Finally, a total of 110 patients with esophageal cancer in this study met the inclusion criteria and were included in this study. The general characteristics of patients with esophageal cancer were showed in Table 1 . A total of 26 women and 84 men were enrolled in this study, and the median age at diagnosis of esophageal cancer was 67.5 years(range 36-79 years). Among them, there were 48 patients with cervical and upper thoracic esophageal cancer, 43 with middle thoracic esophageal cancer, and 19 with lower thoracic esophageal cancer. Among them, there were 2 patients with clinical stage I, 42 patients with stage II, 46 patients with stage III, and 20 patients with stage IV. There were 57 patients who received concurrent chemoradiotherapy and 53 patients who received only radiotherapy. A total of 24 patients had lymphopenia before radiotherapy, accounting for 21.82% of the total patients.The median prescribed dose of PTV was 6000cGy, ranging from (4140-6996)cGy. There were 11 patients with the lowest lymphocyte value with level 1-2 during radiotherapy, accounting for 10% of all the patients, and 110 patients with level 3-4, accounting for 90% of all the patients ( Table 1 ). In all enrolled patients, the incidence of lymphocyte nadir G1, G2, G3 and G4 MinALC during radiotherapy accounted for 0.91%, 9.09%, 62.73% and 27.27%, respectively.

The number of white blood cells in the single radiotherapy group was slightly higher than that in the concurrent chemoradiotherapy group before RT(P=0.049, Table 2 ). During radiotherapy, there was no significant difference in the nadir value of lymphocytes between the two groups, and the white blood cells in the concurrent chemoradiotherapy group were significantly lower than those in the single radiotherapy group (P=0.012, Table 2 ). There was no significant difference in the total number of white blood cells and lymphocytes between the two groups in 1 month after radiotherapy ( Table 2 ).

White blood cells, total lymphocytes, hemoglobin and albumin, and serum prealbumin were significantly decreased during radiotherapy compared with those before radiotherapy ( Table 3 ,P<0.001). All peripheral blood parameters after RT were still significantly lower than those before RT other than albumin and prealbumin. (P<0.001, Table 3 ).

Lymphopenia before RT was defined as ALC<1.1×10⁹/L. Before radiotherapy, whether the peripheral blood lymphocyte count was less than 1.1×10⁹/L was divided into two groups. There was no significant difference between the two groups in KM survival analysis ( Figure 1A , P=0.59). According to the end point of death or not, ROC curve analysis found the cut-off value of the lowest lymphocyte value during radiotherapy (MinALC ≤ 0.41×10⁹/L) and divided all patients into two groups. KM survival analysis was used to compare the survival difference between these two groups. The overall survival of patients in the group (MinALC ≤ 0.41×10⁹/L) was significantly lower than that of the patients in the other group (MinALC>0.43×109/L, Figure 1B , P=0.04).

Univariate COX analysis was performed on each parameter that might be predict survival, and then a multivariate COX proportional hazards survival model was constructed. Univariate COX test showed that the parameter that may predict the overall survival were clinical stage, MinALC(≤0.41 vs >0.41) ( Table 4 , P<0.05). Univariate COX analysis (P<0.1) entered into COX multivariate analysis (forward:wald). In addition to the above 2 factors, MinPA (<200 vs. ≥200mg/L) also entered in multivariate regression analysis. Finally, COX multivariate analysis showed that MinALC may predict the overall survival (see Table 4 , P<0.05).

In order to find whether there is a correlation between the degree of lymphocyte reduction during radiotherapy and nutritional factors, we take MinHb/T1Hb, MinAlb/T1Alb, MinPA/T1PA as the X-axis, and MinALC/T1ALC as the Y-axis.Pearson correlation bivariate statistical analysis was used. The results were shown in Figure 2 , MinHb/T1Hb as the X-axis,P=0.015,MinAlb/T1Alb as the X-axis,P=0.049,MinPA/t1PA as the X-axis,P=0.021.These three nutritional indicators were positively correlated with the decline degree of lymphocytes.

As shown in Figure 3 , taking the occurrence of grade 3-4 radiation esophagitis and radiation pneumonitis as the end points, and the MinALC during radiotherapy as a variable, which is calculated by ROC curve analysis. The results showed that MinALC predicted the occurrence of grade 3-4 radiation pneumonitis (AUC=0.676,P=0.0007, Figure 3A ). However, the prediction of the occurrence of grade 3-4 radiation esophagitis was poor (AUC=0.573,P=0.272, Figure 3B ).

