|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 9 (May 1), 2004: pp. 1572-1582 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.10.042 Prognostic Value of Proliferation, Apoptosis, Defective DNA Mismatch Repair, and p53 Overexpression in Patients With Resected Dukes' B2 or C Colon CancerA North Central Cancer Treatment Group StudyFrom the Mayo Clinic and Mayo Foundation, Rochester; CentraCare Clinic, St Cloud; The Duluth Clinic, Duluth, MN; Cedar Rapids Oncology Project CCOP, Cedar Rapids; Siouxland Hematology-Oncology Associates, Sioux City, IA; University of North Carolina Department of Hematology/Oncology, Chapel Hill, NC; Allegheny General Hospital, Pittsburgh, PA; and Allan Blair Cancer Centre, Regina, Saskatchewan, Canada Address reprint requests to Thomas E. Witzig, MD, 6-28 Stabile Building, 200 First St SW, Rochester, MN 55905; e-mail: witzig{at}mayo.edu
PURPOSE: Molecular studies of colon cancer have provided insights into pathogenesis, yet it is unclear how important these markers are in predicting prognosis. This study investigated the prognostic significance of TUNEL, bcl-2, p53, proliferation marker Ki-67 and DNA mismatch repair (MMR) status in patients with Dukes' stage B2 and C colorectal adenocarcinomas. PATIENTS AND METHODS: Tumor tissue from 366 patients (75% Dukes' C, 25% Dukes' B2) from four randomized North Central Cancer Treatment Group phase III surgical adjuvant trials were used. Eighty-one percent of patients received adjuvant treatment, which was primarily fluorouracil (FU) based (90%). Tumor location was predominantly (87%) the colon. Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL), Ki-67, p53, bcl-2, and MMR were assayed using immunohistochemistry. Stage, grade, MMR, Ki-67, and previously determined flow cytometry markers (ploidy and S phase) were explored for associations with each other and with overall survival (OS) and disease-free survival (DFS). RESULTS: Univariately, stage B2, low grade, diploid, Ki-67 more than 27%, normal p53, and FU-based adjuvant treatment were significantly associated with improved OS and DFS (P < .05). After adjusting for stage, grade, and ploidy in multivariate analysis, Ki-67 remained significantly related to both OS and DFS (P < .01). Active FU-based adjuvant treatment was significant only for OS in this multivariate model. Neither bcl-2 nor TUNEL were significant. CONCLUSION: This retrospective study indicates that Ki-67 and ploidy may have stronger prognostic impact on OS and DFS than other parameters investigated after adjusting for stage and tumor grade. Prospective studies to elucidate the mechanism and prognostic significance of these findings are necessary.
Colorectal carcinoma (CRC) is one of the leading causes of cancer death in the United States.1–3 Patients with stage III (Dukes' C) and a subset of stage II CRC (Dukes' B2) are treated with adjuvant therapies to reduce the risk of recurrence.1,4 Regardless, locoregional recurrence after curative resection remains problematic.5 Therefore, it is a priority to identify potential markers that could predict disease-free survival (DFS) and overall survival (OS) for colorectal cancer patients. Normal tissue homeostasis is maintained by balancing cell proliferation with the physiologic deletion of aberrant or senescent cells via apoptosis or programmed cell death.6–8 Disruption of the apoptotic pathway may confer a selective growth advantage, upsetting this homeostatic state.9 Indeed some studies have shown a decrease in apoptosis in CRCs when compared with both normal and adenoma.6,10 Other studies have shown an increase in proliferation,11–13 indicating that colorectal carcinogenesis might not be solely either a proliferative or apoptotic disorder, but rather might arise through the disruption of the balance of apoptotic and proliferative mechanisms.5,14,15 Potential biomarkers for the control of this balance are bcl-2 and p53. Bcl-2 is an intracellular integral protein localized to the nuclear envelope, the outer mitochondrial membrane, and the endoplasmic reticulum.16,17 It was initially isolated because the translocation, t(14;18), present in follicular and B-cell lymphomas, was noted to result in high levels of bcl-2 protein as detected through immunohistochemistry.18 Other reports have demonstrated that the translocation is not a mandatory precursor to bcl-2 overexpression, which has subsequently been identified by immunohistochemistry in other neoplastic conditions including lung, breast, thyroid, and prostate cancer.11,19 Its association with prognosis seems to be tumor specific. For instance, a correlation between expression and improved prognosis was found with both breast and thyroid carcinomas, but a similar correlation was not identified in prostate tumors.11 Wild-type p53 is a transcriptional regulator at the G1 and G2-M cell cycle checkpoint.15 In the presence of DNA damage, p53 activates effector genes such as p21, causing binding of cyclin-dependent kinases and inhibition of subsequent phosphorylation required for entry into S phase. This G1 arrest allows time for DNA damage repair or for induction of apoptosis to prevent proliferation of cells with deleterious mutations.20 Studies have shown that mutations in p53 usually result in the accumulation of p53 protein (ie, overexpression), which in turn may result in increased proliferation, loss of apoptotic function, and chromosomal instability.15,20,21 A majority of colorectal cancers progress through a multistep process involving a series of genetic changes with phenotypic progression through hyperplasia, adenoma, carcinoma, and metastasis.6,15,22 These tumors are characterized by genomic instability and loss of normal karyotype.23,24 Forty-five percent to 70% of all CRCs harbor p53 abnormalities, suggesting its importance in this progressive accumulation of genetic errors.24,16,18 Interestingly, studies show an inverse relationship between bcl-2 and p53 in CRCs, with the highest levels of bcl-2 in adenomas and high-grade dysplasia. Levels of p53 are lowest at the adenoma stage and peak in carcinomas when bcl-2 expression may be lowest.15,19
