Expression of neuroblastoma related genes in bone marrow at end of high-risk neuroblastoma therapy

Abstract

Background:

Minimal disease quantification may predict Event-Free (EFS) and Overall Survival (OS).

Methods:

We evaluated mRNA expression of five neuroblastoma-associated genes (NB5 assay) in bone marrows (BM) of patients with newly diagnosed high-risk neuroblastoma who received consistent immunotherapy. mRNA expression of CHGA, DCX, DDC, PHOX2B, and TH genes in BM of 479 patients enrolled on immunotherapy arm of Children’s Oncology Group trials ANBL0032 and ANBL0931 was evaluated using real-time polymerase chain reaction (PCR) based TaqMan Low-Density Array. Results from End-Consolidation and End-Therapy were analyzed for association with 5-year EFS/OS, and patient and tumor characteristics. Tests of statistical significance were two-sided.

Results:

NB5 assay detected neuroblastoma-related mRNA in 222/286 (77.6%) of BMs obtained at End-Consolidation and 188/304 (61.8%) at End-Therapy. Any mRNA level detected in End-Therapy BM correlated with significantly worse EFS (57% [49.6%,63.7%] vs. 73.0% [63.5%,80.4%]; p=0.005), but not OS. Analysis limited to patients in complete response at End-Therapy still found a significant difference in EFS with detectable versus not detectable NB5 assay results (58.9% [49.5%, 67.1%] vs. 76.6% [66.1%, 84.2%]; p=0.01). End-Consolidation results did not correlate with EFS or OS. Multivariable analysis determined End-Therapy NB5 assay BM results (p=0.02), age at diagnosis (p=0.002), and Pre-Consolidation response (p=0.02) were significantly associated with EFS independent of other clinical and biologic parameters evaluated, including End-Therapy response.

Conclusions:

If further validated in additional patient cohorts, the NB5 assay ability to independently predict EFS from End-Therapy could improve patient stratification for novel maintenance therapy trials after current End-Therapy to improve outcome.

1|. INTRODUCTION

Neuroblastoma is stratified at diagnosis into risk groups based on patient characteristics and tumor biology^1,2. Current Children’s Oncology Group (COG) high-risk therapy includes induction with chemotherapy and primary tumor resection, consolidation including myeloablative chemotherapy with autologous stem cell transplant and local radiation therapy, and Post-Consolidation isotretinoin and dinutuximab, an antibody directed against disialoganglioside (GD2), plus cytokines (IL-2 and GM-CSF)¹,³. Despite this intensive therapy survival is only 50%¹ with relapses occurring even in patients in a complete response³ by standard response criteria. Identification of timepoints where minimal residual disease (MRD) detection is associated with outcome would facilitate the design of future clinical trials to test whether novel Post-Consolidation immunotherapy and/or additional maintenance therapy beyond current End-Therapy could improve outcome.

Standard response evaluations include computed tomography/magnetic resonance (CT/MRI) imaging, ¹²³I-Metaiodobenzylguanidine (MIBG) scans, and bone marrow (BM) morphology with or without immunohistochemistry. Clinical response is prognostic² but cannot detect MRD. Neuroblastoma tumor mRNA detection in blood and BM provides sensitive and specific MRD detection⁴. Various neuroblastoma mRNA assays of blood, BM, and/or peripheral blood stem cells have demonstrated prognostic value at diagnosis and end of induction^5–8, with minimal data from timepoints before Post-Consolidation Therapy or at End-Therapy⁶.

We developed a five gene assay for neuroblastoma mRNA (NB5 assay) for quantitative assessment of CHGA (chromogranin A), DCX (doublecortin), DDC (dopadecarboxylase), PHOX2B (paired-like homeobox 2b), and TH (tyrosine hydroxylase) expression using TaqMan® Low Density Array (TLDA) methodology^8,9. These five genes are rarely and only weakly expressed by normal bone marrow mononuclear cells but are strongly expressed by neuroblastomas in vivo and by both MYCN amplified and nonamplified neuroblastoma cell lines that are multidrug sensitive or resistant⁸.

NB5 assay results from PBSC collected during induction and blood or BM of patients with relapsed/refractory neuroblastoma^8,9 were associated with lower Event-Free Survival (EFS) and Overall Survival (OS).

In this study, our aim was to determine the prognostic value of the NB5 assay in BM prior to Post-Consolidation immunotherapy and at End-Therapy in a large patient cohort assigned to receive dinutuximab with GM-CSF, IL-2 and isotretinoin on COG ANBL0032 and ANBL0931 Post-Consolidation therapy protocols. NB5 assay results from BM sampled at enrollment (prior to Post-Consolidation immunotherapy) and End-Therapy were analyzed for association with EFS, OS, and patient and tumor biologic variables.

