7 Prognostic Implication of Proliferative Markers MIB-1 and PClO in Esophageal Squamous Cell Carcinoma King-Yin Lam, F.R.c.P.A.’ Simon Ying-Kit Law, F.R.c.s.’ Mike Ka-Pui So, B.SC? Manson Fok, F.R.c.s.’ Lily Tung Ma, F.R.c.P.A.’ John Wong, F.R.C.S? ’ Department of Pathology, University of Hong Kong. Department of Surgery, Queen Mary Hospital, Hong Kong. Supported by a grant from the University of Hong Kong (CRCGiresearch grant 33ti/046/ 0065). The authors would like to thank Mr K.H. Yeung for his technical assistance. Address for reprints: Dr. K.Y. Lam, Department of Pathology, Queen Mary Hospital, Hong IKong. Received July 3, ‘1995; revision received August 29, 1995; accepted September 22, 1995. c 1996 American Cancer Society BACKGROUND. Proliferative markers are related to tumor behavior. The commonly used markers are proliferating cell nuclear antigen (PCNA) and Ki-67. The aim of this study is to evaluate the usefulness of MIB-1 (for Ki-67) and PClO (for PCNA) in the assessment of the clinicopathologic features and prognosis in patients with esophageal squamous cell carcinoma. METHODS. One hundred patients (88 males, 12 females; mean age, 63 years [range, 39 to 83 years]) with surgically resected esophageal squamous cell carcinoma (32 well differentiated, 51 moderately differentiated, and 17 poorly differentiated) were studied. The clinicopathologic features and survival data of these patients were noted. Representative tissue was collected from each tumor and immunohistochemical preparations for MIB-1 and PCl0 were made. RESULTS. The percentages of cells that tested positive for PClO and MIB-1 were much higher in tumor cells than in nonneoplastic cells. The pattern of expression of both markers varied with the differentiation of the tumor. The results observed with MIB-1 staining were better than those with PCIO; because MIB-1 had less background staining, as well as stronger and more uniform positive signals compared with PC10. Thus, further investigation was performed on MIB- 1-stained sections. The tumor cell MIB-1 scores ranged from 169 to 964 positive cells per 1000 cells (mean 598 ? 21 1; median, 636). Although it was significantly associated with the differentiation of the tumor ( P = 0.0001), the score had no significant relationship to the tumor size, location, or stage, or to the patients’ age and sex. The prognosis depended on the size and stage of the lesion. In Stage I11 lesions (n = 83), patients with MIB-1 scores below 300 had longer actual survival rates than those with a score of 300 or above. However, the survival rates of patients in the latter group were better if the greatest dimension of the tumor diameter was 7.5 cm or less. CONCLUSIONS. Proliferative activity in esophageal squamous cell carcinoma, as defined by the MIB-1 immunohistochemical method, is significantly related to tumor differentiation. It is also potentially valuable as a prognostic marker in addition to its use in tumor staging and size. Cancer 1996;77:7-13. 0 I996 American Cancer Socieiy. KEYWORDS: squamous cell carcinoma, proliferating cell nuclear antigen, MIB-1, Ki67, prognosis, differentiation. P roliferative activity of the tumor is a useful parameter in understanding tumor behavior. A correlation between the proliferative activity and overall prognosis has been observed in many cancers.‘ Thus, immunohistochemical determination of cell proliferation-associated antigens has aroused the interest of pathologists in recent years. The commonly used proliferation-associated markers include antibodies to Ki-67 and proliferating cell nuclear antigen (PCNA). The prognostic value of Ki-67 immunostaining in various tumors has 8 CANCER January 1,1996 / Volume 77 / Number 1 been well documented., The most widely used Ki-67 monoclonal antibody was discovered in 1983 by Gerdes and colleagues3 and the gene for the Ki-67 has been located on chromosome This antibody recognizes a labile epitope on a nuclear antigen that is exclusively expressed in proliferating cells, (i.e., in the G, , S, G, phases) and mitosis, but not in Go. The drawback in the use of this antibody is that its application is restricted to fresh material because the epitope does not survive routine histopathologic