Mitotic count reflects prognosis of gallbladder cancer particularly among patients with T3 tumor

The surgical strategy for gallbladder cancer (GBC) depends on the extent of the disease. Thus, the identification of useful prognostic markers exerting strong prognostic impact for each T stage would be beneficial in the development of rational therapeutic strategies. The purpose of this study was to identify useful prognostic markers of GBC for each T stage. CD8+ tumor-infiltrating lymphocytes (TIL), Ki-67 labeling index (LI), p53 nuclear expression and mitotic count (MC) were investigated as candidate prognostic markers. In total, 86 patients with invasive GBC were included. Of the prognostic markers examined, only MC showed a correlation with reduced survival (P=0.0383) in the univariate analysis of overall T stage. In the univariate analysis of T2 stage (n=31), only high p53 expression correlated with survival showing a positive correlation (P=0.0154). In the univariate analysis of T3 stage (n=40), the only factor showing a significant correlation with survival was MC (P=0.0113). Multivariate analysis, including N and M as factors, identified only MC as an independent prognostic factor in T3 stage GBC (P=0.0419). In conclusion, this study demonstrated the strong prognostic impact of MC in GBC, particularly in patients with T3 tumor.


Introduction
The surgical strategy for gallbladder cancer (GBC) depends on the extent of the disease, particularly the T stage from the tumor, node, metastasis (TNM) classification (1). Identification of useful prognostic markers exerting a strong prognostic impact for each T stage would be beneficial in the development of rational therapeutic strategies for each T stage of GBC. Useful histological markers are well-known and easily evaluated in ordinary pathological examinations. This study therefore focused on CD8 + tumor-infiltrating lymphocytes (TIL), Ki-67 labeling index (LI), p53 nuclear expression and mitotic count (MC), all of which have been well investigated in other solid cancers as candidate prognostic markers in GBC.
CD8 + TIL have been considered as manifestations of host immune reactions against cancer cells and strong prognostic impact of CD8 + TIL has been found in a wide variety of solid cancer tissues (2)(3)(4)(5)(6)(7)(8)(9)(10). A gene on chromosome 10 encodes a nuclear protein of 345-395 kDa that is recognized by the antibody of the Ki-67 antigen. Ki-67 protein is expressed during the active phases of the cell cycle (G1, S, G2, and mitosis), but is absent from resting cells (G0). Ki-67 LI is thus considered a marker for cell proliferation and the prognostic impact of Ki-67 LI has been reported in various solid cancer tissues (11)(12)(13)(14). p53 is a well-known tumor suppressor protein that is encoded by the TP53 gene, located on the short arm of chromosome 17. Mutations of the TP53 gene lead to loss of production of the normal p53 protein and synthesis of a mutated protein with an increased half-life which tends to accumulate in the nucleus and can be detected by immunohistochemical staining (15). The prognostic role of p53 nuclear expression as assessed by immunohistochemistry has been reported in various types of solid cancer (16)(17)(18)(19). MC is widely recognized as an indicator of tumor malignancy and the prognostic impact of MC and classification or grading by MC status have been reported for various types of tumor (20)(21)(22)(23)(24).
The purpose of this study was to assess the prognostic impact of CD8 + TIL, Ki-67 LI, p53 nuclear expression and MC in GBC, according to T stage.

Materials and methods
Patients and staging. A total of 101 GBC patients underwent surgical treatment for the primary lesion at the Saga University Hospital between January, 1989 and December, 2011. Of these, 11 patients showing non-invasive intramucosal cancer and 4 patients for whom no tissue samples were preserved were excluded from the study. As a result, a final total of 86 patients with invasive GBC were enrolled in this study. Informed consent for the use of resected tissue was obtained from the patients, and the study protocol was approved by the Ethics Committee of the Faculty of Medicine at the Saga University. Clinical and histopathological staging were based on the TNM Classification of Malignant Tumors established by the International Union Against Cancer (7th edition, 2009) (1).

