* Context.-Traditional morphologic features of tumor aggression are of limited value in predicting the malignant behavior of endocrine neoplasms. We explored the potential value of nuclear proliferative activity (using Ki-67 immunostaining with semiquantitative scoring) in predicting the clinical behavior of pancreatic islet cell tumors (ICTs), and we correlated this characteristic with hormone expression.
Objective.-To evaluate whether Ki-67 immunostaining using a semiquantitative scoring system has value in predicting the clinical behavior of pancreatic ICTs.
Design.-We studied 39 pancreatic ICTs from 39 patients. Twenty-two ICTs did not metastasize in a median follow-up period of 91 months. The remaining 17 neoplasms did produce metastases (8 in liver, 7 in regional lymph nodes, and 2 in peritoneum). Immunohistochemistry was performed using antibodies to Ki-67 and pancreatic hormones (insulin, glucagon, gastrin, somatostatin, pancreatic polypeptide, vasoactive intestinal polypeptide, and corticotropin). A semiquantitative Ki-67 grading system was followed. The nuclear proliferative activity, as deter
mined by a positive reaction for Ki-67, was considered low (25% of cells staining positively).
Results.-The majority of the nonmetastatic ICTs (16 cases, 73%) demonstrated either negative or low staining for Ki-67 (P
Conclusion.-An ICT with low nuclear proliferative activity is unlikely to metastasize, whereas high proliferative activity is associated with a metastatic phenotype. Immunohistochemical assessment of Ki-67 using a semiquantitative scoring system is a simple and reliable detection method of cellular proliferative activity in ICTs of the pancreas.
The histologic criteria for prediction of malignancy in pancreatic islet cell tumors (ICTs) still elude pathologists. Classic architectural and cytologic features of malignancy have failed to predict the biological behavior of these neoplasms reliably.1-3 The only indisputable evidence of malignancy in an ICT is infiltration of adjacent organs or distant metastasis.3 Several ancillary tests for assessment of the metastatic potential of this group of neoplasms have been proposed. These include nuclear morphometry,4 immunohistochemical detection of various hormones,5 alpha-subunit of human chorionic gonadotropin,6,7 progesterone receptor,8 and oncogene expression.9-11 Determination of cell kinetics, such as DNA analysis12,13 and detection of nucleolar organizer regions,14 has also been used. These techniques, however, have yielded inconclusive results. More recently, immunohistochemical detection of cell cycle-related antigens, such as proliferating cell nuclear antigen15 and Ki-67,(16,17) has shown promise in pre-- dieting malignancy in ICTs. Pelosi et al17 demonstrated that the Ki-67 index might be an indicator of metastatic potential in ICTs and an independent predictor of survival. Ferret et al,16 using the monoclonal antibody MIB-1, also showed that the cell proliferation index is a useful predictor of behavior in these tumors. Both studies graded the immunohistochemical reaction after the evaluation of at least 1000 tumor nuclei at x 1000 magnification.
Since a semiquantitative scoring system is easy for pathologists to perform routinely, we explored its application to Ki-67 immunodetection for the assessment of possible metastatic behavior of ICTs. Furthermore, we investigated the relationship of hormone production and metastatic phenotype in these neoplasms.
MATERIALS AND METHODS
We studied 39 pancreatic ICTs from 39 patients. Twenty-two of the patients (56%) had no evidence of metastatic disease in a median follow-up of 91 months (range 3-300 months), and their neoplasms were considered nonmetastatic ICTs. The remaining 17 patients (44%) had metastatic ICTs. In the latter group, immunoperoxidase studies were performed in both the primary pancreatic and the metastatic neoplasm.
Three-micrometer histologic sections of 10% formalin-fixed, paraffin-embedded tissue were deparaffinized in xylene and hydrated in decreasing grades of ethanol. Immersing the slides in 6% hydrogen peroxidase for 3 minutes blocked endogenous peroxidase activity. The slides were then washed in water and phosphate-buffered saline. When antigen retrieval was recommended, the slides were placed in a preheated 10% solution of target retrieval (S1699, Dako Corporation, Carpinteria, Calif) and heated in a vegetable steamer for 20 minutes. After cooling for 20 minutes at room temperature, slides were washed with phosphate-- buffered saline. All slides were then treated with normal horse serum for 5 minutes and incubated with the primary antibody. Primary antibodies used, along with type, code, source, working dilution, incubation period, and antigen retrieval method, are specified in Table 1. Slides were then incubated for 25 minutes with linking solution (LSAB Kit, Universal, K0690, Dako) and for 25 minutes with streptavidin-peroxidase, using phosphate-buffered saline washings between steps. Diaminobenzidine (K3468, Dako) was the chromogen. Cupric sulfate exposure (5 minutes) was used for Ki-67 with fast green counterstain. Harris hematoxylin was the counterstain for the remainder of the antibody reactions. The slides were then rinsed in tap water and dehydrated in increasing grades of isopropyl alcohol, cleared with xylene, and mounted using a synthetic neutral resin.