fixation in formaldehyde. New antibodies MIBMIB-1,2, 3, 5, and IND.64 have been 1 is now commercially available and can be applied to formalin fixed and routinely processed tissues. However, it needs microwave irradiation for retrieval of the antigenicity." The immunostaining pattern, as well as the morphologic and cell cycle distribution of MIB-1 expression, are identical to those of Ki-67, so it can be regarded as a true Ki-67 equivalent.' Antibodies to PCNA also have been widely studied proliferative markers in recent years because they can be detected in formalin fixed tissues. PCNA is a 36KDa acidic nuclear protein that functions as an auxiliary protein for DNA polymerase-delta and has been recognized as an endogenous histologic marker for G, IS-phase in the cell cycle.' The most widely used antibody to PCNA is PC10. Esophageal cancer is the sixth most frequent cause of cancer death worldwide and ranks ninth in world cancer incidence.'" This type of cancer is a major cause of death in certain areas of Asia, including Hong Kong. Various studies have attempted to find markers that act as useful adjuncts to morphologic and clinical features to predict the behavior of esophageal tumors.'','2 In this present study, we examine the immunolocalization and prognostic value of MIB-1 and PCNA in 100 cases of esophageal squamous cell carcinomas. The clinicopathologic information and survival data of all these patients are available for assessment. To the best of our knowledge, the value of MIB-1 has not been reported in human esophageal malignancies. MATERIALS AND METHODS Data Collection Chinese patients in Queen Mary Hospital with surgically resected primary esophageal squamous cell carcinomas between January 1989 and December 1993 were studied. Tissues from the resected specimens of these patients were prospectively collected in a fresh state and then fixed in 10% formalin for less than 48 hours. Macroscopically, the site and size (maximum length) of the tumors were recorded. The histologic sections of these primary esophageal carcinomas were reviewed and classified using the World Health Organization (WHO) criteria.'' Only squamous cell carcinomas were included in the study. A total of 100 patients (88 males and 12 females) was included in the present study. The cancers were staged according to TNM clas~ification.'~ All patients were followed regularly until death and their demographic data reviewed. The actuarial survival rate of the patients was measured from the date of resection of the esophageal cancer to the date of death or April 30, 1995. Staining for Proliferative Markers A representative block from each patient was chosen for the study. The block used was the one in which the most dominant histologic pattern of the carcinoma was identified (because a few carcinomas showed slight variation in tumor differentiation in different foci of the tumor). From the block, 5 p-thick sections were cut and mounted on gelatin-coated glass slides and dried overnight in an oven at 37°C. The sections were dewaxed in xylene, rehydrated through graded concentration of alcohol, and then treated with 0.5% hydrogen peroxide at room temperature for 10 minutes to block endogenous peroxidase activity. They were washed and rinsed with Tris-buffered saline at a pH of 7.6. After pretreatment with 0.5% trypsincalcium chloride (pH = 7.8), the sections were immersed in 10 mM sodium citrate (pH = 6) in a thermoresistant plastic box and were processed in a microwave oven (700 W) for 9 minutes at 95"%C. Ten percent normal sheep serum was then added at room temperature for 10 minutes. The monoclonal antibodies, MIB-1 (150, Immunotech SA, Cedex, France) and PClO (1:20, Dako, Glostrup, Denmark) were applied and incubations performed overnight at 4°C. Immunohistochemistry was performed using the avidin-biotin complex technique with diaminobenzidine development and Mayer's hemalum counterstaining. The sections were washed in water, dehydrated in alcohol, cleared in xylene, and mounted. Negative controls were sections treated the same as above but with omission of the primary antibodies. Assessment of the Proliferative Markers The sections were first scanned at low ( X 40) and medium power ( X 