Clinical and Biological Features Prognostic of Survival After Relapse or Progression of INRGSS Stage MS Pattern Neuroblastoma: A Report from the International Neuroblastoma Risk Group (INRG) Project

Abstract

Background:

Outcomes for patients with INRGSS MS metastatic pattern neuroblastoma at initial diagnosis are well-described. Prognosis after an initial event (relapse, progression, secondary malignancy) is unclear.

Methods:

We investigated characteristics of MS pattern neuroblastoma patients in the International Neuroblastoma Risk Group database who subsequently experienced an event. Post-event overall survival (OS) ± standard error was calculated overall and by diagnosis era: <2000 vs ≥2001. Cox models were used to identify factors prognostic of post-event OS.

Results:

Among 209 patients with an event, 88% were <365 days old at diagnosis; tumors were MYCN amplified in 24% and diploid in 33%. The median (range) time from diagnosis to first event was 8.16 months (7 days-11.24 years). Of 96 patients with known relapse/progression pattern, 75% were metastatic or primary plus metastatic. Five-year post-event OS was 53±3.6% and was higher for ≥2001 (62±5.0%) compared to <2001 (44±4.9%; p=0.0046). In patients diagnosed ≥2001, older age, Hispanic ethnicity, MYCN amplification, 1p LOH, diploidy, high MKI, high LDH, unfavorable histology, and shorter time to first event were prognostic of worse post-event OS. Independent adverse prognostic factors on multivariable testing were non-White race, MYCN amplification, and diploidy.

Summary:

Patients diagnosed ≥2001 have substantially better post-event OS compared to <2001. In those diagnosed ≥2001, most well-accepted prognostic factors for OS at diagnosis are also prognostic of post-event OS. Future studies may evaluate strategies to improve outcomes in this rare population.

INTRODUCTION

Neuroblastoma, the most frequent extracranial solid tumor in children, is known for its clinical and biological heterogeneity. Neuroblastoma presenting with stage MS (metastatic special) metastatic pattern highlights the heterogeneity of this disease. Four neuroblastoma staging systems include a category for this unique subgroup of neuroblastoma: Evan’s stage IV-S [1, 2]; Pediatric Oncology Group (POG) stage DS [3]; International Neuroblastoma Staging System (INSS) 4S [4, 5]; and International Neuroblastoma Risk Group staging system (INRGSS) MS [6]. The most recent staging systems have defined this subtype as diagnosis at <365 days (INSS 4S) or <547 days (INRGSS MS) with metastatic neuroblastoma, and metastatic sites limited to the bone marrow (<10% tumor cells for INSS 4S), liver and/or skin.

The natural history and prognosis of patients with MS pattern disease have previously been documented, including potential for spontaneous regression. A prior systematic review of 1,105 patients with stage 4S/MS disease noted an OS of 84%, but less favorable outcomes for patients with MYCN amplification or chromosome 1p or 11q aberrations [7]. A single institutional review of 31 patients with INSS stage 4S disease patients showed that 25 who were alive >2 years post-treatment had excellent long-term outcomes, with the majority of patients having no clinically or radiologically significant sequalae [8]. After development of INRGSS, a 2011 retrospective study of patients <18 months with metastatic disease found tumor biology to be more critical than age or metastatic pattern for risk stratification [9].

While prognosis and treatment course at initial diagnosis for patients with stage MS disease are known, expectations surrounding prognosis at the time of relapse are less well defined [10, 11]. The goals of the current study were to describe the clinical and biological features, at diagnosis and at first event, of patients with MS pattern neuroblastoma who experience relapse/progression/second malignancy, and to identify factors prognostic of overall survival post-event. We compared OS for those diagnosed <2001 versus ≥2001 to broadly differentiate patients that likely received more modern therapy including reduction in treatment for non-high-risk disease and high-dose chemotherapy with stem cell rescue for high-risk disease.

