Long-term Follow-up of Patients with Intermediate-risk Neuroblastoma Treated with Response- and Biology-Based Therapy: A Report From the Children’s Oncology Group Study ANBL0531

Abstract

Background:

We previously reported excellent 3-year overall survival (OS) for patients with newly diagnosed intermediate-risk neuroblastoma treated with a biology- and response-based algorithm on the Children’s Oncology Group study ANBL0531. We now present the long-term follow-up results.

Methods:

All patients who met the age, stage, and tumor biology criteria for intermediate-risk neuroblastoma were eligible. Treatment was based on prognostic biomarkers and overall response. Event-free survival (EFS) and overall survival (OS) were estimated by the Kaplan-Meier method.

Results:

The 10-year EFS and OS for the entire study cohort (n=404) were 82.0% (95% confidence interval (CI), 77.2%–86.9%) and 94.7% (95% CI, 91.8%–97.5%), respectively. International Neuroblastoma Staging System stage 4 patients (n=133) had inferior OS compared to non-stage 4 patients (n=271; 10-year OS: 90.8% [95% CI, 84.5%–97.0%] vs. 96.6% [95% CI, 93.9%–99.4%], p=0.02). Infants with stage 4 tumors with ≥1 unfavorable biologic feature (n=47) had inferior EFS compared to those with favorable biology (n=61; 10-year EFS: 66.8% [95% CI, 50.4%–83.3%] vs. 86.9% [95% CI, 76.0%–97.8%], p=0.02); OS did not differ (10-year OS: 84.4% [95% CI, 71.8%–97.0%] vs. 95.0% [95% CI, 87.7%–100.0%], p=0.08). Inferior EFS but not OS was observed among patients with tumors with (n=26) versus without (n=314) 11q loss of heterozygosity (LOH) (10-year EFS: 68.4% [95% CI, 44.5%–92.2%] vs. 83.9% [95% CI, 78.7%–89.2%], p=0.03; 10-year OS: 88.0% [95% CI, 72.0%–100.0%] vs. 95.7% [95% CI, 92.8%–98.6%], p=0.09).

Conclusions:

The ANBL0531 trial treatment algorithm resulted in excellent long-term survival. More effective treatments are needed for subsets of patients with unfavorable biology tumors.

Condensed Abstract:

The ANBL0531 trial treatment algorithm resulted in excellent long-term survival. However, more effective treatments are needed for patients with tumors harboring 11qLOH and infants with unfavorable biology stage 4 disease.

INTRODUCTION

Neuroblastoma is notable for its clinical and biological heterogeneity.¹ The survival of the vast majority of children with intermediate-risk (IR) disease is excellent following treatment with moderate-dose chemotherapy and surgery.^2,3 The Children’s Oncology Group (COG) Phase 3 IR disease trial A3961 (NCT00003093) demonstrated that excellent outcomes could be maintained with treatment reductions based on tumor biology.⁴ The subsequent COG Phase 3 trial ANBL0531 (NCT00499616) was designed to determine if therapy could be further reduced in subsets of patients with IR disease based on clinical and biological prognostic markers and treatment response. The IR cohort included children younger than age 12 years with stage 2A/2B tumors that were biopsied only or less than 50% resected or favorable histology stage 3 tumors; infants younger than age 365 days with stage 3, 4 or 4S disease; and toddlers from 365 to younger than 547 days with favorable histology, hyperdiploid stage 4 neuroblastoma, or unfavorable histology stage 3 tumors. Patients with MYCN-amplified tumors were excluded. The 3-year event-free survival (EFS) and overall survival (OS) rates were 83.2% (95% confidence interval (CI), 79.4%–87.0%) and 94.9% (95% CI, 92.7%–97.2%), respectively.² This approach is now considered standard of care for IR patients at COG centers. To investigate if this excellent survival was maintained, we evaluated the long-term outcome of patients enrolled on ANBL0531.

