Background: Central nervous system (CNS) involvement is common in childhood acute lymphoblastic leukemia (ALL). Today all patients receive intensive prophylactic CNS-directed therapy, which is associated with short- and long-term neurotoxicity. The conventional method for diagnosing CNS leukemia is by microscopic examination of cytospin preparations of cerebrospinal fluid (CSF). Both high (CNS3: ≥5 x 109 cells/L) and low level CNS disease (CNS2: <5 x 109 cells/L) indicate intensified CNS-directed therapy. The CNS is involved in 30-40% of relapses, and most occur in patients classified as CNS negative by cytospin (CNS1). We investigated if sensitive flow cytometric (FCM) analysis of CSF at diagnosis improves detection of CNS involvement and prediction of relapse risk.

Methods: Cytospin results were collected from local oncology clinics for patients treated according to the Nordic Society of Pediatric Hematology (NOPHO) ALL2008 protocol from July 2008 to December 2017 (N=1841). CSF samples for FCM were obtained from patients enrolled from September 2012 to December 2017 at 17 participating centers (N=669). CSF samples were collected into Transfix® tubes (Intermedico Ltd., Hellerup, Denmark) and shipped to Rigshospitalet, Copenhagen, Denmark for centralized analysis as previously described (Levinsen at al., Pediatr Blood Cancer 2016). FCM positivity required detection of ≥10 cells with a leukemia-associated phenotype. CSF FCM results were blinded to the treating physician. Time to relapse was analyzed using Cox proportional hazards models with delayed entry at remission and with death and second malignant neoplasm as competing events.

Results: At diagnosis, 241 of 1841 (13.1%; CNS2: 8.9% and CNS3: 4.2%) patients were positive by cytospin, whereas 167 of 669 (25.0%) patients were positive by FCM. FCM positivity was not associated with shipment duration (median: 2 days). In total, 191 of the 669 patients (28.6%) were CNS positive by cytospin and/or FCM. CNS positivity was associated with higher white blood cell (WBC) count at diagnosis, T-cell ALL, lack of t(12;21), and traumatic lumbar puncture (TLP) (all comparisons: p < 0.001). However, patients with CNS positivity did not differ in their end of induction bone-marrow minimal residual disease (MRD) levels compared to the remaining patients (cytospin: median MRD 1.2 x 10-4 vs. 0.9 x 10-4, p = 0.058; FCM: median MRD 2.0 x 10-4 vs. 1.7 x 10-4, p = 0.62). During follow-up, 30 patients relapsed with nine of the relapses involving the CNS and 16 relapses isolated to the bone-marrow. The 4-year cumulative incidence of any relapse was higher for patients who were CNS positive by cytospin (16.1% vs. 8.9%), FCM (18.6% vs. 6.0%), and combined cytospin/FCM (19.0% vs. 5.5%) (Figure 1a). Simple Cox regressions yielded relapse hazard ratio (HR) estimates of CNS positivity at diagnosis of 2.4 for cytospin (95% CI 1.0-5.7, p = 0.04), 3.5 for FCM (95% CI 1.7-7.2, p < 0.001), and 4.1 for combined cytospin/FCM (95% CI 2.0-8.7, p < 0.001). Cytospin and/or FCM was associated with increased risk of both isolated bone-marrow relapse (16 events, HR 3.1, 95% CI 1.2-8.4, p = 0.024) and of any CNS relapse (9 events, HR 4.9, 95% CI 1.2-19.5, p = 0.025). In a multiple Cox model stratified by immunophenotype and risk group, predictors of relapse were age (per year: HR 1.1, 95% CI 1.0-1.2, p = 0.003), WBC at diagnosis (per doubling: HR 1.3, 95% CI 1.1-1.6, p = 0.005), and cytospin and/or FCM (HR 3.2, 95% CI 1.4-7.2, p = 0.006) (Figure 1b). Cytospin and/or FCM was associated with a significantly higher risk of relapse for BCP-ALL (N=581; 23 relapses; HR 4.5, 95% CI 1.9-10.4, p = < 0.001). For T-cell ALL no significant association was observed (N=85; 7 relapses; HR 1.6, 95% CI 0.3-8.2, p = 0.58), however the number of T-cell ALL was too small to allow reliable conclusions. In a simple Cox model, TLP was associated with a significantly higher risk of relapse than no TLP (HR 2.7, 95% CI 1.3-5.9, p = 0.011), but this was only the case for patients who were positive by FCM (FCM pos: HR 2.7, 95% CI 1.0-7.0, p = 0.043; FCM neg: 0.7, 95% CI 0.1-5.0, p = 0.68).

Conclusion: FCM markedly increases the detection rate for CNS involvement at diagnosis, and distinguish between patients at high and low risk of relapse. Diagnosis of CNS leukemia by combined cytospin and FCM analysis should be the standard for accurate classification of CNS status, which will enable better stratification of CNS-directed and systemic therapy.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution