Open Access

ARTICLE

Integrating Spatial Multi-Omics and Machine Learning to Unravel the Role of PANoptosis in Bladder Cancer Prognosis and Immunotherapy Response

Liangju Peng^1,2, Tingting Cai^1,2, Peihang Xu^1,2, Cong Chen³, Qingzhi Xiang^1,2, Yiping Zhu^1,2, Dingwei Ye^1,2,*, Yijun Shen^1,2,*

1 Department of Urology, Fudan University Shanghai Cancer Center, Shanghai, 200032, China
2 Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, 200032, China
3 Department of Nursing, Fudan University Shanghai Cancer Center, Shanghai, 201321, China

* Corresponding Authors: Dingwei Ye. Email: ; Yijun Shen. Email:

Oncology Research 2025, 33(9), 2463-2489. https://doi.org/10.32604/or.2025.064331

Received 12 February 2025; Accepted 20 May 2025; Issue published 28 August 2025

Abstract

Background: Studies have reported the special value of PANoptosis in cancer, but there is no study on the prognostic and therapeutic effects of PANoptosis in bladder cancer (BLCA). This study aimed to explore the role of PANoptosis in BLCA heterogeneity and its impact on clinical outcomes and immunotherapy response while establishing a robust prognostic model based on PANoptosis-related features. Methods: Gene expression profiles and clinical data were collected from public databases. Spatial heterogeneity of cell death pathways in BLCA was evaluated. Consensus clustering was performed based on identified PANoptosis genes. Cell death pathway scores, molecular, and pathway activation differences between different groups were compared. Protein-protein interaction (PPI) network construction was constructed, and immune-related gene sets, tumor immune dysfunction and exclusion (TIDE) scores, and SubMap analysis were used to evaluate immunomodulator expression and immunotherapy efficacy. Ten machine learning algorithms were utilized to develop the most accurate predictive risk model, and a nomogram was created for clinical application. Results: BLCA demonstrated a spatially heterogeneous distribution of pyroptosis, apoptosis, and necroptosis. Notably, T effector cells significantly colocalized with total apoptosis. Two PANoptosis modes were identified: high PANoptosis (high. PANO) and low PANoptosis (low. PANO). High. PANO was associated with worse clinical outcomes and advanced tumor stage, and increased activation of immune-related and cell death pathways. It also showed increased infiltration of immune cells, elevated expression of immunomodulatory factors, and enhanced responsiveness to the immunotherapy. The PANoptosis-related machine learning prognostic signature (PMLS) exhibited strong predictive power for outcomes in BLCA. CSPG4 was identified as a key gene underlying prognostic and therapeutic differences. Conclusion: PANoptosis shapes distinct prognostic and immunological phenotypes in BLCA. PMLS offers a reliable prognostic tool. CSPG4 may represent a potential therapeutic target in PANoptosis-driven BLCA.

---

Keywords

---