PATIENTS AND METHODS

Patient Enrollment

The ANBL0531 study (NCT00499616) opened in October 2007 and closed to enrollment in June 2011. Stratification to initial treatment Group was based on International Neuroblastoma Staging System (INSS) stage⁵, age at diagnosis, International Neuroblastoma Pathology Classification (INPC) histology⁶, ploidy^7,8 and 1p36 and 11q23 allele status⁹ (Supplemental Table S1). Written informed consent was obtained from parents or guardians before the initiation of any study-related treatment or procedures, and the protocol was approved by the Ethics Committees from all participating institutions. The study was conducted in accordance with Good Clinical Practice principles and the Declaration of Helsinki.

Study Design

IR patients were assigned to initially receive two (Group 2), four (Group 3), or eight (Group 4) cycles of baseline chemotherapy based on their age, stage, and tumor histology and genetic markers. Tumors with FH, hyperdiploidy, and normal 1p and 11q alleles were classified as favorable biology. Unfavorable biology tumors exhibited one or more unfavorable features (diploidy, UH, 1p LOH, and or 11q LOH). Details of the ANBL0531 treatment regimen were reported previously.² The drugs included in each cycle of treatment are shown in Supplemental Table S2.

A protocol specific modified 1993 International neuroblastoma response criteria (INRC) was used to define response.⁵ Decisions regarding additional treatment with surgery or chemotherapy following completion of the initially assigned group-based were based on disease response (Figure 1A). The response criteria for treatment endpoints were defined by stage and initial group assignment. For Group 2 and 3 patients with stage 2A/2B or 3 neuroblastoma who achieved a partial response (PR) or better of the primary tumor following assigned chemotherapy, surgery was not required and no further up-front treatment was given. Similarly, Groups 2 and 3 patients who achieved treatment endpoints following the assigned cycles of chemotherapy and surgery did not receive additional treatment. Additional cycles of chemotherapy were administered to patients who did not achieve treatment endpoints following initial group-assigned therapy.² Patients with progressive, non-metastatic disease within 3 years from enrollment were treated with up to six cycles of CPM and TOPO with VGPR as the treatment endpoint. Patients with progressive metastatic disease or those receiving alternative therapies were taken off protocol therapy (Figure 1B).

FIGURE 1.

(A) ANBL0531 treatment overview. (B) Group patient numbers including completion of protocol, discontinuation from protocol and deaths (created with BioRender.com).

Statistical Analysis

Categorical variables were compared using a chi-squared test or Fisher’s exact test (in the case of small sample size). EFS time was calculated from the date of diagnosis to the first occurrence of an event (relapse, progression, secondary malignancy, or death) or, if no event, until the date of last follow-up. OS time was calculated from the date of diagnosis to the occurrence of death from any cause or, if the patient was still alive, until the date of last follow-up. Survival estimates per the Kaplan-Meier method¹⁰ are presented at 10 years, with standard errors (SE) per Peto et al.¹¹ Two-sided log-rank tests were used to compare survival curves by subgroups. A p-value <0.05 was considered statistically significant.

To determine the independent prognostic strength for survival of various prognostic factors, including 11q LOH status (no LOH vs. LOH), tumor biology (favorable vs. unfavorable), Group (2 vs. 3 vs. 4), ploidy (hyperdiploid vs. diploid), INSS stage (non-stage 4 vs. stage 4), and age at diagnosis (<18 vs. ≥18 months), Cox proportional hazards models with the Efron method of handling tied event times were fit. Multivariable Cox models were fit to identify features independently prognostic of EFS and OS (within the context of the standard treatment for intermediate-risk neuroblastoma) among previously established neuroblastoma prognostic features including age (<18 months vs ≥18 months), INSS stage (non–stage 4 vs stage 4), ploidy (hyperdiploid vs. diploid), and 11q status (no LOH vs. LOH).¹² Biology (favorable vs unfavorable) and Group (2 vs. 3 vs. 4) were also included due to their importance for treatment determination in the study.

RESULTS

Patient Characteristics and Treatment

Among the 464 newly diagnosed NBL patients enrolled on ANBL0531,18 were deemed ineligible and 42 were subsequently found to be high-risk and removed from protocol, as previously reported.² Group assignments, patient characteristics, tumor biology, number of cycles of chemotherapy patients actually received in each initial group for the remaining 404 evaluable patients are summarized in Table 1.² The initial group assignment and disease stage for patients with tumors that harbored 11q LOH are summarized in Table 2.² Based on response and/or physician/family preference, the number of treatment cycles differed in a subset of patients from the initial assigned treatment as previously described.²

Table 1.

