Accuracy of modified gleason grading on prostatic biopsies.

Objective: The Modified Gleason Grading System was proposed by the 2005 International Society of Urological Pathology (ISUP) Consensus Conference. In Thailand, prostatic biopsies are mostly diagnosed by general pathologists. The accuracy of the Gleason grading on prostatic biopsies is still questioned. We assessed the accuracy of modified Gleason Grading on prostatic biopsies and detected pitfalls in the grading. Methods: Sixty-nine cases of prostatic biopsies which diagnosed adenocarcinoma were re-examined according to the modified Gleason grading system by an experienced pathologist, four general pathologists and two pathology residents. Results: The accuracy of modified Gleason scores on prostatic biopsies ranged from 79.7% to 98.6% with 68.3 to 100 of 95% CI. Most biopsies with more than one score of difference were missed from Gleason pattern 5. Gleason pattern 5 was found in 25 biopsies. Percents of Gleason pattern 5 were classified into <5% (12 biopsies), 5-10% (6 biopsies), 11-50% (5 biopsies), and >50% (3 biopsies) of the tumor volume. The mean of the number of the observers who fail to detect Gleason pattern 5 in each category was 4 (SD 1.2), 3.6 (SD 1.8), 1.3 (SD 0.95), and 0.33 (SD 0.5), respectively. We found a small amount of Gleason pattern 5 was significantly missed (p = 0.01). Conclusion: The accuracy of Gleason grading on prostate biopsies among general pathologists and pathology residents in this study was fairly good. The missed Gleason pattern 5 is the most common pitfall. Pathologists should be aware of Gleason pattern 5 because it might be present in a prostate biopsy of less than 5 % of the tumor volume.

Pitfalls in classifying lymphomas.

BACKGROUND: Although the WHO classification (2001) requires a great deal of morphologic, immunophenotypic, genetic, and clinical features for classifying lymphomas, it is still feasible to misdiagnose under limited resources, especially a limited panel of antibodies used for immunophenotyping. To identify pitfalls in classifying lymphomas among hematopathologist, general pathologists, and pathology residents under this situation. MATERIAL AND METHOD: Newly diagnosed lymphoma cases from 1 July 2002 to 30 June 2003 at Siriraj Hospital were included for two rounds of individually blinded review by a hematopathologist, two general pathologists, and three pathology residents. Final diagnoses were given by consensus. Pitfalls were determined from misdiagnosis, in each case analyzed in terms of frequency. RESULTS: One hundred and four lymphoma cases included 61 diffuse large B-cell lymphoma (DLBCL, 58.6%), 12 MALT lymphoma (11.5%), eight follicular lymphoma (FL, 7.7%), seven classical Hodgkin lymphoma (HL, 6.7%), four unspecified peripheral T-cell lymphoma (PTCL, 3.8%), three Burkitt lymphoma (BL, 2.9%), two subcutaneous panniculitis-like T-cell lymphoma (SPTCL, 1.9%), and seven other uncommon types (1% each). Pitfalls were low infrequency on diagnosis of DLBCL, nodular sclerosis HL, and SPTCL (8% each), but not different among the participants only in DLBCL. Pitfalls in diagnosis of MALT lymphoma, mixed cellularity HL, BL, unspecified PTCL, and FL were 60%, 50%, 33%, 29%, and 24%, respectively. However, considering hematopathologist and non-hematopathologist groups, pitfalls in the former were lower, especially in the uncommon types of lymphoma. CONCLUSION: Pitfalls in classifying lymphomas are common. Interest in hematopathology reduces misdiagnosis in lymphomas other than DLBCL.