Introduction

Esophageal cancer is one of the most common causes of tumor-related death in the world, and more than half of the patients with esophageal cancer were diagnosed with locally advanced tumors1. In the past few decades, researchers have conducted a series of clinical studies to establish a standard treatment for esophageal cancer2,3. CROSS trial confirmed that preoperative chemoradiotherapy could significantly prolong the overall survival of patients, making neoadjuvant chemoradiotherapy as the recommended treatment for locally advanced esophageal cancer4. Although this clinical trial confirmed that neoadjuvant chemoradiotherapy had the advantage of improving prognosis, nearly half of the patients still had postoperative recurrence or distant metastasis4. This suggests that more systematic and effective treatments are still needed to prevent potential recurrence and metastasis of esophageal cancer.

A series of phase III clinical trials have confirmed that immunotherapy represented by programmed cell death protein-1 (PD-1) inhibitors can be used as first-line treatment for advanced esophageal cancer5,6,7. Subsequently, several clinical trials have evaluated the safety and efficacy of neoadjuvant chemotherapy combined with immunotherapy, and confirmed that immunotherapy could be used as a part of neoadjuvant therapy in patients with locally advanced esophageal cancer8,9. 25–56% of esophageal cancer patients acquired to complete pathological response (pCR) after neoadjuvant chemotherapy combined with immunotherapy, while 50–90% patients acquired to objective response rate (ORR), indicating that preoperative chemotherapy combined with immunotherapy is a feasible neoadjuvant therapy for esophageal cancer9,10,11,12.

It is suggested that pathological response is associated with the prognosis of the patients. Therefore, predicting pathological response rate of the patients who received neoadjuvant therapy is necessary. Several studies have reported the predictive models that can predict pCR and tumor regression grade in patients with esophageal cancer after neoadjuvant immunotherapy13,14. However, there is still no simple and feasible method to predict the ORR of patients. Therefore, the aim of this study was to establish an effective model for predicting the ORR of esophageal cancer patients after neoadjuvant TP (platinum + taxol) regimen combine with PD-1 inhibitors (Pembrolizumab, Camrelizumab, Tislelizumab or Sintilimab).

Methods

Patients

Consecutive esophageal cancer patients who were diagnosed and treated in the First Affiliated Hospital of Xi’an Jiaotong University from January 2020 to February 2022 were enrolled in this study according to the inclusion criteria. The inclusion criteria were as follows: (1) pathological diagnosis: esophageal carcinoma; (2) received neoadjuvant TP regimen combined PD-1 inhibitors; (3) received radical surgery; (4) the clinicopathological and postoperative pathological data were completed; (5) not received any other anti-tumor therapy; (6) no immune system disease. Patients did not meet with the inclusion criteria were excluded in this study. In this study, a total of 42 patients were included as a training cohort. According to the above inclusion and exclusion criteria, we also enrolled 53 consecutive patients treated in the Shaanxi Provincial Cancer Hospital Affiliated to Medical College of Xi’an Jiaotong University from January 2020 to December 2021 as a validation cohort (Fig. 1). We defined the ORR as pCR (no residual tumor cells in the resected specimens and all resected lymph nodes) + major pathological response (≤ 10% viable tumor cells in the resected primary tumor and all resected lymph nodes) + partial pathological response (≤ 50% viable tumor cells in the resected primary tumor and all resected lymph nodes)15,16.

Figure 1
figure 1

Flow chart of patient selection. TP, taxol + platinum; PD-1, programmed cell death protein-1.

Full size image

Data collection and processing

The baseline data, clinicopathological data, treatment-related data and laboratory indexes of the enrolled patients were collected. The data was processed by using Microsoft Excel and SPSS 26.0. The optimal cut-off value of continuous data was calculated by receiver operating characteristic (ROC) curve. Then, the continuous data was converted into binary data according to the optimal cut-off value.

