The morphologic classification of acute lymphoblastic leukemia in childhood. Observations on concordance using a...

The French-American-British (FAB) cooperative group proposed a simple scoring system in 1981 to improve concordance among researchers and clinicians for the subclassification of acute lymphoblastic leukemia (ALL) into subtypes L1 and L2. The authors subtyped 50 consecutively admitted patients with common ALL of childhood according to the FAB scoring system to assess concordance among a member of the FAB group (J.B.) and two pathologists (S.S. and C.C.) who are not FAB members. The initial agreement of 66% was unsatisfactory to the authors prompting the critical evaluation of the major sources of disagreement. Concordance was improved greatly by the use of two technics that are logical extensions of the scoring system: performing 100-200 differential cell counts to determine the correct percentages in borderline cases, and using reference drawings that illustrate blasts with cytoplasmic contents of 15%, 20%, and 25% in determining the cell's nuclear cytoplasmic ratio. The authors believe that the criteria explicitly given greater weight by the FAB scoring system (nuclear cytoplasmic ratio and nucleoli) are the most subjective. By following the criteria carefully and resolving differences together at a multiheaded microscope, a concordance of 92% was achieved.