Exploratory analyses of the associations between Ki-67 expression, lymph node metastasis, and prognosis in patients with esophageal squamous cell cancer

Patient recruitment

A total of 168 patients with stage I–IVa ESCC who consecutively received treatment at the Second Affiliated Hospital of Fujian Medical University from January 2018 to June 2023 were included in this retrospective study. The cases included in this study were patients who met the following criteria: (1) ESCC confirmed by pathology and with any type of radical resection; (2) classified as stage I to stage IVa based on postoperative pathological examination according to the 8th American Joint Committee on Cancer (Donohoe & Phillips, 2017); (3) age > 18 years; (4) an Eastern Cooperative Oncology Group (ECOG) Score < 3 points or Karnofsky’s performance score ≥60; (4) detailed follow-up information to calculate overall survival (OS) and (5) provision of informed consent, either from the patients themselves or their families. Exclusion criteria were: stage IVb disease; serious underlying diseases, or dual or multiple cancers.

Radical surgery was performed in all patients, with or without adjuvant treatment; however, specific surgical details were not limited. Due to the retrospective nature of this study and the potential impact of preoperative treatment on Ki-67 expression status, any patient who received neoadjuvant therapy, including chemotherapy, radiation therapy, and immunotherapy, were excluded. Given that immunohistochemical (IHC) examination of Ki-67 is routinely performed in postoperative pathology of EC in our hospital, additional collection of paraffinized tissue to assess Ki-67 expression status was not required. All IHC stained sections were obtained from Department of Pathology, and relevant clinical and pathological data were retrieved from the electronic medical record system. Written informed consent was obtained from surviving patients, or from the family members of patients who had died. Survival information for each participant was obtained from case records, telephone follow-ups, or official death records. OS was defined as the period from the date of diagnosis to the date of death from any cause or the latest follow-up. The latest follow-up period for surviving cases was up to December 31, 2023. This study was approved by the ethics committee of the Second Affiliated Hospital of Fujian Medical University (Ethics review number: 2023-416).

During the pre-established recruitment period, a total of 236 patients with ESCC underwent postoperative Ki-67 protein IHC examination. Fifty-three of these patients were excluded because they had undergone preoperative treatment. After obtaining the patients’ clinical information from the electronic medical record system, 15 patients were excluded due to a lack of detailed information. Ultimately, 168 patients were included in this study. A flowchart of this study is presented in Fig. 1.

Immunohistochemical examination and grading of Ki-67

Although additional Ki-67 immunohistochemistry experiments were not needed for this study, detailed methodological information is provided. These examinations are routinely performed at our hospital. ESCC tissue sections of four µm thickness were cut from surgically excised ESCC cancer tissue, fixed in formalin, and embedded in paraffin. All samples were mounted on silane-coated slides. IHC staining was performed using an automatic staining machine. A rabbit polyclonal antibody against Ki-67 (ab15580; Abcam) was used at 1:500 dilution for IHC staining. Two researchers simultaneously evaluated Ki-67 expression status to ensure consistent results. The intensity of Ki-67 staining was classified according to the following criteria: 0 points for no staining, 1 point for week intensity, 2 points for moderate intensity, and 3 points for strong intensity. Positive IHC-stained cells were counted for each slice under five high-power field (×200). The proportion of positive cells were classified to five groups according to their positive scores (0 points for <5%, 1 point for 5% to 25%, 2 points for 26% to 50%, 3 points for 51% to 75%, and 4 points for 76% to 100%). The product of the intensity and positive scores was used to determine the final Ki-67 expression. Ki-67 expression scores of were divided into four groups: negative (−) for 0 points, weakly positive (+) for 1–4 points, positive (++) for 5–8 points and strongly positive (+++) for 9–12 points. The subjects were divided into a low-expression group (Ki-67^Low group) and a high-expression group (Ki-67^High group), with a threshold expression score of five points.

Statistical analyses

An independent sample t-test was used to compare differences in continuous variables between the two groups. The chi-square test was used for categorical variables to explore the associations between Ki-67 expression and clinicopathological parameters. Propensity score matching (PSM) analysis was applied to correct for background variables and achieve statistical comparability. PSM were conducted using the “MatchIt” package in R, and the nearest neighbor algorithm is used for matching analysis. The Ki-67 expression level (high or low) was set as the dependent variable, while other available clinicopathological parameters were set as independent variables to achieve PSM. For survival outcomes, hazard ratios (HR) and corresponding 95% confidence intervals (CIs) were used as effect sizes to show prognostic performance of Ki-67. Survival was estimated using the Kaplan–Meier method and the log-rank test. The prognostic performance of Ki-67 expression was explored using a Cox proportional hazards model. R packages “survminer” and “survival” were used to carry out survival analyses. Forest-maps based on univariate analysis and multivariate analysis from a Cox proportional-hazards model were drawn with “forestplot” R package. All statistical analyses were performed using R 4.3.1 software package via various R packages. Statistical significance was set at p < 0.05.