Statistical analysis

The statistical analysis was performed by SPSS 26.0 and Rstudio. The difference between the two groups was analyzed by x2 test. Univariate and multivariate logistic regression analyses were conducted to evaluate the factors that might be related to the ORR of the patients after neoadjuvant therapy. According to the result of multivariate logistic regression analysis, a nomogram for predicting ORR was established. The predictive ability was validated by using C-index, ROC curve, calibration curve, decision curve analysis (DCA) and clinical impact curve (CIC). P-value less than 0.05 was considered to be statistically significant.

Ethics approval

This study was performed in accordance with the Declaration of Helsinki, and was conducted under the approval and supervision of the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University (No. XJTU1AF2022LSK-335). The study was a retrospective study, and written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Therefore, the waiver of informed consent was approved by the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University.

Results

Characteristics of the patients with esophageal cancer

A total of 42 patients with esophageal cancer from the First Affiliated Hospital of Xi'an Jiaotong University were enrolled as the training cohort in this study, while 53 patients from Shaanxi Provincial Cancer Hospital Affiliated to Medical College of Xi'an Jiaotong University were enrolled as the validation cohort (Fig. 1). All the patients received neoadjuvant TP regimen combined with PD-1 inhibitors and radical surgery. The mean age of the patients in training cohort was 62.33 ± 7.39 years, and the mean age of the patients in training cohort was 63.56 ± 5.85 years. Postoperative pathological result showed that 54.76% (23/42) and 49.06% (26/53) patients acquired ORR after neoadjuvant therapy in training cohort and validation cohort, respectively. The clinicopathological features and laboratory findings of these patients before neoadjuvant therapy are shown in Tables 1 and 2. Notably, it was found that neutrophil, platelet, platelet-to-lymphocytes ratio (PLR), systemic immune-inflammation index (SII) [SII = (platelet × neutrophil) / lymphocytes], D-dimer and carcinoembryonic antigen (CEA) were significantly different among ORR group and Non-ORR group (Table 2).

Table 1 Characteristics of esophageal cancer patients received neoadjuvant chemotherapy combined with immunotherapy.
Full size table
Table 2 Laboratory indicators of esophageal cancer patients received neoadjuvant chemotherapy combined with immunotherapy.
Full size table

Factors affecting the ORR of the patients with esophageal cancer after neoadjuvant chemotherapy combined with immunotherapy

To explore the factors that might affect the ORR after neoadjuvant therapy, we performed a univariate logistic regression analysis on all clinicopathological features and laboratory indicators of the patients. The result showed that neutrophil (< 5.22 × 109/L), platelet (< 237.50 × 109/L), PLR (< 166.50), SII (< 477.40), aspartate aminotransferase (AST) (≥ 20.00 U/L), D-dimer (≥ 0.24 mg/L) were associated with the ORR of the patients after neoadjuvant therapy (all P < 0.05) (Table 3). Thus, the above positive indicators were selected for further multivariate logistic regression analysis. Although the P-value of some indicators in the univariate regression analysis was greater than 0.05, in order not to omit the factors that might be related to ORR, we also enrolled the indicators with a P-value less than 0.1 [white blood cell (WBC), neutrophil-to-lymphocytes ratio (NLR), albumin, creatinine, fibrin degradation product (FDP), thrombin time (TT) and CEA)] into the subsequent multivariate logistic regression analysis (Table 3).

Table 3 Univariate logistic regression analysis of ORR in patients.
Full size table

The result of the multivariate logistic regression analysis showed that AST, D-dimer and CEA were the independent predictors of the ORR after neoadjuvant TP regimen combined with PD-1 inhibitors (Table 4). All the three indicators were positively related to the ORR of the patients.