METHODS

Patients

The data source for this retrospective study was the International Neuroblastoma Risk Groups (INRG) Data Commons, comprised of patients enrolled on clinical trials from Europe, North America, Australia, New Zealand, and Japan. This database comprises the largest available set of patients with neuroblastoma (n=22,700), and thus the largest available set of patients with relapsed stage MS pattern disease. The eligibility criteria were: diagnosis date from 1984–2021; initial diagnosis of INRGSS stage MS disease; if INRGSS was unknown, patients <365 days at initial diagnosis with INSS stage 4S were eligible, or age 365–546 days with INSS stage 4 disease and metastasis restricted to the liver, skin, and/or bone marrow at initial diagnosis. Of these patients, a subsequent relapse, progression, or secondary malignancy, with the exclusion of death as first event, was necessary for inclusion.

Variables and Statistical Considerations

Aside from pattern of relapse/progression and time to first event, clinical and biologic features were captured at time of initial diagnosis. Biologic features were not available for patients who did not undergo biopsy at diagnosis. Data were not available in this registry for treatment received. For lactate dehydrogenase (LDH) and ferritin, optimal values of 1400 IU/L and 30 ng/mL, respectively, which maximized the event-free survival hazard ratio between elevated versus not elevated in prior analyses [12], were used to dichotomize the patient cohort. A cut-point of 12 months was applied to dichotomize early versus later event, as this cut-point has previously been identified as prognostic in relapsed patients [9].

The primary endpoint was the time from the first occurrence of an event (relapse, progression, secondary malignancy) until death from any cause or last contact. Kaplan-Meier methods were used to calculate the point estimates of post-event OS±SE at 3- and 5-years, overall and by era of diagnosis: <2001 vs ≥2001. The overall cohort was divided by date of diagnosis, <2001 versus ≥2001 to describe changes in outcome according to treatment era.

More detailed survival analyses were restricted to patients diagnosed ≥2001. Univariate and multivariable Cox proportional hazards models identified factors prognostic of post-event OS. The multivariable models included a dummy variable for the ‘unknown’ category of each factor with missing data, to permit inclusion of the entire sample in the model. Factors that were significant in a univariate Cox model were tested in the multivariable model; the most parsimonious model was created using backwards selection, dropping factors with p>0.05. P-values <0.05 were considered statistically significant. Analyses were performed using SAS version 9.4.

RESULTS

Patient Characteristics

Of the 22,700 patients in the INRG Data Commons, 1,511 met the staging criteria for inclusion (Supplemental Figure 1). Of the 1,511, 1,201 were excluded because they did not have an event as of last contact (n=1,157 who were <365 days old; n=44 who were 365–547 days old). Of the remaining 310, 101 patients were excluded due to death as first event, leaving 209 patients who met study eligibility: 103/209 (49%) were diagnosed <2001 and 106/209 (51%) were diagnosed ≥2001 (Supplemental Figure 1). Relapse or progression was the first event in 203/209 (97%) patients, and 6/209 (3%) were patients whose first event was a secondary malignancy (two <2001; four ≥2001). Proportions of clinical and biological factors at initial diagnosis were similar between eras (Table 1).

Table 1.

Clinical and biological characteristics of patients <547 days of age at diagnosis with MS pattern neuroblastoma who experienced disease relapse, progression, or secondary malignancy (n=209)