Ki-67 immunoreaction was evaluated using light microscopy at x400 magnification by 2 independent observers (M.J. and Z.G.). A reaction for Ki-67 was considered positive when nuclear staining in the tumoral cells was identified; the reaction was graded as low (25% of cells staining positively).
P values were calculated by chi^sup 2^ test using Microsoft Excel 7.0 for Windows 98 (Microsoft Corp, Redmond, Wash).
Immunoperoxidase results for both nonmetastatic and metastatic ICTs are displayed in Tables 2 and 3.
The mean age of patients with nonmetastatic ICT was 50 years (range 20-77 years), with a slight predominance for women (male-female ratio 1:1.2). The mean follow-up was 91 months (range 3-300 months). Nine (41%) of the 22 nonmetastatic ICTs did not show any hormone production detectable by immunohistochemistry or by serum blood test, and therefore these cases were considered nonfunctioning endocrine tumors. Of the remaining 13 (59%) cases, 6 were insulinomas, 6 were glucagonomas, and 1 was a vipoma.
The mean age of patients with metastatic ICT was 51 years (range 20-72 years), with a slight predominance of women (male-female ratio 1:1.1). Eight metastatic ICTs showed metastasis in liver, 7 in regional lymph nodes, and 2 in peritoneum. Seven metastatic ICTs (41%) were functioning tumors. Six of these tumors were gastrinomas, and 1 neoplasm was positive for corticotropin. In the remaining 10 metastatic ICTs (59%), no hormone production was detected by immunohistochemistry or serum sample. No differences in Ki-67 immunostaining results were found between the pancreatic neoplasms and their metastases.
Most of the nonmetastatic ICTs (16 cases, 73%) showed either negative staining or low nuclear proliferative activity (Figure 1) (P
Immunohistochemical detection of cell cycle-related antigen, such as proliferating cell nuclear antigen and Ki-67, is a reliable method for studying cell proliferative activity in a variety of tumors. The Ki-67 protein, which is present during all active phases of the cell cycle but is completely absent in resting cells, is a dependable marker for estimation of growth fraction or cell proliferation.18 This antibody has been shown to be superior to proliferating cell nuclear antigen as a predictor of survival in the study of ICTs.17 Perret et al,16 using MIB-1 antibody immunodetection in paraffin-embedded tissues, also showed that the Ki-67 proliferation index is a useful prognostic parameter in pancreatic endocrine neoplasms and that it can be used to assess the cell proliferative activity of these tumors. Furthermore, it may help establish their prognosis and therefore separate potentially metastatic neoplasms from those with little or no metastatic potential.16,17
Previous reports have graded the immunohistochemical detection of Ki-67 using an index method of quantification.15-17 This method requires counting of at least 1000 tumor nuclei at X1000 magnification. In the current study, the semiquantitative assessment of Ki-67-reactive cells was equally effective in separating nonmetastatic from metastatic tumors; hence, it is a useful marker for malignant behavior in ICTs. This method of quantification is faster and more practical for routine use by pathologists. Our results are in agreement with the findings of Von Herbay et al,19 who demonstrated in a smaller number of patients that Ki-67 immunohistological semiquantitative scores correlated with those obtained with DNA flow cytometry of neuroendocrine tumors of gastroenteropancreatic origin.
Previous studies using the index scoring system for quantification of Ki-67 immunoreaction in ICTs found that indexes greater than 5% are associated with poor prognosis.16,17 Our results support these observations, since the neoplasms that showed positive Ki-67 reaction in less than 5% of the tumor cells (low reaction) did not demonstrate metastatic behavior.
We did not observe statistical differences between the metastatic behavior of ICTs with or without hormone expression. Similarly, there were no statistically significant differences between the Ki-67 proliferative scores of endocrine pancreatic neoplasms with and without hormone production.16-20
Our results support the hypothesis that ICTs with low nuclear proliferative activity are unlikely to demonstrate metastatic behavior. Conversely, tumors that exhibit high nuclear proliferative activity are likely to be associated with a metastatic phenotype. These findings should be useful in predicting the clinical aggressiveness of an ICT and could help in preselecting patients who might benefit from rigorous follow-up. Furthermore, immunoperoxidase detection of Ki-67 with a semiquantitative scoring system has proved to be a simple and reliable method for detection of cell proliferative activity in these neoplasms.
The authors thank the Cancer Registry at Jackson Memorial Hospital, and especially Martha E. Oliva, BS, RHIT, Jackson Memorial Hospital Cancer Registry, for assistance in collecting data.
1. Lam K, Lo C. Pancreatic endocrine tumour: a 22-year clinico-pathological experience with morphological, immunohistochemical observation and a review of the literature. EurJ Surg Oncol. 1997;23:36-42.
