Diagnosing and Staging Central Nervous System Involvement in Pediatric ALL: A Global Survey on Its Heterogeneity

BACKGROUND

In childhood acute lymphoblastic leukemia (ALL), central nervous system (CNS) involvement is associated with increased risk of relapse. Traditionally, CNS involvement has been staged using cell counting and microscopic cytomorphological assessment of centrifuged and stained samples (cytospins) of the cerebrospinal fluid (CSF). More recently CSF flow cytometry has been shown to be more sensitive than cytology and to have prognostic significance. CNS staging is a basis for treatment stratification worldwide. CSF diagnostic protocols of study groups differ in small details, however, larger variability in practice was suspected based on informal discussions within the childhood ALL community.

AIMS

To systematically evaluate real world practice of CSF diagnostics and staging in pediatric ALL patients across hospitals globally.

METHODS

A survey was developed and distributed in 2021-2022 with target hospital categories and hospital numbers per country. Questions focused on numerous details of routine practice at the given hospitals in year 2021.

RESULTS

Eighty-two centers from 47 countries across five continents responded with a median 20 new pediatric ALL patients per year per center. Significant heterogeneity was observed in testing methodologies and CNS staging.

The staging LP is contraindicated by high peripheral WBC, with threshold varying between 50 and 200 x10⁹/L in 46% of the centers, while 54% of the centers set no threshold (LP is performed even in case of extreme leukocytosis). The timing of staging LP relative to the initiation of systemic antileukemic therapy also varies, with 57% of hospitals performing the LP strictly before therapy, while others routinely allowing some (1 and 7 days depending on hospital) systemic steroid and/or chemotherapy before.

There is heterogeneity in the selection of test modalities to analyze the CSF. Manual cell counting is applied in 73%, automated cell counting in 43%, cytospins in 84% and flow cytometry in 33% of the hospitals during routine CNS-staging. Ten percent of the centers use only one modality, 54% of them use 2 modalities. A larger proportion of centers in high income countries analyze WBCs in cytospins routinely as compared to centers in middle income countries (92% versus 65%, p = 0.006). A similar trend was observed regarding CSF flow cytometry (39% vs 17%, p = 0.07). Similarly, a heterogenous set of testing modalities and cut offs/definitions are applied for CSF red cell quantification and for defining traumatic LP (TLP).

Among hospitals which assess cytospins, the CSF volume centrifuged to prepare these specimens varies between 15 and 3000 microliters. Typical CSF volumes analyzed by flow cytometry range between 200 and 4000 microliters. There is also wide variation regarding the cut-off blast numbers to define positive findings in case of cytospins (≥1 to ≥5 cells) and flow cytometry (1 to 50 events characterized as test positivity in different centers). Interestingly, there is no correlation between the CSF volume used and the minimum number of blasts to define the tests’ positivity.

These heterogeneities are seen both within study groups and within individual countries in the same study group.

DISCUSSION

The above inconsistencies and differences in practice must have obvious clinical implications by significantly impacting CNS staging and thus CNS1/2/3/TLP frequencies and therefore treatment decisions for ALL patients worldwide. Moreover, internationally well-connected, larger pediatric oncology centers involved in research are thought to be overrepresented among survey participants. In view of this bias, global diagnostic heterogeneity can be larger and the quality of diagnostics poorer for a large proportion of patients than what our survey reflects.

These results raise concern over published findings on the prognostic value of leukemic CNS involvement, particularly the CNS2 stage, and the comparability of published studies. We propose that details on CNS staging methods and adherence to the diagnostic guidelines should be included in publications of clinical studies. There is an urgent need for standardization, ideally through the development of international consensus guidelines in this field.