Annals of Surgical Oncology Sign the Guestbook
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
10.1245/s10434-006-9188-4
Annals of Surgical Oncology 13:1633-1644 (2006)
© 2006 Society of Surgical Oncology
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Huang, H.-Y.
Right arrow Articles by Hsiung, C.-Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Huang, H.-Y.
Right arrow Articles by Hsiung, C.-Y.

Original Article

Immunohistochemical Expression of p16INK4A, Ki-67, and Mcm2 Proteins in Gastrointestinal Stromal Tumors: Prognostic Implications and Correlations with Risk Stratification of NIH Consensus Criteria

Hsuan-Ying Huang, MD1, Wen-Wei Huang, MD2, Ching-Nan Lin, MD3, Hock-Liew Eng, MD1, Shau-Hsuan Li, MD4, Chien-Feng Li, MD3, David Lu, MD5, Shih-Chen Yu, BS1 and Ching-Yeh Hsiung, MD6

1 Department of Pathology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan
2 Department of Family Medicine, Buddhist Dalin Tzu Chi General Hospital, Chiayi, Taiwan
3 Department of Pathology, Chi-Mei Foundation Medical Center, Tainan, Taiwan
4 Division of Oncology, Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan
5 Department of Pathology, Chi-Mei Foundation Hospital, Liouying Campus, Tainan, Taiwan
6 Department of Radiation Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, College of Medicine, Chang Gung University, 123 Ta-Pei Road, Niao-Sung, Kaohsiung County, Taiwan

Correspondence: Address correspondence and reprint requests to: Ching-Yeh Hsiung, MD; E-mail: a120600310{at}yahoo.com


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background: Inactivation of p16INK4A promotes G1/S progression of cell cycle. Mini-chromosome maintenance protein-2 (Mcm2), a novel cell proliferation marker, is known to better correlate with clinical outcomes than Ki-67 in many carcinomas. Since gastrointestinal stromal tumors (GISTs) sometimes remains challenging in prognostication, we analyzed the utility of these three markers in GISTs.

Methods: Immunohistochemistry was performed in tissue microarrays of 277 primary GISTs and correlated with NIH consensus criteria and clinical outcomes.

Results: The increment of NIH risk levels significantly correlated with increasing labeling indices (LI) of both Ki-67 (P <.001) and Mcm2 (P <.001) and loss of p16INK4A expression (P <.035). However, the latter aberration did occur in 23% of very low/low-risk GISTs. The relationship between Mcm2 and Ki-67 LIs could be modeled as linear (P <.001, r = 0.697), while Mcm2 LI was considerably higher (P <.001) with a stepwise escalation related to risk levels. Ki-67 LI >5% (P <.0001) and Mcm2 LI >10% (P <.0001) were strongly predictive of inferior disease-specific survival (DSS), while aberrant loss of p16INK4A only reached a trend (P = .0954). In multivariate analyses, independent adverse factors of DSS were high-risk category (RR = 16.93, P <.0001), metastatic disease (RR = 4.12, P = .0015), Ki-67 LI >5% (RR = 3.55, P = .001), and presence of epithelioid histology (RR = 2.17, P = .0308).

Conclusions: Prognostic efficacy of NIH consensus criteria is substantiated. P16INK4A deregulation can occur early in GIST tumorigenesis and marginally correlates with patient survival. Despite Ki-67 LI being an independent prognosticator, simultaneous detection of Mcm2 is recommended as a prognostic adjunct of GISTs, given its better sensitivity and stepwise escalation with increasing risk levels.

Key Words: p16INK4A • Ki-67 • Mcm2 • GIST • NIH consensus criteria


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Gastrointestinal stromal tumors (GISTs) constitute the most common primary mesenchymal tumors of the digestive tract and characteristically express CD117 in most cases.1–3 However, approximately 5% of histologically typical GISTs are CD117-negative,3,4 for which protein kinase C-theta (PKC-{theta}) has been recently considered as a helpful diagnostic adjunct.5,6 Central to the tumorigenesis of GISTs are gain-of-function mutations in c-KIT or alternative PDGFRA proto-oncogenes.1,3,4,7–10 The activating mutations of these tyrosine kinase receptor genes confer constitutive, ligand-independent kinase activity that initiates the development of GISTs.1,3,4,7–10 Despite new insights into the pathogenesis, GISTs have been considered as prognostically treacherous with a wide clinical spectrum, ranging from very indolent to highly aggressive behavior.1–3,11,12 Up to 30% of newly diagnosed GISTs are overtly malignant or have features connoting a high malignant potential,3 while a minority of small and mitotically inactive GISTs may still behave aggressively.1,2,11 Recently, in a workshop convened by American National Institutes of Health (NIH), a consensus scheme was proposed to stratify patients with GISTs into four risk categories, based on the combination of mitotic rate and tumor size.1 Few, if any, studies have evaluated GISTs of sufficiently large sample size to substantiate the prognostic efficacy of this consensus system.9 Therefore, it is desirable to search objective prognostic markers to correlate with the actual clinical outcomes of GISTs and the risk categories determined by NIH consensus scheme.

