AMERICAN JOURNAL OF CLINICAL PATHOLOGY Editorial Ki-67 Staining as a Means to Simplify Analysis of Tumor Cell Proliferation 444 of cell proliferation. Sahin and colleagues demonstrate that Ki-67 staining and the mitotic count correlate in some but not all tumor types. Those S-phase cells labeling with 3 H-thymidine or bromodeoxyuridine form a subpopulation of Ki-67-staining cells. This yields labeling indices smaller than Ki-67-staining values. Studies of various tumors5'6 demonstrate a strong correlation between Ki67 staining and thymidine or bromodeoxyuridine-LI. The correlation is equal to or better than that with mitotic figures. There is no need to use these latter technically complex and time-consuming tests outside the research laboratory. Many considerflowcytometry to be a rapid and reliable method for assessment of tumor cell DNA ploidy and proliferation. Its advantage is the ability to analyze large numbers of cells or nuclei in a short period of time. This is complemented by the use of dual-color analysis and light scatter parameters to define better the proliferating cells of interest.7 This analysis can be accomplished without direct pathologist intervention. For a routine pathology practice, its disadvantages appear to outweigh its advantages. The need for a single suspension of cells or nuclei prevents direct assessment of tumor cell heterogeneity. Analysis of DNA histograms is as much an art as a science.8 The "accuracy" of this analysis is influenced by the amount of tissue necrosis and normal stromal components, the presence or absence of DNA aneuploid tumor cells, as well as by instrument variables. The ability of Ki67 to identify directly proliferating tumor cells overcomes several problems associated with DNA aneuploid tumors. These include a significant overlap of DNA aneuploid and DNA diploid peaks, the presence of only a small DNA aneuploid peak, and the presence of multiple DNA aneuploid peaks. Despite these differences, Sahin and colleagues and others910 report a good correlation between Ki-67 staining andflowcytometry results in carefully designed studies. Sahin and colleagues indicate that this correlation appears to be better in tumors with a high rather than a low percentage of S phase cells or percentage of S phase cells plus G2/M value. As the authors explain, this discordance is due to dilution of tumor cells by stromal cells and the variation in the length of Gl between tumor cells with low and high proliferation (S phase) rates. The third point to consider is that techniques to quantify cell proliferation lack precision and accuracy inde- Downloaded from http://ajcp.oxfordjournals.org/ by guest on September 30, 2016 There is an increasing demand by clinicians for pathologists to provide objective data that will help them guide therapy and predict the prognosis of individual cancer patients. Recent reports indicate that measurement of tumor cell proliferation yields such useful data.1'2 The pathologist must choose among several methods used to measure this parameter. These include counting mitotic figures, measuring the labeling index after incorporation of 3H-thymidine or bromodeoxyuridine, cell-cycle analysis by flow cytometry, and Ki-67 immunostaining of a nuclear protein associated with cell proliferation. With each technique having its advantages and disadvantages, the final decision may well be based on which method is the simplest to set up and interpret. In a routine pathology practice, immunohistochemical staining with Ki-67 monoclonal antibody is arguably the best method to measure cell proliferation. Its staining of frozen sections, smears, and imprints provides direct and rapid identification of proliferating cells, without timeconsuming procedures or expensive equipment. Although this equates to a cost-effective test easily established in pathology laboratories already familiar with immunohistochemical procedures, there are several factors to consider before deciding. First, the methods to quantify cell proliferation do not necessarily measure the same thing. The Ki-67 monoclonal antibody recognizes an antigenic epitope, associated with distinct DNA-binding nuclear protein, that is destroyed in fixed, paraffin-embedded tissue.3 Cell staining intensity is low in G1 and increases to peak levels in cells in the G2 and M phases of the cell cycle. There is no staining of quiescent (GO) cells.4 This pattern of cell staining identifies the growth fraction of the tumor. This contrasts with the other techniques. Counting mitotic figures measures only a portion of those cells in the G2 phase and all M-phase cells. Thymidine and bromodeoxyuridine-LI labeling indices directly measure those cells in the S phase only. Flow cytometric analysis of DNA staining quantifies cell proliferation as a function of cells in the S phase alone or in combination with the G2/M phase of the cell cycle. Second, despite