Abstract
Background:
MYCN amplification (MYCN-A) is an established adverse prognostic factor in neuroblastoma. The extent to which the prognostic impact of MYCN-A depends on other factors has not been fully characterized.
Patients and methods:
Using the International Neuroblastoma Risk Group database, we constructed Cox models of overall survival (OS) to obtain hazard ratios of the effect of MYCN-A within subgroups defined by other prognostic factors. Cox models assessed the degree to which the prognostic impact of MYCN-A was modulated by each other covariate. We used absolute hazard ratio (HR) differences to construct classification trees to identify subgroups with greatest differential prognostic effect of MYCN-A.
Results:
In a cohort of 6223 patients with known MYCN status, the OS hazard ratio associated with MYCN-A was 6.3 (95% confidence interval 5.7-7.0, P < .001). Age at diagnosis conferred the largest HR absolute difference for MYCN-A between subgroups (HR absolute difference 16.6; HRs for MYCN-A of 19.6 for <18 months, 3.0 for ≥18 months). MYCN-A remained significantly prognostic of OS after controlling for other factors, abrogating their prognostic strength. Patients whose outcome was most impacted by MYCN status were those who were <18 months, had high mitosis karrhyohexis index (MKI) and low ferritin.
Conclusion:
The prognostic strength of MYCN-A varies depending on which patient subgroup defined by other neuroblastoma risk factors is examined, with greatest strength in patients with otherwise favorable features. MYCN-A has little effect within some subgroups, aiding clinical decision-making if MYCN status cannot be assessed. Subgroups where MYCN-A has large effect may be prioritized for agents targeting Myc family proteins.
Keywords: amplification, context, MYCN, neuroblastoma, prognosis
1 |. INTRODUCTION
The classification and treatment approach to neuroblastoma, the most common extracranial solid tumor of childhood, is complicated by the clinical heterogeneity of the disease. MYCN status, determined at the time of diagnosis, is an important adverse prognostic factor.1 It has been over three decades since the historic discoveries linking amplification of the oncogene MYCN with rapid tumor progression,2–4 resulting in MYCN status as a critical prognostic factor that remains a cornerstone of current risk classification systems.5
Previous work from our group identified associations between other known features of neuroblastoma and differential rates of MYCN amplification (MYCN-A). MYCN-A demonstrates complex and differential associations with many other prognostic factors.6,7 Few studies have examined the prognostic context of these associations in specific subpopulations of neuroblastoma, demonstrating the presence of MYCN-A appears to have a greater adverse prognostic impact in patients with otherwise favorable features (eg, younger age and lower stage).8–11 In contrast, in older patients with higher stage disease, the prognostic impact of MYCN-A has been more modest or even undetectable.5,12 For example, one study found that MYCN status did not significantly impact overall survival (OS) in older patients with stage 4 disease.13 Additionally, two recent studies did not demonstrate a prognostic effect of MYCN-A on patients with high-risk disease.14,15
A comprehensive investigation into the context dependence of MYCN status is needed to provide clinicians a more nuanced understanding of its prognostic impact in the setting of other clinical, biological, and treatment factors. Further, as treatment strategies have evolved, it is unclear whether the prognostic impact of MYCN status has evolved as well. In this study, we utilized the International Neuroblastoma Risk Group (INRG) database to perform a comprehensive evaluation of the variation of the prognostic impact of MYCN-A. This analysis, while primarily serving to provide improved understanding of MYCN status as a prognostic factor in neuroblastoma, may also aid our understanding of the important interactions between MYCN status and other clinical, biological, and treatment factors. Discerning differences in the degree to which MYCN-A adversely impacts prognosis may enable providers to more accurately weigh MYCN status when assessing an individual patient’s risk of treatment failure.
2 |. METHODS
2.1 |. Patients
Patients diagnosed with neuroblastoma between 1990 and 2016 were selected from the INRG database and were eligible for the analysis if they had known outcome and MYCN status (coded as amplified vs nonamplified). There were no other inclusion or exclusion criteria for this analysis.
2.2 |. Covariates
MYCN status was the primary predictor variable of interest for this analysis. MYCN status was dichotomized as amplified (MYCN-A) versus nonamplified (MYCN-NA). MYCN status was determined according to local standards as previously described.6
Clinical factors of interest, at the time of diagnosis, included sex, age, International Neuroblastoma Staging System (INSS) stage (dichotomized as stage 4 vs all other stages including 4S),16 primary site, presence of bone marrow metastases, presence of bone metastases, lactate dehydrogenase (LDH) level (dichotomized using updated cut point as previously17), ferritin level (dichotomized using updated cut point as previously17), and year of diagnosis (dichotomized around the year 1999 when the addition of high-dose therapy with autologous stem cell rescue became routine). Biological covariates of interest included ploidy (hyperdiploidy = any DNA index >1.0), 1p loss of heterozygosity (LOH), 11q aberration (unbalanced LOH),18 presence of any segmental chromosomal aberration (SCA) (either 1p LOH and/or 11q LOH), International Neuroblastoma Pathology Classification histologic classification,19 diagnostic category (neuroblastoma and nodular ganglioneuroblastoma vs all others), MKI, and grade of differentiation.