Table 4 Multivariate logistic regression analysis of ORR in patients.
Full size table

Establishment and validation of a nomogram for predicting the ORR of the patients after neoadjuvant therapy

According to the result of the multivariate logistic regression analysis, a nomogram based on AST, D-dimer and CEA was established to predict the ORR of the esophageal cancer patients who received neoadjuvant TP regimen combined with PD-1 inhibitors (Fig. 2). The C-index of the nomogram was 0.93 (95% CI 0.86–1.00). ROC curve of the nomogram showed that the value of area under curve (AUC) was 0.931 (Fig. 3A). Besides, the nomogram was validated by calibration curve (Fig. 3B) and the mean absolute error was 0.035. In addition, the DCA curve (Fig. 3C) and CIC curve (Fig. 3D) showed that the ability of the nomogram prediction model to predict pCR after NAC was pretty good.

Figure 2
figure 2

The nomogram for predicting the ORR of the patients after neoadjuvant TP regimen combined with PD-1 inhibitors. ORR, objective response rate; TP, taxol + platinum; PD-1, programmed cell death protein-1; AST, aspartate aminotransferase; CEA, carcinoembryonic antigen.

Full size image
Figure 3
figure 3

Internal validation of the nomogram for predicting the ORR after neoadjuvant therapy from training cohort. (A). Receiver operating characteristic curve. (B). Bootstrap validation curve. (C). Decision curve analysis. (D). Clinical impact curve. ORR, objective response rate.

Full size image

We also performed external validation of the nomogram by using validation cohort from Shaanxi Provincial Cancer Hospital Affiliated to Medical College of Xi'an Jiaotong University. The C-index of the validation cohort was 0.86 (95%CI: 0.76–0.95), while the AUC of ROC curve was 0.855 (Fig. 4A). The validation curve (Fig. 4B), DCA (Fig. 4C) and CIC (Fig. 4D) of the validation cohort also showed a good predictive ability of the nomogram.

Figure 4
figure 4

External verification of the nomogram for predicting the ORR after neoadjuvant therapy from validation cohort. (A). Receiver operating characteristic curve. (B). Bootstrap validation curve. (C). Decision curve analysis. (D). Clinical impact curve. ORR, objective response rate.

Full size image

Discussion

This study evaluated the clinicopathological features and the laboratory hematological indexes that might affect the ORR of esophageal cancer patients after neoadjuvant therapy. Based on the positive factors, a nomogram for predicting ORR was established and verified.

The application of neoadjuvant therapy for esophageal cancer has greatly improved the prognosis of the patients3,17. The addition of immunotherapy to neoadjuvant therapy significantly prolonged the 5-year survival of the patients18,19. For locally advanced esophageal cancer, a number of prospective single-arm studies have confirmed the effectiveness and safety of neoadjuvant chemotherapy combined with immunotherapy20,21,22. After neoadjuvant therapy, these patients achieved high R0 resection rate, pCR rate and ORR. The subsequent problem is how to use accurate and simple methods to predict the efficacy of neoadjuvant therapy. Recently, some studies have reported predictive models that can predict the tumor regression grade (TRG) and pCR after neoadjuvant immunotherapy13,14. However, there is no related research on predicting ORR.

Previous studies have confirmed that inflammation and nutrition indexes before treatment were associated with the prognosis of the patients23,24,25. According to these indexes, the pathological response of the patients after neoadjuvant therapy can be effectively predicted23,24,25. For example, a recent study showed that the difference in albumin levels before and after neoadjuvant immunotherapy and WBC count before neoadjuvant therapy were significantly correlated with the TRG in patients13. Besides, inflammatory indicators, including NLR, PLR, LMR and SII, could well predict pCR after neoadjuvant immunization and participated in the development of various cancers23,26,27. In our study, we found that there were significant differences in neutrophil, platelet, PLR and SII between ORR group and non-ORR group. Univariate logistic regression analysis also showed that neutrophil, platelet, PLR and SII were significantly correlated with ORR. These results suggested that the expression level of these indexes before neoadjuvant therapy might affect the ORR of patients. Thus, in clinical practice, we might evaluate the efficacy of neoadjuvant immunotherapy by using these indexes.