The previously described carcinogenesis accounts for approximately 85% of all CRCs. The remaining 15% are thought to arise through a different pathway characterized by the inactivation or mutation of the DNA mismatch repair protein complex,4 which includes hMLH-1, hMSH-2, hMSH-6, hMSH-3, and hPMS2. These tumors tend to remain diploid24 but do exhibit widespread alterations in simple repeat regions of DNA called microsatellites. The current methodology for assessing whether a tumor is microsatellite stable or microsatellite instable (MSI) relies on evaluating normal and tumor DNA for alterations in microsatellite loci. This allows for subtyping of MSI tumors into either MSI-H ( The majority (> 98%) of MSI-H tumors are the result of either the mutation of hMLH-1/hMSH-2 or methylation of hMLH-1. This results in the abrogation of the DNA mismatch repair mechanism. Lindor et al25 recently showed that immunostaining for hMLH-1 and hMSH-2 accurately predicts the molecular finding of MSI-H versus MSI-L/microsatellite stable. Therefore, loss of expression of either protein is an indication of dMMR. Although CRC with dMMR frequently have wild-type p53,24,27 several dMMR cell lines do have mutant p53, which has been shown to cause p53 protein stabilization (ie, overexpression).21 Despite studies showing that p53 overexpression is believed to favor the development of nondiploid tumor expansion,28,29 these dMMR cell lines still maintain chromosome number and structure, indicating a p53-independent pathway.23 In fact, in vitro studies show that bcl-2 overexpression can block p53-induced cell death, perhaps through downregulation of p21, suggesting that intact p53 alone does not prevent carcinogenesis.20,16 Similarly, bcl-2 overexpression has been paradoxically linked with cell growth arrest in a p53-independent fashion in vitro.17 The coregulation of these proteins is complex, yet studies have shown that both seem to have prognostic significance for initial diagnosis and responsiveness to some chemotherapeutic agents in CRC, thus having implications for overall disease-free survival.9,16,19–21 By studying the interplay of apoptotic markers (bcl-2, p53) on both the proliferative (Ki-67, S phase) and apoptotic (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling [TUNEL]) indices in the setting of a large number of MMR and dMMR CRCs, we hoped this study would determine the individual and combined impact of assays for bcl-2, p53, Ki-67, S phase and/or TUNEL on predicting survival.
Patient Population Patients were registered from four North Central Cancer Treatment Group phase III clinical trials investigating surgical adjuvant chemotherapy in colorectal patients (Table 1). Samples were provided for 412 patients from 1,388 patients registered on the clinical trials. Eighty-one percent of patients had been randomly assigned for adjuvant treatment, which was mainly fluorouracil (FU) based (90%). Treating or contributing physicians and laboratory personnel were blinded to clinical information and patient outcome. Serial sections were cut from each sample and used for all laboratory analyses. A total of 366 patients had sufficient tissue available for these analyses, of which, 98% had a complete set of laboratory correlates. The use of all samples was approved by the Institutional Review Board of the Mayo Clinic.
TUNEL Assay All patient samples were stained in duplicate using Apoptag Plus Peroxidase In Situ Apoptosis Detection Kit (Serologicals, Norcross, GA) optimized for formalin-fixed, paraffin-embedded tissue, as previously described.34 Slides were deparaffinized and rehydrated followed by a hydrogen peroxide and methanol quench for 20 minutes. The samples were treated with 25 µg/mL proteinase K at 37°C for 20 minutes. After the sample was washed in phosphate-buffered saline, a 1:3 dilution of terminal deoxynucleotidyl transferase was added and incubated at 37°C for 1 hour. After a stop solution was applied for 15 minutes and the samples were washed, samples were then incubated with antidigoxigenin peroxidase conjugate at 37°C for 30 minutes. Slides were developed with a 1:20 dilution of diaminobenzidine substrate, counterstained with methyl green, dehydrated, and coverslipped. Analysis of TUNEL slides was done using morphometric cell count. Briefly, between 200 and 350 40x histologic images, selected by a pathologist, were captured using a quantitative histomorphometry system (BLISS; Bacus Laboratories, Lombard, IL). These images were downloaded into a WebSlide Browser program and the subsequent images recalled on a computer screen. Apoptotic bodies were counted across the entire scanned area. Intact carcinoma cells were counted across 10% of the captured area and extrapolated across the entire area to derive a normal cell count. The apoptotic index was calculated as follows: (total number of apoptotic bodies/total number of intact carcinoma cells) x 100. This analysis was compared with both manual counting methods and morphometric area analysis and statistically determined to be most representative of overall apoptotic index.34
Immunohistochemistry and Scoring
Ki-67 slides had an area marked as representative of the overall tumor by a pathologist and were read on the CAS 200 (Bacus Laboratories). Its use for quantifying Ki-67 is well documented.35 Briefly, the CAS 200 is a digital morphometry system that yields a percent positive for a given sample based on the total area stained positive (read at a wavelength of 500 nm) divided by the total area stained (read at a wavelength of 620 nm). Bcl-2 slides were read at 40x by a pathologist and visually scored as 0 to 3+ based on total percent area positive. p53 slides were visually scored as a ratio of the total staining intensity (0 to 3+) over the area positive as defined by quadrants (1 to 4). Mismatch repair status was evaluated by staining with hMLH-1 and hMSH-2 antibodies. Slides were scored as either positive or negative based on the presence (+) or absence (–) of expression of hMLH-1 or hMSH-2.