OS was the sole clinical outcome of interest. OS time was calculated from the time from diagnosis to death, with surviving patients censored at time of last follow-up.
2.3 |. Statistical analyses
The INRG cohort of 14501 patients who met eligibility for this analysis was randomly and equally divided into a Test cohort and a Validation cohort. Except where noted, analyses were performed first in the Test cohort and then confirmed in the Validation cohort.
Point estimates of OS were calculated using the methods of Kaplan and Meier,20 with standard errors according to the methods of Greenwood.21
We used absolute hazard ratio differences to construct classification trees to identify subgroups with the strongest and weakest prognostic effect of MYCN-A.
Starting from the overall Test cohort, we used a series of univariate Cox models of OS to calculate the HR for MYCN within each of the subgroups for each factor. For example, the HR for MYCN was calculated within patients <18 months of age at diagnosis, and also within ≥18 months. Then we calculated the absolute difference in the HR (HRdiff) by subtracting the HR of the group with more favorable outcome from the HR of the group with less favorable outcome, for example, HRdiff = (MYCN HR for age ≥18 months) – (MYCN HR for age < 18 months). The factor with the largest HRdiff was selected to form the first split/branch of the classification tree. This procedure was repeated within each branch until the sample size became too small (subgroup <5), HRdiff < 1, or until there were no further subgroups in which MYCN was significant. Data from the Validation cohort were then placed into categories according to the tree created from the Test cohort to determine if there was evidence to support the split/branch decisions made based on the Test cohort.
In addition, we used bivariate Cox models to assess the extent to which each covariate impacted the HR for death associated with MYCN-A. We also used Cox models that included an interaction term of [MYCN status] × [covariate] to formally assess for modification of the prognostic impact of MYCN-A by each covariate.
3 |. RESULTS
3.1 |. Patient characteristics
In the Test cohort (n = 7251), MYCN-A was observed in 928 cases (12.8%), MYCN-NA in 5295 cases (73.0%), and MYCN status was unknown in 1028 cases (14.2%). Cases with unknown MYCN status were excluded from further analysis. Details of the clinical and biological features for the Test and Validation cohorts are shown in Tables 1 and S1, respectively. Features of patients universally considered high risk (patients with metastatic disease diagnosed ≥18 months of age) for the Test and Validation cohorts are shown in Table S2.
TABLE 1.
Characteristics of the Test cohort according to MYCN status
| Variable | All patients n (%) (n = 7251) |
MYCN-A n (%) (n = 928) |
MYCN-NA n (%) (n = 5295) |
MYCN unknowna n (%) (n = 1028) |
|---|---|---|---|---|
| Male | 2535 (53) | 371 (56) | 1895 (52) | 269 (56) |
| Female | 2241 (47) | 296 (44) | 1732 (48) | 213 (44) |
| (Unknown) | 2475 | 261 | 1668 | 546 |
| Age ≥18 months | 3290 (45) | 587 (63) | 2109 (40) | 594 (58) |
| Age <18 months | 3961 (55) | 341 (37) | 3186 (60) | 434 (42) |
| INSS stage 4 | 2567 (37) | 667 (73) | 1401 (27) | 499 (53) |
| All other stages | 4455 (63) | 245 (27) | 3765 (73) | 445 (47) |
| (Unknown) | 229 | 16 | 129 | 84 |
| Adrenal primary | 2777 (43) | 460 (57) | 1927 (40) | 390 (48) |
| Other primary sites | 3640 (57) | 343 (43) | 2868 (60) | 429 (52) |
| (Unknown) | 834 | 125 | 500 | 209 |
