753 HOW RELIABLE IS TUMORGRADING IN PENILE CANCER–RESULTS OF AN INTERNATIONAL GRADING STUDY

INTRODUCTION AND OBJECTIVES: Since reliable molecular prognostic parameters for inguinal lymph metastases are not known, tumor grading is often used as prognostic tool for indication of lymphadenctomy and is now integrated into T category of TNM classification. The reliability of tumor grading remains under discussion. Therefore we examined inter-observer variability in grading in an international grading study. METHODS: 112 patients with histological diagnosed primary penile carcinomas were included into this study. Specimens were transferred from formalin-fixed, paraffin-embedded donor blocks to recipient block via TMA device (Beechers Instruments Inc., Maryland, USA). For assembling the TMA block, each case was represented twice, one core from the central tumor proportion and one the invasion front of the tumor. The H&E stained slights were sent to 7 Uropathologist in different countries. The grading of the tumors was compared and interobserver variability was examined (calculated) for matched pairs with the Cohen’s kappa coefficient. RESULTS: We found great differences in between the observers. The percentage of described G1 tumors ranged from 7.3–39.1%, G2 tumors form 28–67.1 % and for G3 tumors from 20-39.4%. The interobserver variability was found to be very high. Kappa coefficient was low and reached moderate levels only in 46.4% of the matched pairs (range 0.19-0.67). CONCLUSIONS: Due to the great interobserver variability the tumor grading of penile carcinomas does neither appear as a reproducible nor as a reliable predictor for the malignant potential of the tumors. Therefore inclusion of grading data into T classification seems not useful and the inclusion of other prognostic markers is necessary for better characterization of the tumors.