Nasopharyngeal carcinoma (NPC) has a 10–15% recurrence rate, while no long term or durable treatment options are currently available. Single-cell profiling in recurrent NPC (rNPC) may aid in designing effective anticancer therapies, including immunotherapies. For the first time, we profiled the transcriptomes of ∼60,000 cells from four primary NPC and two rNPC cases to provide deeper insights into the dynamic changes in rNPC within radiation fields. Heterogeneity of both immune cells (T, natural killer, B, and myeloid cells) and tumor cells was characterized. Recurrent samples showed increased infiltration of regulatory T cells in a highly immunosuppressive state and CD8 + T cells in a highly cytotoxic and dysfunctional state. Enrichment of M2-polarized macrophages and LAMP3 + dendritic cells conferred enhanced immune suppression to rNPC. Furthermore, malignant cells showed enhanced immune-related features, such as antigen presentation. Elevated regulatory T cell levels were associated with a worse prognosis, with certain receptor-ligand communication pairs identified in rNPC. Even with relatively limited samples, our study provides important clues to complement the exploitation of rNPC immune environment and will help advance targeted immunotherapy of rNPC.

Background Postoperative radiotherapy (PORT) is beneficial in the improvement of local‐regional control and overall survival (OS) for major salivary gland carcinomas (SGCs), and distant metastasis remained the main failure pattern. This study was designed to develop a nomogram model involving immune‐inflammation index to predict distant metastasis‐free survival (DMFS) of major SGCs. Patients and Methods A total of 418 patients with major SGCs following PORT were randomly divided into a training ( n = 334) and validation set ( n = 84). The pre‐radiotherapy neutrophil‐to‐lymphocyte ratio (NLR), and platelet‐to‐lymphocyte ratio (PLR) were calculated and transformed as continuous variables for every patient. Associations between DMFS and variables were performed by univariate and multivariable analysis using Log‐rank and Cox regression methods. A nomogram was constructed based on the prognostic factors identified by the Cox hazards model. The decision curve analysis (DCA) was conducted with the training and validation set. Results The estimated 3‐, 5‐, and 10‐year DMFS were 79.4%, 71.8%, and 59.1%, respectively. The multivariate analysis revealed that age ( p = 0.033), advanced T stage ( p = 0.003), positive N stage ( p < 0.001), high‐risk pathology ( p = 0.011), and high PLR ( p = 0.001) were significantly associated with worse DMFS. The nomogram showed good calibration and discrimination in the training (AUC = 80.9) and validation set (AUC = 87.9). Furthermore, the DCA demonstrated favorable applicability, and a significant difference ( p < 0.001) was observed for the DMFS between the subgroups based on the nomogram points. Conclusion The nomogram incorporating clinicopathological features and PLR presented accurate individual prediction for DMFS of the patients with major SGCs following PORT. Further external validation of the model is warranted for clinical utility.

Background To explore the prognostic role of FDG PET/CT maximal standard uptake values of metastatic lesions (SUVmax-M) in patients with de novo metastatic nasopharyngeal carcinoma (mNPC) following palliative chemotherapy and locoregional radiotherapy (LRRT). Methods We retrospectively collected the information of 86 eligible patients between Jan 2012 and Oct 2020. All the parameters involving SUVmax and serum lactate dehydrogenase (LDH) at diagnosis were evaluated and cutoff values were determined by the maximum log-rank statistic method. The multivariate analysis was performed using Cox proportional hazards regression to identify the independent prognostic factors associated with overall survival (OS). All estimated survival rates were conducted with Kaplan–Meier method. Results Median survival and progression time in the cohort were 38.2 and 13.9 months, respectively. The univariable analysis showed that male, number of metastatic sites ≥ 4, presence of liver, serum LDH ≥ 229, SUVmax-M ≥ 10, SUVmax-M-sum ≥ 10, and SUVmax-M-mean ≥ 8.8 were significant prognostic factors. Five variables were identified after LASSO regression and entered into the multivariate analysis. Furthermore, liver involvement (P = 0.039), elevated LDH (≥ 229) (P = 0.05) and higher SUVmax-M (≥ 10) (P = 0.004) were significantly associated with worse OS. Conclusion The high SUVmax of metastatic lesions (≥ 10), liver involvement, and elevated serum LDH (≥ 229) at diagnosis could independently predict poor survival for de novo mNPC patients treated with palliative chemotherapy following LRRT.