| Thoracicprimary | 1233 (19) | 39 (5) | 1041 (22) | 153 (18) |
| Other primary sites | 5198 (81) | 765 (95) | 3755 (78) | 678 (82) |
| (Unknown) | 820 | 124 | 499 | 197 |
| Bone metastasis | 1024 (29) | 247 (49) | 524 (22) | 253 (40) |
| No bone metastasis | 2547 (71) | 257 (51) | 1912 (78) | 378 (60) |
| (Unknown) | 3680 | 424 | 2859 | 397 |
| Bone marrow metastasis | 1347 (38) | 315 (62) | 728 (30) | 304 (48) |
| No bone marrow metastasis | 2229 (62) | 190 (38) | 1706 (70) | 333 (52) |
| (Unknown) | 3675 | 423 | 2861 | 391 |
| High LDH ≥ 1400 U/L | 588 (16) | 278 (66) | 215 (8) | 95 (15) |
| Low LDH < 1400 U/L | 3024 (84) | 146 (34) | 2351 (92) | 527 (85) |
| (Unknown) | 3639 | 504 | 2729 | 406 |
| High ferritin ≥30 ng/mL | 2055 (58) | 298 (78) | 1433 (55) | 324 (54) |
| Low ferritin <30 ng/mL | 1507 (42) | 85 (22) | 1150 (45) | 272 (46) |
| (Unknown) | 3689 | 545 | 2712 | 432 |
| Diploid | 1350 (31) | 300 (54) | 1001 (28) | 49 (39) |
| Hyperdiploid | 2937 (69) | 253 (46) | 2608 (72) | 76 (61) |
| (Unknown) | 2964 | 375 | 1686 | 903 |
| LOH/aberration at 1p | 373 (19) | 174 (73) | 183 (11) | 16 (28) |
| No 1p LOH/aberration | 1596 (81) | 66 (27) | 1488 (89) | 42 (72) |
| (Unknown) | 5282 | 688 | 3624 | 970 |
| Aberration at 11q | 217 (15) | 16 (10) | 199 (16) | 2 (12) |
| No 11q Aberration | 1184 (85) | 138 (90) | 1032 (84) | 14 (88) |
| (Unknown) | 5850 | 774 | 4064 | 1012 |
| Any SCA | 539 (36) | 176 (83) | 346 (27) | 17 (68) |
| No SCA | 963 (64) | 36 (17) | 919 (73) | 8 (32) |
| (Unknown) | 5749 | 716 | 4030 | 1003 |
| Unfavorable histology | 1630 (36) | 477 (89) | 1018 (28) | 135 (45) |
| Favorable histology | 2882 (64) | 61 (11) | 2655 (72) | 166 (55) |
| (Unknown) | 2739 | 390 | 1622 | 727 |
| High MKI | 491 (15) | 261 (59) | 185 (7) | 45 (19) |
| Low/intermediate MKI | 2874 (85) | 181 (41) | 2506 (93) | 187 (81) |
| (Unknown) | 3886 | 486 | 2604 | 796 |
| Undiff/poorly differentiated | 3035 (85) | 477 (98) | 2316 (82) | 242 (89) |
| Differentiating | 536 (15) | 12 (2) | 493 (18) | 31 (11) |
| (Unknown) | 3680 | 439 | 2486 | 755 |
| Neuroblastoma and GNBL, nodular | 3955 (91) | 521 (99) | 3029 (90) | 405 (89) |
| Ganglioneuroblastoma; intermixed, maturing subtype, well differentiated | 397 (9) | 6 (1) | 342 (10) | 49 (11) |
| (Unknown) | 2899 | 401 | 1924 | 574 |
| Diagnosed <1999 | 3224 (44) | 415 (45) | 2107 (40) | 702 (68) |
| Diagnosed ≥1999 | 4027 (56) | 513 (55) | 3188 (60) | 326 (32) |
Abbreviations: SCA, segmental chromosomal aberration of either 1p and/or 11q; Undiff, undifferentiated.
Two-sided chi-squared test identified a significantly higher proportion of patients with “MYCN unknown” than “MYCN known” (P < .05) for: age ≥18 months; INSS stage 4; presence of bone metastasis; presence of bone marrow metastasis; presence of any SCA; unfavorable histology; and diagnosed <1999.
3.2 |. Prognostic impact of MYCN amplification varies by context
In the overall Test cohort, the hazard ratio for death for patients with MYCN-A was 6.3 (95% confidence interval 5.7-7.0, P < .001) compared to the reference group of MYCN-NA. MYCN-A remained prognostic within each clinical or biological subgroup, though the prognostic impact was not uniform across subgroups (Table 2). MYCN-A generally had the strongest adverse prognostic impact in otherwise more favorable subgroups of patients (eg, younger age or lower stage), while the prognostic impact was generally weaker among unfavorable subgroups of patients (eg, older age or INSS stage 4).
TABLE 2.
