Nomogram incorporating Epstein-Barr virus DNA and a novel immune-nutritional marker for survival prediction in nasopharyngeal carcinoma

Nasopharyngeal carcinoma (NPC), a distinctive head and neck cancer arising from nasopharyngeal epithelium, has a high reported incidence of 5.0/100,000 in Southeast Asia [1]. Despite survival rates have increased with the advance of multidisciplinary management and treatment, approximately 20% of patients still experience disease progression and mortality [2‐4]. Therefore, to improve the treatment effects, identifying key predictors to select patients at different risk levels is required for individualized treatment.

The dynamic interaction between immune response and tumor microenvironment has become a popular focus of attention. Increasing evidence shows that the immune response of a patient contributes to cancer development and progression [5‐7]. Lymphocytes, a crucial type of immune cells, play an essential role in immune monitoring by inhibiting the proliferation and metastasis of cancer cells via lymphocyte-mediated cytotoxicity [8]. And its level can reveal prognosis of cancer. Okadome [9] et al. found that the low tumor-infiltrating lymphocytes (TILs) were associated with poor overall survival (OS) in esophageal carcinoma. In addition, nutritional status is linked to therapeutic response, treatment toxicity and prognosis of various cancers [10, 11]. Serum albumin (ALB), a classical biomarker of the body’s nutritional status, can reliably assess the nutritional status of cancer patients. Decreased albumin level is a sign of poor prognosis for cancer [12‐15]. Li et al. suggested serum albumin as an effective prognostic biomarker of NPC patients [16]. Using a combination of these factors of immune and nutritional status, a novel indicator, LA (Lymphocytes × albumin), caught our interest. The LA, calculated by the product of lymphocytes and albumin, was first proposed in distinguish benign and malignant pancreatic cystic neoplasm [17]. Recently, it had been demonstrated that low LA was particularly related to poor survival outcomes in rectal cancer [18]. However, the prognostic value of LA in NPC is still unclear, and using LA alone is still insufficient for making individualized predictions.

As a specific pathogenic factor of NPC, the contribution of Epstein-Barr virus (EBV) DNA cannot be ignored [19]. The detection of EBV DNA level in plasma is extensively used in population screening, prognosis evaluation and risk stratification [20‐22]. In plasma samples, the sensitivity and specificity of EBV DNA in screening for nasopharyngeal carcinoma were 97.1% and 98.6%, respectively [21]. High plasma EBV DNA concentration was significantly correlated with advanced tumor stage and worse prognosis [20, 23]. Many scholars further revealed that EBV DNA, in combination with other indicators, strengthened the predictive efficacy [24‐26]. For example, Jin [25] combined EBV DNA with the systemic inflammation response index (SIRI) and found that it achieved the largest area under the curve (AUC) to predict survival. And in Huang’s study, the predictive capacity of EBV DNA combined with the C-reactive protein/albumin ratio (CAR) has a higher C-index of 0.693, which was superior to EBV DNA or CAR alone [26]. Therefore, we tried to combine LA and EBV DNA to improve the predictive ability.

Nomogram, a more precise tool for individual prediction of a clinical event than traditional TNM staging system, is commonly used for risk estimation in various types of cancers [27]. In the nomogram developed by Wang et al., we can intuitively and accurately predict the survival probability for individual patients with small-cell lung cancer with a higher C-index of 0.722 [28]. Hence, in the present study, we initially attempted to show the prognostic significance of LA in NPC and then investigated the effect of its combination with EBV DNA on improving prediction. Finally, we established nomograms to further enhance the predictive precision for individual patients in NPC.

Owing to tumor heterogeneity, personalized treatment has become an important developing strategy for cancer treatment. Precise prognosis estimation helps clinicians tailor treatment options to improve survival outcomes. Hence, further refined risk stratification of patients with cancer is indispensable for individual treatment. Our study confirmed that LA was a strong independent factor for predicting clinical outcomes. The combination of LA and EBV DNA provided more detailed information for patient stratification. Moreover, we constructed comprehensive prognostic nomograms incorporating LA and EBV DNA for PFS and OS prediction. The C-index, calibration curve, and DCA curve demonstrated that this model achieved excellent predictive efficiency and clinical benefits.

To the best of our knowledge, this is the first study to focus on the value of LA in NPC. We noticed that a low LA was obviously related to poorer 5-year PFS and OS in both the training and validation cohorts. This suggests that patients with low lymphocyte and low albumin levels had worse survival outcomes, which is consistent with previous studies [16, 31‐33]. Multivariate analysis also showed that low LA levels were an independent risk factor for NPC. Although the underlying mechanism between LA and poor cancer prognosis is unclear, a possible explanation can be proposed.