200) for all the fields with tumor and nontumor esophageal tissues, respectively, to account for the heterogeneity of distribution. The number of cells showing positive nuclear staining and the pattern of staining were recorded. Necrotic tissues, stromal cells, and lymphoid cells were not included in the recording. The distribution of the staining of MIB-1 and PClO in the tumor and the adjacent nonneoplastic epithelium in each case were assessed semiquantitatively and compared. The staining was divided into categories according to the percentage of the cells stained (0-25%, 26-50%, 51-75%, and 76100%). The sections stained with MIB-1 were then counted at X 400 magnification. At least 1000 nuclei were counted in each section. The MIB-1 score was presented as number of positive nuclei per 1000 nuclei counted. Proliferative Markers MIB-1 and PClO in Esophageal Cancerllarn et al. FIGURE 1. Positive staining in the nuclei of the basal and parabasal cells of the nonneoplastic esophageal epithelium (MIB-1, x130). 9 FIGURE 2. Positive staining in the nuclei of the tumor cells in the periphery of the cell nests of a moderately differentiated squamous cell carcinoma. (MIB-1, x130). RESULTS Clinicopathological Features FIGURE 3. The random positive staining in nuclei of the tumour cells in a poorly-differentiated squamous cell carcinoma. (MIB-1, x130). Reproducibility was determined by counting replicates and recounting of the same case to avoid intraobserver variation. All the counting procedures were performed by one of the authors (Lam) in order to eliminate interobservers variations. Statistical Analysis All data were computerized and statistical tests were performed with the aid of the SAS statistical package. The test used for continuous data was correlation analysis. The Kruskal-Wallis test was used for categorical data. The significance of various parameters on survival were analyzed by the multivariate Cox regression. Survival curve was plotted using the Kaplan-Meier rnethod.l5 The tests were considered significant when their overall P values were less than 0.05. The mean age of the 100 patients was 63 years (range, 39 to 83; standard deviation, 9). The esophageal cancers are composed of 32% well differentiated, 51% moderately differentiated, and 17% poorly differentiated squamous cell carcinomas. Ten percent of the tumors were in the upper esophagus, 56% in the middle esophagus, and 34% in the lower esophagus. The lengths of the carcinomas ranged from 1 cm to 12 cm (mean, 5 cm). Six percent of the patients had preoperative chemotherapy. The majority (83%) of these patients had Stage 111 lesions (T3, N1, MO or T4, any N, MO). The other patients had Stage I1 (13% Stage IIA [T2/T3, NO, MO], 2% Stage IIB [Tl/T2, N1, Moll, and Stage I V lesions (2%) (any T, any N, M l ) . Dysplasia or carcinoma was present in either the upper or lower resection margin in 14% of the cases. Staining for Proliferative Markers PClO and MIB-1 immunostaining were confined to the nuclei in both neoplastic and nonneoplastic cells. The mitotic figures were particularly immunoreactive and were highlighted by these antibodies. In all nonpathologic esophageal epithelia, only the nuclei of the basal and parabasal cells were positive for the markers (Fig. 1). There was a heterogeneity of PClO and MIB-1 staining in the esophageal cancers. In well and moderately differentiated squamous cell carcinomas, positive PClO or MIB1 staining was noted in the less differentiated cells in the periphery of the tumor cell nests. The central keratinizing areas and adjacent tumor cells were negative for the markers (Fig. 2). Conversely, the staining was more diffuse in poorly differentiated carcinomas (Fig. 3). In all instances, the percentages of positive cells for either markers in nonneoplastic epithelia (< 10%) were 10 CANCER January 1,1996 I Volume 77 I Number 1 TABLE 1 Correlation of Tumor MIB-1 Score with Clinical and Pathological Parameters in Esophageal Squamous Cell Carcinoma Parameter Finding No. of cases Mean ? standard deviation/ median MIB-1 score Sex Male Female Upper Middle Lower Well Moderate Poor I1 A I1 B III IV 88 12 10 56 34 32 51 li 13 2 83 2 596 5 2171799 608 i I711668 590 ? 1781590 624 5 2191674 558 ? 