Clinical and pathological characteristics of the patients

A total of 168 patients with ESCC were enrolled, including 75 (44.64%) with weakly positive expression, 51 (30.36%) with positive expression and 42 (25.00%) with strong positive expression. No patients were negative for Ki-67 expression in our cohort. Finally, 75 (44.64%) patients were assigned to the Ki-67^Low group, and 93 (55.36%) patients were assigned to Ki-67^High group. Among these, 38 (22.62%) patients were staged at stage I, 53 (31.55%) at stage II, 65 (38.69%) at stage III and 12 (7.14%) at stage Iva; 66 (39.29%) patients were ≥60 years, and 102 (60.71%) were aged < 60 years. A total of 98 (58.33%) patients were male. Among the patients, 101 (60.12%) patients had undergone adjuvant radiotherapy or chemotherapy after surgery. The Ki-67 intensity scores were 1.31 ± 0.46 and 2.45 ± 0.50 points in the Ki-67^Low and Ki-67^High groups, respectively and this difference is statistically significant (t = 15.348, P < 0.001). The Ki-67 positivity scores were 1.95 ± 0.68 and 3.49 ± 0.50 points in the Ki-67^Low and Ki-67^High groups, respectively, with statistically significant difference (t = 16.497, P < 0.001). The Ki-67 positivity ratings were 2.56 ± 1.18 and 8.55 ± 2.07 points in the Ki-67^Low and Ki-67^High groups, respectively, and this difference is statistically significant (t = 23.555, P < 0.001). The IHC staining results of Ki-67 in the two cases of ESCC with different levels of Ki-67 expression are shown in Figs. 2A and 3A. Detailed comparative data between the two groups were shown in Table 1. Owing to significant differences in some clinical features between the Ki-67^Low and Ki-67^High groups, propensity matching analysis was conducted. The clinical information after matching is shown in Table 2. After PSM, the baseline characteristics between the two groups were found to be comparable (P > 0.05). Detailed patient data included in this study are presented in Tables S1 and S2.

Association between Ki-67 expression and Tumor Node Metastasis (TNM) stage, lymph node metastasis, vascular invasion, and perineural invasion

The hematoxylin-eosin staining results of LNM, VI, and PI in the two cases of ESCC with different levels of Ki-67 expression are depicted in Figs. 2B, 3C, 2D, 3B, 3C and 3D. Ki-67 expression increased progressively with advancing TNM tumor stage, as shown in Table 3 and Fig. 4A. The respective proportions of patients in stages T1, T2, and T3 in the Ki-67^Low group were 44.00% (33/75), 42.67% (32/75), and 13.33% (10/75), and the respective proportions of patients in stages T1, T2, and T3 in the Ki-67^High group were 25.81% (24/93), 52.69% (49/93), and 21.51% (20/93), with a statistically significant difference (χ² = 6.468, P = 0.039), as shown in Table 3 and Fig. 4B. In the entire cohort, 73.81% (124/168) of the patients exhibited lymph node metastasis. The respective proportions of patients with lymph node metastasis were 64.00% (48/75) in the Ki-67^Low group, and 81.72% (76/93) in the Ki-67^High group, with a statistically significant difference (χ² = 11.219, P = 0.011), as shown in Table 3 and Fig. 4C. VI was observed in 35.71% (60/168) of total patients. The respective proportions of patients with VI in the Ki-67^Low and Ki-67^High groups were 25.33% (19/75) and 44.09% (41/93), with a statistically significant difference (χ² = 6.359, P = 0.012), as shown in Table 3 and Fig. 4D. A total of 36.31% (61/168) patients had PI. The respective proportions of patients with PI in the Ki-67^Low and Ki-67^High groups were 24.00% (18/75) and 46.24% (43/93), with a statistically significant difference (χ² = 8.877, P = 0.003), as shown in Table 3 and Fig. 4E.

Analysis of the effect of Ki-67 expression status on ESCC prognosis

Median follow-up time was 33.5 months (range 3.0–60.0 months) in our cohort. At the end of follow-up, 63.10% (106/168) of patients had died. Kaplan–Meier survival analysis revealed OS proportions at 3 and 5 years of 74.13% (64.18%–85.62%) and 31.55% (23.22%–42.88%) for the Ki-67^Low group, and 51.07% (39.40%–66.20%) and 11.21% (5.81%–21.63%) for the Ki-67^High group (χ² = 34.833, P < 0.001). In the PSM cohort, Kaplan–Meier survival analysis of OS proportions at 3 and 5 years were 68.57% (56.49%–83.23%) and 33.82% (22.94%–49.86%) for the Ki-67^Low group, and 45.22% (31.72%–64.46%) and 18.40% (9.71%–34.84%) for the Ki-67^High group (χ² = 11.372, P < 0.001). Kaplan–Meier curves for OS are depicted in Figs. 5A and 5B. The univariate Cox proportional hazards model revealed that high Ki-67 expression was significantly associated with poor OS compared to that of the low expression (HR = 3.42, 95% CI [2.22–5.27], P < 0.001), as shown in Table 4 and Fig. 6A.

The univariate Cox proportional hazards model also revealed that TNM stage, VI, PI and postoperative treatment exhibited significant effect on OS (P < 0.05). A multivariable COX model was analyzed using stepwise regression based on the first model with significant univariates in Table 4. The result of multivariable model is shown in Table 5 and Fig. 6B. In the stepwise regression multivariable model, high Ki-67 expression was significantly associated with poor OS in patients with ESCA compared to that of the low expression (HR = 1.98, 95% CI [1.33–2.94], P < 0.001), suggesting that high Ki-67 expression is an independent prognostic factor in patients with ESCC.