Statistical Analysis OS was calculated as the number of days from random assignment to the date of death or last contact. DFS was calculated as the number of days from random assignment to the date of disease recurrence or death. Patients were censored at 5 years for DFS and 8 years for OS. The distribution of DFS and OS was estimated by Kaplan-Meier37 methodology and the log-rank test was used to test for significant differences in DFS and OS. Cox proportional hazards models38 were used to explore the association of clinical and laboratory parameters with OS and DFS. Models were stratified according to the patient's original treatment study for analyses involving OS and DFS. The Score statistic and Likelihood ratio test were used to test for significance in univariate and multivariate models, respectively. Backward and stepwise model procedures were used to identify covariate(s) most strongly associated with DFS and OS. The Wald statistic was used in multivariate models to test and estimate CIs for the significance of a single covariate in the presence of other covariates. Laboratory correlates (Ki-67, TUNEL, p53, bcl-2, S phase, and dMMR) were evaluated both continuously and categorically, with categorizations investigated using Martingale residual analysis39 and classification and regression trees (CART).40 Cut points were determined on the basis of the functional form of a given covariate, suggesting values for which hazard ratios differ, and subgroups were defined on the basis of the risk of an event in a univariate Cox model. Aneuploid and tetraploid tumors, as determined by flow cytometry, were summarily categorized as nondiploid (unpublished data submitted for publication). Patients lacking expression of either hMLH-1 or hMSH-2 were considered dMMR. A cutpoint of 5 for total S phase was determined (unpublished data submitted for publication) and used in the current analyses. Cut points of 1.5% and 27% were established for the TUNEL and Ki-67, respectively. Bcl-2 was categorized as either expressed (> 10% cells positive) or not expressed. Similarly, p53 was subgrouped as either overexpressed (at least 50% of cells staining with 2+ intensity or greater) or normal (less than 50% of cells staining with 2+ intensity or less).
Summary statistics, frequency distributions and graphical methods were used to describe the distributions of the parameters investigated. Correlations were performed within and across each laboratory parameter, including clinical characteristics. Statistical tests were two sided, with P
Clinical Characteristics and Patient Outcome The final sample of 366 patients was representative of the population of 1,388 patients from which they were drawn with regard to Dukes' stage, grade, sex, age, and active FU-based adjuvant treatment (P > .05). However, patients in this study population had a higher rate of colon primaries (P < .001) and tended toward improved survival (P = .053). Overall, 22% of patients had disease recurrence within 5 years and 43% died within 8 years, with a median follow-up of 8.7 years for surviving patients. Five-year estimates of OS and DFS on the basis of clinical characteristics are summarized in Table 3. Univariately, Dukes' stage B2 and low-grade tumors had improved DFS (P < .001 and P = .002, respectively) and OS (P < .001 and P < .001, respectively). Patients receiving active FU-based adjuvant treatment also had improved DFS and OS (P = .031 and 0.001, respectively).
Correlation Between Laboratory Parameters The majority of tumors exhibited p53 overexpression (76%), lack of bcl-2 expression (73%), and low TUNEL (86%). Lack of bcl-2 expression was significantly associated with low Ki-67 (P < .011) and marginally associated with p53 overexpression (P = .074). Overexpression of p53 occurred more frequently in tumors that were stage C (P = .062) or nondiploid (P < .001). dMMR tumors were more commonly diploid (P < .001), high grade (P < .001), and p53 normal (P = .03). A weak positive correlation (  = 0.17) was observed between total S phase and Ki-67 (Fig 1).