Factor	All (n=209)		Diagnosed <2001 (n=103)		Diagnosed ≥2001 (n=106)
	n	OS ± SE (%)	n	OS ± SE (%)	n	OS ± SE (%)
Post-event OS 3-year OS 5-year OS	209	58 ± 3.5 53 ± 3.6	103	46 ± 4.95 44 ± 4.9	106	69 ± 4.6 62 ± 5.0
Features at Initial Diagnosis
	n	%	n	%	n	%
Age  <365 days  365–547 days of age	183 26	87.6% 12.4%	89 14	86.4% 13.6%	94 12	88.7% 11.3%
Sex  Female  Male  Unknown	76/160 84/160 49	47.5% 52.5%	36/74 38/74 29	48.7% 51.4%	40/86 46/86 20	46.5% 53.5%
Race  White  Black  Asian  Other  Unknown	82/97 13/97 2/97 0/97 112	84.5% 13.4% 2.1% 0%	30/37 7/37 0/37 0/37 66	81.1% 18.9% 0% 0%	52/60 6/60 2/60 0/60 46	86.7% 10% 3.3% 0%
Ethnicity  Hispanic  Not Hispanic  Unknown	7/121 114/121 88	5.8% 94.2%	0/42 42/42 61	0% 100%	7/79 72/79 27	8.9% 91.1%
Primary Site┼  Adrenal  Non-adrenal abdominal  Thoracic  Pelvic  Neck  Other  Unknown	133/189 41/189 11/189 5/189 3/189 1/189 20	70.4% 21.7% 5.8% 2.6% 1.6% 0.5%	62/87 19/87 5/87 1/87 0/87 0/87 16	71.3% 21.8% 5.8% 1.2% 0% 0%	71/102 22/102 6/102 4/102 3/102 1/102 4	69.6% 21.6% 5.9% 3.9% 2.9% 0.98%
MYCN status  Not amplified  Amplified  Unknown	144/190 46/190 19	75.8% 24.2%	63/88 25/88 15	71.6% 28.4%	81/102 21/102 4	79.4% 20.6%
11q  Balanced or no aberration  Deletion, imbalance or unbalanced  Unknown	28/37 9/37 172	75.7% 24.3%	6/9 3/9 94	66.7% 33.3%	22/28 6/28 78	78.6% 21.4%
17q  No gain  Gain  Unknown	5/8 3/8 201	62.5% 37.5%	3/6 3/6 97	50% 50%	2/2 0/2 104	100% 0%
1p  No loss or aberration  LOH, deletion, or imbalance  Unknown	39/65 26/65 144	60% 40%	16/30 14/30 73	53.3% 46.7%	23/35 12/35 71	65.7% 34.3%
Ploidy  Hyperdiploid  Diploid  Unknown	64/95 31/95 114	67.4% 32.6%	20/31 11/31 72	64.5% 35.5%	44/64 20/64 42	68.8% 31.3%
Grade of differentiation  Differentiating  Undifferentiated  Unknown	4/123 119/123 86	3.3% 96.7%	2/33 31/33 70	6.1% 93.9%	2/90 88/90 16	2.2% 97.8%
MKI  Low  Intermediate  High  Unknown	75/116 22/116 19/116 93	64.7% 18.97% 16.4%	27/36 4/36 5/36 67	75% 11.1% 13.9%	48/80 18/80 14/80 26	60.0% 22.5% 17.5%
LDH (U/L)  <1400  ≥1400  Unknown	82/122 40/122 87	67.2% 32.8%	53/80 27/80 23	66.3% 33.8%	29/42 13/42 64	69.05% 30.95%
Serum ferritin (ng/mL)  <30  ≥30  Unknown	29/96 67/96 113	30.2% 69.8%	11/58 47/58 45	19% 81%	18/38 20/38 68	47.4% 52.6%
Histologic Classification  Favorable  Unfavorable  Unknown	102/125 23/125 84	81.6% 18.4%	35/46 11/46 57	76.1% 23.9%	67/79 12/79 27	84.8% 15.2%
Features at First Event
Site of First Relapse / Progression  Primary  Metastatic  Primary + metastatic  Unknown	24/96 46/96 26/96 113	25.0% 47.9% 27.1%	8/36 17/36 11/36 67	22.2% 47.2% 30.6%	16/60 29/60 15/60 46	26.7% 48.3% 25%
If metastatic relapse, anatomical site(s) of relapse  MS pattern  Non-MS pattern  Unknown	8/33 25/33 176	24% 76%	1/3 2/3 100	33% 66%	7/30 23/30 76	23% 77%
Time from Initial Diagnosis to First Event  < 6 months  6–12 months  >12 months	81 54 74	38.8% 25.8% 35.4%	39 28 36	37.9% 27.2% 35%	42 26 38	39.6% 24.5% 35.9%
Median (range) time from diagnosis to first event(days)	209	249 (7–4106)	103	238 (7–3072)	106	251 (7–4106)

^┼not mutually exclusive: 5 patients have reported more than one primary site.

Overall, 183/209 (88%) were diagnosed <365 days of age. Of those with known data, 84/160 (53%) were males, 82/97 (85%) were white, and 114/121 (94%) were non-Hispanic. Primary tumor site was the adrenal gland in 133/189 (70%), and 46/190 (24%) patients had MYCN amplified tumors. Of those with known data, 9/37 (24%) had aberration of 11q, 26/65 (40%) had aberration of 1p, and 102/125 (82%) patients had favorable histology tumors. We infer that tissue was obtained from a minimum of 190 patients as at least one tissue-based marker was available.