During stepwise evolution of cancer, sequential deregulation of multiple cell cycle regulators represents a common mode that promotes progression and survival advantage of aggressive tumor subsets. These genetic alterations can confer adverse prognostic impact that is particularly striking in cancers with disease-specific initiating oncogenic events and simple karyotypes, such as translocation-associated sarcomas.13 GISTs can also be segregated into this type of neoplasms by virtue of the initiating c-KIT or PDGFRA mutations and relatively simple cytogenetic aberrations in the early stage of tumorigenesis.13

P16INK4A is key regulator at the G1-S checkpoint of the cell cycle, and alteration of its function plays a critical role in tumorigenesis.14,15 In a recent comparative genomic hybridization (CGH) study on GISTs, the chromosomal deletion of 9p21, spanning the p16INK4A gene locus, was found associated with malignant and metastatic diseases in GISTs.16 However, few series addressing the prognostic role of p16INK4A immunohistochemical expression in GISTs provided contradictory results.17–21 For instance, Schneider-Stock et al.19 found that aberrant loss of p16INK4A expression was predictive of poor patient survival, but Nakamura et al.21 failed to validate its prognostic implication in Japanese patients. These discrepant data raised a practical concern about whether, for prognostication, p16INK4A immunostain can be indiscriminately used as a surrogate marker for various inactivating mechanisms of p16INK4A gene without validation by a large, multivariate study.

The proliferation of tumor cells is highly related to the rate of DNA synthesis and may provide prognostic information. Although Ki-67, the traditional marker of growth fraction, was previously shown to correlate with aggressiveness of GIST,21–24 conflicting opinions challenged whether its expression level provides better prognostic utility than mitotic rate.2,25 This might be mostly ascribed to its suboptimal reproducibility in assessing tumors with a low proliferation capacity.26–28 It has also been suggested that Ki-67 plays a role in the ribosome biosynthesis rather than being directly responsible for cell proliferation.29 Therefore, an ideal biomarker for evaluating proliferation should be essential for genomic replication and exhibit a broader range of expression that allows a quick, objective assessment of individual case. The minichromosome maintenance proteins (Mcm), consisting of six members (Mcm2–7), are a family of highly conserved proteins, which form a hexameric complex for regulating eukaryotic DNA duplication.26,30 Mcm2 has been found superior to Ki-67 in defining proliferative compartments of tumor cells and associated with prognosis in a wide variety of carcinomas.26–28,31 In contrast, remarkably little information regarding its prognostic role is available in mesenchymal neoplasms. Recently, we have reported that high Mcm2 expression is an adverse factor of metastasis-free survival in myxofibrosarcomas on univariate analysis.32 Nevertheless, there has been thus far no investigation examining the role of Mcm2 in GISTs and its correlation with NIH consensus criteria. By using tissue microarray (TMA)-based immunohistochemistry, we aimed at analyzing the expression patterns, associations with NIH risk categories, and prognostic implications of p16INK4A, Ki-67, and Mcm2 proteins in a large cohort of 277 GISTs.


   MATERIAL AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Study Group, Clinicopathological Evaluation, and Follow-up
We retrieved 339 cases of gastrointestinal mesenchymal tumors consecutively resected between 1985 and 2002 from the archives of two tertiary medical centers in Southern Taiwan. After histopathological review, other entities morphologically simulating GISTs were excluded, including 8, 12, 7, and 5 cases of mesenteric or pelvic desmoid-type fibromatoses, gastric schwannomas, gastric inflammatory fibroid polyps, and leiomyomas of the esophagus or colorectal muscularis mucosae, respectively. A total of 307 GISTs with typical histology were verified by strong reaction to c-kit (DAKO, Carpenteria, CA, USA) and/or positive PKC-{theta} staining (27, BD Pharmagen, San Diego, CA, USA) for a minor subset with no or weak c-KIT expression.5,6 The following clinicopathological variables were evaluated and recorded, including age, sex, location, size, histologic type, cellularity, nuclear pleomorphism, mitotic count per 50 high power fields (HPF), and risk category of NIH consensus criteria. Institutional review boards of both centers approved retrospective clinical data collection and procurement of archival tissue blocks. Clinical follow-up was obtained from the medical charts, the hospital tumor registries, or direct contact with patients themselves or family. Follow-up data were available in 277 patients, who formed the study cohort for statistical analyses. The end of follow-up time was as of 30 June 2005, and the duration ranged from 6 to 231 months (mean: 57.3; median: 40.7). Only two patients with gastric GISTs and another two with small intestinal lesions had recently received imatinib after July 2004 for tumor relapses. In addition, 16 patients had hepatic or peritoneal metastases at the initial operation that were all excised together with the primary tumors.

Construction of TMAs and Immunohistochemistry
For immunostains of p16INK4A, Ki-67, and Mcm2, a hematoxylin/eosin-stained slide was recut from each donor block and used as a template to define representative areas for constructing TMA blocks. To circumvent tissue heterogeneity and to increase interpretable cases, six tissue cylinders (0.6 mm in diameter) for each specimen were punched from selected areas using a precision instrument (Beecher Instruments, Silver Spring, MD, USA) and arrayed into eight recipient blocks. In addition, we also arrayed three each carcinomas of the uterine cervix and breast for the purposes of orientation and external positive controls in each TMA. These cervical and breast carcinomas were previously stained positive for the p16INK4A and Mcm2 proteins, respectively. Sections of TMA blocks were cut onto an adhesive-coated glass slide system (instrumedicus, Hackensack, NJ, USA) at 3-µm thickness.

The primary antibodies used for prognostic evaluation were as follows: p16INK4A (6H12, 1:20, Novocastra), Ki-67 (MIB-1, 1:100, DAKO), and Mcm2 (CRCT2.1, 1:25, Novocastra). For antigen retrieval, slides were heated by microwave in 10 mM citrate buffer (pH 6) for 7 min. Primary antibodies were detected using the ChemMate DAKO EnVision kit (DAKO, K5001). The slides were incubated with the secondary antibody for 30 min and developed with 3,3-diaminobenzidine for 5 min. Slides were then counterstained with Gill’s Hematoxylin. Coverslips were applied with CureMount mounting medium (Instrumedicus, Hackensack, NJ, USA). Incubation without the primary antibody was used as a negative control.

