Abstract

Abstract 2499

Background and Methods:

Multiparameter flow cytometry (MFC) is increasingly used in AML diagnostics as a supplement to morphologic examination. Here we retrospectively examine the concordance between MFC and morphologic examination for the detection of AML in CSF. Between 2005 and 2012, CSF from 604 sequential patients with AML or high-risk MDS was obtained at diagnosis or in follow-up of treatment for AML involving the central nervous system. Morphologic examination was done using conventional Wright-Giemsa stained cytospin preparations. MFC was done using a standard single-tube myeloid panel (HLA-DR/CD15/CD33/CD19/CD117/ CD13/ CD38/CD34/CD71/CD45) and/or a panel targeting the prior AML immunophenotype.

Results:

Results detecting any evidence of AML in the CSF by both techniques is noted in the Table below:

Flow PositiveFlow NegativeFlow SuspiciousTotals
Morphology positive 70 (12%) 1 (0%) 0 (0%) 71 (12%) 
Morphology negative 13 (2%) 469 (78%) 5 (1%) 487 (81%) 
Morphology atypical 15 (2%) 30 (5%) 1 (0%) 46 (8%) 
Totals 98 (16%) 500 (83%) 6 (1%) 604 
Flow PositiveFlow NegativeFlow SuspiciousTotals
Morphology positive 70 (12%) 1 (0%) 0 (0%) 71 (12%) 
Morphology negative 13 (2%) 469 (78%) 5 (1%) 487 (81%) 
Morphology atypical 15 (2%) 30 (5%) 1 (0%) 46 (8%) 
Totals 98 (16%) 500 (83%) 6 (1%) 604 

The concordance between MFC and morphology (excluding atypical morphology) was high (97%; 539/558), in large part due to the predominance of uninvolved samples using both techniques. However, of the samples positive by MFC, 13% were negative by morphology and another 15% showed atypical cells insufficient for positivity by morphology. Of samples positive by morphology, only a single sample (1.4%) was negative by MFC. In addition, 65% of cases in which morphology was judged atypical were MFC negative, suggesting suboptimal specificity. Patients typically received intrathecal treatment based on positive MFC making it difficult to know whether the disease in the CSF would have expanded without treatment in patients who were MFC positive/morphology negative.

Conclusion:

Subject to the limitation that the morphology may not have been interpreted entirely independently of the MFC result, the results suggest morphology yields a false negative result in between 13% and 28% of samples and may not be sufficiently sensitive for routine screening of CSF. To follow-up our findings we are examining whether there are features that distinguish patients who were MFC positive/ morphology negative or vice versa and performing a study in which morphology and MFC are being independently assessed.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.