Pattern and Timing of First Relapse / Progression

96/203 patients (47%) had known data for pattern of first relapse or progression. Of these 96, 24/96 (25%) had first relapse or progression exclusively at the primary site, 46/96 (48%) at a metastatic site, and 26/96 (27%) at primary and metastatic sites. Of 72 patients with metastatic site at relapse, data were known for the relapse pattern in 33. Of these 33, 8/33 (24%) remained stage MS pattern, and 25/33 (76%) had metastatic sites beyond liver, skin, and bone marrow (i.e., not MS pattern), though skin was not coded as a specific potential site of relapse in the database.

The median [range] time from diagnosis to first event was 8.16 months (7 days to 11.24 years) for the overall cohort (n=209), 7.80 months (7 days to 8.4 years) for those diagnosed <2001 (n=103), and 8.21 months (7 days to 11.24 years) for ≥2001 (n=106). Overall, 81/209 (39%) of patients had their first event in <6 months after initial diagnosis, 54/209 (26%) had an event between 6–12 months after diagnosis, and 74/209 (35%) had an event >12 months after diagnosis. This pattern was similar regardless of treatment era. Specifically, of 103 patients diagnosed <2001, 39/103 (38%) patients had their first event in <6 months after initial diagnosis, 28/103 (27%) had an event between 6–12 months after diagnosis, and 36/103 (35%) had an event >12 months after diagnosis. Of 106 patients diagnosed ≥2001, 42/106 (40%) patients had their first event in <6 months after initial diagnosis, 26/106 (25%) had an event between 6–12 months after diagnosis, and 38/106 (36%) had an event >12 months after diagnosis.

Post-event Survival

For the entire cohort of 209 patients, 5-year post-event OS was 53±3.6% (Figure 1A). Post-event OS was higher for those diagnosed ≥2001 (5-year post-event OS 62±5.0%; n=106) compared to those diagnosed <2001 (5-year post-event OS 44±4.9%; n=103; p=0.0046; Figure 1B).

Figure 1A.

Kaplan-Meier curve of OS from time of first event for all patients diagnosed with MS pattern neuroblastoma (n=209).

Figure 1B.

Kaplan-Meier curves of OS from time of first event for patients diagnosed with MS-pattern disease <2001 (n=103) versus ≥2001 (n=106).

Univariate Survival Analyses

Given the more modern therapy likely administered to patients diagnosed ≥2001 and their improved post-event OS, prognostic survival analyses focused on this cohort. In univariate analyses, statistically significant factors associated with inferior post-event OS were: age at initial diagnosis 365–547 days (hazard ratio [HR]=3.85); Hispanic ethnicity (HR=3.62); MYCN amplification (HR=5.81); 1p loss/aberration (HR=3.86); diploidy (HR=4.59); high MKI (HR=3.48); LDH ≥1400 U/L (HR=3.58); unfavorable INPC histology (HR=2.66); and ≥12 months from initial diagnosis to first event (HR=2.09; Table 2 and Figure 2A-2I).

Table 2.

Testing to identify factors from initial diagnosis that are prognostic of post-event overall survival, in patients diagnosed ≥2001 at <547 days of age with MS pattern neuroblastoma (n=106)