Assessment of Immunohistochemical Staining
Two pathologists (CFL, CNL) blinded to clinicopathological data and patient outcomes independently assessed the TMA slides. All cells of each tissue cylinder included in the TMAs were evaluated, and the percentage of tumor cells with definite moderate to intense nuclear immunoreactivity was recorded. Only cases containing two or more preserved tissue cores were scored, and scores from multiple cores for each marker in the same patient were averaged to obtain a mean labeling index (LI). Whenever the two pathologists disagreed (e.g., difference in mean LI >10%), the third pathologist (HYH) reviewed the TMA slides and the majority decision was applied. By testing a series of different values (see statistical methods), the cutoffs of mean LIs to define high expression of Ki-67 and Mcm2 were determined as >5% and >10% of tumor nuclei stained, respectively. According to the cutoff previously adopted by Powell et al.,33 complete loss of p16INK4A nuclear labeling was considered as aberrant deregulation. Any nuclear labeling of tumor cells, with or without cytoplasmic staining, was considered positive, with non-tumoral stromal or inflammatory cells serving as internal controls.

Statistical Analyses
Statistical analyses were performed using the SPSS 10 software package. Associations or comparisons between various parameters were assessed using Pearson correlation coefficient, {chi}2, one-way ANOVA, and Student’s or Paired t-test as appropriate. The endpoint analyzed was disease-specific survival (DSS), for which patients were censored at the time of their last clinical follow-up appointment or at their date of death not related to GISTs. A series of cutoff values were tested for continuous variables, such as age and mean LIs of Ki-67 and Mcm2. The cutoffs giving the best P values were adopted to construct Kaplan–Meier curves and compare prognostic differences by univariate log-rank test. In a stepwise forward fashion, parameters with P values (0.1 at univariate level were in principle entered into Cox regression model to analyze their relative prognostic importance. However, as component factors of NIH consensus scheme, tumor size and mitoses were not introduced in multivariate analyses. For all analyses, two-sided tests of significance were used with P <.05 considered significant.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Clinicopathological Findings and Follow-up
The median age for the study group of 277 patients, 131 males and 146 females, was 62 years (range, 24–85 years), with 71 patients (25.6%) aged greater double equals70 years. There were 164 gastric and 113 intestinal tumors, including 87 and 26 located in the small bowel and rectum, respectively. The tumor size ranged from 0.6 to 30 cm (median, 6 cm), with 112 (40%), 109 (39%), and 56 (21%) cases being less double equals5 cm, >5–10 cm, and >10 cm, respectively. Histologically, 216 tumors were pure spindle type (78%), and 61 tumors (22%) showed at least focal epithelioid histology. High cellularity, defined as nuclear overlapping with scanty visible intervening stroma, was observed in 52 cases (19%), while prominent nuclear pleomorphism was present in 24 cases (9%). Mitotic counts revealed a wide variation from 0 to 184 (median, 3) per 50 HPF and were 0–5/50 HPF in 188 cases (68%), 6–10/50 HPF in 40 (14%), and >10/50 HPF in 49 (18%). Based on the NIH consensus scheme, 13 (4.5%), 80 (29%), 85 (30.5%), and 99 (36%) cases were classified as very low-risk, low-risk, intermediate-risk, and high-risk groups, respectively (Fig. 1A–CGo). When last seen, 196 patients were alive without evidence of disease (70.7%), 21 were alive with relapsed disease (7.6%), 40 patients died of GISTs (14.4%: gastric, n = 20; intestinal, n = 20), and 20 died of unrelated causes (7.2%). Of the 261 patients with primary localized GISTs, 49 experienced tumor relapses during the follow-up period, including local recurrences in 23 and metastases in 32 to the liver (n = 23), peritoneum (n = 7), lymph node (n = 1), and bone (n = 1). For survivors at the time of analysis, the mean and median durations of follow-up were 64.7 and 50.1 months for the entire cohort, 63.4 and 50.6 months for gastric GISTs, and 66.8 and 57.5 months for intestinal lesions. In addition, these durations (mean and median) among survivors were 64.9 and 56.0 months, 71.0 and 61.3 months, and 56.1 and 38.2 months for the very low/low-risk, intermediate-risk, and high-risk cases, respectively.


Figure 1
View larger version (115K):
[in this window]
[in a new window]

  		FIG. 1. Representative examples of GIST tissue microarray, stained with hematoxylin and eosin, p16INK4A, Ki-67, and Mcm2. Photomicrographs of GISTs classified as low-risk (A), intermediate-risk (B), and high-risk (C) according to NIH consensus criteria with increasing cellularity and mitotic rate. P16INK4A immunostaining showed positive staining in a low-risk GIST (D) and aberrant loss of expression in intermediate-risk (E) and high-risk counterparts (F). Ki-67 labeling was sparse in the low-risk (G) and intermediate-risk (H) GISTs but preferentially expressed in the high-risk GIST (I). Mcm2 staining showed a gradual stepwise increase in labeling index in low-risk (J), intermediate-risk (K), and high-risk (L) GISTs.

 
Immunohistochemical Findings
TMA-based immunohistochemical data were interpretable and scored for p16INK4A, Ki-67, and Mcm2 in 268, 230, and 245 cases, respectively. The nuclear expression of p16INK4A (Fig. 1D–FGo) displayed a wide variation in LI from 0 to 100% (mean, 13.49 ± 21.56%), and aberrant loss of p16INK4A was identified in 80 cases (29.6%). To validate the TMA findings for GISTs with complete negativity, whole tissue sections for p16INK4A immunohistochemistry were also performed for cases proportionally selected from various risk groups and yielded concordant results (data not shown). Ki-67 LI (Fig. 1G–IGo) ranged from 0 to 35% (mean, 3.79 ± 5.69%), and high expression (LI >5%) was identified in 45 cases (19.6%). As compared to Ki-67, Mcm2 LI (Fig. 1J–LGo) displayed a considerably broader spectrum, ranging from 0 to 73% (mean, 8.17 ± 12.59%), and 56 cases (22.9%) exhibited Mcm2 overexpression (LI >10%).

