Nasopharyngeal carcinoma (NPC) is an epithelial carcinoma originating from the inner lining of the nasopharyngeal mucosa. It is highly aggressive and most commonly occurs in the pharyngeal recess of the nasopharynx (1). Based on the International Agency for Research on Cancer (IARC), there were about 133 thousand new cases of NPC, representing 0.7 percent of the amount of all cancer cases diagnosed in 2020 (2). NPC has unique geographical distribution characteristics; 70% of cases are diagnosed in East and Southeast Asia, especially in southern China. It is recognized that the epidemic of NPC is connected with the infection of Epstein-Barr virus (EBV) (3–5), and other risk factors include dietary habits, environmental influences (6), and genetic susceptibility (7). Recently, Luo proposed the refreshingly NPC ecology theory, which posits cancer as a multidimensional spatiotemporal “unity of ecology and evolution “pathological ecosystem. The cancer ecological theory can help us to open a new perspective and better understand the complex process of NPC development and metastasis (8). In the era of intensity-modulated radiotherapy (IMRT), the therapeutic effect of NPC patients with low clinical stage has been significantly improved (9–11). However, due to the early clinical symptoms being vague and difficult to distinguish and the invasion and metastasis of tumor cells being strong, some patients are already in the advanced stage when they are diagnosed. About 4-15% of patients already had metastases when they were first diagnosed with NPC (12), and the survival of these patients is often highly variable and unsatisfactory. Currently, TNM staging is used to predict the survival of NPC. However, patients with distant metastases were classified as stage M1 (13).The same stage may have different survival outcomes. TNM staging has limited predictive value, and more effective and accurate predictive tools may be needed.

NPC is prone to lung metastasis (14). Some studies have shown that NPC with metastasis limited to the lung may have better overall survival (OS) than those with other sites of metastasis (15, 16). However, this conclusion has not been confirmed by large-scale studies, and the majority of patients with lung metastasis also have metastasis to other sites. In 2011, Cao et al. (17) developed risk subsets containing some clinical variables to predict the survival rate of lung metastasis of NPC patients. However, the results need to be visualized more, and clinical detection methods were limited at that time. Some patients with small metastases of lung may be missed, which may bias the study results. Therefore, the survival prognosis of these patients needs further study.

Surveillance, Epidemiology, and End Results (SEER) is a public database and research resource created by the National Cancer Institute (18). This database collects and stores data on cancer incidence, survival, and therapies across America to support research and epidemiological investigations. In this study, a prognostic model of NPC with lung metastasis was established through the data of the SEER database, and the model was externally verified by patients from the Affiliated Tumor Hospital of Guangxi Medical University to help doctors predict the survival of these patients individually and provide guidance for decision-making.

The treatment of distant metastasis patients is essential to improve the efficacy of NPC, which substantially affects the NPC patients ‘survival rate (19). The TNM staging system can only forecast the survival of some patients. It has been shown that patients in the same TNM stage could have distinct outcomes after getting a similar treatment, while patients with metastasis are mainly at the same stage. Therefore, developing an individualized and accurate model for prognosis judgment and therapy is highly desirable. Studies have shown that patients with different metastatic sites and states may have different prognoses. The median OS in patients with single-organ metastatic NPC was better than in patients with extensive metastasis (24.8 months vs 12.8 months, P < 0.001) (20). Huang et al. showed that the prognosis of NPC patients with brain metastases was poor, and the 3-year OS rate was 16.3% (21). Qu et al. proposed that liver metastases were associated with poor cancer-specific survival (CSS) in patients (22). Hui et al. found that nasopharyngeal carcinoma patients with distant metastatic limited to the lung had good OS compared with patients with metastasis from other sites (16). Therefore, we focused on lung metastasis of NPC. This study is the first nomogram for NPC with lung metastasis. Nowadays, nomograms are widely used. Using Cox regression analysis, Wu et al. developed a nomogram for predicting OS in patients with low-grade endometrial stromal sarcoma (23). Zheng et al. developed a nomogram for predicting OS in lung cancer bone metastases patients (24). Nomogram can visualize the results of our study and has good practicability (25, 26).This study indicated that being older was related to worse OS in patients, the same as the conclusions of many researchers. Lin et al. (27) suggested age was linked to poor OS in locoregionally advanced NPC patients. Li et al. (28) also pointed out that NPC patients’ CSS gradually decreases with age, and there is a gender difference, which peaks at 55-60 years of age. This conclusion has also been confirmed in other cancer, and Zheng et al. (24) confirmed that increasing age is related to worse OS in lung cancer bone metastasis. Cancer is considered an age-related disease, and it is generally thought that people over the age of 65 are more likely to have cancer (29). At the same time, due to aging, physiological reserves are reduced, which may lead some patients unable to tolerate anti-tumor therapy. For example, reduced clearance of some cytotoxic chemotherapeutic drugs that need to be excreted by the kidneys will make it difficult for elderly patients to undergo anti-tumor therapy, which will also shorten their OS (30). Mashiro et al. found that chemotherapy-induced neutropenia (CIN) has been related to age > 65 (31). Therefore, additional attention may need to be paid to older NPC patients, and further research in the future may focus on specific treatment options for such patients to improve their prognosis.