2051570 464 t 1961462 623 ? 1861673 776 t 1461819 577 i 2211654 741 i 91741 599 i 2131619 567 ? 2601567 Site Differentiation Stage LOO 80 P value 0.91 0.25 6 > > .d 0.0001 0.84 40 20 0 0 less than those in the cancer cells. The staining pattern and percentage of positive tumor cells was similar in each case using either PClO or MIB-1 by semiquantitative method. However, the staining in MIB-1 was easier to interpret because it had less background staining and more uniform and stronger positive signals when compared with PCIO. Thus, we performed the scoring based on MIB- 1 positivity. MIB-1 Staining and Clinicopathologic Parameters The tumor MIB-1 scores of the whole group of patients ranged from 169 to 964 (mean, 598 z 211; median, 636). The scores had no significant relationship to the age of the patients and the size of the tumors ( P = 0.64 and 0.67, respectively, by correlation analysis). The relationship of other clinicopathologic features to the scores is noted in Table 1. The MIB-1 score was not related to the sex of the patients, the location of the tumors, or the stage of the lesions ( P = 0.91, 0.25, and 0.84, respectively, by KruskalWallis test). Conversely, the score was strongly correlated with the differentiation of the squamous cell carcinomas ( P = 0.0001 by Kruskal-Wallis test). The median score in the poorly differentiated squamous cell carcinomas was much higher than the well differentiated squamous cell carcinomas. By statistical manipulation, in order to transform the score to a variable that is normally distributed, a value x could be calculated from the MIB- 1 scores (x = sin^' square root of [MIB-1 score11000]). The tumors could be classified as well differentiated squamous cell carcinomas if they had an X value greater than or equal to 0.9, whereas poorly differentiated squamous cell carcinomas had an x value of less than 0.9. Survival Analysis The mean follow-up period of the whole group was 16.5 months. Eighty-five percent of the patients had already 10 20 30 40 50 60 70 months FIGURE 4. The survival rate of patients with Stage I1 and Stage Ill esophageal squarnous cell carcinoma. died by the time of completion of the study. The median survival of the whole group was 11 months. The survival was not related to age, sex, site, tumor differentiation, or MIB-1 score ( P = 0.49, 0.84, 0.99, and 0.77, respectively, by the Cox regression). Conversely, survival of patients was related to the size ( P = 0.003) and stage ( P = 0.022) of the lesions (Fig. 4). In order to evaluate the relationship between MIB-1 scores and survival rates within a given stage, the tumors in Stage 111 (n = 85) were stratified into those with scores greater than 300 and those with scores less than or equal to 300. Patients whose tumors had a MIB-1 score of less than 300 (n = 11) had a median survival of 15 months, whereas those above that score ( n = 72) had a median survival of 9.7 months. The former group had a longer median survival, although the results were marginally insignificant ( P = 0.068 by the Cox regression analysis). We found the cutoff points for MIB-1 scores to start from 300, with stepwise increments of 50 to 750 in each of the 2 groups containing at least 10 patients. The value of 300 gave the largest value of the likelihood ratio statistics, thus defining the 2 most distinguishable groups (Fig. 5). The survival rates of patients with Stage I11 lesions and an MIB-1 score greater than 300 could be further stratified according to the size of the tumors. The cutoff points were derived from the same statistical method described above. The starting point was 2.5 cm, with increments of 0.5 cm. It was observed that patients with a maximum tumor diameter of less than or equal to 7.5 cm had a better survival than those with a maximum tumor Proliferative Markers MIB-1 and PClO in Esophageal CancerlLam et al. 11 100 ...... MIBl s c o r e > 300 MIBl score 300 60 N v 01 4 60 0 d 0 .a 2 40 a v) P 20 2o 0 0 10 20 30 40 50 60 70 months FIGURE 5. The relationship of the MIB-1 score and the survival rate in patients with Stage 111 esophageal squamous cell carcinomas. diameter of greater than 7.5 cm. The former group had a medium survival of 11.6 