Associations and Comparisons Among Variables
The correlations of NIH risk categories with clinicopathological factors and immunohistochemical expression are summarized in Table 1Go. The increment of risk levels was significantly associated with epithelioid histology (P <.001), high cellularity (P <.001), and prominent nuclear pleomorphism (P = .012). In addition, almost all cases (15/16, 94%) with liver or peritoneal metastases at presentation were high-risk (P <.001), except the remaining one classified as intermediate-risk. However, risk levels were unrelated to gender, age, and tumor location. GISTs of higher risk categories also showed significantly increasing tumor size (P <.001), mitotic activity (P <.001), Ki-67 LI (P <.001, Fig. 2AGo), and Mcm2 LI (P <.001, Fig. 2AGo), while Mcm2 LI tended to display a better stepwise escalation with increasing risk level as compared to Ki-67 (Fig. 2AGo). The relationship between Mcm2 and Ki-67 LIs could be modeled as linear (P <.001, r = 0.697, Fig. 2BGo), while Mcm2 LI was considerably higher for both the entire cohort and various risk categories (Fig. 2AGo). Furthermore, the LIs of Mcm2 and Ki-67 were both found positively correlated with tumor size (P <.001, r = 0.270 for Mcm2; P <.001, r = 0.287 for Ki-67) and mitotic count (P <.001, r = 0.343 for Mcm2; P <.001, r = 0.287 for Ki-67). Although loss of p16INK4A staining preferentially affected intermediate-risk and high-risk cases (P = .035), this aberration did occur in approximately one-fourth (23%) of very low/low-risk GISTs, suggesting its role in early tumorigenesis. However, neither Ki-67 (P = .743) nor Mcm2 (P = .889) showed significant difference in LIs between p16INK4A-postive and p16INK4A-negative subgroups.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Clinicopathological data of 277 patients in relation to the risk categories of NIH scheme
 

Figure 2
View larger version (16K):
[in this window]
[in a new window]

  		FIG. 2. A Comparisons between NIH risk categories and labeling indices of Mcm-2 (green) and Ki-67 (blue) in GISTs. For each risk category and the entire cohort, Mcm-2 is detected in substantially more tumor cells than Ki-67 (P = .021 for the very low/low-risk group; P <.001 for the intermediate-risk, high-risk, and whole study groups). Both makers are positively related to increasing risk levels, while Mcm2 LI shows a better stepwise escalation. Line median; box interquartile range; whiskers expected range; circles outliers; stars extremes. B A scatter plot shows a close correlation between the labeling indices of Mcm2 and Ki-67 (P <.001, r = 0.697).

 
Univariate Survival Analyses
The cumulative 5-year rates of DSS for very low/low-risk, intermediate-risk, high-risk, and all 277 cases were 100, 95.4, 49.8, and 80.5%, respectively. Correlations of clinicopathological and immunohistochemical factors to DSS are shown in Table 2Go. and Fig. 3A–JGo. With respect to DSS, metastasis at presentation (P <.0001, Fig. 3AGo), larger tumor size (P <.0001, Fig. 3BGo), higher mitotic count (P <.0001, Fig. 3CGo), higher NIH risk levels (P <.0001, Fig. 3DGo), presence of epithelioid histology (P <.0001, Fig. 3EGo), high cellularity (P = .0004), Ki-67 LI >5% (P <.0001, Fig. 3FGo), and Mcm2 LI >10% (P <.0001, Fig. 3GGo) all strongly correlated with poor patient survival. In addition, old age (greater double equals70 years, P = .0363) was also identified as a significant adverse factor, while aberrant loss of p16INK4A expression only showed an expected trend toward significance for inferior DSS (P = .0954, Fig. 3HGo).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Univariate analyses for disease-specific survival
 

Figure 3
View larger version (17K):
[in this window]
[in a new window]

  		FIG. 3. Kaplan–Meier plots to predict disease-specific survival for the entire study cohort according to status at presentation (A), tumor size (B), mitotic count (C), NIH consensus scheme (D), histologic type (E), Ki-67 labeling index (F), Mcm2 labeling index (G), and p16INK4A expression pattern (H). In addition, Ki-67 (I) and Mcm2 (J) labeling indices are still significantly predictive of disease-specific survival for the high-risk subset.

 
Given no or very few outcome events in the very low/low-risk and intermediate-risk GISTs, subset analysis were only preformed for high-risk cases. We found that the following parameters still remained significant to predict inferior DSS, including metastases at presentation (P = .0003), Ki-67 LI >5% (P <.0001, Fig. 3IGo), Mcm2 LI>10% (P = .0045, Fig. 3JGo), higher mitotic count (P = .0201), old age (P = .0279), and epithelioid histology (P = .0364). However, high cellularity and larger tumor size lost statistical significance.