Factor	n	Post-event OS				p-value
		3-year OS ± SE (%)	5-year OS ± SE (%)	HR	95% CI on HR
Features at Initial Diagnosis
Age  <365 days*  365–547 days of age	94 12	75 ± 4.6 25 ± 12.5	66 ± 5.3 25 ± 12.5	3.85	1.80, 8.24	0.0005
Sex  Female*  Male  Unknown	40 46 20	75 ± 6.9 61 ± 7.4	69 ± 7.5 51 ± 7.7	1.59	0.78, 3.24	0.2
Race@  White*  Black  Asian  Unknown	52 6 2 46	76 ± 6.1 67 ± 19.3 50 ± 35.4	73 ± 6.4 50 ± 20.4 50 ± 35.4	2.51 2.67	0.71, 8.81 0.35, 20.45	0.2 0.3
Ethnicity  Not Hispanic*  Hispanic  Unknown	72 7 27	78 ± 5.0 34 ± 19.5	74 ± 5.4 17 ± 15.6	3.62	1.33, 9.82	0.01
Primary Site┼  Adrenal  Non-adrenal abdominal  Thoracic  Pelvic  Neck  Other	71 22 6 4 3 1	71 ± 5.5 50 ± 10.7 83 ± 15.2 100 ± 0 100 ± 0 100 ± 0	59 ± 6.3 50 ± 10.7 83 ± 15.2 100 ± 0 100 ± 0 100 ± 0	1.07 1.88 2.68 0 0 0	0.53, 2.16 0.93, 3.78 0.37, 19.52 0, 0 0, 0 0, 0	0.9 0.08 0.3 0.99 0.99 0.99
Primary Site┼  Non-adrenal*  Adrenal  Unknown	34 71 4	68± 8.0 71 ± 5.5	68 ± 8.0 59 ± 6.3	1.07	0.53, 2.16	0.9
MYCN status  Not amplified*  Amplified  Unknown	81 21 4	82 ± 4.3 21 ± 9.2	72 ± 5.4 21 ± 9.2	5.81	2.96, 11.39	<0.0001
11q  Balanced or no aberration*  Deletion, imbalance or unbalanced  Unknown	22 6 78	72 ± 9.8 80 ± 17.9	67 ± 10.4 80 ± 17.9	0.49	0.06, 4.009	0.5
1p  No loss or aberration*  LOH, deletion, or imbalance  Unknown	23 12 71	86 ± 7.4 42 ± 14.2	75 ± 9.6 33 ± 13.6	3.86	1.26, 11.85	0.02
Ploidy  Hyperdiploid*  Diploid  Unknown	44 20 42	86 ± 5.3 45 ± 11.1	80 ± 6.4 40 ± 10.95	4.59	1.87, 11.25	0.0009
Grade of differentiation  Undifferentiated*  Differentiating  Unknown	88 2 16	71 ± 5.0 50 ± 35.4	62 ± 5.6 50 ± 35.4	1.34	0.18, 9.84	0.8
MKI  Low/Int*  High  Unknown	66 14 26	76 ± 5.4 34 ± 13.1	64 ± 6.4 34 ± 13.1	3.48	1.58, 7.70	0.002
LDH (U/L)  <1400*  ≥1400  Unknown	29 13 64	77 ± 8.5 38 ± 13.5	77 ± 8.5 38 ± 13.5	3.58	1.23, 10.39	0.02
Serum ferritin (ng/mL)  <30*  ≥30  Unknown	2 20 84	0 ± 0 42 ± 11.5	0 ± 0 34 ± 11.9	1.65	0.21, 12.86	0.6
Histologic Classification  Favorable*  Unfavorable  Unknown	67 12 27	74 ± 5.5 39 ± 14.7	63 ± 6.3 39 ± 14.7	2.66	1.13, 6.25	0.03
Features at First Event
Site of First Relapse / Progression  Primary*  Metastatic  Primary + metastatic  Unknown	16 29 15 46	80 ± 10.2 54 ± 9.4 66 ± 12.4	80 ± 10.2 42 ± 9.6 66 ± 12.4	3.27 2.003	0.95, 11.22 0.48, 8.39	0.06 0.3
Time from Initial Diagnosis to First Event  <12 months*  ≥12 months	68 38	74 ± 5.4 60 ± 8.1	72 ± 5.6 44 ± 8.6	2.09	1.11, 3.95	0.02

^*reference group

^┼not mutually exclusive: 5 patients have reported more than one primary site

^@For non-white vs white*: HR=2.138, 95%CI: (1.092, 4.183); p=0.0265

Abbreviations: OS – overall survival, SE – standard error, HR – hazard ratio, CI – confidence interval

Figure 2.

