Objectives: To evaluate concordance, using the kappa statistic, between tumour histotypes assigned at initial diagnosis and those after slide revision, and to assess the independent predictive value of reviewed tumour histotype.
Materials and methods: We analysed records of 491 patients who had undergone partial or radical nephrectomy for renal cell carcinoma (RCC) in our department from 1986 to 2000. Pathologic slides were reviewed by one uropathologist blinded to original diagnoses and follow-up data, who reassigned histologic subtype and nuclear grades.
Results: Concordance between original and reviewed histotypes was substantial (kappa = 0.52). Data were affected by the year of initial diagnosis: Concordance was moderate (kappa = 0.43) in patients treated from 1986 to 1997 and substantial (kappa = 0.73) in those observed thereafter. The original histotype did not stratify the cancer-related outcome. According to the reviewed histotypes, the 5-year cancer-specific survival probabilities of conventional, papillary and chromophobe RCC patients were 81.3%, 90.1% and 100%, respectively (p = 0.009). Patients with conventional RCC showed significantly lower cancer-specific survival than those with papillary or chromophobe RCC, even in the cases with locally advanced (p = 0.04) or high-grade (p = 0.02) neoplasms. The reviewed histologic subtype was not an independent prediction of cancer-specific survival on multivariate analysis.
Conclusion: Concordance between histologic subtype assigned before 1997 and after slide revision was only moderate. In univariate analysis, reviewed tumour histotype stratified cancer-related outcome, but the data of multivariate analysis were strongly influenced by the prevalent number of conventional RCC.