Patient Characteristics and Outcomes

Characteristics	Patients, No. (%)	10-year EFS, % (95% CI)	10-year OS, % (95% CI)	EFS P	OS P
Overall	404 (100)	82.0 (77.2, 86.9)	94.7 (91.8, 97.5)	N/A	N/A
Age
<18 months	361 (89.4)	81.7 (76.4, 86.9)	94.1 (90.9, 97.2)	0.5047	0.1110
≥18 months	43 (10.6)	85.1 (72.0, 98.3)	100.0 (100.0, 100.0)
INSS Stage
2A/2B	58 (14.4)	83.8 (70.6, 97.0)	100.0 (100.0, 100.0)	0.0599*	0.0171*
3	164 (40.6)	86.1 (79.5, 92.7)	100.0 (100.0, 100.0)
4S	49 (12.1)	79.5 (62.9, 96.2)	81.4 (65.6, 97.2)
4 (infants)	125 (30.9)	78.1 (68.8, 87.3)	90.2 (83.4, 96.9)
4 (toddlers)	8 (2.0)	62.5 (19.2, 100.0)	100.0 (100.0, 100.0)
Biologic Features
All favorable	211 (62.6)	86.1 (80.2, 92.1)	98.1 (95.7, 100.0)	0.0779	0.0031
At least one unfavorable	126 (37.4)	78.4 (68.9, 87.8)	90.9 (84.4, 97.4)
Unknown	67	75.8 (62.2, 89.4)	91.0 (82.3, 99.7)
Initial Assignment
Group 2	175 (43.3)	85.3 (78.6, 92.0)	99.4 (98.0, 100.0)	0.0016	<.0001
Group 3	141 (34.9)	85.6 (78.0, 93.2)	94.3 (89.2, 99.3)
Group 4	88 (21.8)	69.6 (56.9, 82.3)	85.9 (76.6, 95.1)
Histology
Favorable	363 (93.8)	82.9 (77.9, 87.9)	94.9 (92.0, 97.8)	0.6343	0.8748
Unfavorable	24 (6.2)	77.8 (56.1, 99.4)	95.5 (84.8, 100.0)
Unknown	17	69.3 (35.7, 100.0)	88.2 (65.8, 100.0)
Ploidy (infants)
Hyperdiploid	237 (85.9)	84.2 (78.1, 90.2)	95.7 (92.4, 99.1)	0.0747	0.0006
Diploid	39 (14.1)	73.2 (54.0, 92.4)	81.1 (64.3, 97.9)
Unknown	21	70.7 (40.1, 100.0)	80.4 (55.7, 100.0)
MKI
Low	262 (70.2)	82.6 (76.5, 88.7)	94.9 (91.4, 98.3)	0.4324	0.6613
Intermediate	99 (26.5)	78.7 (68.9, 88.6)	93.9 (87.9, 99.8)
High	12 (3.2)	91.7 (72.1, 100.0)	100.0 (100.0, 100.0)
Unknown	31	83.6 (65.9, 100.0)	93.5 (81.5, 100.0)
Tumor Differentiation
Differentiating	24 (6.3)	79.2 (61.5, 96.9)	100.0 (100.0, 100.0)	0.7439	0.2456
Poorly Differentiated/Totally Undifferentiated	357 (93.7)	81.9 (76.7, 87.2)	94.5 (91.5, 97.6)
Unknown	23	87.0 (69.2, 100.0)	91.3 (76.1, 100.0)
1p/11q Allele Status
Normal 1p and 11q	277 (81.2)	84.1 (78.6, 89.6)	96.3 (93.4, 99.2)	0.2135	0.0586
1p and/or 11q LOH	64 (18.8)	77.2 (64.1, 90.4)	90.5 (81.5, 99.5)
Unknown	63	77.8 (63.4, 92.2)	92.1 (83.2, 100.0)
1p Allele Status
Normal 1p	300 (88.0)	82.6 (77.1, 88.1)	95.5 (92.5, 98.5)	0.6435	0.4411
1p LOH	41 (12.0)	84.4 (70.5, 98.3)	92.7 (82.9, 100.0)
Unknown	63	77.8 (63.4, 92.2)	92.1 (83.2, 100.0)
11q Allele Status
Normal 11q	314 (92.4)	83.9 (78.7, 89.2)	95.7 (92.8, 98.6)	0.0252	0.0884
11q LOH	26 (7.7)	68.4 (44.5, 92.2)	88.0 (72.0, 100.0)
Unknown	64	78.1 (64.1, 92.2)	92.2 (83.5, 100.0)
Received CPM/TOPO
Yes	27 (6.7)	51.9 (25.2, 78.5)	92.3 (77.8, 100.0)	<.0001	0.6346
No	377 (93.3)	84.2 (79.5, 89.0)	94.9 (92.0, 97.7)
Primary Tumor Site
Abdominal/retroperitoneum	138 (36.8)	85.1 (77.1, 93.0)	97.0 (93.2, 100.0)	0.9684	0.2819
Adrenal	100 (26.7)	82.8 (72.8, 92.9)	91.8 (84.8, 98.9)
Neck	27 (7.2)	80.8 (57.6, 100.0)	96.2 (85.5, 100.0)
Pelvis	17 (4.5)	82.4 (62.8, 100.0)	100.0 (100.0, 100.0)
Thoracic	93 (24.8)	81.4 (72.0, 90.7)	96.8 (92.5, 100.0)
Unknown/other	29	69.0 (45.2, 92.8)	81.9 (61.3, 100.0)