these differences, studies, such as that by Sahin and colleagues (pages 517-524) in this issue of the American Journal of Clinical Pathology demonstrate that Ki-67 staining correlates well with the other measures 445 HITCHCOCK The last point to consider is whether Ki-67 staining can provide clinically useful information. Several reports indicate that it provides information useful to the clinician. Sahin and colleagues demonstrate a good correlation between Ki-67 staining and the histopathologic features of malignant conditions. The ability of Ki-67 staining to measure cell proliferation in small samples has been combined with sequential in situ sampling of rectal carcinomas to determine the effect of treatment.17 A review of the literature indicates that it is the selection of the cut-off point rather than the evaluation method that influences the prognostic significance of Ki-67 staining. This also is true for the other methods of measuring tumor cell proliferation. The level of Ki-67 staining has an exponential pattern in most tumors. Thus the median rather than the mean value is a more representative place to start determining the cut-off point. Ki-67 staining has been used to determine the growth fraction in non-Hodgkin's lymphomas,1018 melanoma,19 non-small cell carcinoma of the lung,16 and breast carcinomas.6-913'20 The level of expression parallels tumor grade and the clinical behavior of these tumors. All efforts must be made to standardize the methods and instrumentation used to measure tumor cell proliferation. This is needed to minimize both interlaboratory and intralaboratory variation in results, and must be accomplished before the true clinical significance of this technology is known. Of the methods described, quantification of Ki-67 staining should be the simplest to standardize. CHARLES L. HITCHCOCK M.D., Ph.D. Department of Cellular Pathology Armed Forces Institute of Pathology Washington, D.C. REFERENCES 1. Merkel DE, McGuire WL. Ploidy, proliferative activity and prognosis. DNAflowcytometry of solid tumors. Cancer 1990;65:11941205. 2. Meyer JS. Cell kinetics in selection and stratification of patients for adjuvant therapy of breast carcinoma. NCI Monogr 1986; 1:2528. 3. McGurrin JF, Doria Ml, Dawson PJ, et al. Assessment of tumor cell kinetics by immunohistochemistry in carcinomas of the breast. Cancer 1987;59:1744-1750. 4. Gerdes J, Lemke H, Baisch H, et al. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody KJ-67. J Immunol 1984;133:1710-1715. 5. Sasaki K, Matsumura K, Tsuji T, Shinozaki F, Takahashi M. Relationship between labeling indices of Ki-67 and BrdUrd in human malignant tumors. Cancer 1988;62:989-993. 6. Kamel OW, Franklin WA, Ringus JC, Meyer JS. Thymidine labeling index and Ki-67 growth fraction in lesions of the breast. Am J Pathol 1989;134:107-113. 7. Zarbo RJ, Visscher DW, Crissman JD. Two-color multiparametric method for flow cytometric DNA analysis of carcinomas using staining for cytokeratin and leukocyte-common antigen. Anal Quant Cytol Histol 1989,11:391-402. Vol. 96 No. 4 Downloaded from http://ajcp.oxfordjournals.org/ by guest on September 30, 2016 pendent of the laboratory or methodology. Ki-67 staining methodology is less rigorous than the others, and thus a better candidate for standardization. Good-to-excellent staining intensity is obtained from frozen sections, smears, or touch imprints fixed with acetone or 1% paraformaldehyde alone or in combination with methanol." Prolonged storage of frozen tissue or fixed slides does not affect staining.1213 Detection varies with the threshold of sensitivity of the antibody used. Cells with relatively low Ki-67 content could appear negative if the immunocytochemical system is titered to detect cells with high levels. Therefore, all antibody dilutions should be determined using a checkerboard titration technique. The peroxidaseanti-peroxidase and alkaline phosphatase-anti-alkaline phosphatase staining techniques yield equivalent results.14 Dual staining with another antibody-chromogen technique can also be used for those difficult cases in which stromal and tumor cells cannot be distinguished clearly.1213 Ki-67 staining, as well as counting mitoticfiguresand 3 H-thymidine- and bromodeoxyuridine-labeled cells, suffer from a lack of standardized sampling of microscopic fields and expression of results. In their article, Sahin and colleagues note that Ki-67 staining is heterogenous in most tumors; oftentimes restricted to specific areas. This requires a pathologist to undertake a systematic approach to sampling the microscopic slide. This is time consuming and costly and is a significant disadvantage of this technique. Because all staining of nuclei is considered positive, interpretation of Ki-67 staining does not vary with the level of observer experience as noted for counting mitotic figures.15 There is no standard method for sampling