2. Cubilla A, Hadju S. Islet cell carcinoma. Arch Pathol. 1975;99:204-207.
3. Kloppel G, Heitz P. Pancreatic endocrine tumors. Pathol Res Pract. 1988; 183:155-168.
4. Kenny B, Sloan J, Hamilton P, et al. The role of morphometry in predicting prognosis in pancreatic islet cell tumors. Cancer. 1989;64:460-465.
5. Solcia E, Capella C, Buffa R. The contribution of immunochemistry to the diagnosis of neuroendocrine tumors. Semin Diagn Pathol. 1984;1:285-296.
6. Heitz P, Kasper M, Kloppel G, Polak J, Vaitukaitis J. Glycoprotein alpha-- chain production by pancreatic endocrine tumors: a specific marker for malignancy: immunocytochemical analysis of tumors of 155 patients. Cancer. 1983; 51:277-282.
7. Graeme-Cook F, Nardi G, Compton C. Immunocytochemical staining for human chorionic gonadotrophin subunits does not predict malignancy in insulinomas. Am I Clin Pathol. 1990;93:273-276.
8. Viale G, Doglioni C, Gambacorta M, Zamboni G, Coggi G, Bordi C. Progesterone receptor immunoreactivity in pancreatic endocrine tumors: an immunocytochemical study of 156 neuroendocrine tumors of the pancreas, gastrointestinal and respiratory tracts and skin. Cancer. 1992;70:2268-2277.
9. Lohmann D, Funk A, Niedermeyer H. Identification of P53 gene mutations
in gastrointestinal and pancreatic carcinoids by nonradioisotopic SSCA. Virchows Arch B Cell Pathol Incl Mol Pathol. 1993;64:293-296.
10. Roncalli M, Springall D, Varndell I, et al. Oncoprotein immunoreactivity in human endocrine tumors. J Pathol. 1991;163:117-127.
11. Shimizu M, Saitoh Y, Itoh H. Immunohistochemical staining of Ha-ras oncogene product in normal, benign and malignant human pancreatic tissues. Hum Pathol. 1990;21:607-612.
12. Alanen K, Joensuu H, Klemi P, Marin S, Alavaikko M, Nevalainen T. DNA ploidy in pancreatic neuroendocrine tumors. Am J Clin Pathol. 1990;93:784-788. 13. Donow C, Baisch H, Heitz P, Kloppel G. Nuclear DNA content in 27 pancreatic endocrine tumours: correlation with malignancy, survival and expression of glycoprotein hormone alpha chain. Virchows Arch A Pathol Anat Histopathol. 1991;419:463-468.
14. Ruschoff J, Willemer S, Brunzel M, et al. Nucleolar organizer regions and glycoprotein-hormone alpha-chain reaction as markers of malignancy in endocrine tumors of the pancreas. Histopathology. 1993;22:51-57.
15. Pelosi G, Zamboni G, Doglioni C, et al. Immunodetection of proliferating cell nuclear antigen assesses the growth fraction and predicts malignancy in endocrine tumors of the pancreas. Am J Surg Pathol. 1992;16:1215-1225.
16. Perret A, Mosnier J, Buono J, et al. The relationship between MIB-1 proliferation index and outcome in pancreatic neuroendocrine tumors. Am J Clin Pathol. 1998;109:286-293.
17. Pelosi G, Bresaola E, Bogina G, et al. Endocrine tumors of the pancreas: Ki-67 immunoreactivity on paraffin sections is an independent predictor for malignancy: a comparative study with proliferating-cell nuclear antigen and progesterone receptor protein immunostaining, mitotic index, and other clinicopathologic variables. Hum Pathol. 1996;27:1124-1134.
18. Scholzen T, Gerdes J. The Ki-67 protein: from the known and the unknown. j Cell Physiol. 2000;182:311-322.
19. Von Herbay A, Sieg B, Schurmann G, Hofmann WJ, Betzler M, Otto HF. Proliferative activity of neuroendocrine tumours of the gastroenteropancreatic endocrine system: DNA flow cytometric and immunohistological investigations. Gut, 1991;32:949-953.
20. Venkatesh S, Ordonez N, Ajani J, et al. Islet cell carcinoma of the pancreas. Cancer. 1990;65:354-357.
Merce Jorda, MD, PhD; Zeina Ghorab, MD; Gustavo Fernandez, MD; Mehdi Nassiri, MD; Andrew Hanly, MD; Mehrdad Nadji, MD
Accepted for publication August 29, 2002.
From the Departments of Pathology (Drs Jorda, Ghorab, Nassiri, Hanly, and Nadji) and Clinical Oncology (Dr Fernandez), University of Miami, Jackson Memorial Medical Center, Miami, Fla.
Reprints: Merce Jorda, MD, PhD, Department of Pathology, University of Miami, 1611 NW 12th Ave, Holtz Center, 2nd Floor, 2042-E, Miami, FL 33136 (e-mail: email@example.com).
Copyright College of American Pathologists Feb 2003
Provided by ProQuest Information and Learning Company. All rights Reserved