Kaplan-Meier curves of OS from time of first event for patients diagnosed with MS-pattern disease ≥2001 according to the following variables

significantly associated with post-event OS on univariate analyses:

A. age at diagnosis: 365–547 days (n=12) vs <365 days (n=94)

B. ethnicity: Hispanic (n=7) vs non-Hispanic (n=72)

C. MYCN: amplified (n=21) vs not amplified (n=81)

D. 1p deletion: LOH, deletion, or imbalance (n=12) vs no loss or aberration (n=23)

E. ploidy: diploid (n=20) vs hyperdiploid (n=44)

F. MKI: high (n=14) vs low and intermediate (n=66)

G. LDH (U/L): ≥1400 (n=13) vs <1400 (n=29)

H. histologic classification: unfavorable (n=12) vs favorable (n=67)

I. first event: ≥12 months (n=38) vs <12 months (n=68)

Multivariable Survival Analysis

The multivariable approach included a category for ‘unknown’ for each factor with missingness, allowing a parsimonious model to be generated based on the cohort diagnosed ≥2001 (n=106). MYCN amplification [HR=5.3; 95% confidence interval (CI) (2.5, 11.2) vs. non-amplified], non-White race [HR=3.5; 95% CI (1.4, 8.5) vs White race], and diploidy [HR=3.0; 95% CI (1.2, 7.7) vs. hyperdiploid] were independently prognostic of lower post-event OS (Table 3).

Table 3.

Multivariable Cox PH model to identify factors^** at initial diagnosis independently prognostic of overall survival in patients <546 days of age at diagnosis with MS-pattern neuroblastoma who experienced an event after initial diagnosis ≥2001 (n=106)

Factor*	Adjusted HR	95% CI on HR	p-value
MYCN non-amplified	Reference
MYCN amplified	5.3	(2.5, 11.2)	<0.0001
MYCN unknown	4.1	(0.9, 19.7)	0.0763
White race	Reference
Non-White race	3.5	(1.4, 8.5)	0.0066
Hyperdiploid	Reference
Diploid	3.0	(1.2, 7.7)	0.0244
Ploidy unknown	0.8	(0.3, 2.3)	0.6364

^*For each binary factory, this category has an increased risk of death compared to the reference level

^**Factors tested in the model but not found independently statistically significant were: ethnicity, age, time from diagnosis to relapse/progression, LDH, and MKI.

DISCUSSION

In this comprehensive analysis, we provide new data on post-event outcomes as well as clinical and biological features prognostic of post-event survival in patients with MS pattern neuroblastoma. Among patients with known data for site of relapse/progression, we found that a majority of events were metastatic failures, and these events more commonly occurred at sites outside liver, skin, and bone marrow. Patients diagnosed ≥2001 had substantially better post-event OS compared to those diagnosed <2001.

We show that key features identified at initial diagnosis were prognostic of post-event OS in patients with stage MS pattern disease. Well-accepted prognostic factors associated with risk of initial event remain prognostic of OS post-event within this subgroup of patients with MS pattern disease, including age at diagnosis, MYCN status, ploidy, MKI, LDH, and INPC histology. We also add to the growing literature highlighting survival disparities according to race and/or ethnicity in this disease[13, 14].

In our study population, we show that most patients with relapsed MS pattern disease not only had metastatic relapse, but importantly this may include sites outside liver, skin, and bone marrow. Despite this pattern, the 5-year post-event OS of 53±3.6% of the overall (n=209) cohort, and 62±5.0% for the 106 patients diagnosed ≥2001, is favorable when compared to a recent analysis of high-risk patients with relapsed neuroblastoma [15]. That analysis reported the 5-year post-relapse OS as 7.4%, with cases diagnosed ≤2000 having a 5-year post-relapse OS of 2.4% while patients diagnosed >2000 had a 5-year post-relapse OS of 12.7% [15]. An Italian review of 268 infants with initial INSS stage 4S neuroblastoma that included 29 of 57 patients who experienced disease progression with disease confined to 4S metastatic site reported a 5-year post-relapse OS of 69% compared to a 5-year post-relapse OS of 47% for the 28 patients who experienced disease progression outside 4S metastatic sites [16]. Likewise, outcomes for this relapsed/progressing MS population (5-year OS of 62±5.0% for the cohort ≥2001) are similar to outcomes for patients with newly diagnosed high-risk disease (5-year OS of 62.5±1.3% OS as reported from a review of a phase 3 study[17]. The improvement in 5-year post-event OS from <2001 to ≥2001 of nearly 20% in our analysis is likely influenced by more precise biological risk stratification and improvement in treatment over time. For example, therapeutic improvements during this time that would be available to patients had they been initially classified as low-risk but reclassified as high-risk after relapse, include tandem consolidative transplant and immunotherapy with anti-GD2 antibody therapy during post-consolidation care [18, 19].

