OC, by activating the PANoptosis of epithelial cells through TNFSF12-TNFRSF12A and TNF-TNFRSF1A, which sheds light on potential therapeutic interventions involving STAT4+ T cells.