In order to study the correlation between dose and units which related to the degree of lymphopenia, we collected and recorded dosimetric parameters through DVHs. We take MinALC/T1ALC as the Y-axis, and physical factors such as the average dose of PTV, Mus as X-axis, and Pearson bivariate analysis was used to analysis. The degree of MinALC/T1ALC reduction is significantly negatively correlated with the following dose factors(heart max dose, heart mean dose and lung mean dose)(P<0.05, Table 5 ).

Esophageal cancer has the characteristics of high incidence and high mortality, and has an extremely poor prognosis, with a median 5-year survival rate of about 15%-25% (8). Peripheral blood lymphocytes are one of the important members of the human immune system and can reflect the immune function of the human body. Radiation therapy induces lymphopenia, which is associated with a radiation-induced immunosuppressive effect and is common in patients with malignancies (9, 10). One study had reported that grade 4 lymphopenia during chemoradiation was occurred in approximately 31 percent of all the patients (9). Studies have shown that about 37% of patients have grade 4 lymphopenia during CRT, and patients with grade 3-4 lymphopenia are more likely, about 91% (10). This study also reported that patients with grade 4 lymphopenia during CRT had significantly shorter PFS and OS than those without grade 4 lymphopenia (PFS: median 19.1 months vs. 61.7 months and OS: median 34.7 and 63.1 months, respectively) (10). A study had reported that radiotherapy combined with or without chemotherapy can induce severe lymphopenia, which is closely related to the prognosis of patients with various malignant tumors (7). The authors of Davuluri et al. found that the probability of grade 1, 2, 3, and 4 lymphopenia during CRT for esophageal cancer was 2%, 12%, 59%, and 27%, respectively. However, only grade 4 lymphopenia was associated with worse OS (7). An article found that patients with grade 4 lymphopenia had a poorer clinical prognosis, including OS, PFS, and distant metastasis-free survival than patients without grade 4 lymphopenia (11). A study of radiotherapy for esophageal cancer showed that the tumor progression rate and cancer-related mortality were significantly higher in the post-treatment lymphopenia group than in the post-treatment ALC count≥200 cells mm³ group (76.4% vs. 52.8%, P<0.001; 58.4% vs. 39.6%,P=0.003) (9). The mechanisms underlying the association of lymphopenia with poorer survival in malignancies are not fully understood. Animal experiments have found that radiation can promote the release of antigens to stimulate lymphocytes, which can activate lymphocyte-mediated anti-tumor immune responses, resulting in anti-tumor effects (12). Some authors have found that tumor-infiltrating lymphocytes increase in patients after radiotherapy and chemotherapy, which can recognize non-natural antigens and then lead to tumor cell death (13). The correlation between lymphocyte reduction and prognosis may be related to the inhibition of immune function by radiotherapy. First, radiotherapy rays that cause suppression of bone marrow activity which leading to a decrease in the number of lymphocytes generated. And rays can also damage lymphoid organs such as the thymus or spleen, which may lead to immunosuppression (14). In addition, radiotherapy rays can directly damage the normal function of peripheral blood lymphocytes and directly inhibit immune function (9). Our study found that there was no significant difference in the nadir value of lymphocytes in patients with or without combined chemotherapy during radiotherapy ( Table 2 ). In addition, WBC, hemoglobin, albumin, and serum prealbumin were significantly decreased during radiotherapy compared with those before radiotherapy( Table 3 , P<0.001). After radiotherapy, all parameters recovered,but ALC recovered more slowly ( Table 3 ). The OS of patients in the MinALC ≤ 0.41×10⁹/L group was significantly lower than that of the patients in the MinALC>0.41×10⁹/L group ( Figure 2B , P=0.04). In addition, COX multivariate analysis showed that MinALC(≤0.41 vs.>0.41×10⁹/L) was significantly associated with OS ( Table 4 , P=0.001).

In addition to immunity status, the nutritional status of individuals is closely related to the prognosis of various malignancies (15, 16). The most typical clinical symptom of esophageal cancer patients is a progressive feeling of blocking eating. Because of direct involvement in eating, the probability of malnutrition in patients with esophageal cancer ranks first among all malignant tumors (17). Low albumin is a common malnutrition expression in patients with malignant tumors, which is a typical manifestation of cachexia and is related to the prognosis of patients with various malignant tumors, including esophageal cancer (18, 19). Many studies have shown that albumin can prevent tumorigenesis by stabilizing cell growth and inhibiting DNA replication (20). Basic experiments show that the activation of T and B lymphocytes in vitro requires the presence of serum albumin (21). High-protein dietary intervention in cancer patients may stimulate the body’s immune response.A study had showed that a low-carbohydrate, high-protein combination diet (10.6% carbohydrate, 63.5% protein) slows down the rate of tumor growth in mice compared to a traditional diet (55.2% carbohydrate, 23.2% protein) (22). And mice with a high-protein diet had less chromosomal damage in the bone marrow and reduced oxidative damage in the liver and spleen compared with mice with a low-protein diet (23). A study found that a lower level of albumin was an independent predictor of early death (less than 6 months) in esophageal cancer (18). Lower albumin levels not only reflect poorer nutritional status, but also reflect tumor aggressiveness status (18). Low hemoglobin, or anemia, is the main determinant of whether human tumor cells are hypoxic and can directly affect the sensitivity of tumor cells to radiotherapy (24). Studies have shown that malnutrition such as anemia causes many adverse clinical consequences, including reduced sensitivity to treatments such as RT, increased risk of treatment toxicity during anti-tumor periods, and reduced survival (25). At present, whether there is a correlation between the changes of hemoglobin and albumin and the decrease of lymphocytes during radiotherapy for esophageal cancer has not been reported. Patients with cachexia usually experience lymphopenia due to decreased production of cell-stimulating factors (26). Our study showed that MinHb/T1Hb,MinAlb/T1Alb,MinPA/t1PA were significantly positively correlated with the ratio of MinALC/T1ALC( Figure 2 , P<0.05). Our study showed that the nutritional status of patients is closely related to the degree of lymphopenia, so it is important to strengthen nutritional support for patients with esophageal cancer during radiotherapy.

Zhou et al. showed that the decrease in lymphocyte count in lung cancer patients reflected the severity of radiation pneumonitis. Values of lymphocytes and CD4+ T lymphocyte subsets proved as independent predictors of radiation pneumonitis. The lower peripheral blood levels of lymphocytes and CD4+ T lymphocyte were associated with an increased risk of radiation pneumonitis, which was validated by this mice model (27). Yang et al. showed that the platelet-to-lymphocyte ratio during treatment (P=0.027), and neutrophil-to-lymphocyte ratio at the end of treatment (P=0.001) were the independent predictors for symptomatic radiation pneumonitis in patients with Esophageal Cancer (28). Zhang et al. revealed that Higher CD8+ T cell count after radiotherapy in lung cancer patients was associated with an increased risk of radiation pneumonitis (29). The neutrophil-lymphocyte ratio(NLR) was higher in patients who developed symptomatic radiation pneumonitis (p=0.012).The NLR is a useful biomarker for predicting symptomatic radiation pneumonitis development after RT in NSCLC patients (30). And a study showed that lymphocyte percentage was related to radiation pneumonia in patients with lung cancer after RT(P<0.05) (31). Recent a study showed that Pre- and post-RT percentage of CD8+ T cell were the independent factors of ≥grade 2 radiation pneumonia in patients with esophageal squamous cell carcinoma (32). In our study, we found that MinALC could predict the occurrence of grade 3-4 radiation pneumonitis (P=0.0007, Figure 3A ). This result may be related to the decrease of lymphocytes that caused by radiation rays, which leads to a decline in immune function and is more likely to cause pneumonia. Due to the limited sample size, this result still needs to be confirmed by a large sample of clinical trials. Severe pneumonia can directly lead to the death of patients. Therefore, for patients with poor lung function before radiotherapy, the irradiated volume and dose of bilateral lungs should be strictly controlled.

This study has its own limitations. Firstly, this study is a retrospective study, maybe there are many selective biases when selecting patients (for example, more elderly patients were enrolled, because elderly patients are more inclined to refuse surgery). Secondly, due to the small number of patients enrolled, the subgroup analysis is not conducted for different groups of chemotherapy schemes and chemotherapy doses. Thirdly, due to the limitation of the retrospective study, the factors that may cause lymphopenia cannot be fully included, such as whether the patient uses other drugs that may cause lymphopenia during radiotherapy. Therefore, all the findings of this article still need to be further confirmed by large-scale prospective research.