2 |. METHODS

2.1 |. Eligible patients

Eligibility included patients with high-risk neuroblastoma per COG risk classification¹ enrolled by December 31, 2011 on ANBL0032(NCT00026312)³ or ANBL0931(NCT01041638)¹⁰ for Post-Consolidation therapy, and assigned to isotretinoin and dinutuximab with GM-CSF and IL-2 (based on intention to treat) who submitted at least one BM sample that yielded NB5 results.

ANBL0032 and ANBL0931 eligibility included age ≤31 years at neuroblastoma diagnosis and induction (Pre-Consolidation) response for primary site and soft tissue/bone metastases of complete (CR), very good partial (VGPR) or partial (PR) response per International Neuroblastoma Response Criteria (INRC)-1993¹¹. See Supporting Information file for BM response requirements and response assessment methods. Pre-Consolidation analyses utilized overall response excluding BM response. Analyses of End-Consolidation (at ANBL0032/ANBL0931 enrollment) and End-Therapy (Post-Consolidation immunotherapy completion) overall response included BM response. End-Consolidation response was not captured for ANBL0931 patients. No ANBL0931 patient had PD at End-Consolidation, which was an eligibility exclusion for ANBL0931.

Informed consent was obtained from patients and/or guardians prior to ANBL0032/ANBL0931 enrollment. Both studies were approved by Institutional Review Boards (IRBs) where patients received therapy in accord with an assurance approved by U.S. Department of Health and Human Services and with the Declaration of Helsinki. IRB approval for NB5 assay analyses was obtained at Children’s Hospital Los Angeles.

2.2 |. Treatment Regimens

See Supplementary Table S1 for induction and consolidation regimens allowed. Completion of induction and consolidation was defined as End-Consolidation.

Patients enrolled on ANBL0032 from 10/18/2001–4/20/2009 were randomized to Post-Consolidation Therapy with six cycles of either: isotretinoin (regimen A), or isotretinoin and dinutuximab with GM-CSF and IL-2 (regimen B)³. Patients with biopsy documented persistent neuroblastoma were non-randomly assigned to regimen B. After interim analysis showed superior outcome for regimen B, patients enrolled after 4/20/2009 were non-randomly assigned to regimen B. Our analyses included ANBL0032 patient subset assigned to regimen B randomly or non-randomly and all ANBL0931 patients (all received regimen B) as of December 31, 2011. End-Therapy was defined as regimen B completion.

2.3 |. NB5 Assay

See Supporting Information file Methods/Supplemental Table S2 for details. Bilateral BM aspirate specimens were required for NB5 assay^8,9 at End-Consolidation (2–4 weeks before ANBL0032/ANBL0931 enrollment), after 3 cycles of Post-Consolidation therapy, and within 14 days of End-Therapy. Samples included in End-Therapy analyses were obtained within 60 days of starting cycle 6 of Post-Consolidation therapy.

RNA from 10 million MNC from BM specimens was subjected to reverse transcription and quantification (cycle threshold value [Ct]) of CHGA, DCX, DDC, PHOX2B, and TH and housekeeping genes B2M, GAPDH, HPRT1, and SDHA with pre-designed Applied Biosystems TaqMan Low Density Array (TLDA) cards (ThermoFischer Scientific). The normalized geometric mean of Ct for the five detection genes was used as the NB5 score. If any one gene had a Ct <40 then the geometric mean would be <40 and this was considered a “detectable” NB5 assay result. Based on dilution experiments (See Supporting Information file) the NB5 assay sensitivity was determined to be 1 tumor cell/million MNC.

2.4 |. Statistical Analysis

See also Supporting Information file/Methods. Log-rank test, product-limit estimates, and Cox regression analysis¹², were used to examine influence of NB5 results and other assessments on EFS/OS from End-Consolidation or End-Therapy. In Table 3, all variables of interest were initially included in multivariable Cox regression model regardless of univariable significance. Individual variables were removed from Cox regression model in stepwise fashion in reverse order of significance. Variables significant at p<0.05 were retained in final selected model. P-values were based on Wald test¹² for both univariable and multivariable comparisons. Departure from the proportional hazards assumption was assessed by visual inspection of Kaplan-Meier plots¹³ of the many univariable analyses performed. There was no evidence of systematic qualitative nor profound quantitative departure from proportional hazards.

TABLE 3.