months, whereas the latter group had a medium survival of 5 months ( P = 0.016 by the Cox regression). The impact of MIB-1 score and size on the survival rate is summarized in Figure 6. Conversely, the MIB-1 scores in Stage 111 lesions could not be further stratified into groups if they were first stratified into 2 groups according to the size of the tumors. Thus, the size of the tumors was more important than the MIB1 scores in determining the behavior of the tumors in Stage 111 esophageal squamous cell carcinomas. For patients with Stage I1 and Stage IV lesions, the numbers were too few to be stratified into groups for proper statistical analysis. In addition, the results were not significant in these patients even if we relaxed the criteria of having at least 10 patients in each group for the analysis. In addition, it was noted that in patients with either Stage I1 or Stage 111 lesions, the survival was not significantly related to the differentiation of the tumors ( P = 0.42 for Stage 2; P = 0.27 for Stage 111). DISCUSSION Antibodies to PCNA have been widely used in recent years to study the kinetics of various human lesions.' The expression of PCNA, as identified by PC10, has been found to be correlated with other markers like mitotic counts, flow cytometric analysis, nucleolar organizer region-associated proteins (AgNORs), transforming growth factor, epidermal growth factor receptor, and bromodeoxyuridine (BrdU), as well as Ki-67.'"'* However, when com- 0 10 20 30 40 50 60 70 months FIGURE 6. The relationship of the MIB-1 score and tumor size on the survival rate of patients with Stage Ill esophageal squamous cell carcinomas. pared with MIB-1, PClO suffers from certain disadvantages: (1)the long half-life of PCNA makes staining persist in noncycling cells; (2) the staining may be positive in DNA repair in addition to DNA synthesis; and (3) the staining varies with antibody dilution and lacks a clear plateau.' In the present study, we observed that staining in MIB- 1 was easier to interpret because it had less background staining and stronger and more uniform positive signals when compared with PC10. In addition, consistent with the findings of other authors,' the staining pattern and percentage of positivity of PClO and MIB-1 analyzed by semiquantitative method were similar. Thus, we agree with the current thinking that MIB-1 is a more robust marker than PClO and use the MIB-1 score for asse~sment.',*,~ The limitations of clinical studies using proliferative markers include heterogeneity in proliferation within tumors and variability in factors such as case selections, number of cells counted, size of samples, fixation time, etc.' We employed strict criteria in this study in order to provide high quality data for interpretation. All the samples in the study were prospectively collected so the tissue fixation time and the number of blocks taken were standardized and the clinical/survival data were all adequately collected. The number of patients involved in the study was also large so as to allow more representative statistical testing results. In addition, the choice of representative block for staining, choice of representative fields for counting, counting of at least 1000 cells in each sec- 12 CANCER January 1,1996 / Volume 77 / Number 1 tion, and repetition of counting increased the chances of reproducibility of the MIB-1 scores and minimized the heterogeneity in proliferation within the tumors. Reports of proliferative activities of esophageal squamous cell carcinomas using Ki-67 on cryostat sections are relatively r a ~ e . ' ~ -The ' ~ number of cases reported in these studies are usually limited. The largest group of patients was presented by Porschen and colleague^.'