Breast cancer prognosis and immunological characteristics are predicted using the m6A/m5C/m1A/m7G-related long noncoding RNA signature.

According to statistics, breast cancer (BC) has replaced lung cancer as the most common cancer in the world. Therefore, specific detection markers and therapeutic targets need to be explored as a way to improve the survival rate of BC patients. We first identified m6A/m5C/m1A/m7G-related long noncoding RNAs (MRlncRNAs) and developed a model of 16 MRlncRNAs. Kaplan-Meier survival analysis was applied to assess the prognostic power of the model, while univariate Cox analysis and multivariate Cox analysis were used to assess the prognostic value of the constructed model. Then, we constructed a nomogram to illustrate whether the predicted results were in good agreement with the actual outcomes. We tried to use the model to distinguish the difference in sensitivity to immunotherapy between the two groups and performed some analyses such as immune infiltration analysis, ssGSEA and IC50 prediction. To explore the novel anti-tumor drug response, we reclassified the patients into two clusters. Next, we assessed their response to clinical treatment by the R package pRRophetic, which is determined by the IC50 of each BC patient. We finally identified 11 MRlncRNAs and based on them, a risk model was constructed. In this model, we found good agreement between calibration plots and prognosis prediction. The AUC of ROC curves was 0.751, 0.734, and 0.769 for 1-year, 2-year, and 3-year overall survival (OS), respectively. The results showed that the IC50 was significantly different between the risk groups, suggesting that the risk groups can be used as a guide for systemic treatment. We regrouped patients into two clusters based on 11 MRlncRNAs expression. Next, we conducted immune scores for 2 clusters, which showed that cluster 1 had higher stromal scores, immune scores and higher estimated (microenvironment) scores, demonstrating that TME of cluster 1 was different from cluster 2. The results of this study support that MRlncRNAs can predict tumor prognosis and help differentiate patients with different sensitivities to immunotherapy as a basis for individualized treatment for BC patients.