Cox regression analysis (raw data) of Event-Free Survival (EFS) calculated End of Therapy showing univariable, full multivariable, and selected models of clinical and biologic factors

Variable	Value	Number	Univariable Analysis		Multivariable Analysis		Selected models analysisi
			HRa (95% CI)b	Global P-Value	HRa (95% CI)b	Global P-Value	HR a (95% CI) b		Global P-Value
Age at diagnosis	< 18 months	41	0.456 (0.210 0.992)	0.003	0.364 (0.161 0.825)	0.002	0.438 (0.201 0.955)		0.002
	18 to 59 months (ref)	197	1.000		1.000		1.000
	60+ months	65	1.653 (1.096 2.493)		1.692 (1.085 2.640)		1.684 (1.115 2.545)
INSSc stage at diagnosis	Stage 1/2/3/4S (ref)	46	1.000	0.08	1.000	0.35
	Stage 4	203	2.091 (1.084 4.033)		1.685 (0.824 3.443)
	Missing/ Unknown	54	1.690 (0.780 3.662)		1.526 (0.570 4.087)
MYCN Status	Amplified (ref)	99	1.000	0.23	1.000	0.62
	Not amplified	106	1.494 (0.938 2.378)		0.908 (0.543 1.519)
	Unknown	98	1.207 (0.738 1.975)		1.421 (0.535 3.774)
Ploidy	Diploid (ref)	91	1.000	0.45	1.000	0.09
	Hyperdiploid	98	0.773 (0.485 1.233)		0.704 (0.428 1.157)
	Unknown	114	0.781 (0.499 1.221)		0.389 (0.160 0.946)
NB5 in bone marrow at End-Therapy	Detectable	188	1.819 (1.195 2.767)	0.01	1.495 (0.956 2.337)	0.08	1.696 (1.110 2.591)		0.02
	Not detectable (ref)	115	1.000		1.000		1.000
Response Pre-Consolidation	CRd (ref)	84	1.000	0.01	1.000	0.11	1.000		0.02
	VGPRe	100	1.891 (1.093 3.272)		1.746 (0.908 3.358)		1.861 (1.073 3.226)
	PRf	119	2.210 (1.306 3.739)		2.164 (1.051 4.456)		2.083 (1.225 3.543)
Response End- Consolidation	CRd (ref)	102	1.000	0.03	1.000	0.41
	VGPRe	69	1.866 (1.109 3.140)		1.094 (0.578 2.070)
	PRf	60	1.690 (0.981 2.912)		0.806 (0.384 1.692)
	PDg/NRh	5	4.347 (1.518 12.448)		2.733 (0.860 8.689)
	Unknown	67	1.450 (0.837 2.513)		1.055 (0.553 2.013)
Response End-Therapy	CRd (ref)	209	1.000	0.07	1.000	0.22
	VGPRe	56	1.153 (0.708 1.878)		0.922 (0.542 1.569)
	PRf	28	1.513 (0.837 2.736)		1.402 (0.704 2.793)
	PDg/NRh	6	4.122 (1.503 11.303)		3.467 (1.105 10.883)
	Unknown	4	1.262 (0.309 5.147)		1.502 (0.316 7.132)
Tandem transplant	No (ref)	243	1.000	0.13	1.000	0.20
	Yes	60	0.671 (0.400 1.125)		0.697 (0.403 1.207)

^aHR=Hazard Ratio,

^bCI= Confidence Interval,

^cINSS=International Neuroblastoma Staging System,

^dCR=Complete Response,

^eVGPR=Very Good Partial Response,

^fPR=Partial Response,

^gPD=Progressive Disease,

^hNR=No Response

ⁱFinal model resulting from reverse stepwise procedure described in Supplemental Methods

3 |. RESULTS

3.1 |. Patient and Tumor Characteristics

BM samples from 479 patients among 542 enrolled on ANBL0032 or ANBL0931 are included in our analyses (Table 1, Supporting Information file Methods/Supplemental Figure S1). NB5 assay results were obtained from 421 of 479 patients from End-Consolidation (n=286), End-Therapy (n=304), or both (n=168). Median (range) years follow-up for patients without events for EFS is 5.2 (0.1–14.2) from Post-Consolidation and 4.8 (0.1–13.7) from End-Therapy, and for OS is 5.1 (0.1–14.2) from Post-Consolidation and 4.7 (0.1–13.7) from End-Therapy. Insufficient numbers of blood samples were obtained from End-Consolidation (n=32) and End-Therapy (n=72) for valid analyses.

TABLE 1.

Five-year Event-Free Survival (EFS) and Overall Survival (OS) from enrollment (End-Consolidation) on ANBL0032/ANBL0931 for patients assigned to dinutuximab arm who submitted at least one bone marrow for NB5 assay.

