Table 1 summarizes the clinical characteristics of cases included herein. Approximately 90% of the cases were male, with 31 (40%) receiving DCF. Moreover, 60 cases (71%) received two cycles of chemotherapy, whereas 25 (29%) received only one cycle due to severe adverse events.
Response to neoadjuvant chemotherapy
Among the 77 patients included herein, 58 had swollen lymph nodes larger than 10 mm, with the response to chemotherapy being determined using the rate of change of the target lesions. In other patients, the response was determined using the primary esophageal lesion. Accordingly, 37 cases were classified as PR, 39 as SD, and 1 as PD, respectively. None of the cases were classified as CR. This study divided cases into two groups: the effective response group (PR) and ineffective response group (SD and PD).
The effective response group tended to have more cases receiving DCF, although no significant difference was noted (p = 0.067). However, the effective response group had significantly more cases who received two chemotherapy cycles than one cycle (p < 0.001).
Relationship between chemotherapy response and serum IgG level
Patients included herein had a serum IgG level ranging from 460 to 2223 mg/dL. The median [25th percentile (Q1), 75th percentile (Q3)]) serum IgG level in the effective (n = 37) and ineffective (n = 40) response groups were 976 (871, 1087) and 1227.5 (1087, 1466), respectively (Fig. 2). The effective group had a significantly lower serum IgG level than the ineffective group (p < 0.001).
The ROC curve was created using “effective” as the target, with the area under curve being 79.8%. The cutoff serum IgG value at which the Youden index was maximized was determined to be 1087 mg/dL (sensitivity 0.757, specificity 0.750) (Fig. 3A).
Using the determined cutoff value, cases were then divided into high (39 cases) and low (38 cases) IgG groups. Although the high IgG group was older than the low IgG group, no significant differences in gender, final T factor, final N factor, chemotherapy regimen, and number of chemotherapy cycles were observed between both groups. The low IgG group had significantly more cases who had effective response to chemotherapy compared to the high IgG group (Fig. 3B).
Univariate and multivariate analyses
To identify independent predictors for the response to chemotherapy, clinicopathological factors were assessed using univariate and multivariate logistic regression analyses. Accordingly, univariate analysis identified age (odds ratio = 0.237, 95% CI = 0.086–0.649; p = 0.005), neoadjuvant chemotherapy cycles (odds ratio = 9.120, 95% CI = 2.720–30.600; p < 0.001), neutrophil count (odds ratio = 0.210, 95% CI = 0.067–0.658; p =0.007), lymphocyte count (odds ratio = 0.095, 95% CI = 0.159–0.454; p = 0.003), and IgG levels (odds ratio = 0.107, 95% CI = 0.038–0.302; p < 0.001) as significant predictors for the response to chemotherapy. However, multivariate analysis showed that only neoadjuvant chemotherapy cycles (odds ratio = 16.700, 95% CI = 2.990–93.300; p = 0.001) and IgG levels (odds ratio = 0.085, 95% CI = 0.019–0.382; p = 0.001) were independent predictors for the response to chemotherapy in ESCC (Table 2).
All cases underwent esophagectomy, after which the relationship between histopathological response and serum IgG value was investigated. The number of cases according to each grade is presented in Figure 4A. Effective pathological response was defined as Grade 1b or higher (less than two-thirds residual tumor cells) was defined to be pathologically effective. Accordingly, cases with effective pathological response had significantly lower pretreatment serum IgG levels compared to those who did not (p = 0.006) (Fig. 4B).