Immunophenotyping Versus Morphological Evaluation Of Fresh and Stabilized Cerebrospinal Fluid As Diagnostic Tool For CNS Involvement In Childhood Acute Lymphoblastic Leukemia

Presence of malignant cells in cerebrospinal fluid (CSF) is a risk factor in pediatric acute lymphoblastic leukemia (ALL). Consequently, these patients receive extra intrathecal treatment. We evaluated the concordance between morphological and flow-cytometric (FCM) results, both on freshly analyzed and on stabilized, overnight transported samples.

Diagnostic CSF samples of 61 newly diagnosed pediatric ALL patients were divided in two aliquots. One was sent to the local laboratory and processed within a few hours after sampling (lab1). A second aliquot was 1:1 diluted with sterile medium and sent to the reference laboratory (lab2). For all samples, a MGG (May-Grünwald-Giemsa) stained cytocentrifuged slide was morphologically evaluated and two 6-color FCM stainings were performed. Samples were considered positive (CNS2) by MGG if at least one blast was seen and by immunophenotyping if a cluster (≥ 10 events) of ALL cells was detected.

Comparison of morphological data between both laboratories showed concordance in 33 samples (53%). In 20 of the 28 discordant samples only 1 or 2 blasts were reported by a single laboratory. Comparison of FCM data between laboratories was concordant in 58 samples (95%). Three samples (tumor load 15%-18%) were reported positive by one laboratory only; in two of these <100 cells could be acquired per tube. Comparison between FCM and morphology showed discordant results in 25 samples (41%) for lab1 and 19 (31%) in lab2. One discordant case was positive by FCM but negative by morphology, in all others only morphology was reported positive. The latter cases mostly had 1 or 2 blasts reported by morphology, generally not confirmed by the other laboratory.

This study shows that morphological analysis of CSF samples is non-reproducible if only 1 or 2 blasts are suspected. In contrast, FCM analysis of CSF is highly reproducible between fresh and stabilized samples. Although immunophenotyping still seems less sensitive than morphology, discordant samples generally have only 1 or 2 blasts reported by MGG, which is highly irreproducible. Nevertheless, by using a single 8-color tube higher cell numbers may be acquired and the FCM sensitivity may be increased.

No relevant conflicts of interest to declare.