Variable	Category	Number patients (%)	5-year EFS (95%CIa)	EFS p-value	5-year OS (95%CIa)	OS p-value
All patients assigned to dinutuximab arm with NB5 assay results		479 (100%)	55.8 (51.2, 60.2)	-----	70.2 (65.6, 74.3)	----
Age at diagnosis N=479	< 18 months	66 (13.8%)	72 (59.3, 81.3)	0.02	79.5 (67.2, 87.6)	0.19
	18 – 59 months	316 (66.0%)	56.3 (50.4, 61.8)		70.2 (64.5, 75.1)
	60+ months	97 (20.3%)	44.3 (34.0, 54.1)		64.8 (53.8, 73.8)
Sex N=479	Male	292 (61.0%)	53.4 (47.3, 59.1)	0.27	67.9 (62.1, 73.0)	0.09
	Female	187 (39.0%)	59.5 (51.9, 66.3)		73.7 (66.1, 79.9)
Tandem transplant N=479	No	390 (81.4%)	53.3 (48.1, 58.2)	0.02	68.3 (63.1, 72.9)	0.04
	Yes	89 (18.6%)	66.6 (55.3, 75.7)		79.2 (68.7, 86.5)
INSSa stage at diagnosis N=479	Stage 1/2/4S	19 (4.0%)	89.1 (63.4, 97.1)	< 0.001	94.7 (68.5, 99.2)	< 0.001
	Stage 3	48 (10.0%)	75.4 (59.8, 85.6)		93.5 (81.3, 97.8)
	Stage 4	341 (71.2%)	49.7 (44.1, 55.0)		64.7 (59.1, 69.7)
	Unknown	71 (14.8%)	63.2 (50.2, 73.6)		74.9 (62.4, 83.8)
MYCN N=479	Amplified	160 (33.4%)	61.6 (53.5, 68.7)	0.12	73.5 (65.7, 79.8)	0.82
	Not amplified	181 (37.8%)	48 (40.4, 55.2)		67.5 (59.9, 74.0)
	Unknown	138 (28.8%)	59.3 (50.4, 67.2)		70 (60.9, 77.4)
DNA Ploidy n=479	Diploid	155 (32.4%)	51.2 (42.9, 58.9)	0.25	67.1 (58.8, 74.1)	0.25
	Hyperdiploid	160 (33.4%)	57.8 (49.6, 65.2)		73 (65.0, 79.5)
	Unknown	164 (34.2%)	58.2 (49.9, 65.6)		70.5 (62.3, 77.2)
Response Pre-Consolidation N=479	Complete	133 (27.8%)	69.6 (60.8, 76.8)	< 0.001	81.3 (73.3, 87.1)	0.002
	Very Good Partial	166 (34.7%)	54 (45.8, 61.5)		68.2 (60.1, 75.0)
	Partial	180 (37.6%)	47.3 (39.7, 54.5)		64.1 (56.4, 70.8)
b Response End- Consolidation n=479	Complete	142 (29.7%)	65.1 (56.4, 72.5)	0.01 (0.002 excluding unknown)	78 (69.7, 84.3)	0.14 (0.08 excluding unknown)
	Very Good Partial	98 (20.5%)	56.3 (45.7, 65.6)		67.4 (56.1, 76.4)
	Partial	91 (19.0%)	44.4 (33.9, 54.4)		66.4 (55.4, 75.3)
	No response/PDa	7(1.5%)	14.3 (0.7, 46.4)		38.1 (6.2, 71.6)
	Unknown	141b(29.4%)	55.8 (47.0, 63.7)		68.8 (60.2, 75.9)
c Response End-Therapy n=383	Complete	265 (69.4%)	66.9 (60.9, 72.2)	0.03 (0.06 excluding unknown)	78 (71.9, 82.9)	0.10 (0.20 excluding unknown)
	Very Good Partial	71 (18.6%)	62.7 (50.0, 73.0)		80.5 (68.7, 88.2)
	Partial	31 (8.1%)	49.7 (30.9, 66.0)		70.7 (49.3, 84.4)
	No Response	6(1.6%)	33.3 (4.6, 67.6)		50 (11.1, 80.4)
	Unknown	9(2.4%)	27.7 (4.6, 58.6)		47.6 (12.3, 76.9)

^aCI=Confidence Interval, INSS=International Neuroblastoma Staging System, PD=Progressive Disease

^bIncludes the 73 patients enrolled on ANBL0931 which did not collect this data

^cFor Response End-Therapy only, EFS/OS are calculated from End-Therapy

A comparison of characteristics of patients enrolled on ANBL0032 (n=568) or ANBL0931 (n=105) who submitted NB5 BM samples (n=479) versus patients without BM samples (n=194) (Supplemental Table S3) found significant differences in tandem versus single transplant (18.6% versus 9.3%; p=0.003) and Pre-Consolidation response of CR/VGPR/PR (100% versus 61.4%; p=0.000). EFS/OS were the same for these two cohorts.

Univariable analysis of association of patient and tumor characteristics with EFS/OS (Table 1) found younger age, tandem versus single transplant, lower INSS stage, and better overall response at Pre-Consolidation, End-Consolidation, and End-Therapy were significantly associated with higher EFS. Histology (defined by International Neuroblastoma Pathology Classification¹⁴ was not evaluated due to inadequate numbers in favorable (n=14) versus unfavorable (n=325) groups, and 140 patients with unknown histology. Lower INSS stage and better response at Pre-Consolidation, End-Consolidation, and End-Therapy were significantly associated with higher OS.