Multivariate Survival Analyses
As for the entire cohort, high-risk category of NIH scheme was the strongest independent negative factor of DSS [risk ratio (RR) = 16.93, P <.0001), followed by metastatic disease at presentation (RR = 4.12, P = .0015), Ki-67 LI >5% (RR = 3.55, P = .001), and presence of epithelioid histology (RR = 2.17, P = .0308). Nevertheless, high cellularity, old age, Mcm2 LI >10%, and aberrant loss of p16INK4A were not statistically independent. With respect to the high-risk subset, only Ki-67 LI>5% (RR = 4.01, P = .0005) and metastatic disease at presentation (RR = 3.94, P = .0017) independently correlated with worse DSS. These findings indicated that NIH scheme represented the most important prognosticator, as compared to other clinicopathological and immunohistochemical markers. In addition, Mcm2 LI, albeit significant by univariate log-rank test, was not as effective as Ki-67 LI in multivariate comparison.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The development of a reliable method for GIST prognostication is a major aim of clinical management, given the relevance of proper selection of eligible patients for imatinib mesylate therapy.34 Numerous studies have attempted to delineate criteria to distinguish benign from malignant GISTs and led to the development of various classification schemes.1,2,11,12 The NIH consensus scheme proposed by Fletcher et al.1 did not consider the influence of tumor location, although other investigators had reported a more aggressive course in intestinal lesions.7,10,35 This contention appeared to be an area of divergence in viewpoints. Based on the extremely large cohort of over 2,600 GISTs, Miettinen et al.11,36 recently asserted that the tumor-related mortality of small intestinal GISTs was two times higher than gastric GISTs. Although others and we could not reproduce this finding, it must be added that none of these series enrolled more than 300 cases,19,21,37 including one population-based study.37 As compared to the series of Miettinen et al.,11,36 other potential sources of bias in our cohort might be ascribed to the relatively shorter time of follow-up and a smaller proportion of high-risk cases in intestinal GISTs (37% vs. 42%).11,36 Nevertheless, we have substantiated the robust prognostic efficacy of NIH consensus scheme in multivariate analysis by identifying high-risk category as the strongest independent adverse factor of DSS, with a 16.93-fold increased risk. Despite assessing different endpoints, this figure was in keeping with the finding of Martin et al.,9 who reported a 12.36-fold increased risk of relapse-free survival for high-risk GISTs. Our data regarding other classical prognosticators were generally consistent with prior reports, showing that large tumor size, high mitotic count, and metastasis at presentation were constantly predictive of poor outcomes for the entire spectrum of GISTs.2,3,19,38 When considering high-risk cases alone, we found that the prognostic value of tumor size was no longer significant, unlike mitotic activity and status at presentation. However, in multivariate analysis for this subset, metastatic disease represented the sole independent classical parameter to predict inferior DSS, in parallel with high Ki-67 LI. As to hypercellularity and old age, the predictive power of these two parameters appeared conflicting among earlier studies but proved significant in our univariate analysis.2,9,18,23,24,35,39 For histologic type, it had been recognized as prognostically relevant for GISTs but almost only at the univariate level.39–41 However, pure spindle histology was independently associated with increased DSS in the current series.

The p16 INK4A protein, encoded by the INK4A gene located at 9p21, is a critical negative regulator of G1/S progression, which inhibits CDK4/CDK6 kinases and thus prevents phosphorylation of the Rb protein.14,15 Abrogation of the p16INK4A gene has been documented in limited series on GISTs through various inactivating mechanisms, including homozygous deletion, promoter hypermethylation, loss of heterozygosity, and point mutation with varying frequencies.17,18,20,42 Although these aberrations can theoretically result in down-regulation of p16INK4A protein, it seems questionable whether immunohistochemical scoring can reliably aid in assessment of mutant protein for GIST prognostication. Apart from varied clones of antibodies tested, this may be at least in part ascribed to the primary difficulty in evaluating reduction of p16INK4A expression within tissues in an already low abundance of proteins,14 thereby resulting in heterogeneous methods to define tumors with p16INK4A deregulation.17–19,33,43,44 By using a cutoff at 20% positivity, Schneider-Stock et al.19 demonstrated that aberrant underexpression of p16INK4A was an independent adverse predictor for GISTs and might represent a reliable marker for routine evaluation of aggressiveness. However, Nakamura et al.21 did not prove this alteration as prognostically valid by adopting 50% reactivity as the cutoff. By considering only complete negativity as aberrant, we at best found that p16INK4A-postive GISTs approached but did not reach significance to have an improved DSS rate (P = .0954), while other cutoffs, including 20, 50, and 80% positivity, etc., did not provide better predictive values. Although G1/S checkpoint regulation is supposed to be deranged in p16INK4A-negative tumors with a theoretically consequent increase in cell proliferation, relatively few studies had shown this association in clinical tumor samples.43–46 Likewise, we could not significantly substantiate higher expression of Ki-67 and Mcm2 in p16INK4A-negative GISTs, which was in agreement with the observation of Nakamura et al.21

Despite p16 INK4A deregulation preferentially affecting GISTs of higher risk-categories in our series (P = .035), we agree with Schneider-Stock et al.19 in the notion that p16INK4A deregulation can indeed occur at a relatively early stage of GIST tumorigenesis. For very low/low-risk GISTs, Schneider-Stock et al.19 and we reported aberrant loss of p16 INK4A in 43 and 23% of cases, respectively, which clearly argued against two much smaller series, stating that p16INK4A loss was exclusively detected in "malignant" GISTs.17,18 Accordingly, it is tempting to speculate that the timing of p16INK4A gene abrogation in the evolution of GISTs may be different among various inactivating mechanisms, thereby culminating in different clinical impact. On the one hand, promoter hypermethylation of p16INK4A gene had been detected in both precursor lesions and tumors of pulmonary non-small cell carcinomas,47 gastric adenocarcinomas,48 and oral squamous cell carcinomas,49 etc. These studies suggested that epigenetic silencing of p16INK4A gene is a critical early event that precedes apparent malignant changes in evolution of certain types of cancers.47–49 On the other hand, chromosomal deletion at 9p21, spanning the locus of p16INK4A gene, was a late cytogenetic alteration preferentially affecting advanced GISTs,3,8,16 which could not cogently explain a significant proportion of very low/low-risk GISTs with aberrant p16INK4A expression, as seen in Schneider-Stock’s report and our data.19 These contentions allude to whether epigenetic hypermethylation chiefly accounts for p16INK4A deregulation in very low/low-risk GISTs, which is not necessarily equivalent to homozygous deletion in terms of prognostic impact. In the above context, it necessitates a great precaution to apply p16INK4A immunostaing for risk assessment of GISTs until standardized scoring criteria are established by integrating comprehensive analyses of methylation and genetic status of p16INK4A gene.

