Interobserver Reproducibility of Gleason Histological Grading of Prostatic carcinoma in a Teaching Hospital Setting in Sri Lanka

Objectives

To assess the interobserver reproducibility of Gleason grading of prostatic cancer among three pathologists

Method

Retrospective analysis of 50slides of previously diagnosed adenocarcinoma of prostate from radical prostatectomy and TURP specimens were subjected to Gleason grading on routinely processed haemotoxylin and eosin stained sections by three pathologists who did the grading independantly. The same 50 slides were reviewed and re graded by the same three pathologists after a time interval of 4 weeks after they have studied literature highlighting more descriptive criteria on the grading system.

The interobserver reproducibility was analysed using generalised kappa values.

Results

The generalised kappa value for interobserver variation among the three pathologists in the first instance was 0.306 indicating a fair level of agreement.

In the second instance the generalised kappa values for interobserver agreement was 0.488 indicating a moderate level of agreement.

The lowest level of interobserver agreement was observed in the Gleason score 2-4 and 7 categories.

Conclusion

The results of this study are comparable to several other studies conducted in UK and USA. It emphasises the importance of refinement of diagnostic criteria in Gleason grading and continuous education of pathologists on this grading system to reduce the inter observer variation in this grading system.

Key words

Interobserver variation, Gleason grading, Prostate carcinoma