[Agreement among observers in the classification of acute leukemias]

Rev Invest Clin. 1991 Jul-Sep;43(3):223-8.
[Article in Spanish]

Abstract

This study was carried out to establish the level of concordance between two observers from two different health institutions in Mexico City, in the diagnosis of acute leukemias and their different varieties. We studied 73 consecutive cases of adults with these diseases. Each one of the two observers established their diagnosis on two occasions at least 15 days apart. They first made their diagnosis taking as a base the neoplastic cells morphology in bone marrow smears, and after that, with morphology plus specific cytochemistry. The outcomes of the two observers were also compared with the official diagnosis. Kappa test was performed to know interobserver and intraobserver concordance. The kappa values for the diagnosis myeloid/lymphoid were found among the highest (51 to 91). Weighted kappa was also applied to know the level of concordance in the diagnosis of the different varieties of acute leukemia, myeloid and lymphoid. In these cases the weighted kappa values were lower compared with the previous values (lymphoid, from 47 to 82; myeloid, from 30 to 66). Cytochemistry paradoxically was a confusing factor when it was used: in these cases the kappa values were lower (32 to 84) than morphology alone (39 to 91). The outcomes showed the subjective level in the diagnosis of the myeloid subtypes was more important in them than in the lymphoid subtypes.

Publication types

  • Comparative Study
  • English Abstract
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Acute Disease
  • Adult
  • Biomarkers, Tumor / analysis
  • Bone Marrow / pathology
  • Bone Marrow Examination / statistics & numerical data*
  • Hematology
  • Humans
  • Leukemia / classification
  • Leukemia / diagnosis*
  • Leukemia / pathology
  • Neoplastic Stem Cells / chemistry
  • Neoplastic Stem Cells / pathology
  • Observer Variation
  • Prospective Studies
  • Single-Blind Method
  • Staining and Labeling

Substances

  • Biomarkers, Tumor