Most events occurred within 12 months of initial diagnosis. For both cohorts (diagnosis <2001 and ≥2001), close to 40% of patients had an event within 6 months of initial diagnosis, another quarter of patients within the next 6 months, and the remaining 35% after 1 year. In a prior study of 2,266 relapsed or progressive neuroblastoma patients of all stages, the median time from diagnosis to first relapse was 13.2 months [10], which is longer than the 8.2 months in our study of patients with relapse, progression, or secondary malignancy and MS pattern disease. Since not all patients with stage MS pattern disease receive initial chemotherapy, the time from initial diagnosis to first event may have been influenced by receipt of treatment versus observation at initial presentation, with patients who received treatment upfront possibly experiencing a longer time to first event. Further, it is possible that stage MS patients may experience disease progression but not receive therapy and subsequently experience spontaneous regression. In our analysis, patients with early event within 12 months of initial diagnosis had superior 5-year post-event OS compared to those with first event >12 months from diagnosis. In a report of 233 neuroblastoma patients who were eligible for/enrolled on a COG phase 1 or 2 trial from 2002–2014, the median time to first relapse (TTFR) was 18.7 months, but it was not possible to identify an optimal TTFR cutoff value that was prognostic of post-relapse OS [20].

A major strength of the current study was the use of the largest available cohort with follow-up over a 38-year period. All 209 patients with stage MS neuroblastoma who had an event also had available data regarding time from initial diagnosis to first event and post-event OS. The inclusion of patients with secondary malignancy as a first event is considered both a strength and a limitation. In prior INRG analyses of first event cohorts [10], data were unavailable to distinguish or exclude secondary malignancy from relapse/progression events. Therefore, it is a strength that the results of our study are comparable to the prior publications. However, the clinical interpretation of our outcome results for ≥2001 is limited and diluted by the inclusion of four patients with secondary malignancy, who likely received different treatment than patients with relapse/progression. A second limitation of this analysis was missing data, particularly in the sub cohort of patients diagnosed <2001. Further, we do not have data regarding second and subsequent episodes of relapse or progression. Third, data regarding risk classification and treatment approaches utilized in the upfront setting or relapse setting were not available for analysis, which make episodes of progression difficult to interpret. As discussed, treatment for MS pattern disease can vary from observation to intensive multi-modality therapy, and thus post-event survival may possibly have been affected by upfront treatment. Fourth, most tumor characteristics reflect tumor pathologic and biological features present at initial diagnosis, and these features were not reassessed at the time of relapse. Moreover, we lack data on newer genomic features, such as ALK status, at diagnosis or relapse. We were limited in our ability to determine whether a patient remained as MS pattern at the time of relapse/progression. Site of relapse was collected as: primary, bone, bone marrow, liver, lymph nodes, lung, CNS, and other; unfortunately, “skin” site was not collected. Therefore, patients may have been misclassified as to whether they remained as MS pattern at the time of relapse/progression. Future studies could evaluate whether patients with adverse clinical and biological features may benefit from earlier treatment or help justify biopsy of the relapsed/progressive tumor for understanding additional biological characteristics (e.g., ALK mutations) and provide information on possible use of appropriate targeted therapy.

In summary, this study provides new data regarding clinical and biological characteristics as well as outcomes post-event from the largest known cohort of patients with MS pattern neuroblastoma. These results provide an informative and prognostic tool for patients and clinicians with initial or relapsed/progressive stage MS neuroblastoma. We strongly encourage ongoing detailed and prospective clinical, biological, and treatment data collection for this rare subgroup of patients to facilitate our ability to improve their outcomes.

Abbreviations Key

CI

Confidence Interval

COG

Children’s Oncology Group

CNS

Central Nervous System

HR

Hazard Ratio

INPC

International Neuroblastoma Pathology Classification

INRG

International Neuroblastoma Risk Group (INRG) Project

INRGSS

International Neuroblastoma Risk Group Staging System

INSS

International Neuroblastoma Staging System

LDH

Lactate Dehydrogenase

MKI

Mitotic Karyorrhexis Index

MS

Metastatic Special

OS

Overall Survival

POG

Pediatric Oncology Group

SE

Standard Error

TTFR

Time To First Relapse