Differential effects of MYCN amplification on overall survival (OS) according to other clinical and biological variables for Test cohort
| Variable | 5-year OS for MYCN-A (%) | 5-year OS for MYCN-NA (%) | OS hazard ratio (95% CI) | P-value | Hazard ratio absolute difference Δa | |
|---|---|---|---|---|---|---|
| Overall | 37 ± 1.7 | 85 ± 0.5 | 6.3 (5.7, 7.0) | <0.0001 | NA | |
| Sex | Male | 37 ± 3 | 84 ± 1 | 6.0 (5.0, 7.2) | <.00001 | 1.3 |
| Female | 37 ± 3 | 87 ± 1 | 7.3 (6.0, 8.9) | <0.0001 | ||
| Age | ≥18 months | 34 ± 2 | 67 ± 1 | 3.0 (2.7, 3.4) | <0.0001 | 16.6 |
| <18 months | 41 ± 3 | 96 ± 0.4 | 19.6 (15.6, 24.5) | <0.0001 | ||
| INSS stage | Stage 4 | 30 ± 2 | 58 ± 1 | 2.5 (2.2, 2.8) | <0.0001 | 9.9 |
| All other stages | 56 ± 3 | 96 ± 0.4 | 12.4 (9.7, 15.8) | <0.0001 | ||
| Primary site adrenal | Adrenal | 36 ± 2 | 80 ± 1 | 5.0 (4.3, 5.8) | <0.0001 | 2.0 |
| Other sites | 44 ± 3 | 89 ± 1 | 7.0 (5.9, 8.4) | <0.0001 | ||
| Primary site thoracic | Thoracic | 61 ± 8 | 92 ± 1 | 6.1 (3.5, 10.5) | <0.0001 | 0.5 |
| Other sites | 38 ± 2 | 83 ± 1 | 5.6 (4.9, 6.3) | <0.0001 | ||
| Bone metastasis | Present | 28 ± 3 | 49 ± 2 | 2.1 (1.7, 2.5) | <0.0001 | 4.8 |
| Absent | 44 ± 3 | 89 ± 1 | 6.9 (5.6, 8.5) | <0.0001 | ||
| Bone marrow metastasis | Present | 29 ± 3 | 53 ± 2 | 2.3 (1.9, 2.7) | <0.0001 | 5.9 |
| Absent | 49 ± 4 | 92 ± 1 | 8.2 (6.3, 10.7) | <0.0001 | ||
| LDH | ≥1400 U/L | 29 ± 3 | 52 ± 4 | 1.9 (1.5, 2.4) | <0.0001 | 3.6 |
| <1400 U/L | 47 ± 4 | 86 ± 1 | 5.5 (4.3, 7.1) | <0.0001 | ||
| Ferritin | ≥30 ng/mL | 34 ± 3 | 78 ± 1 | 4.7 (3.9, 5.6) | <0.0001 | 6.6 |
| <30 ng/mL | 43 ± 6 | 92 ± 1 | 11.3 (7.9, 16.2) | <0.0001 | ||
| Ploidy | Diploid | 38 ± 3 | 77 ± 2 | 4.2 (3.5, 5.2) | <0.0001 | 4.3 |
| Hyperdiploid | 42 ± 3 | 90 ± 1 | 8.5 (6.9, 10.4) | <0.0001 | ||
| 1p aberration | Present | 39 ± 4 | 71 ± 4 | 3.0 (2.1, 4.2) | <0.0001 | 3.6 |
| Absent | 44 ± 7 | 87 ± 1 | 6.6 (4.5, 9.5) | <0.0001 | ||
| 11q aberration | Present | 43 ± 16 | 63 ± 4 | 2.3 (1.1, 4.9) | 0.03 | 8.0 |
| Absent | 41 ± 5 | 91 ± 1 | 10.3 (7.4, 14.2) | <0.0001 | ||
| SCA | Any | 39 ± 4 | 67 ± 3 | 2.7 (2.1, 3.6) | <0.0001 | 8.0 |
| No | 47 ± 10 | 92 ± 1 | 10.7 (6.1, 18.7) | <0.0001 | ||
| INPC histology | Unfavorable | 38 ± 2 | 66 ± 2 | 2.6 (2.2, 3.0) | <0.0001 | 13.2 |
| Favorable | 62 ± 6 | 97 ± 0.4 | 15.8 (9.9, 25.3) | <0.0001 | ||
| MKI | High | 38 ± 3 | 75 ± 3 | 3.7 (2.6, 5.1) | <0.0001 | 3.4 |
| Low / Intermediate | 43 ± 4 | 88 ± 1 | 7.1 (5.6, 9.0) | <0.0001 | ||
| Grade of differentiation | Undiff./poorly differentiated | 39 ± 2 | 85 ± 1 | 5.9 (5.0, 6.9) | <0.0001 | 9.5 |
| Differentiating | 48 ± 15 | 95 ± 1 | 15.4 (6.2, 38.5) | <0.0001 | ||
| Diagnostic category | Neuroblastoma and GNBL, nodular | 39 ± 2 | 86 ± 1 | 6.4 (5.5, 7.4) | <0.0001 | 2.6 |
| Otherb | 83 ± 15 | 97 ± 1 | 9.0 (1.1, 75.3) | 0.0417 | ||
| Year of diagnosis | <1999 | 30 ± 2 | 82 ± 1 | 6.0 (5.1, 6.9) | <0.0001 | 0.5 |
| ≥1999 | 43 ± 2 | 87 ± 1 | 6.5 (5.8, 7.3) | <0.0001 | ||
Note. Hazard ratios reflect outcome for patients with MYCN amplified (MYCN-A) tumors relative to patients with MYCN non-amplified (MYCN-NA) tumors (reference level).
Abbreviations: CI, confidence interval; GNBL, ganglioneuroblastoma; INPC, International Neuroblastoma Pathology Classification; SCA, segmental chromosome aberration; Undiff., undifferentiated.