Associations With Patient Outcomes Five-year survival rates on the basis of univariate associations of laboratory parameters with OS and DFS are listed in Table 4. High Ki-67 and normal p53 were predictive of increased DFS (P < .003 and P = .042, respectively) and OS (P < .001 and P = .035, respectively). Patients with diploid rather than nondiploid tumors also had improved DFS and OS (P = .007 and .003, respectively). Bcl-2, TUNEL, dMMR, and total S phase did not reach statistical significance. In a multivariate model adjusted for Dukes' stage and grade, ploidy and Ki-67 remained significant for predicting both DFS and OS (Table 5). In a larger multivariate model including grade, stage, and ploidy, Ki-67 continued to be significant (P < .05; data not shown). In addition, patients with diploid and high Ki-67 tumors had improved OS in comparison with patients with nondiploid and low Ki-67 tumors (P < .0001; Fig 2A).
Subgroup Analysis Given the small number of patients with rectal carcinoma (13%), these patients were excluded from subgroup analysis. Univariately within stage C patients, ploidy and normal p53 were significant for both DFS (P = .001 and .014, respectively) and OS (P = .001 and .008, respectively). Ki-67 more than 27% and dMMR were significant for OS (P = .009 and .030, respectively) but only marginally significant for DFS (P = .085 and .095, respectively; Table 6). No marker was significant in a univariate model for either DFS or OS within stage B2 patients (data not shown).
In a multivariate model within stage C patients, ploidy remained significant for predicting both DFS and OS in the presence of tumor grade (P < .001). High Ki-67 and FU-based adjuvant treatment were significant only for OS in this multivariate model (P < .05, data not shown). Patients with stage C tumors that were high Ki-67 and diploid had improved OS in comparison with low Ki-67 and nondiploid tumors (P = .0003; Fig 2B). Consistent with these findings, there remained a statistically significant decrease in survival for stage B2 patients with low Ki-67 and nondiploid tumors (P = .0018; Fig 2C).
This retrospective analysis of stage B2 and C colorectal adenocarcinomas investigated the changes in several pro- and antiapoptotic laboratory parameters with relation to OS and DFS. The strength of this study is three-fold. First, the results presented are based on a large sample of patients, recruited through the North Central Cancer Treatment Group, who participated in phase III randomized clinical trials investigating what is considered as active and standard FU-based CRC treatment regimens. Second, complete clinical patient outcome data were available with a median of 8.7 years follow-up from the date of enrollment, allowing a comprehensive evaluation of the associations of laboratory parameters with OS and DFS. Finally, the patient population allowed investigation of the effects of different markers in relation to disease pathogenesis and active FU-based treatment versus no (or nonactive) treatment. The data from this study indicate that a high Ki-67 rate (> 27%) is significantly associated with increased OS and DFS (P < .001 and .003, respectively). Previous reports in both colon and other cancers have reported that an increase in the rate of proliferation is correlated with a less favorable prognosis.9,41–45 If one considers that a tumor results from the previously hypothesized imbalance between apoptosis and proliferation wherein mutant cells are allowed to proliferate without apoptotic interference, a higher rate of cell growth would be associated with more cells containing genetic abnormalities, and therefore potentially a worse prognosis. However, no such association was seen in this study.
Because most of the patient population (81%) in our study received adjuvant chemotherapy, theoretically the increased survival associated with higher Ki-67 might represent a secondary affect. Previous studies have shown that low proliferative indices can be indicative of reduced response to treatment.46–48 The corollary of such findings would be that tumors with greater proliferation would have better response to treatment. To ascertain if the improved OS and DFS observed in the current study were merely indicative of an increased response to treatment, tumors from patients who received active FU-based chemotherapy were analyzed separately from those who did not. Within treatment classification (treatment v no treatment), patients with tumors exhibiting Ki-67 more than 27% showed improved OS regardless of whether they received active treatment. Patients receiving an active FU-based treatment whose tumors showed high Ki-67 (> 27%) had a 5-year survival rate of 75% (95% CI, 68 to 80) versus 61% for those whose tumors had low Ki-67 ( Interestingly, when the two most significant parameters for predicting OS and DFS, ploidy and Ki-67, were investigated in an additive multivariate model, an even more dramatic association with improved prognosis (Fig 2A) was observed. This observation also was seen in the stage-dependent subgroup analysis (Fig 2B). Among stage C patients, patients with high Ki-67 and diploid tumors had improved OS compared with patients with low Ki-67 and nondiploid tumors. The subgroup of stage B2 patients was smaller (n = 81), but there remains a significant decrease in OS among patients with low Ki-67 and nondiploid tumors (Fig 2C). Statistical testing for an interaction between Ki-67 and ploidy revealed no significant interaction, indicating that high Ki-67 and diploid are additive factors. Although both have been reported previously as independently significant, this represents the first examination of the two parameters in conjunction, and suggests that there might be genes that potentially modulate the deleterious effects of proliferation in tumor cells. It was recently reported that growth factors that are initially mitogenic might in fact upregulate genes involved in cell cycle arrest downstream even in the presence of aberrant p53, resulting in a more