3.2 |. NB5 assay

Neuroblastoma mRNA was detected in majority of BM samples at End-Consolidation (222/286; 77.6%) and End-Therapy (188/304; 61.8%). At End-Consolidation (Supplemental Table S4), NB5 assay was detectable in patients with any INSS stage¹¹, including stage 1–2 (6/8; 75%), stage 3 (15/30; 50%), stage 4 (165/201; 82.1%), and stage 4S (1/4; 25%). NB5 assay results from End-Therapy remained detectable across all stages; including stage 1–2 (5/9: 55.6%), stage 3 (17/34; 50%), stage 4 (131/203: 64.5%) and stage 4S (2/3; 66%).

Comparison of patient cohorts with detectable versus not detectable NB5 results (Supplemental Table S4) at End-Consolidation showed the detectable cohort had more stage 4 tumors (p=0.00), more MYCN non-amplified tumors (p=0.01), and fewer patients in CR Pre-Consolidation (p=0.02).

3.3 |. NB5 Assay Correlation with Outcome

EFS/OS from End-Consolidation was not significantly different between patients with detectable versus not detectable NB5 results at this timepoint (Fig. 1A, B). Although significance was observed for EFS stratified by delta Ct (Table 2, Fig. 2A), the two middle delta Ct strata had better outcome than for not detectable results, suggesting quantitative mRNA results are not significant at this timepoint. OS stratified by delta Ct at End-Consolidation was also not significant (Fig. 2B). For patients in CR at End-Consolidation, detectable NB5 assay results were not significantly associated with EFS (66.2%[52.3%,77.0%] versus 75%[52.6%,87.9%];p=0.33) or OS (81.6%[68.4%,89.6%] versus 82.4%[59.3%,93.0%];p=0.71) from End-Consolidation. For patients in CR at End-Therapy, NB5 results from End-Consolidation analyzed either as dichotomous results or stratified by delta Ct were not significantly associated with EFS from End-Therapy (p=0.69 and p=0.25; Table 2).

Figure 1.

Kaplan-Meier estimates of Event-Free Survival (EFS) and Overall Survival (OS) stratified by BM NB5 assay results. P-values are for comparison of detectable versus not detectable. A) EFS and B) OS calculated from date of study entry for patients with BM NB5 results from End-Consolidation; C) EFS and D) OS calculated from End-Therapy for patients with BM NB5 results from End-Therapy. All P-values are based on log rank test.

TABLE 2.

NB5 bone marrow results from End-Consolidation (n=286) or End-Therapy (n=303) correlation with 5-year Event-Free Survival (EFS) calculated from timepoint of NB5 assay

NB5 assay time-point	Response at End-Therapy	Age at diagnosis	NB5 Assay Result	NB5 Assay delta Cta	Number patients (%)	5-year EFS (95%CIa)	p-value
b End- Consolidation	Any response	Any age	Not Detectable		64 (22.4%)	60.3 (47.1, 71.2)	0.77
			Detectable	Any	222 (77.6%)	61.7 (54.7, 68.0)	0.02d
				20.5+	36 (12.6%)	68.8 (50.8, 81.3)
				19.5-<20.5	58 (20.3%)	78.8 (64.9, 87.7)
				<19.5	128 (44.8%)	52.5 (43.3, 60.9)
	Complete response	Any age	Not Detectable		35 (24.5%)	71.4 (53.5, 83.4)	0.69
			Detectable	Any	108 (75.5%)	74.4 (64.8, 81.7)	0.25a
				20.5+	16 (11.2%)	81.3 (52.3, 93.6)
				19.5-<20.5	30 (21%)	85.7 (65.9, 94.4)
				<19.5	62 (43.4%)	67.1 (53.8, 77.3)
cEnd-Therapy	Any response	Any age	Not detectable		115 (38%)	73 (63.5, 80.4)	0.01
			Detectable	Any	188 (61.8%)	57 (49.6, 63.7)	0.03a
				20.5+	49 (16.2%)	53.8 (38.6, 66.8)
				19.5-<20.5	87 (28.7%)	60.7 (49.5, 70.2)
				<19.5	52 (17.2%)	53.8 (39.5, 66.2)
	Complete Response	Any age	Not detectable		87 (41.6%)	76.6 (66.1, 84.2)	0.01
			Detectable	Any	122 (58.4%)	58.9 (49.5, 67.1)	0.08
				20.5+	37 (17.7%)	52.9 (35.4, 67.7)
				19.5-<20.5	51 (24.4%)	60.8 (46.2, 72.6)
				<19.5	34 (16.3%)	61.8 (43.5, 75.7)
c End-Therapy	Any response	Age <18 months	Not Detectable	Any	13 (4.3%)	92.3 (56.6, 98.9)	0.002
			Detectable	Any	28 (9.2%)	78.4 (58.1, 89.7)	0.002
		Age 18–59 months	Not Detectable	Any	80 (26.4%)	73.2 (61.9, 81.6)	0.002
			Detectable	Any	117 (38.6%)	58.7 (49.1, 67.1)	0.002
		Age 60+ months	Not Detectable	Any	22 (7.3%)	62.5 (38.5, 79.3)	0.002
			Detectable	Any	43 (14.2%)	38.4 (23.8, 52.8)	0.002
	Any response	CR Pre-Consolidation	Not detectable	Any	41 (48.8%)	87.4 (72.2, 94.6)	0.02
			Detectable	Any	43 (51.2%)	66.6 (50.2, 78.7)	0.02
		CR Post- Consolidation	Not Detectable	Any	44 (43.1%)	81.1 (65.8, 90.0)	0.12
			Detectable	Any	58 (56.9%)	67.1 (53.4, 77.6)	0.12