^*P-value compares stage 4 vs not stage 4.

TABLE 2.

11q LOH status by international neuroblastoma staging system (INSS) stage and initial group assignment

Characteristic	11q LOH (%)^	Normal 11q (%)^	P (Chi-Square Test)
INSS Stage
2A/2B	1 (3.9)	4 (13.7)	<0.0001*
3	5 (19.2)	138 (44.0)
4S	2 (7.7)	37 (11.8)
4 (infants)	16 (61.5)	91 (29.0)
4 (toddler)	2 (7.7)	5 (1.6)
Initial Assignment
Group 2	0 (0.0)	175 (55.7)	0.0001**
Group 3	8 (30.8)	98 (31.2)
Group 4	18 (69.2)	41 (13.1)

^{^}Percentage denominator is total patients with 11q LOH and total patients with normal 11q respectively

^*P-value compares stage 4 vs not stage 4

^**P-value compares Group 3 vs. Group 4, since, patients with 11q LOH could not be assigned to Group 2

Survival Outcomes

Survival: Entire Cohort, According to Initial Group Assignment, Clinical Factors, and Tumor Biology

The median follow-up from time of diagnosis for the 333 patients without event was 10.1 years (range: 55 days-12.4 years). The 10-year EFS and OS rates for the entire cohort (n=404) were 82.0% (95% CI, 77.2%–86.9%) and 94.7% (95% CI, 91.8%–97.5%), respectively (Table 1; Figure 2A). Patients initially assigned to Group 2 (n=175) had a 10-year EFS and OS of 85.3% (95% CI, 78.6%–92.0%) and 99.4% (95% CI, 98.0%–100%), respectively. Patients assigned to Group 3 (n=141) at diagnosis had a 10-year EFS and OS of 85.6% (95% CI, 78.0%–93.2%) and 94.3% (95% CI, 89.2%–99.3%), respectively. Group 4 patients (n=88) had a 10-year EFS and OS of 69.6% (95% CI, 56.9%–82.3%) and 85.9% (95% CI, 76.6%–95.1%), respectively. EFS and OS were statistically significantly worse among Group 4 patients compared to those assigned to Groups 2 and 3 (P=0.002 and P=<0.001, respectively) (Figures 2B and 2C). EFS and OS were not significantly different among Group 2 and 3 patients who achieved greater than a PR versus a PR (P=0.2035 and P=0.7122 respectively) (Supplemental Figures S1A and S1B).

FIGURE 2.

(A) Event-free survival (EFS) and overall survival (OS) for intermediate risk (IR) patients (n=404). (B) EFS for IR patients according to initial treatment assignment (n=404; P=0.002). (C) OS for IR patients according to initial treatment assignment (n=404; P<0.001). (D) EFS for evaluable, intermediate risk patient according to 1p loss of heterozygosity (LOH) status (n=341; P=0.64). (E) OS for evaluable, intermediate risk patient according to 1p LOH status (n=341; P=0.44). (F) EFS for evaluable, intermediate risk patient according to 11q LOH status (n=340; P=0.03). (G) OS for evaluable, intermediate risk patient according to 11q LOH status (n=340; P=0.09). Life tables data is available in Supplemental Table S6.