well-differentiated tumor type.7,50 Studies investigating this potential link are underway. Although the weak correlation between the two proliferative indices, total S phase and Ki-67, may seem counterintuitive, this finding is not novel. Previous large-scale studies in breast cancer by Bergers et al51 and Barzanti et al52 (n = 932 and 330, respectively) have shown a similar weak correlation between S phase and either mitotic cell counts or Ki-67. It is important to note that although both technologies presented herein ultimately estimate proliferative rate, the methodologies used are quite different. This study used the Ki-67 antibody that reacts with cells in the late G1, S, and G2-M phase of the cell cycle, whereas the flow cytometric measure of S phase measures only the cells in one portion of the cycle.53 Furthermore, the Ki-67 immunohistochemical staining in this study was analyzed using computer-aided imaging that allowed a technologist to evaluate only intact carcinoma cells and was therefore less susceptible to contamination from either normal or inflammatory cells. Although an attempt was made to control for normal cell contamination in determining S phase values by only selecting tumors that had more than 20% tumor by hematoxylin and eosin staining, contaminating effects of either normal or inflammatory cell populations could not be eliminated completely. Leonardi et al54 recently reported that, across a series of 238 bladder tumors, the presence of inflammatory cells increased the risk of underestimating S phase values compared with those obtained by immunohistochemistry. Whether these differences actually reflect biologically unique proliferative markers in colon cancer requires additional study. Data from this study also demonstrate the complex interaction between bcl-2 and p53 in the initiation and progression of colorectal carcinogenesis. An inverse relationship of marginal significance between lack of bcl-2 and p53 overexpression (P = .074) was identified. This apparent coregulation of bcl-2 and p53 generally agrees with previous reports15,18,21 and with the current genetic model for tumor initiation and progression.22,55,56 This model postulates that CRC represents the culmination of a multistep process facilitated by a series of genetic alterations that might convey a selective growth advantage to aberrant cell populations. Bcl-2, which has been shown to occur earlier in the adenoma-to-carcinoma sequence, might represent a marker involved in tumor initiation. Its overexpression generally decreases apoptosis and thus allows genetic alterations to evade mechanisms of cellular deletion. The observation by Bosari et al57 that bcl-2 expression localized mainly to dysplastic cells in adenomas also implicates its role in initiation. Therefore, the finding in the current study that the majority (73%) of colorectal adenocarcinomas lacked bcl-2 expression generally correlates with previous studies of bcl-2 in colon cancer.11,15,18 It is important to note that although other groups have referred to the lack of bcl-2 expression in carcinomas as a loss, a similar conclusion could not be made in this study because adenomas were not evaluated.18 Although p53 overexpression was found to have a significant inverse relationship to survival in our study, this is an inconsistent finding among various patient groups.4,58–60 Within the colon, the significance of p53 overexpression can vary depending on tumor location (ie, proximal v distal). This might reflect the differing carcinogenic mechanism associated with each location. Proximal tumors are more commonly MSI and diploid, whereas distal tumors have a higher frequency of allelic loss, p53 accumulation, and aneuploidy.61 Sun et al62 and Bosari et al63 demonstrated that cytoplasmic p53 overexpression using the CM1 antibody independently predicted a worse prognosis for patients with distal versus proximal tumors. Subsequent p53 mutational studies in colorectal adenocarcinoma by Goh et al64 and Borresen-Dale et al65 have shown that specific mutations are associated with more aggressive tumors and decreased survival in patients with distal tumors but not proximal tumors. A similar stratification was identified in our study, with distal tumors having significantly decreased survival compared with proximal tumors (P = .03). It is important to note that this study investigated p53 dysregulation via immunohistochemistry rather than mutation analysis of the p53 gene. Cripps et al66 reported that approximately 33% of CRCs containing immunohistochemically stable p53 do not have a detectable p53 mutation. The DO7 monoclonal antibody was used for this analysis and has been shown to have strong clinical correlation in other studies.18,19 However, Leahy et al67 recently reported that detection of p53 protein accumulation using DO7 may not be completely concordant with mutation analysis. Regardless, the finding that the majority of tumors (76%) overexpressed p53 and that this correlated with decreased survival agrees with the preponderance of studies investigating p53 in colorectal cancer.19,21,49,67–70 In conclusion, although these data confirmed many of the previously reported results with regard to p53 and bcl-2, they also identified an additive relationship between the proliferative index and ploidy. The identification of the stage B2 group (low Ki-67 and nondiploid) with lower OS warrants additional prospective study to determine if this group of patients might benefit from novel adjuvant chemotherapy regimens. Finally, data with regard to Ki-67 and OS and DFS, especially in light of the active versus nonactive treatment comparison, warrant additional study to elucidate the complex role of proliferative genes in colorectal carcinogenesis. To this end, extensive correlative studies are planned using specimens from other trials that had no treatment arms as well as from the recently completed stage II study (C9581) to expand on the data currently presented.
The authors indicated no potential conflicts of interest.