^aCt=Cycle threshold, CI=Confidence intervals

^bEFS is calculated from End-Consolidation timepoint

^cEFS is calculated from End-Therapy timepoint

^dp-values include “Not Detectable” group

Figure 2.

Kaplan-Meier estimates of Event-Free Survival (EFS) and Overall Survival (OS) for patients with bone marrow NB5 assay results stratified by NB5 delta Ct score. A) EFS and B) OS calculated from date of study entry for patients with NB5 results from End-Consolidation. C) EFS and D) OS calculated from End-Therapy for patients with NB5 assay results from End-Therapy. All P-values are based on log-rank test.

Five-year EFS of patients with neuroblastoma mRNA in BM at End-Therapy (57% [49.6%,63.7%]) was significantly worse (p=0.005) than those with no detectable mRNA (73% [63.5%,80.4%]) (Table 2, Fig. 1C). Any level of neuroblastoma mRNA was significantly associated with lower EFS (p=0.03) (Table 2, Fig. 2C). There was no association of NB5 results with OS either with dichotomous or stratified analysis of delta Ct (Figs. 1D and 2D).

Paired specimens from End-Consolidation and End-Therapy were available for 168 patients. For patients with no detectable neuroblastoma mRNA at End-Consolidation (n=39/168; 23.2%), there was no significant difference in 5-year EFS (p=0.80) or OS (p=0.60) if End-Therapy result was detectable (n=24; EFS 70.6% [47.9%, 84.8%] and OS 86.5% [63.8%, 95.5%]) versus not detectable (n=15; EFS 66.7% [37.5%, 84.6%] and OS 80.0% [49.9%, 93.1%]). However, for patients with detectable neuroblastoma mRNA at End-Consolidation (n=129/168;76.8%), 5-year EFS (p=0.04) and OS (p=0.04) were significantly higher if End-Therapy result was not detectable (n=39;EFS 81.4% [65.0%, 90.7%] and OS 94.5% [79.8%, 98.6%]) versus detectable (n=90; EFS 64.3% [53.5%, 73.3%] and OS 76.3% [63.0%, 85.3%]).

3.4 |. NB5 assay stratified by other factors

When End-Therapy BM NB5 assay results were stratified by response, better response was significantly associated with higher EFS for response time points at End-Therapy (Fig. 3A, B; p=0.01 for CR, p=0.02 for VGPR/PR) and Pre-Consolidation (Fig. 3C; p=0.03 for CR). Five-year EFS for 209 patients in CR at End-Therapy was 66.9% [60.9%, 72.2%].

Figure 3.

Kaplan-Meier estimates of Event-Free Survival (EFS) calculated from End-Therapy for patients with bone marrow NB5 assay results from End-Therapy stratified by detectable versus not detectable in following groups: A) Patients with a Complete Response (CR) at End-Therapy, B) Patients with a Very Good Partial Response (VGPR) or Partial Response (PR) at End-Therapy, C) Patients with a CR Pre-Consolidation, and D) Patients stratified by age at diagnosis (months). The symbol (+) indicates a detectable NB5 assay result and (−) indicates a not detectable result.

Patients in CR with not detectable NB5 results had 5-year EFS of 76.6% [66.1%, 84.2%] versus patients with detectable results (5-year EFS 58.9% [49.5%, 67.1%] p=0.01) (Table 2). Five-year OS was 83.2% (72.4%, 90.0%) and 74.4% (64.7%, 81.8%) for not detectable versus detectable (p=0.10). Within each response category at both time points, detectable neuroblastoma mRNA was associated with lower EFS (Fig. 3A-C, Table 3).

