A total of 20 patients who had received neoadjuvant CRT followed by surgery were enrolled in this study. All the formalin-fixed, paraffin-embedded (FFPE) specimens of the patients were available for evaluation. The study protocol was approved by the Ethics Review Board of Mie University Hospital and all the included patients provided written informed consent for their tissues to be used in this study.

All the patients received systemic 5-FU and CDDP chemotherapy with concurrent radiotherapy. The regimen included 4 cycles of 5-FU (600 mg/m² administered intravenously over 24 h), plus tegafur and uracil (400 mg/kg of body weight administered orally for 5 days) and CDDP (4 mg/day administered intravenously for 5 days) with concurrent 40 Gy radiation followed by surgery. Preoperative radiotherapy was delivered to the primary tumor as well as the peritumoral area at a dose of 40 Gy in 20 fractions within 4 weeks. The time interval between neoadjuvant CRT and surgery was 2–3 weeks.

The clinical response following preoperative CRT was evaluated by barium esophagography, endoscopy and computed tomography. The results were graded as complete response (CR), partial response (PR), no change (NC) or progressive disease. The pathological response to CRT was evaluated using the Mandard tumor regression grade (TRG) (16). The tumors were classified according to the Mandard system into 5 grades as follows: i) TRG1, CR with absence of residual cancer and fibrosis extending through the wall; ii) TRG2, presence of residual tumor cells scattered through the fibrotic area; iii) TRG3, increased numbers of residual cancer cells, with predominant fibrosis; iv) TRG4, residual cancer outgrowing the fibrotic area; v) TRG5, absence of regressive changes. We categorized patients in categories TRG1 and 2 as responders and those in categories TRG3-5 as non-responders.

The FFPE specimens were cut into 2- to 3-µm sections. Following deparaffinization and dehydration, the sections were placed in a 10 mmol/l sodium citrate buffer (pH 6.0) and autoclaved at 121°C for 10 min for antigen retrieval. The sections were incubated in 3% hydrogen peroxide for 10 min, blocked and incubated with a primary antibody overnight at 4°C. Monoclonal mouse anti-human Fas antibody (B-10, catalog no. sc-8009; Santa Cruz Biotechnology, Santa Cruz, CA, USA; dilution 1:100), polyclonal rabbit anti-human FasL antibody (C-20, catalog no. sc-957; Santa Cruz Biotechnology; dilution 1:100) and monoclonal mouse anti-human Ki67 antibody (MIB-1, code M7240; Dako Cytomation, Glostrup, Denmark; dilution 1:100) were the primary antibodies used in a labeled streptavidin-biotin system (EnVision™ + Dual Link System-HRP; Dako Cytomation). Antibody binding was visualized using 3,3′-diaminobenzidine (Dako Cytomation). All the sections were counterstained with hematoxylin prior to being dehydrated and mounted. At least 2 sections per specimen were stained to confirm reproducibility. Negative controls were prepared simultaneously with pre-immune immunoglobulin.

The sections were observed under a light microscope (BX50, Olympus, Tokyo, Japan). We calculated IHC scores by multiplying the percentage of positive epithelial cells (0–100%) by the staining intensity, as previously described (17). The staining intensities were scored as follows: 0, negative; 1, weak; and 2, strong for Fas and Ki67; and 0, negative; 1, weak; 2, moderate; and 3, strong for FasL. The IHC scores ranged between 0 and 200 for Fas and Ki67; and between 0 and 300 for FasL. Each sample was scored by two investigators (K.T. and Y.O.) who were blinded to the clinicopathological information regarding the origin of the samples.

We previously recorded the numbers of peritumoral CD8- and Foxp3-positive cells using light microscopy and the median count was calculated for each sample. The median values for the number of CD8- and Foxp3-positive cells were 70 (range, 18–250) at a magnification x200 and 13 (range, 1–57) at a magnification x100 (18). We analyzed the correlation of Fas/FasL expression with peritumoral CD8- and Foxp3-positive cells.

Statistical analysis was performed using StatView v5.0 software (SAS Institute Inc., Cary, NC, USA). Significant differences were analyzed using the Chi-square test. The Pearson's correlation test was used to determine statistical correlations. Recurrence-free and overall survival probabilities were calculated from the date of surgery to the date of recurrence or death, using the Kaplan-Meier product limit method; intergroup differences were determined using the log-rank test. P<0.05 was considered to indicate a statistically significant difference.

The patient characteristics are summarized in Table I. The median age of the patients was 68 years (range, 52–77 years) and the male:female ratio was 9:1. The median follow-up period was 16 months (range, 4–91 months). Of the 20 tumors, 12 were located in the lower, 7 in the middle and 1 in the upper part of the esophagus. The mean size of the tumors was 40 mm (range, 20–82 mm). The post-CRT pathological T stages were as follows: pT1 (n=1), pT2 (n=9), pT3 (n=9) and pT4 (n=1). A total of 10 patients (50%) presented with lymph node metastases. The median number of total dissected lymph nodes was 14 (range, 0–39) and the median value of lymph node density was 5.45% (range, 0–55.6%). A total of 17 tumors (85%) exhibited well or moderately differentiated squamous cell carcinoma histology. R0 resection was performed in 80% of the cases. The clinical response was classified as follows: NC, 9 patients; PR, 10 patients; and CR, 1 patient. The Mandard TRG grades were as follows: TRG1, no patients; TRG2, 9 patients; TRG3, 7 patients; TRG4, 4 patients; and TRG5, no patients (non-responders, n=11; responders, n=9).

Fig. 1 shows the IHC findings for Fas, FasL and Ki67 in ESCC following neoadjuvant CRT. Fas and FasL were expressed in the cytoplasm and nuclei of cancer cells. Additionally, FasL expression was diffusely detected in stromal cells. Ki67 expression was detected in cancer cell nuclei. There was no correlation between the IHC findings for Fas/FasL and Ki67. The median values of the IHC scores for Fas, FasL and Ki67 were 80 (range, 1–200), 190 (range, 30–300) and 10 (range, 1–180), respectively. We classified those cases with values above the median IHC score as the high-expression group and the remainder as the low-expression group.

High FasL expression was significantly correlated with disease recurrence in patients treated with curative intent (P=0.0134). However, there were no correlations between Fas/FasL expression and other clinicopathological characteristics (Table II). High Ki67 expression was significantly correlated with lymph node metastasis and lymphatic invasion (P=0.007 and 0.025, respectively) (data not shown).

The recurrence-free survival and overall survival according to Fas and FasL expression using the Kaplan-Meier product limit method are shown in Fig. 2. High expression of FasL was found to be significantly associated with poor recurrence-free and overall survival (P=0.0102 and 0.0385, respectively). Patients with low Fas and high FasL expression exhibited a poorer recurrence-free survival (P=0.0035).

Although the differences did not achieve statistical significance, Fas expression was inversely correlated with peritumoral Foxp3-positive cells, whereas FasL expression appeared to be inversely correlated with peritumoral CD8-positive cells and directly correlated with Foxp3-positive cells (Fig. 3).