Supported by the National Cancer Institute Grant (CA78899) and conducted as a collaborative trial of the North Central Cancer Treatment Group and Mayo Clinic. Presented as an abstract for oral presentation at the 38th Annual Meeting of the American Society of Clinical Oncology, Orlando, FL, May 18-21, 2002. Work for this study was completed at the Mayo Clinic, Rochester, MN. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Shulman K, Schilsky RL: Adjuvant therapy of colon cancer. Semin Oncol 22:600–610, 1995[Medline] 2. Lanza G, Maestri I, Ballotta MR, et al: Relationship of nuclear DNA content to clinicopathologic features in colorectal cancer. Mod Pathol 7:161–165, 1994[Medline] 3. Greenlee RT, Murray T, Bolden S, et al: Cancer statistics. CA Cancer J Clin 50:7–33, 2000[Abstract]
4. Watanabe T, Wu T-T, Catalano PJ, et al: Molecular predictors of survival after adjuvant chemotherapy for colon cancer. N Engl J Med 344:1196–1206, 2001 5. Seong J, Chung EJ, Kim H, et al: Assessment of biomarkers in paired primary and recurrent colorectal adenocarcinomas. Int J Radiat Oncol Biol Phys 45:1167–1173, 1999[CrossRef][Medline]
6. Bedi A, Pasricha PJ, Akhtar AJ, et al: Inhibition of apoptosis during development of colorectal cancer. Cancer Res 55:1811–1816, 1995 7. Takano Y, Saegusa M, Ikenaga M, et al: Apoptosis of colon cancer: Comparison with Ki-67 proliferative activity and expression of p53. J Cancer Res Clin Oncol 122:166–170, 1996[CrossRef][Medline] 8. Ueda M, Kumagai K, Ueki K, et al: Growth inhibition and apoptotic cell death in uterine cervical carcinoma cells induced by 5-fluorouracil. Int J Cancer 71:668–674, 1997[CrossRef][Medline]
9. Sinicrope FA, Hart J, Hsu H-A, et al: Apoptotic and mitotic indices predict survival rates in lymph node-negative colon carcinomas. Clin Cancer Res 5:1793–1804, 1999 10. Tsujitani S, Shirai H, Tatebe S, et al: Apoptotic cell death and its relationship to carcinogenesis in colorectal carcinoma. Cancer 77:1711–1716, 1996[Medline] 11. Flohil CC, Janssen PA, Bosman FT: Expression of bcl-2 protein in hyperplastic polyps, adenomas, and carcinomas of the colon. J Pathol 178:393–397, 1996[CrossRef][Medline] 12. Sinicrope FA, Roddey G, McDonnell TJ, et al: Increased apoptosis accompanies neoplastic development in the human colorectum. Clin Cancer Res 2:1999–2006, 1996[Abstract]
13. Sugao Y, Koji T, Yao T, et al: The incidence of apoptosis during colorectal tumorigenesis. Int J Surg Pathol 8:123–132, 2000 14. Yang H-B, Chow N-H, Sheu B-S, et al: The role of bcl-2 in the progression of the colorectal adenoma-carcinoma sequence. Anticancer Res 19:727–730, 1999[Medline] 15. Hao XP, Ilyas M, Talbot IC: Expression of bcl-2 and p53 in the colorectal adenoma-carcinoma sequence. Pathobiology 65:140–145, 1997[Medline] 16. Miyashita T, Krajewski S, Krajewska M, et al: Tumor suppressor p53 is a regulator of bcl-2 and bax gene expression in vitro and in vivo. Oncogene 9:1799–1805, 1994[Medline]
17. Pietenpol JA, Papadopoulos N, Markowitz S, et al: Paradoxical inhibition of solid tumor cell growth by bcl-2. Cancer Res 54:3714–3717, 1994 18. Kaklamanis L, Savage A, Mortensen N, et al: Early expression of bcl-2 protein in the adenoma-carcinoma sequence of colorectal neoplasia. J Pathol 179:10–14, 1996[CrossRef][Medline] 19. Kaklamanis L, Savage A, Whitehouse R, et al: Bcl-2 protein expression: Association with p53 and prognosis in colorectal cancer. Br J Cancer 77:1864–1869, 1998[Medline] 20. Bukholm IK, Nesland JM: Protein expression of p53, p21 (waf1/cip1), bcl-2, bax, cyclin D1 and pRb in human colon carcinomas. Virchows Arch 436:224–228, 2000[CrossRef][Medline] 21. Buglioni S, D'Agnano I, Cosimelli M, et al: Evaluation of multiple bio-pathological factors in colorectal adenocarcinomas: Independent prognostic role of p53 and bcl-2. Int J Cancer 84:545–552, 1999[CrossRef][Medline] 22. Dumont N: Genetic and epigenetic contributions to colorectal cancer. APMIS 107:711–722, 1999[Medline] 23. Eshleman JR, Casey G, Kochera ME, et al: Chromosome number and structure both are markedly stable in RER colorectal cancers and are not destabilized by mutation of p53. Oncogene 17:719–725, 1998[CrossRef][Medline] 24. Curtis LJ, Georgiades IB, White S, et al: Specific patterns of chromosomal abnormalities are associated with RER status in sporadic colorectal cancer. J Pathol 192:440–445, 2000[CrossRef][Medline]