Only 5 patients experienced events more than 3 years from diagnosis, including 2 patients with local relapse/progression, 1 patient with metastatic only progression, 1 patient who developed a second malignancy (rhabdomyosarcoma) 8 years after completing therapy, and 1 patient who died following multiple local relapses with the first relapse being ~28 months after diagnosis.

Among patients with known 1p allele status (n=341), EFS and OS did not differ among those with versus without 1p LOH (10-year EFS: 82.6% [95% CI, 77.1%–88.1%] versus 84.4% [95% CI, 70.5%–98.3%], P=0.64, Figure 2D; 10-year OS: 95.5% [95% CI, 92.5%–98.5%] versus 92.7% [95% CI, 82.9%–100%], P=0.44, Fig 2E). 11q allele status was known for 340 patients. Significantly inferior EFS was observed among patients with 11q LOH (n=26) compared to those without 11q LOH (n=314; 10-year EFS: 68.4% [95% CI, 44.5%–92.2%] versus 83.9% [95% CI, 78.7%–89.2%], P=0.03, Figure 2F). However, OS did not significantly differ according to 11q status (10-year OS: 88.0% [95% CI, 72.0%–100%] versus 95.7% [95% CI, 92.8%–98.6%], P=0.09, Figure 2G). A trend associating worse OS for patients with either or both 1p and 11q LOH compared to those with normal 1p and 11q allele status was observed (10-year OS: 90.5% [95% CI, 81.5%–99.5%] versus 96.3% (95% CI, 93.4%–99.2%], respectively), although this did not reach statistical significance (P=0.0586).

A backward-selected multivariable Cox model indicated that only Group was predictive of EFS (p=0.0041). Patients in Group 4 had a significantly increased risk of event compared to Group 2 (HR: 2.716) and Group 3 (HR: 2.673) (Supplemental Table S3). The order of removal was age at diagnosis, tumor biology, ploidy, INSS stage, and 11q status, with the least statistically significant term with p >0.05 dropping out at each step. For OS, using the same factors as for EFS, a backward-selected Cox model indicated that Group (p=0.0430) and ploidy (p=0.0305) were predictive of OS. Compared to patients in Group 2, patients in Group 4 (HR: 15.614) and Group 3 (HR: 10.525) had significantly increased risk of death (Supplemental Table S4). In addition, patients with diploid tumors had a significantly increased risk of death compared to those with hyperdiploid neuroblastoma (HR: 3.285). The order of removal was age at diagnosis, 11q status, tumor biology, and INSS stage.

Patients Treated with CPM/TOPO

There was no significant difference in the proportion of patients in Groups 2 and 3 vs Group 4 who achieved treatment end points with CPM/TOPO. Specifically, a PR or better was achieved in 8 of 10 (80%) Group 2 or 3 patients while a VGPR was achieved in 7 of 17 (41.2%) of Group 4 patients (P=0.1071). Patients who required CPM/TOPO to reach ≥PR had an inferior EFS compared to patients who did not require CPM/TOPO to achieve ≥PR (P=0.0001) (Supplemental Figure S2A). However, no significant difference in OS was observed (P=0.29) (Supplemental Figure S2B).

Survival According to Stage and Age

EFS and OS according to stage (2A/B, 3, 4S, 4 [infants], and 4 [toddlers] are summarized in Table 1. INSS stage 4 patients (n=133) had inferior OS compared to non-stage 4 patients (stage 2,3,and 4S, n=271) (p=0.02). Infants with stage 4 tumors with one or more unfavorable biologic features (n=47) had significantly inferior EFS compared to those with favorable biology (n=61) (p=0.02); although OS was not significantly different (p=0.08). Of patients with known tumor 11q status, 10-year EFS and OS for stage 4 patients with 11q LOH (n=18) were 65.5% (95% CI, 34.7%–96.3%) and 88.2% (95% CI, 68.4%–100%), respectively compared to 80.0% (95% CI, 70.0%–90.0%) and 90.4% (95% CI, 83.0%–97.8%) for those without 11qLOH (n=96) (EFS p=0.14; OS p=0.77).