Mcm proteins stand at the convergence point of many signaling pathways involving cell proliferation and have recently been promoted as markers of cancer screening, surveillance, and prognosis.26–28,30,31 Their expression is seen during all phases of cell cycle, including early G1 phase, and may thus better represent the rate of cell proliferation.28,30 To assemble the pre-replication complex for initiating DNA synthesis, Cdt1 and Cdc6, two replication cofactors, are required in G1 phase to load Mcm proteins onto chromatin.28,30 This process is coined as "licensing", which ensures that DNA replicates only once in the subsequent S-phase during each cell cycle.28,30 One main scope of this study was to compare the prognostic utility of Ki-67 versus Mcm2 in GISTs. Several groups indicated that Ki-67 LI is useful in predicting the aggressive behavior of GISTs,22–24 while others could not consistently reproduce its prognostic superiority over conventional histologic parameters.25 According to our results, LIs of Ki-67 and Mcm2 were both positively associated with increasing NIH risk level and strongly predictive of DSS by log-rank tests. Nevertheless, only expression of Ki-67, instead of Mcm2, retained prognostic value in multivariate comparison, which held true for both the entire cohort and the high-risk subset. Despite a good linear correlation between Mcm2 and Ki-67, a considerably higher proportion of proliferative cells were detected using Mcm2 antibody, probably reflecting cells in the early G1 phase that failed to be labeled by Ki-67. Accordingly, we consider that Mcm2, albeit not independent, can still be exploited as a helpful prognostic adjunct for GISTs, given its higher sensitivity than Ki-67 and a more stepwise escalation with increasing risk level.

In summary, the robustness of NIH consensus criteria for GIST prognostication is reaffirmed in this series, which is consistent with recent studies adopting this scheme,9,21 with rare exceptions.19 An approximately 17-fold increased risk of tumor-associated mortality is identified in high-risk GISTs. Of the three ancillary immunohistochemical markers tested, we only demonstrate that high Ki-67 expression to be an independent adverse factor, but it does not provide superior prognostic utility over NIH consensus scheme. Furthermore, assessment of Ki-67 LI is sometimes limited by its suboptimal sensitivity in slower-proliferating tumors, as shown by sparse reactivity in very low/low-risk and even intermediate-risk GISTs. In this regard, simultaneous detection of Mcm2 expression may provide a more objective and easier judgment in predicting clinical aggressiveness. Nevertheless, for p16INK4A immunostaing as a surrogate prognostic marker, it should be applied very cautiously in GISTs before standardized methodologies are established on the basis of thorough elucidation of prognostic implications of various molecular inactivating mechanisms.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Multivariate analyses for disease-specific survival
 

   ACKNOWLEDGMENTS
 
This work was supported in part by grants from National Science Council, Taiwan (NSC93-2320-B-182A-011) and Chang Gung Memorial Hospital (CMRPG83019).