The reference level for calculation of the hazard ratios is MYCN not amplified. The HR absolute difference (HRdiff) is calculated by subtracting the HR for the prognostic factor subgroup that has the better outcome from the subgroup with the worse outcome.
Ganglioneuroblastoma; intermixed, maturing subtype, well differentiated.
Age at diagnosis was the factor with the largest HRdiff (16.6; 19.6 [<18 months] vs 3.0 [≥18 months]) (Table 2). Thoracic primary site was the factor with the smallest HRdiff (0.5; 6.1 [thoracic] vs 5.6 [other sites]). These analyses were repeated in the Validation cohort and similar patterns were observed (Table S3).
We repeated this analysis focused exclusively on the subset of patients in the Test cohort with high-risk disease (Table 3). In this subgroup, MYCN-A was again significantly prognostic in each clinical or biological subgroup, though HRs were somewhat attenuated compared to the overall Test cohort. The HRdiffs were likewise attenuated compared to analyses in the overall Test cohort. These analyses were repeated in the Validation cohort and similar patterns were observed (Table S4).
TABLE 3.
Differential effects of MYCN amplification on overall survival (OS) according to other clinical, biological, and treatment variables for high-risk patients in the Test cohort
| Variable | MYCN-A n |
MYCN-NA n |
5-year OS for MYCN-A (%) | 5-year OS for MYCN-NA (%) | OS hazard ratio (95% CI) | P-value | Hazard ratio absolute difference Δa | |
|---|---|---|---|---|---|---|---|---|
| Overall | 443 | 880 | 29 ± 2.3 | 40 ± 1.8 | 1.6 (1.4, 1.8) | <0.0001 | NA | |
| Sex | Male | 189 | 375 | 33 ± 3.6 | 44 ± 2.7 | 1.6 (1.3, 2.0) | <0.0001 | 0.2 |
| Female | 150 | 254 | 28 ± 3.9 | 44 ± 3.3 | 1.8 (1.4, 2.3) | <0.0001 | ||
| Age | <18 months | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
| ≥18 months | 443 | 880 | 29 ± 2.3 | 40 ± 1.8 | 1.6 (1.4, 1.9) | <0.0001 | ||
| INSS stage | Stage 4 | 443 | 880 | 29 ± 2.3 | 40 ± 1.8 | 1.6 (1.4, 1.9) | <0.0001 | N/A |
| All other stages | N/A | N/A | N/A | N/A | N/A | N/A | ||
| Primary site adrenal | Adrenal | 227 | 382 | 27 ± 3.1 | 35 ± 2.6 | 1.5 (1.2, 1.8) | <0.0001 | 0 |
| Other sites | 151 | 359 | 34 ± 4.1 | 44 ± 2.8 | 1.5 (1.1, 1.9) | 0.0024 | ||
| Primary site thoracic | Thoracic | 18 | 77 | 41 ± 13.1 | 43 ± 6.0 | 1.1 (0.5, 2.2) | 0.8228 | 0.4 |
| Other sites | 360 | 665 | 30 ± 2.5 | 39 ± 2.1 | 1.5 (1.3, 1.8) | <0.0001 | ||
| Bone metastasis | Present | 158 | 366 | 24 ± 3.5 | 33 ± 2.6 | 1.6 (1.3, 2.0) | <0.0001 | 0 |
| Absent | 98 | 174 | 31 ± 4.8 | 42 ± 3.9 | 1.6 (1.2, 2.2) | 0.0024 | ||
| Bone marrow metastasis | Present | 209 | 437 | 24 ± 3.0 | 32 ± 2.3 | 1.6 (1.3, 1.9) | <0.0001 | 0.1 |
| Absent | 49 | 107 | 35 ± 7.0 | 54 ± 5.0 | 1.7 (1.1, 2.6) | 0.023 | ||
| LDH | ≥1400 U/L | 137 | 88 | 21 ± 3.6 | 23 ± 4.9 | 1.3 (0.9, 1.7) | 0.13 | 0.6 |
| <1400 U/L | 48 | 303 | 19 ± 6.3 | 36 ± 3.0 | 1.9 (1.3, 2.7) | 0.0009 | ||
| Ferritin | ≥30 ng/mL | 135 | 313 | 18 ± 3.4 | 33 ± 2.8 | 1.8 (1.5, 2.3) | <0.0001 | 1.1 |
| <30 ng/mL | 24 | 62 | 21 ± 9.5 | 45 ± 7.0 | 2.9 (1.6, 5.4) | 0.0006 | ||
| Ploidy | Diploid | 148 | 230 | 31 ± 4.0 | 39 ± 3.5 | 1.6 (1.3, 2.1) | 0.0003 | 0.2 |