Infants <365 days, with stage 4 disease (n=125; 30.9%) had a 10-year EFS and OS of 78.1% (95% CI, 68.8%–87.4%) and 90.2% (95% CI, 83.4%–96.9%), respectively (Figure 3A). Among those with available tumor biology data (n= 108), EFS was significantly better for stage 4 infants with favorable biology compared to those with unfavorable biology (10-year EFS: 86.9% [95% CI, 76.0%–97.8%] versus 66.8% [95% CI, 50.4%–83.3%], P=0.02, Figure 3B). However, OS was not significantly different for stage 4 infants with favorable versus unfavorable biology tumors (10-year OS: 95.0% [95% CI, 87.7%–100%] versus 84.4% [95% CI, 71.8%–97.0%], P=0.08, Figure 3C).

FIGURE 3.

(A) Event-free survival (EFS) and overall survival (OS) for evaluable infants with stage 4 disease younger than 365 days (n=125). (B) EFS for evaluable infants with stage 4 disease younger than 365 days of age according to tumor biology* (n=108; P=0.0154). (C) OS for evaluable infants with stage 4 disease younger than 365 days of age according to tumor biology* (n=108; P=0.08). Life tables data is available in Supplemental Table S6.

* Favorable biology constituted hyperdiploid tumors with favorable histology and normal 1p and 11q alleles. Unfavorable biology tumors had at least one unfavorable feature (diploidy, unfavorable histology, 1p loss of heterozygosity, or 11q loss of heterozygosity)

Pattern of Relapse

A total of 59 patients developed relapse as of the last documented follow-up. As shown in Table 3, the pattern of first relapse (local versus metastatic versus combined local and metastatic) differed significantly by INSS disease stage (P=0.001). Nine (25.7%) of the 35 patients with isolated local relapse had stage 2A/2B and 17 (48.6%) had stage 3 disease. Relapse in only metastatic sites was most frequent among infants with stage 4 disease (n=9; 64.3%). Relapse in both local and metastatic sites were observed in patients with stage 3 disease (n=2; 20.0%), and infants with stage 4S (n=3; 30.0%) or stage 4 neuroblastoma (n=5; 50.0%). Local combined with metastatic relapse was not observed in any of the toddlers with stage 4 disease or patients with stage 2A/B tumors in the study cohort. Pattern of relapse also differed significantly by initial group assignment (P=0.0293). Local relapse alone was observed in 20 (57.1%) of the Group 2 patients while Group 4 patients more frequently developed relapse in metastatic only sites (n=8, 57.1%) or combined metastatic and local sites (n=6, 60.0%). Although the pattern of relapse was not associated with 1p allele status (P=1.00) or combined 1p/11q allele status (0.0893), a significant difference in relapse pattern was observed according to 11q status (P=0.0422). Among the 59 patients with known 11q status who relapsed, 42 (71.2%) had tumors with normal 11q. Isolated local relapse was observed in 26 (61.9%) of these 42 patients. Of the 26 patients with tumors that harbored 11q LOH (stage 2A/2B=1, stage 3=5, stage 4S=2, stage 4 [infants]=16, stage 4 [toddlers]=2), 6 relapsed in metastatic sites (+/− local sites) and 1 had a local relapse (Table 2 and Supplemental Table S5). All 5 stage 4 patients with 11q LOH who relapsed had tumors with favorable histology. Two of these patients had diploid tumors, and the remaining 3 patients had hyperdiploid neuroblastoma.

TABLE 3.

