Development of a centrosome amplification-associated signature in kidney renal clear cell carcinoma based on multiple machine learning models.

Abstract

Background: Centrosome amplification (CA) has been shown to be capable of initiating tumorigenesis with metastatic potential and enhancing cell invasion. We were interested in discovering how centrosome amplification-associated signature affects the prediction of prognosis and response to therapy in kidney renal clear cell carcinoma (KIRC).

Methods and materials: The TCGA-KIRC dataset was used to construct a centrosome amplification-associated signature using the random survival forest analysis and Cox regression analysis, and the ICGC and GEO datasets were employed for signature validation. Mutation and immune landscapes were outlined and the response to immunotherapy was evaluated. The expression of the screened hub gene was profiled by analyzing single-cell RNA sequencing from GSE159115.

Results: In the TCGA-KIRC cohort, 22 centrosome amplification-associated prognostic genes were discovered. According to the optimal consistency index (0.91), the random survival forest algorithm was selected to determine 7 hub prognostic genes, which were used to construct a centrosome amplification-associated prognostic index (CAAPI). It was discovered that it is connected to high mortality rates, high mutation rates, immunosuppressive cell infiltration, and immune dysfunction. For patients in the high CAAPI group, immunotherapy was not as effective. Single-cell RNA sequencing revealed a high expression of CDK5RAP3 in the tumor cells.

Conclusion: Centrosome amplification played a significant role in regulating tumor microenvironment and responding to immunotherapy, emphasizing its crucial importance in the development and treatment of KIRC. Patients with KIRC may benefit from using CAAPI as a biomarker to predict individual prognosis and assess a response to immunotherapy.