| Hyperdiploid | 119 | 286 | 32 ± 4.5 | 40 ± 3.1 | 1.4 (1.1, 1.8) | 0.0137 | ||
| 1p aberration | Present | 86 | 57 | 28 ± 5.1 | 36 ± 7.2 | 1.6 (1.1, 2.5) | 0.0252 | 0 |
| Absent | 30 | 185 | 26 ± 9.3 | 35 ± 3.9 | 1.6 (0.99, 2.6) | 0.057 | ||
| 11q aberration | Present | 10 | 87 | 23 ± 18.5 | 34 ± 6.2 | 1.6 (0.7, 3.8) | 0.2640 | 0.1 |
| Absent | 75 | 90 | 34 ± 6.0 | 43 ± 5.8 | 1.5 (1.0, 2.3) | 0.0425 | ||
| SCA | Any | 87 | 124 | 28 ± 5.1 | 33 ± 5.0 | 1.6 (1.1, 2.2) | 0.0097 | 0.1 |
| No | 18 | 73 | 36 ± 14.4 | 47 ± 6.5 | 1.7 (0.8, 3.5) | 0.1823 | ||
| Histology | Unfavorable | 254 | 476 | 34 ± 3.2 | 44 ± 2.5 | 1.5 (1.3, 1.9) | <0.0001 | 2.9 |
| Favorable | 4 | 33 | 25 ± 21.7 | 64 ± 9.1 | 4.4 (1.2, 16.4) | 0.0288 | ||
| MKI | High | 108 | 66 | 36 ± 4.9 | 50 ± 6.4 | 1.6 (1.1, 2.5) | 0.0219 | 0.1 |
| Low/intermediate | 102 | 308 | 32 ± 4.9 | 41 ± 3.1 | 1.5 (1.1, 2.0) | 0.0038 | ||
| Grade of differentiation | Undiff./poorly differentiated | 236 | 387 | 34 ± 3.3 | 40 ± 2.7 | 1.4 (1.1, 1.7) | 0.0042 | 0.1 |
| Differentiating | 2 | 32 | 50 ± 35.4 | 63 ± 9.3 | 1.3 (0.2, 10.5) | 0.077 | ||
| High-dose chemotherapy | Yes | 192 | 458 | 30 ± 3 | 36 ± 2 | 1.5 (1.2, 1.8) | 0.0003 | 2.6 |
| No | 36 | 94 | 13 ± 6 | 57 ± 5 | 4.1 (2.6, 6.6) | <0.0001 | ||
| Year of diagnosis | <1999 | 258 | 464 | 35 ± 3 | 44 ± 3 | 1.5 (1.2, 1.9) | <0.0001 | 0.3 |
| ≥1999 | 185 | 416 | 21 ± 3 | 37 ± 2 | 1.8 (1.5, 2.2) | <0.0001 | ||
Note. Hazard ratios reflect outcome for patients with MYCN amplified tumors (MYCN-A) relative to patients with MYCN non-amplified (MYCN-NA) tumors (reference level).
Abbreviations:; CI, confidence interval; SCA, segmental chromosome aberration; Undiff. = undifferentiated.
The reference level for calculation of the hazard ratios is MYCN not amplified. The HR absolute difference (HRdiff) is calculated by subtracting the HR for the prognostic factor subgroup that has the better outcome from the subgroup with the worse outcome.
Within patients with high-risk disease, we assessed the influence of treatment on the prognostic impact of MYCN-A. We observed a differential prognostic impact of MYCN-A in patients treated with (HR = 1.5 for MYCN-A vs MYCN-NA) and without (HR = 4.1 for MYCN-A vs MYCN-NA) high-dose chemotherapy (HRdiff = 2.6). Too few patients had received GD2-directed therapy for meaningful analysis (n = 50).
3.3 |. Effect modification of the prognostic impact of MYCN-A by covariates
Compared to the univariate HR for MYCN-A of 6.3, HRs for MYCN-A were lower after controlling separately for each other factor, except in the case of 11q, controlling for which increased the MYCN-A hazard ratio to 7.3 (Table 4). The covariates with the greatest attenuation of the prognostic effect of MYCN-A included LDH (MYCN-A HR = 3.1 after controlling for LDH), stage (HR = 3.2), and SCA (HR = 3.3). Similar results were seen in the Validation cohort (Table S5). Among patients with high-risk disease in the Test cohort, the HR for MYCN was already attenuated (univariate hazard ratio of 1.6) and did not vary appreciably after controlling for each other factor (Table S6).
TABLE 4.