^ They reported in their study of 33 German patients a mean Ki-67 labelling index (using cryostat sections) of 35.1% (ranging from 14.2 to 64.1). The authors could find no other data on the use of MIB-1 in esophageal squamous cell carcinomas in the English literature. In this study of 100 patients, all tumors were positive for MIB1. The mean MIB-1 score was high (598 2 211, range, 169 to 964). In concurrence with the results of Jaskiewicz and De G r ~ o t , 'Ki-67 ~ expression was noted in the normal epithelium, which could signify an early step of malignant transformation. It is worth noting that the MIB-1 scores in the nonneoplastic epithelia were much lower than in the tumor tissue in all the cases under study. This study showed no statistically significant relationship between MIB- 1score and clinicopathologic variables such as age, sex, tumor location, tumor size, and TNM stage. The results were consistent with the previous studies using Ki-67 (using cryostat sections).23,24 Conversely, the location of the positive staining in the periphery of tumor cell nests was consistent with the reports by most other authors."^" This study also demonstrated a highly significant correlation between the MIB-1 score and the differentiation of the tumors. The results suggested that the score could be used as an objective measurement for grading of the squamous cell carcinoma. Another important finding was that tumor stage and size of esophageal cancers were important parameters in assessing the prognosis of patients with esophageal squamous cell carcinoma. Most of the patients with esophageal squamous cell carcinomas in Hong Kong had Stage I11 lesions. The tumors in this group of patients could be stratified into two groups, those with MIB-1 scores of greater than 300 and those with scores less than or equal to 300, in the current study by statistical evaluation. It was noted that patients in the former group had a poorer prognosis than patients in the latter group (median survival of 9.7 months vs. 15 months). In addition, the former group could also be further subdivided into those with a tumor diameter of greater than 7.5 cm and those with a tumor diameter of less than or equal to 7.5 cm, with the former group having a poor prognosis (median survival of 11.6 months vs. 5.3 months). Thus, the MIB-1 score may be a useful adjunct in assessing tumor behavior, but the usefulness of the scores was not as important as the stage and size of the tumors in this study. In conclusion, our results indicate that cellular prolif- eration in esophageal squamous cell carcinomas, as defined by MIB- 1 immunohistochemical study, correlate with cellular differentiation and, in addition to the tumor size, may be a good indicator in predicting the prognosis in Stage 111 cancers. REFERENCES 1. Yu CCW, Filipe MI. Update on proliferation-associated antibodies applicable to formalin fixed paraffin embedded tissue and their clinical applications. Histochern J 1993;25~843-53. 2. Yu CCW, Woods AL, Levison DA. The assessment of cellular proliferation by immunohistochemistry: a review of currently available methods and their applications. Histochem J1992;24:121-31. 3. Gerdes 1, Schwab U, Lemke H, Stein H. Production of a monoclonal antibody reactive with a human nuclear antigen associated with cell proliferation. Int J Cancer 1983;31:13-20. 4. Fonatsch C, Duchrow M, Rieder H, Schluter C, Gerdes J. Assignment of the human Ki-67 gene (MK167) to 10q25qter. Genonzics 1991;11:476-7. 5. Key G, Becker MHG, Baron B, Duchrow M, Schluter C, Flad HD, et al. New Ki-67 equivalent murine monoclonal antibodies (MIB 1-3) generated against bacterially expressed parts of the Ki-67 cDNA containing three 66 bp repetitive elements encoding for the Ki-67 epitope. Lab Invest 1993;68:629-36. 6. Schluter C, Duchrow M, Wohlenberg C, Becker MHG, Key G, Flad HD, et al. The cell proliferation-associated antigen of antibody Ki-67: a very large, ubiquitous nuclear protein with numerous repeated elements, representing a new kind of cell cycle-maintaining proteins. J Cell BioZ 1993;123:513-22. 7. Meggetto F, Al Saati T, Cohen-Knafo E, Roubinet F, Selves J, Bouche G, et al. Production of a monoclonal antibody (IND.64) identifymg a cell cycle-associated antigen using spleen cells from nude mice bearing Ichikawa tumour. J Path01 1992; 168:187-96. 8. Cattoretti G, Becker MHG, Key G, Duchrow M, Schluter C, Galle J, et al. Monoclonal antibodies against recombinant parts of the K1-67 antigen (MIB 1 and MIB 3) detect proliferating cells in microwave-processed formalin-fixed paraffin sections. J Pathol 1992;168:357-63. 9. McCormick D, Chong H, Hobbs C, Datta C, Datta C, Hall PA. Detection of the Ki-67 antigen in fixed and wax-embedded sections with the monoclonal antibody MIB 1. HistopatholO ~ Y1993;22:355-60. 