Immunohistochemical staining and evaluation of MC.
Sections cut from formalin-fixed paraffin-embedded tissue blocks were used. The primary antibodies used were CD8 (dilution 1:50, clone C8/144B; DakoCytomation, Glostrup, Denmark), p53 (prediluted, clone DO-7; Nichirei Biosciences, Tokyo, Japan) and Ki-67 (dilution 1:30, clone MIB-1; DakoCytomation). The slides were heated in ethylenediaminetetraacetic acid (EDTA) (pH 9.0) in a microwave oven for antigen retrieval. The EnVision™+ system (DakoCytomation) was used as the secondary antibody. Slides were visualized by diaminobenzidine tetrahydrochloride (DAB 4HCl) and nuclei were counterstained with hematoxylin. An Autostainer Plus ® automatic stainer (DakoCytomation) was used for staining. Ki-67 LI was determined using the ratio of positive nuclear staining of Ki-67 and classified as ≤10% (low group) or >10% (high group). Assessment of p53 was also determined by positive nuclear staining and classified as ≤30% (low group) or >30% (high group). The cut-off value of p53 and Ki-67 LI in previous studies varied greatly. The cut-off value of p53 and Ki-67 was determined to divide the cohort into two comparable groups effectively. CD8 + lymphocytes within the cancer cell nest were regarded as CD8 + TIL, according to a previous report (2). CD8 + TIL were counted in the three fields showing the most abundant distribution of CD8 + TIL using a x10 objective lens. The number of CD8 + TIL was then determined as the mean count for these three fields. CD8 + TIL was analyzed separately on the tumor surface and the invasion front and was categorized as ≤10 (low group) and >10 (high group). Mitoses were counted in 10 high-power fields (HPF; magnification, x400) on slides stained using hematoxylin and eosin (H&E) and categorized as ≤10/10 HPF or >10/10 HPF. Assessments of immunohistochemical staining and mitoses were performed and confirmed by consensus decision by two pathologists (M.M. and K.K. or Y.T. and K.K.).
Statistical analysis. Statistical analysis was performed using the JMP software version 8 (SAS Institute, Cary, NC, USA). Statistical analysis to compare the two groups was performed using the Student's t-test, the χ 2 test and Fisher's exact test, as appropriate. The survival analyses were performed as Table I. Analysis of prognostic factors in 86 GBC patients.

Univariate analysis
Multivariate analysis  Fig. 1. The P-value

Discussion
The prognosis following surgery for GBC is markedly different according to the results for the T stage, as are the therapeutic strategies (25,26). Thus, studies for GBC according to T stage better reflect actual prognosis after surgery and provide more Table III. Analysis of prognostic factors in T3 GBC patients.  useful information for clinical treatment compared with studies on overall GBC. Generally, survival of patients with T1 lesions is particularly good and simple cholecystectomy with or without lymphadenectomy is thus widely accepted as sufficient for T1 lesions (27). By contrast, survival of patients with T4 lesions is extremely poor and chemotherapy or palliative therapy is typically appropriate, except in rare cases where en bloc resection of multiple organs is applicable. This study therefore focused on patients with T2 and T3 tumors. This study investigated correlations between survival after surgery and the status of CD8 + TIL, Ki-67 LI, p53 expression and MC as potential prognostic markers for GBC. However, results for these candidates were insufficient for use as markers, with the exception of the results for MC. Concerning CD8 + TIL, only one study that investigated the prognostic impact of CD8 + TIL, reporting that CD8 + TIL correlated with prolonged survival in univariate analysis was available (28). However, no prognostic impacts of surface or invasion front CD8 + TIL were observed in our cohort. Therefore, we consider the prognostic impact of CD8 + TIL in GBC to be controversial and suggest that further investigation is required. Several previous studies have reported no prognostic impact of p53 overexpression in GBC (29)(30)(31)(32)(33), whereas reports of poor prognosis with p53 overexpression are also available in the literature (34)(35)(36). Of note, the present study showed an association between p53 overexpression and favorable prognosis in T2 GBC. Taken together, the correlation between p53 and prognosis in GBC remains controversial, although gain of abnormalities in p53 protein is generally considered an early event in the progress of carcinogenesis and is likely to be the usual route for GBC development (37)(38)(39)(40). A previous study reported that patients with GBC showing high Ki-67 exhibited worse postoperative prognosis compared with those showing low Ki-67 (41). Several previous studies reported that Ki-67 LI of cancer cells did not correlate with patient survival, supporting our results (30,32,33), which is in agreement with findings of the present study. The possible reason for Ki-67 not correlating with survival, despite the MC significant correlation of MC with survival, is that Ki-67 LI involves the G1, S and G2 phases of the cell cycle, while MC only involves the mitotic phase and might therefore reflect the rapidity of cell proliferation more sensitively compared with Ki-67 LI.