Received for publication June 5, 2006. Accepted for publication June 5, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Fletcher CD, Berman JJ, Corless C, et al. Diagnosis of gastrointestinal stromal tumors: a consensus approach. Hum Pathol 2002; 33(5):459–65.[CrossRef][Medline]
  2. Miettinen M, El-Rifai W, L HLS, Lasota J. Evaluation of malignancy and prognosis of gastrointestinal stromal tumors: a review. Hum Pathol 2002; 33(5):478–83.[CrossRef][Medline]
  3. Corless CL, Fletcher JA, Heinrich MC. Biology of gastrointestinal stromal tumors. J Clin Oncol 2004; 22(18):3813–25.[Abstract/Free Full Text]
  4. Corless CL, Schroeder A, Griffith D, et al. PDGFRA mutations in gastrointestinal stromal tumors: frequency, spectrum and in vitro sensitivity to imatinib. J Clin Oncol 2005; 23(23):5357–64.[Abstract/Free Full Text]
  5. Duensing A, Joseph NE, Medeiros F, et al. Protein kinase C theta (PKCtheta) expression and constitutive activation in gastrointestinal stromal tumors (GISTs). Cancer Res 2004; 64(15):5127–131.[Abstract/Free Full Text]
  6. Blay P, Astudillo A, Buesa JM, et al. Protein kinase C theta is highly expressed in gastrointestinal stromal tumors but not in other mesenchymal neoplasias. Clin Cancer Res 2004; 10(12 Pt 1):4089–95.[Abstract/Free Full Text]
  7. Antonescu CR, Sommer G, Sarran L, et al. Association of KIT exon 9 mutations with nongastric primary site and aggressive behavior: KIT mutation analysis and clinical correlates of 120 gastrointestinal stromal tumors. Clin Cancer Res 2003; 9(9):3329–37.[Abstract/Free Full Text]
  8. Heinrich MC, Rubin BP, Longley BJ, Fletcher JA. Biology and genetic aspects of gastrointestinal stromal tumors: KIT activation and cytogenetic alterations. Hum Pathol 2002; 33(5):484–95.[CrossRef][Medline]
  9. Martin J, Poveda A, Llombart-Bosch A, et al. Deletions affecting codons 557–558 of the c-KIT gene indicate a poor prognosis in patients with completely resected gastrointestinal stromal tumors: a study by the Spanish Group for Sarcoma Research (GEIS). J Clin Oncol 2005; 23(25):6190–8.[Abstract/Free Full Text]
  10. Lasota J, Kopczynski J, Sarlomo-Rikala M, et al. KIT 1530ins6 mutation defines a subset of predominantly malignant gastrointestinal stromal tumors of intestinal origin. Hum Pathol 2003; 34(12):1306–12.[CrossRef][Medline]
  11. Miettinen M, Sobin LH, Lasota J. Gastrointestinal stromal tumors of the stomach: a clinicopathologic, immunohistochemical, and molecular genetic study of 1765 cases with long-term follow-up. Am J Surg Pathol 2005; 29(1):52–68.[CrossRef][Medline]
  12. Franquemont DW. Differentiation and risk assessment of gastrointestinal stromal tumors. Am J Clin Pathol 1995; 103(1):41–7.[Medline]
  13. Borden EC, Baker LH, Bell RS, et al. Soft tissue sarcomas of adults: state of the translational science. Clin Cancer Res 2003; 9(6):1941–56.[Abstract/Free Full Text]
  14. Ruas M, Peters G. The p16INK4a/CDKN2A tumor suppressor and its relatives. Biochim Biophys Acta 1998; 1378(2):F115–77.[Medline]
  15. Huang HY, Illei PB, Zhao Z, et al. Ewing sarcomas with p53 mutation or p16/p14ARF homozygous deletion: a highly lethal subset associated with poor chemoresponse. J Clin Oncol 2005; 23(3):548–58.[Abstract/Free Full Text]
  16. El-Rifai W, Sarlomo-Rikala M, Andersson LC, Knuutila S, Miettinen M. DNA sequence copy number changes in gastrointestinal stromal tumors: tumor progression and prognostic significance. Cancer Res 2000; 60(14):3899–903.[Abstract/Free Full Text]
  17. Sabah M, Cummins R, Leader M, Kay E. Loss of heterozygosity of chromosome 9p and loss of p16INK4A expression are associated with malignant gastrointestinal stromal tumors. Mod Pathol 2004; 17(11):1364–71.[CrossRef][Medline]
  18. Ricci R, Arena V, Castri F, et al. Role of p16/INK4a in gastrointestinal stromal tumor progression. Am J Clin Pathol 2004; 122(1):35–43.[CrossRef][Medline]
  19. Schneider-Stock R, Boltze C, Lasota J, et al. Loss of p16 protein defines high-risk patients with gastrointestinal stromal tumors: a tissue microarray study. Clin Cancer Res 2005; 11(2 Pt 1):638–45.[Abstract/Free Full Text]
  20. Schneider-Stock R, Boltze C, Lasota J, Miettinen M, Peters B, Pross M, Roessner A, Gunther T. High prognostic value of p16INK4 alterations in gastrointestinal stromal tumors. J Clin Oncol 2003; 21(9):1688–97.[Abstract/Free Full Text]
  21. Nakamura N, Yamamoto H, Yao T, et al. Prognostic significance of expressions of cell-cycle regulatory proteins in gastrointestinal stromal tumor and the relevance of the risk grade. Hum Pathol 2005; 36(7):828–37.[CrossRef][Medline]
  22. Carrillo R, Candia A, Rodriguez-Peralto JL, Caz V. Prognostic significance of DNA ploidy and proliferative index (MIB-1 index) in gastrointestinal stromal tumors. Hum Pathol 1997; 28(2):160–5.[CrossRef][Medline]
  23. Hasegawa T, Matsuno Y, Shimoda T, Hirohashi S. Gastrointestinal stromal tumor: consistent CD117 immunostaining for diagnosis, and prognostic classification based on tumor size and MIB-1 grade. Hum Pathol 2002; 33(6):669–76.[CrossRef][Medline]
  24. Rudolph P, Gloeckner K, Parwaresch R, Harms D, Schmidt D. Immunophenotype, proliferation, DNA ploidy, and biological behavior of gastrointestinal stromal tumors: a multivariate clinicopathologic study. Hum Pathol 1998; 29(8):791–800.[CrossRef][Medline]
  25. Wong NA, Young R, Malcomson RD, et al. Prognostic indicators for gastrointestinal stromal tumours: a clinicopathological and immunohistochemical study of 108 resected cases of the stomach. Histopathology 2003; 43(2):118–26.[CrossRef][Medline]
  26. Dudderidge TJ, Stoeber K, Loddo M, Atkinson G, Fanshawe T, Griffiths DF, Williams GH. Mcm2, Geminin, and KI67 define proliferative state and are prognostic markers in renal cell carcinoma. Clin Cancer Res 2005; 11(7):2510–7.[Abstract/Free Full Text]
