Abstract
BACKGROUND: Liquid biopsies offer less burdensome sensitive disease monitoring. Bone marrow (BM) metastases, common in various cancers including neuroblastoma, is associated with poor outcomes. In pediatric high-risk neuroblastoma most patients initially respond to treatment, but in the majority the disease recurs with only 40% long-term survivors, stressing the need for more sensitive detection of disseminated disease during therapy. METHODS: To validate sensitive neuroblastoma mRNA RT-qPCR BM testing, we prospectively assessed serial BM samples from 345 international high-risk neuroblastoma patients, treated in trials NB2004 (GPOH) or NBL2009 (DCOG), using PHOX2B, TH, DDC, CHRNA3, and GAP43 RT-qPCR mRNA markers and BM GD2-immunocytology. Association between BM-infiltration levels and event-free survival (EFS) and overall survival (OS) was estimated by using Cox regression models and Kaplan-Meier's methodology. RESULTS: BM infiltration >10% by RT-qPCR at diagnosis was prognostic for survival (adjusted hazard ratio (HR) 1.82 [95%CI 1.25-2.63] and 2.04 [1.33-3.14] for EFS and OS, respectively). Any post-induction RT-qPCR positivity correlated with poor EFS and OS, with a HR of 2.10 [1.27-3.49] and 1.76 [1.01-3.08] and 5-years EFS of 26.6% [standard error 5.2%] versus 60.4% [6.7] and OS of 43.8% [5.9] versus 65.7% [6.6] for RT-qPCR-positive patients versus RT-qPCR-negative patients. In contrast, post-induction immunocytology positivity was not associated with EFS or OS (HR 1.22 [0.68-2.19] and 1.26 [0.54-2.42]). CONCLUSION: This study validates the association of not clearing of BM metastases by sensitive RT-qPCR detection with very poor outcome. We therefore propose implementation of RT-qPCR for minimal residual disease testing in neuroblastoma to guide therapy.