Refining the Limits of Borderline Lymphoproliferative Disorders.

Background The concept of borderline lymphoproliferative disorder (LPD) has not been clearly defined. Methods This study aimed to classify patients with leukemic LPD (n = 597, excluding hairy cell leukemia, mantle cell lymphomas, and CD10-positive LPDs) into CLL or non-CLL applying three diagnostic strategies (the D'Arena and CLLflow scores and CD43 expression) and to better characterize unclassified patients. Results Patients with concurring CLL-like (n = 441) or non-CLL like (n = 99) results with the three diagnostic strategies were determined to have CLL and non-CLL, respectively. Patients with discordant results (n = 57) were analyzed taking into consideration each individual cytometric marker and cytogenetic data: 41 were classified (11 CLL, 30 non-CLL) and 16 (2.7% of the entire series) could not and were considered borderline LPD. Excluding borderline LPD, the CLLflow score had the highest accuracy of the three strategies. With the addition of CD43 no patient was misclassified. With the aid of hierarchical clustering, 12 of the 16 borderline patients seemed to fall into two well-defined antigenic groups. None of the diagnostic strategies could reliably pick out borderline LPD. Conclusion The combination of the CLLflow score and CD43 generally has a high diagnostic accuracy for leukemic LPD but it is not reliable to identify or diagnose borderline LPD. This latter group needs further study to determine its underlying biology. (c) 2018 International Clinical Cytometry Society