25. Lindor NM, Burgart LJ, Leontovich O, et al: Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol 20:1043–1048, 2002
26. Ribic CM, Sargent DJ, Moore MJ, et al: Tumor microsatellite status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 349:247–257, 2003
27. Ward R, Meagher A, Tomlinson I, et al: Microsatellite instability and the clinicopathological features of sporadic colorectal cancer. Gut 48:821–829, 2001 28. Carder PJ, Cripps KJ, Morris R, et al: Mutation of the p53 gene precedes aneuploid clonal divergence in colorectal carcinoma. Br J Cancer 71:215–218, 1995[Medline] 29. Carder PJ, Wyllie AH, Purdie CA, et al: Stabilised p53 facilitates aneuploid clonal divergence in colorectal carcinoma. Oncogene 8:1397–1401, 1993[Medline] 30. Moertel CG, Fleming TR, Macdonald JS, et al: Levamisole and fluorouracil for adjuvant therapy of resected colon carcinoma. N Engl J Med 322:352–358, 1990[Abstract]
31. O'Connell MJ, Martenson JA, Wieand HS, et al: Improving adjuvant therapy for rectal cancer by combining protracted-infusion fluorouracil with radiation therapy after curative surgery. N Engl J Med 331:502–507, 1994
32. Wiesenfeld M, O'Connell MJ, Wieand HS, et al: Controlled clinical trial of interferon-gamma as postoperative surgical adjuvant therapy for colon cancer. J Clin Oncol 13:2324–2329, 1995
33. O'Connell MJ, Mailliard JA, Kahn MJ, et al: Controlled trial of fluorouracil and low-dose leucovorin given for 6 months as postoperative adjuvant therapy for colon cancer. J Clin Oncol 15:246–250, 1997 34. Garrity MM, Burgart LJ, Riehle DL, et al: Identifying and quantifying apoptosis: Navigating technical pitfalls. Mod Pathol 16:389–394, 2003[CrossRef][Medline]
35. Cheng L, Pisansky TM, Sebo TJ, et al: Cell proliferation in prostate cancer patients with lymph node metastasis: A marker for progression. Clin Cancer Res 5:2820–2823, 1999 36. Witzig TE, Loprinzi DC, Gonchoroff NJ, et al: DNA ploidy and cell kinetic measurements as predictors of recurrence and survival in stages B2 and C colorectal cancer. Cancer 68:879–888, 1991[CrossRef][Medline] 37. Kaplan E, Meier P: Nonparametric estimation for incomplete observations. J Am Stat Assoc 53:457–481, 1958[CrossRef] 38. Cox D: Regression models and life tables (with discussion). J R Stat Soc (B) 34:187–220, 1972
39. Grambash PM, Therneau TM: Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 81:515–526, 1994 40. Therneau TM, Atkinson EJ: An introduction to recursive partitioning using the RPART routines. Technical Report Series 61. Rochester, MN, Mayo Clinic, Department of Health Sciences Research, 1997 41. King K, Hwang J, Chau G, et al: Ki-67 expression as a prognostic marker in patients with hepatocellular carcinoma. J Gastroenterol Hepatol 13:273–279, 1998[Medline] 42. Trihia H, Murray S, Price K, et al: Ki-67 expression in breast carcinoma. Cancer 97:1321–1331, 2003[CrossRef][Medline] 43. Stavropoulos NE, Filiadis I, Ioachim E, et al: Prognostic significance of p53, bcl-2 and Ki-67 in high risk superficial bladder cancer. Anticancer Res 22:3759–3764, 2002[Medline] 44. Toquet C, Le Neel JC, Guillou L, et al: Elevated (> or = 10%) MIB-1 proliferative index correlates with poor outcome in gastric stromal tumor patients: A study of 35 cases. Dig Dis Sci 47:2247–2253, 2002[CrossRef][Medline] 45. Jones S, Clark G, Koleszar S, et al: Low proliferative rate of invasive node-negative breast cancer predicts for a favorable outcome: A prospective evaluation of 669 patients. Clin Breast Cancer 1:310–314, 2001[Medline]
46. Itamochi H, Kigawa J, Sugiyama T, et al: Low proliferation activity may be associated with chemoresistance in clear cell carcinoma of the ovary. Obstet Gynecol 100:281–287, 2002 47. Suzuki M, Tsukagoshi S, Saga Y, et al: Assessment of proliferation index with MIB-1 as a prognostic factor in radiation therapy for cervical cancer. Gynecol Oncol 79:300–304, 2000[CrossRef][Medline] 48. Imdahl A, Jenkner J, Ihling C, et al: Is MIB-1 proliferation index a predictor for response to neoadjuvant therapy in patients with esophageal cancer? Am J Surg 179:514–520, 2000[CrossRef][Medline]