Univariate and bivariate Cox models of overall survival for patients in the Test cohort
| Univariate model | n |
P-value |
HR (95% CI on HR) |
|||
|---|---|---|---|---|---|---|
| MYCN: MYCN-A vs MYCN-NA | 6223 |
<0.0001 |
6.3 (5.7, 7.0) |
|||
| Bivariate models: MYCN + additional covariate | Additional covariate |
MYCN |
||||
| n | P-value | HR (95% CI on HR) | P-value | HR (95% CI on HR) | ||
| Age (≥18 months vs <18 months) | 6223 | <0.0001 | 3.7 (3.2, 4.2) | <0.0001 | 4.7 (4.2, 5.2) | |
| INSS stage (stage 4 vs others) | 6078 | <0.0001 | 6.7 (5.8, 7.7) | <0.0001 | 3.2 (2.8, 3.6) | |
| Primary site adrenal (yes vs no) | 5598 | <0.0001 | 1.5 (1.3, 1.7) | <0.0001 | 5.8 (5.1, 6.5) | |
| Primary site thoracic (no vs yes) | 5600 | <0.0001 | 2.0 (1.6, 2.5) | <0.0001 | 5.6 (5.0, 6.3) | |
| Bone metastasis (yes vs no) | 2940 | <0.0001 | 3.4 (2.9, 3.9) | <0.0001 | 3.4 (2.9, 3.9) | |
| Bone marrow metastasis (yes vs no) | 2939 | <0.0001 | 4.4 (3.8, 5.2) | <0.0001 | 3.1 (2.7, 3.6) | |
| LDH (≥1400 U/L vs < 1400 U/L) | 2990 | <0.0001 | 3.0 (2.4, 3.7) | <0.0001 | 3.1 (2.5, 3.8) | |
| Ferritin (≥30 ng/mL vs < 30 ng/mL) | 2966 | <0.0001 | 2.4 (1.9, 2.9) | <0.0001 | 5.5 (4.6, 6.5) | |
| Ploidy (diploid vs hyperdiploid) | 4162 | <0.0001 | 1.8 (1.5, 2.0) | <0.0001 | 5.9 (5.0, 6.8) | |
| 1p aberration (yes vs no) | 1911 | <0.0001 | 1.8 (1.3, 2.4) | <0.0001 | 4.3 (3.2, 5.7) | |
| 11q aberration (yes vs no) | 1385 | <0.0001 | 3.0 (2.2, 3.9) | <0.0001 | 7.3 (5.6, 9.7) | |
| SCA (yes vs no) | 1477 | <0.0001 | 3.5 (2.6, 4.7) | <0.0001 | 3.3 (2.5, 4.3) | |
| Histology (unfavorable vs favorable) | 4211 | <0.0001 | 9.6 (7.7, 12.1) | <0.0001 | 2.9 (2.5, 3.4) | |
| MKI (high vs intermediate/low) | 3133 | 0.0015 | 1.4 (1.1, 1.8) | <0.0001 | 5.8 (4.7, 7.1) | |
| Grade of differentiation (undifferentiating vs differentiating) | 3298 | <0.0001 | 2.9 (2.0, 4.4) | <0.0001 | 6.0 (5.1, 7.1) | |
SCA, segmental chromosome aberration.
In Cox models of interaction of MYCN with a covariate, each other covariate had a significant interaction with MYCN status, with the exception of thoracic primary site (Table 5). The effects of these interactions appeared to be bidirectional, with differential prognostic impact of MYCN status depending on the covariate subgroup (as also seen in Table 2) and differential prognostic impact of the covariate depending on MYCN status. HRs for the covariates were generally attenuated in the context of MYCN-A, with some prognostic factors losing statistical significance altogether. A similar pattern was observed in the Validation cohort (Table S7).
TABLE 5.
Bivariate Cox models of overall survival including interaction terms for patients in the Test cohort
| HR (95% confidence interval) |
|||||||
|---|---|---|---|---|---|---|---|
|
P-value |
HR for death for patients with MYCN-A compared to MYCN-NA, in subgroups defined by covariate |
HR for death for patients with unfavorable covariate result compared to favorable covariate result in subgroups defined by MYCN status |
|||||
| Additional covariate (reference group shown in parentheses and represents the more favorable subgroup) | Additional covariate | MYCN status | Interaction term (MYCN * covariate) | Covariate reference group (favorable) | Covariate nonreference group (unfavorable) | MYCN-A | MYCN-NA |
| Age (< 18 months) | <0.0001 | <0.0001 | <0.0001 | 22.7 (18.1, 28.4) | 2.9 (2.6, 3.3) | 1.1 (0.9, 1.3) | 8.7 (7.2, 10.5) |
| INSS stage (not stage 4) | <0.0001 | <0.0001 | <0.0001 | 12.9 (10.1, 16.5) | 2.4 (2.1, 2.7) | 2.0 (1.6, 2.5) | 10.8 (9.1, 12.9) |
| Primary site adrenal (not adrenal) | <0.0001 | <0.0001 | 0.004 | 7.1 (5.9, 8.5) | 5.0 (4.2, 5.8) | 1.2 (1.0, 1.5) | 1.7 (1.5, 2.0) |
| Primary site thoracic (yes thoracic) | <0.0001 | <0.0001 | 0.74 | 6.1 (3.5, 10.6) | 5.6 (4.9, 6.3) | 1.9 (1.1, 3.1) | 2.1 (1.6, 2.6) |