Age at diagnosis was significantly associated with EFS (p=0.02) but not OS (p=0.19) in univariable analysis (Table 1). Across all age cohorts, with any response, End-Therapy detectable results were associated with significantly lower EFS (Fig. 3D, Table 2: p=0.002).

3.5 |. Cox Regression Analysis

Table 3 presents Cox Regression analysis of EFS calculated from End-Therapy. Univariable analysis found younger age at diagnosis (p=0.003), not detectable NB5 results (p=0.01) and better response Pre-Consolidation and End-Consolidation (p=0.01 and p=0.03) were associated with higher EFS. In multivariable analysis, only age remained significant (p=0.002). Using selected models analysis, detectable neuroblastoma mRNA at End-Therapy (p=0.02) was independently significantly associated with worse EFS (Hazard ratio 1.696; 95% Confidence Interval 1.110, 2.591). Other significant factors were age (p=0.002) and response Pre-Consolidation (p=0.02).

4 |. DISCUSSION

Our prospective analysis performed in a large cohort of patients who received identical dinutuximab-cytokines as Post-Consolidation immunotherapy on two multi-center COG trials demonstrated that neuroblastoma mRNA detection in BM with NB5 assay at End-Therapy is significantly associated with lower EFS in patients with high-risk neuroblastoma, including patients in complete clinical response. Any amount of neuroblastoma mRNA was associated with lower EFS from End-Therapy. NB5 results remained independently significant in Cox regression analyses including known prognostic factors. This is the largest prospective biomarker analysis in patients with high-risk neuroblastoma receiving Post-Consolidation dinutuximab immunotherapy performed to date.

Age at diagnosis and Pre-Consolidation response, both known prognostic factors², were also significantly associated with EFS in Cox regression analysis. The significance of Pre-Consolidation response combined with End-Therapy NB5 results suggests earlier response to Induction remains prognostic despite additional myeloablative consolidation therapy and Post-Consolidation immunotherapy.

NB5 results at End-Therapy were not significantly associated with OS. This could be due to ongoing immune response that leads to tumor dormancy^15–17 but not complete tumor elimination and/or other biologic factors that affect the growth potential of minimal numbers of residual tumor cells¹⁸. The NB5 assay was significantly associated with EFS, with the majority of events consisting of tumor recurrence/relapse. The standard approach for tumor recurrence/relapse is to offer patients salvage therapy, and this may have prolonged overall survival so that no significant difference in OS was detected. Data on salvage therapy after patients went off protocol therapy was not part of data collection for ANBL0032 or ANBL0931. BM samples may also not assess all residual tumor sites. NB5 analysis of blood could be hypothesized to improve prognostic value and optimize serial sampling. However, we found lower sensitivity of the NB5 assay in blood versus BM in refractory/recurrent neuroblastoma patients⁹.

Detectable neuroblastoma mRNA in BM at End-Consolidation was not significantly associated with EFS/OS and thus the NB5 assay may not be a useful biomarker at this timepoint to stratify patients at initiation of novel Post-Consolldation immunotherapies. This may be due to the impact of immunotherapy and isotretinoin on residual tumor.

We did not assess early response during immunotherapy. Neuroblastoma mRNA detection using a four gene quantitative RT-PCR assay after two cycles of 3F8 in heterogenous cohort of refractory or first/second remission patients was associated with lower OS¹⁹. Future NB5 assay studies could include earlier immunotherapy timepoints.

Serial samples before and after Post-Consolidation Therapy were associated with better EFS only for patients with detectable neuroblastoma mRNA before and none afterward. These analyses were done in a small number of patients with serial data available. Validating these findings requires larger patient numbers, however they support dinutuximab efficacy against tumor resistant to chemotherapy and radiation.

Half of patients with stage 3 tumors had detectable mRNA in BM sampled at End-Consolidation and End-Therapy, suggesting localized tumors may have occult metastases. This is supported by metastases seen in 41% of patients with high-risk stage 3 neuroblastoma for whom frontline CCG 3891 therapy was ineffective²⁰. Among stage 1–2 tumors at original diagnosis, 56% had detectable NB5 results at End-Therapy. These patients either had MYCN amplification and/or metastatic progression prior to enrollment on ANBL0032 or ANBL0931 per eligibility requirements. Detectable NB5 results across all stages at diagnosis may explain why stage was not significant in multivariable analysis.

A potential limitation to our findings is differences between the patient cohort with NB5 assay samples submitted and the cohort without samples (Supplemental Table S3). More of the cohort with NB5 samples submitted received tandem transplants and had CR/VGPR/PR Pre-Consolidation. Pre-Consolidation response was also unknown among more patients without NB5 samples. Both tandem transplant and better Pre-Consolidation response are associated with better EFS^2,21. However, the cohort with NB5 samples and cohort without samples had the same EFS/OS, supporting the applicability of our NB5 assay findings to both cohorts.

