Immune signature of pheochromocytoma and paraganglioma in context of neuroendocrine neoplasms associated with prognosis

Purpose To understand prognostic immune cell infiltration signatures in neuroendocrine neoplasms (NENs), particularly pheochromocytoma and paraganglioma (PCPG), we analyzed tumor transcriptomic data from The Cancer Genome Atlas (TCGA) and other published tumor transcriptomic data of NENs. Methods We used CIBERSORT to infer immune cell infiltrations from bulk tumor transcriptomic data from PCPGs, in comparison to gastroenteropancreatic neuroendocrine tumors (GEPNETs) and small cell lung carcinomas (SCLCs). PCPG immune signature was validated with NanoString immune panel in an independent cohort. Unsupervised clustering of the immune infiltration scores from CIBERSORT was used to find immune clusters. A prognostic immune score model for PCPGs and the other NENs were calculated as a linear combination of the estimated infiltration of activated CD8(+)/CD4(+) T cells, activated NK cells, and M0 and M2 macrophages. Results In PCPGs, we found five dominant immune clusters, associated with M2 macrophages, monocytes, activated NK cells, M0 macrophages and regulatory T cells, and CD8(+)/CD4(+) T cells respectively. Non-metastatic tumors were associated with activated NK cells and metastatic tumors were associated with M0 macrophages and regulatory T cells. In GEPNETs and SCLCs, M0 macrophages and regulatory T cells were associated with unfavorable outcomes and features, such as metastasis and high-grade tumors. The prognostic immune score model for PCPGs and the NENs could predict non-aggressive and non-metastatic diseases. In PCPGs, the immune score was also an independent predictor of metastasis-free survival in a multivariate Cox regression analysis. Conclusion The transcriptomic immune signature in PCPG correlates with clinical features like metastasis and prognosis.

Automated summary

In this study, we analyzed transcriptomic data from NENs to identify immune cell infiltration signatures, particularly in PCPG. We validated the immune signature of PCPG using an independent cohort. We identified five dominant immune clusters in PCPG, associated with different immune cell types. Furthermore, we developed a prognostic immune score model that could predict non-aggressive and non-metastatic diseases.