Diamond L W, Mishka V G, Seal A H, Nguyen D T
Institut für Pathologie der Universität zu Köln, Germany.
Int J Biomed Comput. 1994 Nov-Dec;37(3):211-24. doi: 10.1016/0020-7101(94)90120-1.
Accurate diagnosis and classification of hematologic malignancies (acute leukemias and chronic lymphoproliferative disorders) requires a multiparameter approach including peripheral blood analysis, bone marrow examination, and immunophenotyping. We have designed knowledge-based computer systems for interpretation of the hemogram and peripheral blood smears, analysis of flow cytometric immunophenotyping panels, and morphologic assessment of bone marrow specimens. The 3 modules share a relational database which includes pertinent clinical history in addition to the laboratory results. The bone marrow module automatically writes a complete interpretative report with a final diagnosis by searching all of the databases for appropriate clinical, peripheral blood, and immunophenotyping information. The ability of the 3 modules to interact, and the quality of the interpretative reports were tested on 100 consecutive patients with leukemia. The final diagnosis made by the bone marrow system agreed with the hospital diagnosis in 94 cases and the authors' interpretation in 99 of 100 cases.
