ABSTRACT
Background: Tumor genetic anomalies and immune dysregulation are pivotal in the progression of multiple myeloma (MM). Accurate patient stratification is essential for effective MM management, yet current models fail to comprehensively incorporate both molecular and immune profiles. Methods: We examined 776 samples from the MMRF CoMMpass database, employing univariate regression with LASSO and CIBERSORT algorithms to identify 15 p53-related genes and six immune cells with prognostic significance in MM. A p53-TIC (tumor-infiltrating immune cells) classifier was constructed by calculating scores using the bootstrap-multicox method, which was further validated externally (GSE136337) and through ten-fold internal cross-validation for its predictive reliability and robustness. Results: The p53-TIC classifier demonstrated excellent performance in predicting the prognosis in MM. Specifically, patients in the p53low/TIChigh subgroup had the most favorable prognosis and the lowest tumor mutational burden (TMB). Conversely, those in the p53high/TIClow subgroup, with the least favorable prognosis and the highest TMB, were predicted to have the best anti-PD1 and anti-CTLA4 response rate (40 %), which can be explained by their higher expression of PD1 and CTLA4. The three-year area under the curve (AUC) was 0.80 in the total sample. Conclusions: Our study highlights the potential of an integrated analysis of p53-associated genes and TIC in predicting prognosis and aiding clinical decision-making in MM patients. This finding underscores the significance of comprehending the intricate interplay between genetic abnormalities and immune dysfunction in MM. Further research into this area may lead to the development of more effective treatment strategies.