the slide. Care must be taken to insure that selection of fields does not bias evaluation of the tumor cells. The pathologist can randomly sample areas or select areas with the highest labeling density. There appears to be a good correlation of results from either approach as long as enough fields are sampled. Wintzer and co-workers13 recommended that at least 10 high-power fields be sampled and report that sampling of 40 high-power fields yields statistically representative results regardless of the total number of tumor cells counted per section. The influence of section planes on their results was small. The number of positive cells can be expressed relative to the number of high-power fields, per number of tumor cells, or per square millimeter. The use of an ocular reticule overcomes the problem of variation in the area of a single high-power field among different microscopes, and permits counting of the number of stained cells per square millimeter. Quantification of Ki-67 staining can be adapted easily to image analysis techniques. The results are associated with total nuclear area staining, which is not the same as the percentage of positive nuclei.1016 446 AMERICAN JOURNAL OF CLINICAL PATHOLOGY Editorial 15. 16. 17. 18. 19. 20. the monoclonal antibody Ki-67. Am J Dermatopathol 1990;12: 458-461. Donhuijsen K. Mitosis counts: Reproducibility and significance in grading malignancy. Hum Pathol 1986; 17:1122-1125. Simony J, Pujol J-L, Radal M, et al. In situ evaluation of growth fraction determined by monoclonal antibody Ki-67 and ploidy in surgically resected non-small cell lung cancers. Cancer Res 1990;50:4382-4387. Porschan R, Lohe B, Hengels K-J, Borchard F. Assessment of cell proliferation in colorectal carcinomas using the monoclonal antibody Ki-67. Correlation with pathohistologic criteria and influence of irradiation. Cancer 1989;64:2501-2505. Slymen DJ, Miller TP, Lippman SM, et al. Immunobiologic factors predictive of clinical outcome in diffuse large-cell lymphoma. J Clin Oncol 1990;8:986-993. Kaudewitz P, Braun-Falco O, Ernst M, et al. Tumor cell growth fractions in human malignant melanomas and the correlation to histopathologic tumor grading. Am J Pathol 1989; 134:1063-1068. Brown RW, Allred DC, Clark GM, Tandon AK, McGuire WL. Prognostic significance and clinical-pathologic correlations of cellcycle kinetics measured by Ki-67 immunocytochemistry in axillary node-negative carcinoma of the breast. Abstract presented to the 13th Ann San Antonio Breast Cancer Symposium, 1990. A.J.C.P. • October 1991 Downloaded from http://ajcp.oxfordjournals.org/ by guest on September 30, 2016 8. Frierson HF. The need for improvement inflowcytometric analysis of ploidy and S-phase fraction. Editorial. Am J Clin Pathol 1991;95:439-441. 9. Isola JJ, Helin HJ, Hella MJ, Kallioniemi O. Evaluation of cell proliferation in breast carcinoma. Comparison of Ki-67 immunohistochemical study, DNAflowcytometric analysis, and mitotic count. Cancer 1990;65:1180-1184. 10. Schwartz BR, Pinkus G, Bacus S, Toder M, Weinberg DS. Cell proliferation in non-Hodgkin's lymphomas. Digital image analysis of Ki-67 antibody staining. Am J Pathol 1989;134:327-336. 11. van Dierendonck JH, Wijsman JH, Keijzer R, van de Velde CJH, Cornelisse CJ. Cell-cycle-related staining patterns of anti-proliferating cell nuclear antigen monoclonal antibodies. Comparison with BrdUrd labeling and Ki-67 staining. Am J Pathol 1991,138: 1165-1172. 12. Porschan R, Kriegel A, Langan C, et al. Assessment of proliferative activity in carcinomas of the human alimentary tract by Ki-67 immunostaining. Int J Cancer 1991;47:686-691. 13. Wintzer H-O, Zipfel I, Schulte-Monting J, Hellerich U, von Kieist S. Ki-67 immunostaining in human breast tumors and its relationship to prognosis. Cancer 1991;67:421-429. 14. Hofmann-Wellenhof R, Smolle J, Kerl H. The influence of staining procedures on the assessment of cell proliferation as defined by 447 IMAGEDISCS™ START W H E R E TEXTBOOKS A N D LECTURES STOP r Y ou've read die textbooks, heard the lectures and seen die slides. Now you can see thousands more images with the world's best teaching files on Videodisc, IMAGEDISCS: • Instant access and display • Up to 54,000 images/side. • Many examples for complete understanding. ticCJL <- *J -ff+e- I - ft.- • Multiple views of examples. • Superior image quality JOIN THE LEADERS: Leading medical societies • Motion sequences (real time). and institutions have chosen to publish on ImageDiscs™ including: • Rapid, stimulating, enjoyable ASCI? ARI? AFIR Mayo Clinic, Massachusetts General HospitalHarvard Medical School, ACR Institute, American Board of Radiology and die RSNA. To learn more call anytime, 1-800-848-4912, extension 2348. To order use die convenient postage paid order card to your left. learning. Downloaded from http://ajcp.oxfordjournals.org/ by guest on September 30, 2016 ImageDisc Library™ some of