  27. Gonzalez MA, Pinder SE, Callagy G, et al. Minichromosome maintenance protein 2 is a strong independent prognostic marker in breast cancer. J Clin Oncol 2003; 21(23):4306–13.[Abstract/Free Full Text]
  28. Tachibana KE, Gonzalez MA, Coleman N. Cell-cycle-dependent regulation of DNA replication and its relevance to cancer pathology. J Pathol 2005; 205(2):123–9.[CrossRef][Medline]
  29. MacCallum DE, Hall PA. The location of pKi67 in the outer dense fibrillary compartment of the nucleolus points to a role in ribosome biogenesis during the cell division cycle. J Pathol 2000; 190(5):537–44.[CrossRef][Medline]
  30. Blow JJ, Dutta A. Preventing re-replication of chromosomal DNA. Nat Rev Mol Cell Biol 2005; 6(6):476–486.[CrossRef][Medline]
  31. Kato H, Miyazaki T, Fukai Y, et al. A new proliferation marker, minichromosome maintenance protein 2, is associated with tumor aggressiveness in esophageal squamous cell carcinoma. J Surg Oncol 2003; 84(1):24–30.[CrossRef][Medline]
  32. Huang HY, Kang HY, Li CF, Eng HL, Chou SC, Lin CN, Hsiung CY. Skp2 overexpression is highly representative of intrinsic biological aggressiveness and independently associated with poor prognosis in primary localized myxofibrosarcomas. Clin Cancer Res 2006; 12(2):487–98.[Abstract/Free Full Text]
  33. Powell EL, Leoni LM, Canto MI, et al. Concordant loss of MTAP and p16/CDKN2A expression in gastroesophageal carcinogenesis: evidence of homozygous deletion in esophageal noninvasive precursor lesions and therapeutic implications. Am J Surg Pathol 2005; 29(11):1497–504.[CrossRef][Medline]
  34. Saund MS, Demetri GD, Ashley SW. Gastrointestinal stromal tumors (GISTs). Curr Opin Gastroenterol 2004; 20(2):89–94.[CrossRef][Medline]
  35. Emory TS, Sobin LH, Lukes L, Lee DH, O’Leary TJ. Prognosis of gastrointestinal smooth-muscle (stromal) tumors: dependence on anatomic site. Am J Surg Pathol 1999; 23(1):82–7.[CrossRef][Medline]
  36. Miettinen M, Makhlouf H, Sobin LH, Lasota J. Gastrointestinal stromal tumors of the jejunum and ileum: a clinicopathologic, immunohistochemical, and molecular genetic study of 906 cases before imatinib with long-term follow-up. Am J Surg Pathol 2006; 30(4):477–89.[CrossRef][Medline]
  37. Nilsson B, Bumming P, Meis-Kindblom JM, et al. Gastrointestinal stromal tumors: the incidence, prevalence, clinical course, and prognostication in the preimatinib mesylate era—a population-based study in western Sweden. Cancer 2005; 103(4):821–9.[CrossRef][Medline]
  38. DeMatteo RP, Lewis JJ, Leung D, Mudan SS, Woodruff JM, Brennan MF. Two hundred gastrointestinal stromal tumors: recurrence patterns and prognostic factors for survival. Ann Surg 2000; 231(1):51–8.[CrossRef][Medline]
  39. Cooper PN, Quirke P, Hardy GJ, Dixon MF. A flow cytometric, clinical, and histological study of stromal neoplasms of the gastrointestinal tract. Am J Surg Pathol 1992; 16(2):163–70.[Medline]
  40. Feakins RM. The expression of p53 and bcl-2 in gastrointestinal stromal tumours is associated with anatomical site, and p53 expression is associated with grade and clinical outcome. Histopathology 2005; 46(3):270–9.[CrossRef][Medline]
  41. Singer S, Rubin BP, Lux ML, Chen CJ, Demetri GD, Fletcher CD, Fletcher JA. Prognostic value of KIT mutation type, mitotic activity, and histologic subtype in gastrointestinal stromal tumors. J Clin Oncol 2002; 20(18):3898–905.[Abstract/Free Full Text]
  42. Perrone F, Tamborini E, Dagrada GP, et al. 9p21 locus analysis in high-risk gastrointestinal stromal tumors characterized for c-kit and platelet-derived growth factor receptor alpha gene alterations. Cancer 2005; 104(1):159–69.[CrossRef][Medline]
  43. Straume O, Sviland L, Akslen LA. Loss of nuclear p16 protein expression correlates with increased tumor cell proliferation (Ki-67) and poor prognosis in patients with vertical growth phase melanoma. Clin Cancer Res 2000; 6(5):1845–53.[Abstract/Free Full Text]
  44. Talve L, Sauroja I, Collan Y, Punnonen K, Ekfors T. Loss of expression of the p16INK4/CDKN2 gene in cutaneous malignant melanoma correlates with tumor cell proliferation and invasive stage. Int J Cancer 1997; 74(3):255–9.[CrossRef][Medline]
  45. Salvesen HB, Das S, Akslen LA. Loss of nuclear p16 protein expression is not associated with promoter methylation but defines a subgroup of aggressive endometrial carcinomas with poor prognosis. Clin Cancer Res 2000; 6(1):153–9.[Abstract/Free Full Text]
  46. Klein WM, Hruban RH, Klein-Szanto AJ, Wilentz RE. Direct correlation between proliferative activity and dysplasia in pancreatic intraepithelial neoplasia (PanIN): additional evidence for a recently proposed model of progression. Mod Pathol 2002; 15(4):441–7.[CrossRef][Medline]
  47. Belinsky SA, Nikula KJ, Palmisano WA, et al. Aberrant methylation of p16(INK4a) is an early event in lung cancer and a potential biomarker for early diagnosis. Proc Natl Acad Sci USA 1998; 95(20):11891–6.[Abstract/Free Full Text]
  48. Lee JH, Park SJ, Abraham SC, et al. Frequent CpG island methylation in precursor lesions and early gastric adenocarcinomas. Oncogene 2004; 23(26):4646–54.[CrossRef][Medline]
  49. von Zeidler SV, Miracca EC, Nagai MA, Birman EG. Hypermethylation of the p16 gene in normal oral mucosa of smokers. Int J Mol Med 2004; 14(5):807–11.[Medline]



This article has been cited by other articles:


Home page
 Clin. Cancer Res.Home page
C.-F. Li, W.-W. Huang, J.-M. Wu, S.-C. Yu, T.-H. Hu, Y.-H. Uen, Y.-F. Tian, C.-N. Lin, D. Lu, F.-M. Fang, et al.
Heat Shock Protein 90 Overexpression Independently Predicts Inferior Disease-Free Survival with Differential Expression of the {alpha} and {beta} Isoforms in Gastrointestinal Stromal Tumors
Clin. Cancer Res., December 1, 2008; 14(23): 7822 - 7831.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Huang, H.-Y.
Right arrow Articles by Hsiung, C.-Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Huang, H.-Y.
Right arrow Articles by Hsiung, C.-Y.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS