A total of 157 patients who underwent hepatectomy for ICC were eligible for inclusion in the present study at the Second Affiliated Hospital of Chongqing Medical University between January 2006 and January 2011. ICC was confirmed by histopathology and the anatomic location of the tumor was determined by review of histopathology, radiology and operation notes. After review, 47 of the 157 patients with ICC were excluded (15 lacked integrated clinical data, 13 lacked prospectively collected follow-up data and 19 had insufficient tissue for investigation). The remaining 110 patients (mean age, 53 years; age range, 28–79 years; 62 males and 48 females; 28 with type 2 DM and 82 without DM) were enrolled in the study. Cancerous and paracancerous tissues were obtained for both the DM and non-DM groups. Paracancerous tissue was defined as normal liver tissue >5 cm away from the tumor.

DM was diagnosed as a fasting plasma glucose level of >7.0 mmol/l (126 mg/dl), or a plasma glucose level of >11.1 mmol/l (200 mg/dl) at 2 h in a 75-g oral glucose tolerance test, or typical DM symptoms (polydipsia, polyuria, overeating, emaciation, fatigue or obesity) together with a casual plasma glucose level of >11.1 mmol/l (200 mg/dl; the normal fasting plasma glucose level is 3.9–6.1 mmol/l) (21).

All the patients were followed up until death or the end of the study in December 2016, with a median follow-up time of 55 months. After hepatectomy, all the patients were followed up every 3 months for 1 year and every 6 months thereafter. Liver function (based on the levels of alanine aminotransferase, aspartate aminotransferase, total bilirubin and albumin), prothrombin time (PT), abdominal ultrasound, chest film and enhanced CT or MRI were collected to monitor the patients. Diagnosis of tumor recurrence (intrahepatic and extrahepatic recurrence) was based on typical imaging findings. Tumor tissues were collected immediately upon resection from patients with ICC, and were subsequently fixed and paraffin-embedded, as described below, for immunohistochemistry.

The present retrospective study was conducted in accordance with the Declaration of Helsinki 2013 edition and national and international guidelines, and was approved by the Ethical Review Committee of the Second Affiliated Hospital of Chongqing Medical University. Written informed consent was provided by all patients.

The ICC tissues were fixed in 4% paraformaldehyde at room temperature for 24 h and embedded in paraffin. Immunohistochemical staining of paraffin sections was performed using a two-step protocol. Briefly, the ICC sections (3-µm-thick) were deparaffinized in xylene I for 15 min and xylene II for 15 min at 37°C, and rehydrated in a graded ethanol series (100, 95, 80 and 75% ethanol for 5 min each). Subsequently, antigen retrieval was performed in 10 mmol/l sodium citrate solution (pH 6.0) at 100°C for 15 min, and the samples were cooled for 30 min at room temperature. Endogenous peroxidase activity was inhibited using 3% hydrogen peroxide for 30 min at 37°C, and 5% goat serum (Origene Technologies, Inc.) was used to block non-specific binding for 15 min at 37°C, followed by incubation with a primary rabbit monoclonal anti-PKM2 antibody (1:1,000; cat. no. ab137852; Abcam) at 4°C overnight. Subsequently, the sections were incubated with a secondary anti-rabbit biotin-labelled IgG antibody (1:100; cat. no. SAP-9100; OriGene Technologies, Inc.) at 37°C for 30 min. After washing with PBS, the visualization signal was detected using 3,3′-diaminobenzidine (Boster Biological Technology) and counterstaining was performed using hematoxylin at room temperature for 5 sec. To evaluate PKM2 expression, the slides were assessed independently by two experienced pathologists with minimal interobserver variability. The slides were assessed using an orthotopic light microscope (magnification, ×100; Zeiss AG).

A semi-quantitative assessment method score was used. Scoring parameters included the staining intensity (range, 0–3; 0, negative; 1, weak; 2, moderate; and 3, strong) and the percentage of positive cells (range, 1–4; 0, negative or ≤5%; 1, 6–25%; 2, 26–50%; 3, 51–75%; and 4, 76–100%). The staining intensity was based on the color of the positive markers: Light yellow indicated weak staining, brown-yellow indicated moderate staining and dark brown indicated strong staining. The percentage of positive cells and the intensity scores were added to determine the final staining scores. A total score <4 was defined as low PKM2 expression, while a score ≥4 was defined as high PKM2 expression.

Statistical analyses were performed using SPSS 22.0 (IBM Corp.). The primary endpoint of the present study was overall survival (OS) after hepatectomy. OS was recorded as the time from the disease diagnosis to death due to any cause. Tumor recurrence-free survival (RFS) was recorded as the time from tumor resection to tumor recurrence. The significance of intergroup differences in continuous data was assessed using an unpaired Student's t-test to analyze the difference of PKM2 expression between ICC patients with DM and without DM, while the significance of differences in categorical data was assessed using the χ² test or Fisher's exact test (two-tailed). Survival analysis was performed using the Kaplan-Meier method and the log-rank test. Univariate and multivariate analyses were performed using the Cox proportional hazards model to identify independent prognostic factors (multivariate analysis was based on significant results from the univariate analysis). The TNM staging system (22) was used to determine the tumor stage in the analysis of the clinicopathological features and prognosis of patients. Categorical variables are expressed as frequencies (%). The results of the survival analysis are described as hazard ratios (HRs) and 95% CIs. P<0.05 (two-sided) was considered to indicate a statistically significant difference.

Between January 2006 and January 2011, 110 patients (28 with type 2 DM and 82 without DM) were included in the present study. The baseline characteristics were similar between the two groups, except that patients in the DM group had a higher frequency of vascular invasion (46.4 vs. 23.2%; P=0.019). There were no significant differences in sex, age, TNM stage, tumor diameter, R0 resection, differentiation degree, lymph node metastasis, intrahepatic metastasis, multiplicity, total bilirubin, alanine aminotransferase, aspartate aminotransferase, albumin and PT (Table I).

At the end of the follow-up period, 61 (55.5%) patients had died. The median follow-up time was 55 months for all patients. The OS rate of patients with DM was significantly worse than that of patients without DM (P=0.004; Fig. 1). The 1-, 3- and 5-year OS rates for patients with DM were 42.6, 23.0 and 23.0%, which were lower than 83.8, 48.2 and 41.5% in patients without DM, and the median survival time in patients with and without diabetes was 10 and 36 months, respectively (Fig. 1). As shown in Table II, univariate analysis revealed that advanced TNM stage (HR, 0.474; 95% CI, 0.245–0.918; P=0.027), >5 cm tumor diameter (HR, 0.460; 95% CI, 0.260–0.814; P=0.008), R1 resection (HR, 0.515; 95% CI, 0.310–0.857; P=0.011), lymph node metastasis (HR, 2.865; 95% CI, 1.658–4.952; P<0.001), intrahepatic metastasis (HR, 9.266; 95% CI, 4.863–17.657; P<0.001), multiplicity (HR, 6.125; 95% CI, 3.482–10.774; P<0.001), vascular invasion (HR, 5.875; 95% CI, 3.252–10.614; P<0.001), high PKM2 expression (HR, 1.984; 95% CI, 1.177–3.344; P=0.010) and DM (HR, 2.152; 95% CI, 1.255–3.691; P=0.005) were adverse prognostic factors that affected OS in patients with ICC. In addition, multivariate analysis identified the following factors as independent predictors for poor OS: DM (HR, 1.989; 95% CI, 1.084–3.65; P=0.026), high PKM2 expression (HR, 1.364; 95% CI, 1.048–2.948; P=0.007), intrahepatic metastasis (HR, 2.826; 95% CI, 1.288–6.021; P=0.010), multiplicity (HR, 4.004; 95% CI, 1.923–8.336; P<0.001) and vascular invasion (HR, 3.187; 95% CI, 1.516–6.701; P=0.002).

During follow-up, 78 (70.9%) patients experienced tumor recurrence. The RFS rate in patients with DM was significantly worse than that in patients without DM (P=0.004; Fig. 2). The 1-, 3- and 5-year RFS rates for patients with DM were 32.1, 16.3 and 10.9%, which were lower than 60.5, 38.8 and 31.7% in patients without DM, and the median RFS time in patients with and without diabetes was 5 and 17 months, respectively (Fig. 2). As shown in Table III, univariate analysis revealed that lymph node metastasis (HR, 2.664; 95% CI, 1.638–4.331; P<0.001), intrahepatic metastasis (HR, 5.640; 95% CI, 3.143–10.123; P<0.001), multiplicity (HR, 4.427; 95% CI, 2.718–7.212; P<0.001), vascular invasion (HR, 5.100; 95% CI, 3.017–8.619; P<0.001), high PKM2 expression (HR, 2.048; 95% CI, 1.289–3.254; P=0.002) and DM (HR, 1.985; 95% CI, 1.222–3.225; P=0.006) were adverse prognostic factors that affected RFS in patients with ICC. In addition, multivariate analysis identified the following factors as independent predictors for poor OS: DM (HR, 1.784; 95% CI, 1.042–3.053; P=0.035), high PKM2 expression (HR, 1.567; 95% CI, 1.057–3.012; P=0.006), multiplicity (HR, 2.898; 95% CI, 1.564–5.369; P=0.001) and vascular invasion (HR, 2.655; 95% CI, 1.340–5.259; P=0.005).

PKM2 expression was mainly concentrated in the cytoplasm and nucleus. Representative images of immunohistochemical staining are shown in Fig. 3A. High PKM2 expression was observed in 61 (55.5%) patients with ICC; among these patients, 85.7% patients had DM, while 45.1% patients did not have DM (Fig. 3B). Similarly to HCC (23), high PKM2 expression was associated with poor OS and RFS (Fig. 3C and D). Notably, the combination of low PKM2 expression and no DM had a favorable prognostic value, while patients with high PKM2 expression and DM had the shortest OS ad RFS time (Fig. 4A and B).