Aim-To carry out an objective assessment of two systems of scoring immunohistochemical staining, evaluating interobserver and intraobserver error. Methods-92 cervical tumours underwent immunohistochemical staining for p53 and epidermal growth factor receptor. Staining was assessed using two methods: a standard 4 point scale and a descriptive method, performed by three observers. Interobserver and intraobserver error were assessed for both scoring methods. Results-In terms of interobserver error between three observers, no diVerence was found between a simple 4 point scale method of evaluation and the use of a highly circumscribed method. In all evaluations, interobserver error was scored as moderate ( w 0.48-0.49). However, evaluation of immunohistochemical staining by a panel of observers led to a marked improvement in the interobserver error scores ( w 0.63). Conclusions-There should be standardisation of immunohistochemical staining and scoring methods. More attention should be paid to measurement of interobserver and intraobserver error in studies. Use of a panel of tissue control slides and consensus scoring by several observers can lead to improvement in reproducibility. (J Clin Pathol 1999;52:75-77)

COX-1 and COX-2 are members of the cyclooxygenase (COX) family, which influence tumor invasion and apoptosis. The objective of the study was to assess the relationship between COX-1 and COX-2 expression in early-stage disease and subsequent disease relapse and long-term survival. Women with FIGO stage I and II cervical carcinoma, younger than 50 years, treated between 1981 and 1990 were included. COX-1 and COX-2 expressions in the tumors were assessed by immunohistochemistry. COX-1 and COX-2 were expressed in 61% (17/28) and 57% (16/28) of tumors, respectively. COX-1 nonexpressers showed an improved overall survival compared to expressers (log-rank test, P= 0.09). There was no significant difference in the overall survival in COX-2 nonexpressers compared to expressers (P= 0.6). Out of eight women with disease relapse, COX-1 or COX-2 expression was noted in six of eight tumors, and both were expressed in five of eight tumors. Our preliminary data suggest an adverse prognosis with COX-1 expression in early-stage cervical carcinoma and a trend toward COX-1 expression in disease relapse. The association between COX-2 expression and a worse prognosis was not proven in this study.

No abstract