49. Allegra CJ, Paik S, Colangelo LH, et al: Prognostic value of thymidylate synthase, Ki-67, and p53 in patients with Dukes' B and C colon cancer: A national cancer institute-national surgical adjuvant breast and bowel project collaborative study. J Clin Oncol 21:241–250, 2003 50. Ewton DZ, Kansra S, Lim S, et al: Insulin-like growth factor-I has a biphasic effect on colon carcinoma cells through transient inactivation of the forkhead1, initially mitogenic, then mediating growth arrest and differentiation. Int J Cancer 98:665–673, 2002[CrossRef][Medline] 51. Bergers E, van Diest PJ, Baak JPA: Cell cycle analysis of 932 flow cytometric DNA histograms of fresh frozen breast carcinoma material. Cancer 77:2258–2266, 1996[CrossRef][Medline] 52. Barzanti F, Dal Susino M, Volpi A, et al: Comparison between different cell kinetic variables in human breast cancer. Cell Prolif 33:75–89, 2000[CrossRef][Medline] 53. Ostrowski ML, Pindur J, Laucirica R, et al: Proliferative activity in invasive breast carcinoma: A comprehensive comparison of MIB-1 immunocytochemical staining in aspiration biopsies to image analytic, flow cytometric and histologic parameters. Acta Cytol 45:965–972, 2001[Medline] 54. Leonardi E, Dalla Palma P, Reich A, et al: Biological characterisation of superficial bladder cancer by bivariate cytokeratin 7/DNA analysis, flow cytometric assessment of MIB-1, and an immunohistochemical study. Anal Cell Pathol 21:21–33, 2000[Medline] 55. Fearon ER, Vogelstein B: A genetic model for colorectal tumorigenesis. Cell 61:759–767, 1990[CrossRef][Medline] 56. Watanabe T, Muto T: Colorectal carcinogenesis based on molecular biology of early colorectal cancer, with special reference to nonpolypoid (superficial) lesions. World J Surg 24:1091–1097, 2000[CrossRef][Medline] 57. Bosari S, Moneghini L, Graziani D, et al: Bcl-2 oncoprotein in colorectal hyperplastic polyps, adenomas, and adenocarcinomas. Hum Pathol 26:534–540, 1995[CrossRef][Medline]
58. Allegra CJ, Parr AL, Wold LE, et al: Investigation of the prognostic and predictive value of thymidylate synthase, p53, and Ki-67 in patients with locally advanced colon cancer. J Clin Oncol 20:1735–1743, 2002 59. Soong R, Grieu F, Robbins P, et al: p53 alterations are associated with improved prognosis in distal colonic carcinomas. Clin Cancer Res 3:1405–1411, 1997[Abstract] 60. Bosari S, Viale G, Roncalli M, et al: p53 gene mutations, p53 protein accumulation and compartmentalization in colorectal adenocarcinoma. Am J Pathol 147:790–798, 1995[Abstract] 61. Kapiteijn E, Liefers GJ, Los LC, et al: Mechanisms of oncogenesis in colon versus rectal cancer. J Pathol 195:171–178, 2001[CrossRef][Medline] 62. Sun XF, Carstensen JM, Zhang H, et al: Prognostic significance of p53 nuclear and cytoplasmic overexpression in right and left colorectal adenocarcinomas. Eur J Cancer 32A:1963–1967, 1996
63. Bosari S, Viale G, Bossi P, et al: Cytoplasmic accumulation of p53 protein: An independent prognostic indictor in colorectal adenocarcinomas. J Natl Cancer Inst 86:681–687, 1994 64. Goh H, Elnatan J, Low CH, et al: p53 point mutation and survival in colorectal cancer patients: Effect of disease dissemination and tumor location. Int J Oncol 15:491–498, 1999[Medline] 65. Borresen-Dale AL, Lothe RA, Meling GI, et al: TP53 and long term prognosis in colorectal cancer: Mutations in the L3 zinc-binding domain predict poor survival. Clin Cancer Res 4:203–210, 1998[Abstract] 66. Cripps KJ, Purdie CA, Carder PJ, et al: A study of stabilisation of p53 protein versus point mutation in colorectal carcinoma. Oncogene 9:2739–2743, 1994[Medline] 67. Leahy DT, Salman R, Mulcahy H, et al: Prognostic significance of p53 abnormalities in colorectal carcinoma detected by PCR-SSCP and immunohistochemical analysis. J Pathol 180:364–370, 1996[CrossRef][Medline] 68. Bouzourene H, Gervaz P, Cerottini JP, et al: p53 and Ki-ras as prognostic factors for Dukes' stage B colorectal cancer. Eur J Cancer 36:1008–1015, 2000 69. Giatromanolaki A, Stathopoulos GP, Tsiobanou E, et al: Combined role of tumor angiogenesis, bcl-2, and p53 expression in the prognosis of patients with colorectal carcinoma. Cancer 86:1421–1430, 1999[CrossRef][Medline]
70. Zeng ZS, Sarkis AS, Zhang ZF, et al: p53 nuclear overexpression: An independent predictor of survival in lymph node-positive colorectal cancer patients. J Clin Oncol 12:2043–2050, 1994 Submitted October 8, 2003; accepted February 9, 2004. This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
30% loci instable) or MSI-L (between 1% and 29% loci instable),
.05 considered significant. P values were not adjusted for multiple comparisons. Statistical analyses were performed using SAS (Version 6; SAS Institute, Cary, NC) and S-Plus (Version 3; Statistical Sciences, Seattle, WA). 