D-dimer is a degradation product of fibrin and an indicator of hypercoagulability and endogenous fibrinolysis. It was reported that the plasma concentration of D-dimer was related to neoadjuvant chemotherapy efficacy and the prognosis in cancer28. The increase of plasma D-dimer level is related to the progression, increased lymph node metastases and poor prognosis of esophageal cancer29,30. Besides, plasma D-dimer was regarded as an independent prognostic factor for resectable esophageal cancer patients. The 5-year cancer-specific survival of patients with D-dimer ≤ 5.0 μg/mL was significantly better than that of patients with D-dimer > 0.5 μg/mL (35.5% vs. 21.1%)31. However, a previous study pointed out that a low D-dimer level was significantly and independently associated with better overall survival in lung cancer27. Similarly, in this study, we observed that elevated D-dimer level was related to ORR, suggesting the role of D-dimer in improved outcome of esophageal cancer patients with neoadjuvant immunotherapy. Moreover, serum AST is a biomarker of systemic inflammation and immune activation, which can be used to evaluate liver function32. AST was also associated with worse overall survival and a higher 90-day mortality rate after surgery in cancer patients33. But in esophageal cancer, serum AST/aspartate aminotransferase (ALT) level is a significant predictor of overall survival (OS). The 5-year OS of patients with high AST/ALT levels was longer than that of patients with low AST/ALT levels34. Besides, in non-virus-related hepatocellular carcinoma, AST was enrolled for constructing a prognostic model to evaluate the OS of patients, and the model showed a good predictive ability35. In this study, elevated AST was confirmed to contribute the improved ORR after neoadjuvant immunotherapy, which was consistent with the previous study34. As an accurate biomarker for occult advanced disease, preoperative serum CEA level could be used to predict the resectability of patients with esophageal cancer36. It was suggested that CEA could be used to predict the sensitivity of esophageal cancer to chemoradiotherapy37. In the present study, we found that CEA, AST and D-dimer were independently related to the ORR after neoadjuvant immunotherapy, indicating that these laboratory indicators before neoadjuvant therapy played an important role in predicting ORR of patients with esophageal cancer. Clinicians could combine these indicators to estimate the efficacy of neoadjuvant immunotherapy in patients with esophageal cancer.

The nomogram is a simple and useful tool for predicting outcome38. The nomogram model established according to the results of regression analysis can well predict the pathological response of the patients after neoadjuvant therapy13,14,39,40. The prediction effect of these models is pretty good. Based on the results of multivariate regression analysis, we also established a nomogram that could predict ORR. The C-index was 0.93 (95% CI 0.86–1.00), and the AUC value of ROC for the nomogram was 0.931, indicating a high predictive ability. Also, the results of the calibration curve, DCA curve and CIC curve showed that the prediction of the nomogram was pretty good. In addition, the external validation of the nomogram by using validation cohort from Shaanxi Provincial Cancer Hospital Affiliated to Medical College of Xi'an Jiaotong University showed similar results. All the results indicated that our nomogram had a good ability to predict ORR after neoadjuvant immunotherapy.

However, there are still some limitations that cannot be ignored in this study. First, this study is a retrospective analysis, so it is difficult to obtain data such as genetic testing (including PD-1 expression levels) of patients. Therefore, we only analyzed the clinicopathological factors and common laboratory indicators. Second, the sample size of this study is small. Thus, in the next step, we plan to conduct a prospective study, include more indexes and enlarge the sample size through multi-center cooperation to improve the prediction model.

Conclusion

This study analyzed the clinicopathological factors and laboratory indexes that might affect the ORR of the patients with locally advanced esophageal cancer after neoadjuvant chemotherapy combined with immunotherapy. Based on the results of logistic regression analysis, a nomogram model for predicting ORR was established, and the model had good predictive ability. Our study provides a simple and feasible predictive model for ORR in patients with resectable locally advanced esophageal cancer after neoadjuvant TP regimen combined PD-1 inhibitors, which might be used in clinical practice for clinicians.