Outcome biomarkers at specific timepoints provide important tools to stratify patients for trials that will assess efficacy of novel therapy. The occurrence of relapse in patients with high-risk neuroblastoma despite achievement of a complete response³ highlights the need to consider novel maintenance phase therapy. Such clinical trials are underway in single or limited institutions, including anti-GD2 vaccine (NCT04936529) or difluoromethylornithine (DFMO) (NCT02395666). We demonstrated that NB5 assay BM results at End-Therapy (after Post-Consolidation immunotherapy completion) are predictive for EFS from that timepoint, including for patients in clinical CR. If validated in additional large prospective patient cohorts the NB5 assay, Pre-Consolidation response, and age could potentially be utilized to identify patients eligible for novel maintenance therapy administered after completion of current standard immunotherapy (End-Therapy timepoint) and/or to stratify patients for randomized trials in this setting. Serial NB5 assays after End-Therapy during observation or during novel Maintenance therapy could be explored as an additional tool to monitor response. This study was not designed to test the impact of additional therapy for patients with detectable neuroblastoma mRNA at End-Therapy.

Additional data are needed to consider neuroblastoma mRNA detection inclusion into International Neuroblastoma Response Criteria²². Optimal methodology, choice/number of genes, therapy timepoints, and sample source have not been defined⁴. We previously reported NB5 assay detects neuroblastoma mRNA in more patients than two gene assay using TH and PHOX2B⁸. Multivariable prospective analyses of neuroblastoma mRNA and MIBG score, another known prognostic marker^23,24, are needed, although MIBG score prognostic significance at End-Therapy is not established. Pre-Consolidation response, strongly associated with outcome in our Cox regression model, incorporates MIBG response. Studies of the significance of circulating tumor DNA (ctDNA), which also identifies MRD^4,25, are also ongoing.

Specific BM morphology, MIBG score, and soft tissue disease data at the time of the NB5 assay was not collected for patients on ANBL0032 or ANBL0931, therefore we could not analyze their correlations. We previously reported⁹ that the NB5 assay had detectable results in 69% of bone marrows with negative morphology in patients with recurrent/refractory high-risk neuroblastoma. In univariate analysis BM NB5 Ct correlated with the percentage of neuroblastoma tumor cells by routine morphology and the MIBG Curie score. Tumor detected by bone marrow morphology and MIBG imaging was independently associated with stronger NB5 Ct in bone marrow, while soft tissue disease by CT/MRI was not associated with NB5 Ct in bone marrow independently of bone marrow morphology and MIBG imaging.

NB5 assay significance needs to be evaluated in the context of other induction, consolidation, and Post-Consolidation regimens. Significant time points for NB5 results may change with standard therapy modifications, such as dinutuximab inclusion into Induction (NCT03786783)²⁶. Randomized frontline trials could also evaluate which arm has a lower incidence of detectable NB5 results at End-Therapy.

We demonstrated that NB5 assay detectable neuroblastoma mRNA in BM at End-Therapy, diagnosis age, and Pre-Consolidation response are independently associated with EFS from End-Therapy. The NB5 assay remained significant for patients in complete response. If validated in additional patient cohorts with inclusion of other prognostic biomarkers not known for patients in our analysis, the ability of the NB5 assay to independently predict EFS from End-Therapy could improve patient stratification for novel maintenance therapy trials.

ABBREVIATIONS KEY

ASCT

Autologous Stem Cell Transplant

B2M

Beta-2 microglobulin

BM

bone marrow

CHGA

Chromogranin A

COG

Children’s Oncology Group

CR

Complete Response

CT

Computed Tomography

Ct

Cycle threshold value

DCX

doublecortin

DDC

dopadecarboxylase

DFMO

difluoromethylornithine

EFS

Event-free Survival

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

GM-CSF

Granulocyte-Macrophage Colony Stimulating Factor

HPRT1

Hypoxanthine Phosphoribosyltransferase 1

INRC

International Neuroblastoma Response Criteria

INSS

International Neuroblastoma Staging System

MRI

Magnetic Resonance Imaging

MIBG

Metaiodobenzylguanidine

MRD

Minimal Residual Disease

MNC

Mononuclear cells

mRNA

Messenger Ribonucleic Acid

NR

No Response

OS

Overall Survival

PCR

Polymerase Chain Reaction

PD

Progressive Disease

PHOX2B

Paired-like Homeobox 2b

PR

Partial Response

RT-PCR

Reverse Transcriptase Polymerase Chain Reaction

SD

Stable Disease

SDHA

Succinate Dehydrogenase Complex Flavoprotein Subunit A

TH

Tyrosine Hydroxylase

TLDA

TaqMan® Low Density Array

VGPR

Very Good Partial Response