10 Pisani P, Parkin DM, Ferlay 1. Estimates of the worldwide mortality from eighteen major cancers in 1985. Implications for prevention and projections of future burden. I n t J Cancer 1993;55891 -903. 11. Furihata M, Ohtsuki Y, Ogoshi S, Takahashi A, Tamiya T, Ogata T. Prognostic significance of human papillomavirus genomes (type-16,-18) and aberrant expression of p53 protein in human esophageal cancer. Int J Cancer 1993;54:22630. 12. Nishihira T, Hashimoto Y , Katayama M, Mori S, Kuroki T. Molecular and cellular features of esophageal cancer cells. J Cancer Res Clin Oncol 1993;119:441-9. 13. Watanabe H, ]ass JR, Sobin LH, in collaboration with pathologists in eight countries. Definition and explanatory notes. In: World Health Organization International histological classification of tumours: histological typing of oesophageal and gastric tumours. 2nd edition. Berlin: Springer-Verlag, 199O:ll-18. Proliferative Markers MIB-1 and PClO in Esophageal Cancer/Lam et al. 14. Beahrs OH, Henson DE, Hui:ter R I P , Kennedy BJ. Manual for staging of cancer. 4th edition. Philadelphia: JB Lippincott, 199257-59. 15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assa: 1958;158:23-39. 16. Wang LD,Shi ST, Zhou Q, Goldstein S, Hong JY,Shao P, et al. Changes in p53 and cyclin D1 protein levels and cell proliferation in different stages of human esophageal and gastric-cardia carcinogenesis. In2 J Cancer 1994;59:514-9. 17. Hall PA, Levison DA, Woods AL, Yu CCW, Kellock DB, Watkins ]A, et al. Proliferation cell nuclear antigen (PCNA) immunolocalization in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some neoplasms. J Puthol 1980;162285-94. 18. Yu CCW, Fletcher CDM, Kewman PL, Goodlad JR, Burton JC, Levison DA. A comparison of proliferative cell nuclear antigen (PCNA)immunostaining, nucleolar organizer region (AgNOR) staining, and histological grading in gastrointestinal stromal tumours. J Puthol 1992; 166:147-52. 19. Woods AL, Hall PA, Shepherd NA, Handby AM, Waseem NH, Lane DP, et al. The assessment of proliferating cell nuclear antigen (PCNA) immunostaining in primary gastrointestinal lymphomas and its relationship to histological grade, S + G, + M phase fraction (flow cytometric analysis) and prognosis. Histopathology 1991;19:21--27. 20. Hickey K, Grehan D, Reid IM, O'Briain S, Walsh TN, Hennessy TPJ. Expression of epidermal growth factor receptor and proliferating cell nuclmear antigen predicts response of esophageal squamous cell carcinoma to chemoradiotherapy. Cancer 1994;74:1693-98. 13 21. Miyazaki S, Sasano H, Suzuki T, Sawai T, Nishihira '1, hlori S. Nucleolar organizer regions in human esophageal disnrders: comparison with proliferating cell nuclear antigen by ininiiinostaining. Mod Puthol 1992;5:396-401. 22. Jankowski J, McMenemin R, Yu C, Hopwood D, Wornisicy KG. Proliferating cell nuclear antigen in oesophageal diseases: correlation with transforming growth factor alpha expression. Gut 1992;33:587-91. 23. Porschen R, Classen S, Piontek M, Borchard F. Vasculnriziition of carcinomas of the esophagus and its correlation with tumor proliferation. Cancer Res 1994;54:587-9 1. 24. Porschen R, Kriegel A, Langen C, Classen S, H i k e M , I.ohe B, et al. Assessment of proliferative activity in carcinomas of the human alimentary tract by Ki-67 immunostaining. / / , I J Cancer 1991;47:686-91. 25. Jaskiewicz K, DeGroot KM. p53 gene mutants expression. cellular proliferation and differentiation in oesophageal catcinoma and non-cancerous epithelium. Aiiticnrzrer Kr,s 1994; 14:137-40. 26. Matsumura K, Tsuji T, Shinozaki F, Sasaki K, Takahaslii M. Immunohistochemical determination of growth fraction i i i human tumors. Pathol Res Pruct 1989;184:609-13. 27. Sasano H, Miyazaki S, Nishihira T, Sawai T, Nagura H . 'l'hr proliferative cell fraction in cytology speciniens: a sriitly ol human esophageal carcinoma. A m J Clin f'nf/io/ 1992;98: 161-6. 28. Sasano H, Miyazaki S, Gooukon Y, Nishihira T, Sawai .I,, Nagura H. Expression of p53 in human esophageal carcinoma: an immunohistochemical study with correlation t o proliferating cell nuclear antigen expression. H i r m I'trilrol 1992;23:1238-43.