The interactions between immune environment and cancer cells are dynamic and complex, and the immune biomarkers can reveal the prognosis of cancer patients [7]. Lymphocytes are one of the critical cells in the immune response. When tumor growth, its tumor antigens can be recognized by lymphocytes, then the secretion of cytokines activate different kinds of immune cells to exhibit cancer inhibitory effects such as CD4 + and CD8 + T cells, B cells and so on [34‐37]. In this regard, a low peripheral lymphocyte level may reflect a poor lymphocyte-mediated antitumor immune reaction and worse immune surveillance, indicate a poor prognosis. Previous study also showed that circulating lymphocytes can improve cancer patient prognosis by enhancing cancer immune regulation and inhibiting cancer cell proliferation [38].

Moreover, proinflammatory cytokines can not only promote tumor invasion but also inhibit the synthesis of albumin and lead to its leakage by increasing microvascular permeability [10, 39, 40]. Thus, as the disease progresses, albumin levels decrease significantly and lead to malnutrition. Since the key roles of nutrition in determining the fate and functions of immune cells, poor nutritional status could induce an impaired immune response, which promotes cancer progression and causes a worse survival outcomes [41]. Given these findings, a low LA level, multiplied by a low lymphocyte count and low albumin level, might reflect the status of tumor progression, poor immune response and malnutrition at the same time, with a higher predictive accuracy. Therefore, patients with low LA require more aggressive treatment regimens.

Unlike LA, EBV DNA has been recognized as an important indicator of the prognosis of NPC. Studies indicated that EBV DNA was originated from NPC tumor cells, is a fragment of tumor cells necrosis and lysis, and its plasma load can reflect tumor load [42‐44]. Similar to our results, the subgroups analyses of TNM stage revealed that positive EBV DNA was obviously with worse PFS and OS in the III-IVa TNM stage patients. To strengthen the predictive value, Xiong [29] et al. combined systemic immune-inflammation index (SII) and EBV DNA. The study revealed that patients with a high SII and positive EBV DNA had a higher risk of death and disease progression. These results are in line with our research. We combined EBV DNA and LA, and found that the C-index of EBV DNA in combination with LA for PFS and OS was 0.67 and 0.68, respectively, which was higher than that of EBV DNA (PFS: 0.63, OS: 0.60), and even higher than the lactate dehydrogenase/ albumin ratio (LAR) in the research by Zhu et al. [30] and SII in the study by Xiong et al. [29]. Moreover, Kaplan–Meier curves showed that compared with patients with high LA and negative EBV DNA, patients with low LA and positive EBV DNA were associated with a higher risk of tumor progression. This suggests that EBV DNA combined with LA has a superior prognostic value than EBV DNA alone. This integrated biomarker indeed improves the accuracy of outcome prediction. The above results strongly indicate that EBV DNA combined with LA is more effective for further risk stratification in NPC patients. They should be given much more attention, and more chemotherapy cycle, nutrition improvement, targeted therapy, and even immunotherapy need to be considered early to improve the clinical outcomes. Then, we developed nomograms based on LA and EBV DNA and other clinical characteristics.

Recently, nomograms, with their simple, intuitive, and individual characteristics, have been widely used in cancer prognosis prediction. In studies conducted by Tang et al. [45] and Zhang et al. [46], nomograms were a convenient and reliable tool to estimate the survival of NPC patients, and they provided excellent discrimination capacity compared to the current TNM staging system. Consistent with our results, we found that the C-indexes (PFS: 0.69, OS: 0.77) of the nomograms outperformed the TNM staging system (PFS: 0.58, OS: 0.6), which meant that the nomograms were superior in risk stratification. Furthermore, the DCA analysis results further supported that it was significantly better than the TNM staging system in clinical application. Through nomograms, clinicians could more accurately and efficiently identify patients with a high risk of poor clinical outcomes and make the personalized treatment plans. For high-risk patients, more aggressive therapy may be considered to improve their survival.

However, there are some limitations to our study. First, this was a retrospective study. Although we conducted an internal validation to make the results more convincing, it is still far from practical clinical application. Second, in addition to LA, many other indicators such as LDH, CAR, SII and SIRI were reportedly associated with the prognosis of NPC patients. On account of the limited sample size, these relations have not been fully explored. Third, it was a single-center study with the absence of external validation. Hence, a large-scale and multi-center prospective study is required to further validate the value of LA.

In conclusion, our study reported that LA was a simple and easily available biomarker, and when combined with EBV DNA, it was a stronger predictor of PFS and OS in NPC. The nomogram based on LA and EBV DNA is more effective and precise in predicting individual survival outcomes for NPC patients. It can serve as a good supplement to the TNM staging system to improve the identification of high risk of disease progression in NPC patients and guide more aggressive treatments for these patients to prolong their survival.