Brain metastasis of NPC is not common. It is still in the sporadic reporting stage, and the related metastasis pathway has not been fully clarified. Tumor cells enter the Cerebro-Spinal Fluid (CSF) by corroding the skull and destroying the dural membrane, which may be one of the ways of brain metastasis of NPC (32). Brain metastases in tumors are generally regarded as a marker of poor prognosis (33), and in our nomogram, it is also associated with poor OS in NPC patients. Unfortunately, there is currently no consensus on diagnosing brain metastasis in NPC. In some cases, individual patients have been reported to have achieved a good survival prognosis through chemoradiotherapy (34) (35). Although brain metastases from NPC are rare, they cannot be ignored, and these patients may need more mature treatment options to help them survive for a long time.

We found that NPC patients with lung metastases who received chemotherapy and radiotherapy had better survival than those who did not. Currently, according to the results of several multi-center clinical trials, including the CAPTAIN 1ST, JUPITER-02, and RATIONALE 309, PD-1 inhibitor combined with chemotherapy has become the first-line standard treatment for recurrent/metastatic nasopharyngeal cancer (R/M NPC) (36–38). However, these patients have lost the opportunity for radical radiotherapy because of distant metastases. You et al. initiated a phase 3 Randomized Clinical Trial, which indicated that the 2-year OS was longer in the chemotherapy combined with radiotherapy group than in the chemotherapy alone group (54.5% vs. 76.4%) (HR=0.42; 95%CI: 0.23-0.77; p = 0.004) (39). This finding has also been seen in retrospective studies. Liu et al. (40) found that 1-year progression-free survival (80.6% vs. 65.1%, P<0.001) and OS (98.3% vs. 89.5%, P=0.001) were significantly better in patients with immunochemotherapy and radiotherapy than with metastatic who received immunotherapy and chemotherapy. The coalescence of chemotherapy and radiotherapy has a significant synergistic effect. In the cell cycle, the action point of radiotherapy to kill cancer cells is the late G2, M, and G1, and does not affect the S stage. In contrast, the main action point of some chemotherapy drugs, such as gemcitabine, is the S stage so that radiotherapy can play a complementary killing role on chemotherapy-resistant cancer cells (41, 42). In addition, radiation therapy can also produce a “distant effect”, that is the phenomenon of regression of metastatic tumors far from the irradiation field after local irradiation (43). The distant effect may be related to immune activation, such as increased PD⁃L1 expression, lymphocytopenia, and induced aggregation of immunosuppressive cells (44). Unfortunately, further analysis was not possible due to the lack of information on patient immunotherapy.

Overall, our model is successful, as it is composed of 5 common factors, including T-stage, radiation, chemotherapy, age, and brain metastases, and achieves good survival prediction in NPC patients with lung metastases compared to traditional TNM staging. Although the training cohort was based on patients from America, our results show that the model is also suitable for patients in China. However, there is no denying that this study has some shortcomings. At first, more detailed baseline data of patients, such as EBV DNA copy number, chemoradiotherapy regimen, immunotherapy, comorbidities, and Karnofsky performance status (KPS) score of patient status, could not be obtained, so further detailed analysis could not be carried out. Second, the age in the two cohorts has a statistical difference, in which patients in the training cohort were more concentrated in age. However, this may not have obviously biased the results. Finally, with a fraction of patients in the validation cohort diagnosed in 2021 and a follow-up cutoff of less than three years, predictions of three-year long-term survival have yet to have much chance to prove their value.

In this study, a survival and prognosis model for patients with NPC with lung metastasis was established, which can be used as a supplement to TNM staging, help clinicians analyze the prognosis of these patients individually, and support a simple tool for treatment decision-making.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material.

The studies involving humans were approved by Affiliated Tumor Hospital of Guangxi Medical University. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because This is a retrospective study. Some of the data comes from public databases, and these data are public resources. The other part of the data is the collection or study of previously archived data, documents, materials, and information that is recorded in a way that the investigator cannot contact the patient.

ZX: Data curation, Formal analysis, Methodology, Software, Writing – original draft. KL: Data curation, Formal analysis, Writing – original draft. FS: Data curation, Formal analysis, Writing – original draft. XY: Data curation, Formal analysis, Writing – original draft. HZ: Data curation, Formal analysis, Writing – original draft. SQ: Writing – review & editing.