Patterns of First Relapse by Stage, Group Assignment, and 1p and/or 11q Allele Status

Characteristic*	Local Relapse (#, %)	Metastatic Relapse (#, %)	Local and Metastatic Relapse (#, %)	P (Fisher’s Exact test)
INSS Stage
2A/2B	9 (25.7)	0 (0.0)	0 (0.0)	0.0011**
3	17 (48.6)	2 (14.3)	2 (20.0)
4S	1 (2.9)	1 (7.1)	3 (30.0)
4 (infants)	7 (20.0)	9 (64.3)	5 (50.0)
4 (toddler)	1 (2.9)	2 (14.3)	0 (0.0)
Initial Assignment
Group 2	20 (57.1)	3 (21.4)	2 (20.0)	0.0293
Group 3	8 (22.9)	3 (21.4)	2 (20.0)
Group 4	7 (20.0)	8 (57.1)	6 (60.0)
1p/11q allele status
Normal 1p and 11q	24 (88.9)	8 (61.5)	6 (66.7)	0.0893
1p and/or 11q LOH	3 (11.1)	5 (38.5)	3 (33.3)
Unknown	8	1	1
1p allele status
Normal 1p	25 (92.6)	12 (92.3)	8 (88.9)	1.0000
1p LOH	2 (7.4)	1 (7.7)	1 (11.1)
Unknown	8	1	1
11q allele status
Normal 11q	26 (96.3)	9 (69.2)	7 (77.8)	0.0422
11q LOH	1 (3.7)	4 (30.8)	2 (22.2)
Unknown	8	1	1

^*1 additional patient with unknown site of relapse has been excluded

^**P-value compares stage 4 vs not stage 4.

DISCUSSION

The COG phase 3 study ANBL0531 sought to reduce therapy for subsets of IR patients using a biology- and response-based algorithm. Three-year EFS of 83.2% and OS of 94.9% were achieved with this approach.² Here, we build on and extend the conclusions of the initial publication and show 10-year EFS and OS were maintained at 82.0% and 94.7%, respectively. The low rate of relapse 2 years after diagnosis raises questions regarding the benefit of the ANBL0531 surveillance imaging schedule, which included primary tumor site imaging and MIBG scans for 3 years after therapy completion. For select groups, such as patients with 2A/2B disease (who had no reported metastatic relapses on this study), the use of MIBG scans to evaluate recurrent metastatic disease may not be required during off therapy follow-up. A reduction in the required surveillance scans in this patient population would decrease long-term radiation exposure and limit cumulative sedation risks.

Based on the favorable outcome observed for stage 4 toddlers treated with high-risk therapy in previous studies,^13,14 ANBL00531 evaluated EFS and OS for subgroups of toddlers with stage 3 and 4 disease treated with less intensive, IR therapy. Although stage 4 toddlers treated with IR therapy had inferior EFS compared to patients with stages 2A/2B/3 disease and infants with stage 4 disease, survival was 100% at 10 years, supporting this treatment approach. A recent retrospective study similarly showed excellent survival among subsets of toddlers with neuroblastoma assigned to reduced treatment after reclassification of risk group from high to intermediate on the basis of 18 month age cutoff.¹⁵

Consistent with the original ANBL0531 results, our long term follow-up shows that patients with tumors harboring 11q LOH had inferior EFS compared to patients with normal 11q allele status, although OS did not differ. However, 1p LOH was not prognostic of outcome in this cohort. Survival tree regression analyses of a large cohort of patients used to establish the International Neuroblastoma Risk Group (INRG) risk classification similarly demonstrated associations between 11q aberration and worse EFS and OS among patients <18 months of age with stage 2 and 3 MYCN-NA tumors and patients with stage 4S MYCN-NA tumors.¹⁶ Others have also reported inferior outcome and higher risk of relapse among patients with localized neuroblastoma tumors and 4S disease with 11q deletion.¹⁷ However, both 1p LOH and 11q LOH were found to be associated with inferior progression-free survival (PFS) in a cohort patients with low-risk (LR) and IR disease analyzed by Attiyeh and colleagues.⁹ The SIOPEN group reported that segmental chromosomal profiles (which included 11q loss) were associated with inferior PFS in patients with localized unresectable or metastatic (stage 4 and 4S) MYCN-NA neuroblastoma.¹⁸ In contrast, among patients with relapsed MS disease treated after 2001, Campbell et al., reported that 1p LOH was associated with worse OS, whereas 11q LOH status was not associated with survival.¹⁹ Among high-risk (HR) patients, Pinto and colleagues recently reported that 11q LOH was associated with poor response to current era induction chemotherapy whereas 1p LOH was not prognostically significant.²⁰ In an additional study of HR patients ≥18 months of age with stage 3 unfavorable histology MYCN-NA neuroblastoma, 11q loss/LOH was shown to be independently associated with inferior survival compared to patients without these aberrations.²¹

In this IR patient cohort, several established biologic markers were associated with outcome.¹² In univariate analyses, EFS was significantly inferior among patients with tumors that harbored 11q LOH, and patients with diploid tumors had worse OS compared to those with hyperdiploid neuroblastomas. Only initial group assignment was associated with both EFS and OS in univariate and multivariable analyses. The prognostic strength of Group reflects the multiple clinical and biological prognostic markers used to assign initial treatment in this study.

