Automated Cytochemistry in Acute Leukemia

Automated Cytochemistry in Acute Leukemia G. Giuseppe d’Onofrio G. Giorgio Mango Giuseppe d’Onofrio, Giorgio Mango, Hematology Service, Università Cattolica del Sacro Cuore, Rome (Italy) We wish to make several remarks about the paper by Cranendonk et al.: Evaluation of the use of the Hemalog D in acute lymphoblastic leukemia and disseminated non-Hodgkin’s lymphoma in children [Acta haemat. 71: 18–24 (1984)], which deals with automated cytochemistry performances in acute leu-kemias. The subject, indeed, is not a worthless one: automated leukocyte differential devices, in fact, are more and more diffused among hematology laboratories, so that the awareness of their qualities and defects deserves considerable interest. In that paper, the authors report their results concerning Hemalog D differential counts in 79 children with acute lymphoblastic leukemia (ALL) and 18 children with disseminated nonHodgkin’s lymphoma; only 25 of them, however, were studied since the onset of their disease. The authors generalize their conclusions, claiming that the Hemalog D seems inadequate in diagnosis and follow-up of leukemic patients. Over the last 4 years we have been able to analyze a large number of leukemic blood samples with the Hemalog D [1, 3]. We would like to report very briefly our experience, which is indeed quite different. As far as reliability and accuracy are concerned, the Hemalog D appears entirely capable of detecting and signaling pathological samples: with one exception, in 43 cases of acute leukemia adequate flags (increase in LUCs or HPX, abnormal Remainder, LR or LPX warnings) were constantly provided, allowing an immediate microscope check. The only exception was represented by a patient with acute promy-elocytic leukemia, in which no alarm signals were reported in the printout: the peroxidase display plot, however, was clearly abnormal, and the association with pseudo-eosinophilia, marked leukocytosis, decreased hemoglobin and platelet count would have pointed out that sample as abnormal in any case. The negative results obtained by Cranendonk et al. may be explained in part by considering several details of their study. (1) These authors were interested only in lymphoblastic leukemias; the Hemalog D, owing to the peroxidase staining it carries out automatically in each blood sample analyzed for routine differential, is obviously cleverer in recognizing and characterizing non-lymphoid leukemias. (2) The great majority of the patients studied at diagnosis (19 out of 25) had ALL-LI subtype: the small LI lympho142 Correspondence