The pattern of relapse among the IR patients with tumors that harbored 11q LOH differed from those with normal 11q alleles. Among the 42 patients with normal 11q status who relapsed, isolated local failures were observed in 26 (61.9%) patients and 16 (38.1%) had metastatic relapse (+/− local relapse). However, only 1 of the 7 patients with 11q LOH had an isolated local failure. The other 6 (85.7%) patients relapsed in metastatic sites (+/− local relapse). A recent analysis of an INRG cohort largely comprised of high-risk patients with relapsed disease, similarly showed that favorable prognostic markers were enriched in patients with isolated local relapse.²²

Stage 4 infants <365 days with unfavorable biology neuroblastoma had significantly worse EFS compared to those with favorable biology tumor in our long-term analysis. The 10-year EFS was 66.8% (95% CI, 50.4%–83.3%), which is similar to the previously reported 3-year EFS of 66.8% (95% CI, 53.1%–80.6%), reflecting the low number of late events.² OS was not significantly different for stage 4 infants with favorable versus unfavorable biology tumors, indicating that salvage treatments were successful for many of the patients who developed relapsed neuroblastoma.

Although the biology and response-based algorithm used to assign treatment on ANBL0531 led to excellent long-term OS, subsets of IR patients with unfavorable biology tumors required salvage therapy to treat relapsed disease. The prognostic value of genomic and biological biomarkers is known to depend on the treatment received. Thus, a deeper understanding of the drivers of neuroblastoma growth and therapy resistance may provide insight for the development of more effective, individualized treatments for patients with intermediate risk tumors, ultimately reducing the significance of the prognostic factors analyzed in this study. Recent studies have shown that a subset of IR tumors harbor molecular features including telomere maintenance mechanisms (TMM),^23,24 mutations in ALK^25,26 or other genes in the RAS/MAPK pathway, and/or p53 mutations.²⁷ Ackermann et al., reported that outcome was significantly worse for non-high-risk patients with tumors that harbored TMM compared to those with neuroblastoma that lacked TMM.²³ Further, survival was lowest among patients with tumors that harbored TMM combined with RAS/MAPK or p53 pathway mutations.²³ If these findings are validated in prospective studies, genomic analysis of IR neuroblastoma may inform treatment decisions and identify patients who may benefit from alternative strategies including novel drug combinations.

Overall, excellent long-term survival was observed for IR NBL patients enrolled on ANBL0531 and this treatment has become the standard of care for the treatment of IR NBL patients at COG institutions. Our results support decreasing surveillance imaging and associated radiation exposure for subsets of IR patients in the future. Inferior EFS, however, was observed among subsets with unfavorable biologic features. To improve the outcome of these subsets of IR patients, nimble, cost-effective approaches are needed to precisely molecularly classify neuroblastoma tumors and to develop new biomarkers and more effective biologically-based treatment strategies.

Abbreviations

INRC

International Neuroblastoma Response Criteria

OS

Overall Survival

EFS

Event Free Survival

HR

High Risk

IR

Intermediate Risk

CI

Confidence Interval

LOH

Loss of Heterozygosity

BM

Bone Marrow

INSS

International Neuroblastoma Staging System

VGPR

Very Good Partial Response

PR

Partial Response

INPC

International Neuroblastoma Pathology Classification

UH

Unfavorable Histology

CPM

Cyclophosphamide

TOPO

Topotecan

PD

Progressive Disease

MYCN-NA

MYCN Non-Amplified

PFS

Progression-Free Survival

LR

Low Risk

COG

Children’s Oncology Group

INRG

International Neuroblastoma Risk Group

TMM

Telomere Maintenance Mechanism