| Bone metastasis (no mets) | <0.0001 | <0.0001 | <0.0001 | 7.2 (5.8, 8.8) | 2.0 (1.7, 2.4) | 1.5 (1.2, 1.9) | 5.4 (4.5, 6.5) |
| Bone marrow metastasis (no mets) | <0.0001 | <0.0001 | <0.0001 | 8.5 (6.6, 11.1) | 2.2 (1.9, 2.6) | 1.8 (1.4, 2.3) | 6.9 (5.7, 8.4) |
| LDH (<1400 U/L) | <0.0001 | <0.0001 | <0.0001 | 5.4 (4.2, 7.0) | 1.9 (1.5, 2.5) | 1.6 (1.2, 2.1) | 4.5 (3.6, 5.7) |
| Ferritin (<30 ng/mL) | <0.0001 | <0.0001 | <0.0001 | 11.7 (8.2, 16.8) | 4.6 (3.9, 5.6) | 1.2 (0.9, 1.7) | 3.1 (2.4, 4.0) |
| Ploidy (hyperdiploid) | <0.0001 | <0.0001 | <0.0001 | 8.5 (6.9, 10.4) | 4.2 (3.4, 5.2) | 1.2 (0.9, 1.5) | 2.3 (1.9, 2.8) |
| 1p aberration (no aberration) | <0.0001 | <0.0001 | 0.0021 | 6.6 (4.6, 9.6) | 3.0 (2.1, 4.2) | 1.1 (0.8, 1.6) | 2.5 (1.8, 3.4) |
| 11q aberration (no aberration) | <0.0001 | <0.0001 | 0.0001 | 10.5 (7.6, 14.5) | 2.2 (1.1, 4.6) | 0.9 (0.4, 1.9) | 4.4 (3.2, 6.2) |
| SCA (no SCA) | <0.0001 | <0.0001 | <0.0001 | 11.3 (6.5, 19.7) | 2.7 (2.0, 3.5) | 1.1 (0.6, 1.9) | 4.6 (3.3, 6.4) |
| Histology (favorable) | <0.0001 | <0.0001 | <0.0001 | 16.8 (10.5, 26.8) | 2.5 (2.2, 3.0) | 1.9 (1.3, 3.0) | 12.8 (9.9, 16.5) |
| MKI (intermediate/low) | <0.0001 | <0.0001 | 0.0029 | 7.0 (5.5, 8.8) | 3.7 (2.6, 5.2) | 1.1 (0.9, 1.5) | 2.1 (1.5, 3.0) |
| Grade of neuroblastic differentiation (differentiating) | <0.0001 | <0.0001 | 0.049 | 14.9 (6.0, 36.8) | 5.9 (5.0, 6.9) | 1.4 (0.6, 3.1) | 3.4 (2.2, 5.3) |
Abbreviations: SCA, segmental chromosome aberration.
3.4 |. Classification trees define subgroups with differential prognostic effect of MYCN-A
In a classification tree based on HRdiff (Figure 1), age had the largest differential impact on OS given MYCN status (HRdiff = 16.6), and age subgroups formed the first branch. Within patients <18 months, the second branch was MKI (HRdiff = 61.4). Within patients ≥18 months, the second branch was grade (HRdiff = 8.6). Third and fourth branch points are further depicted in Figure 1 and revealed that patients with all of the following features had outcomes most impacted by MYCN status: <18 months, high MKI, and low ferritin (Figure S1A). In contrast, patients with all of the following features had outcomes least impacted by MYCN status: ≥18 months, undifferentiated grade, stage 4, high ferritin (Figure S1B). Application of data from patients from the Validation cohort to these same subgroups yielded similar findings (Figure S2).
FIGURE 1.
Classification tree of the Test cohort that defines subgroups with greatest differential effect of MYCN status on hazard for death. Each node has a minimum MYCN-A or MYCN-NA subgroup sample size ≥5, MYCN statistically significant for overall survival, and an absolute HR difference > 1
4 |. DISCUSSION
In this comprehensive analysis of the prognostic impact of MYCN-A, we provide a refined understanding of MYCN status as a prognostic factor in neuroblastoma. While MYCN-A remained a significant prognostic factor after controlling for other clinical and biological factors, we found the prognostic value of MYCN-A to be highly context-dependent, with specific clinical and biological factors serving to augment or diminish its prognostic impact. We were able to identify groups with survival most and least impacted by MYCN status. This work complements prior work assessing the interaction between other prognostic factors and the incidence of MYCN-A. While these previous studies show MYCN-A has differential associations with other prognostic factors,5–12 the current analysis provides a deeper understanding of the prognostic value of MYCN-A that may guide clinicians as they assess an individual patient’s prognosis.
The feature demonstrating the greatest absolute difference in OS relative to MYCN-A was age. The difference in the MYCN HRs for OS for patients <18 months compared to ≥18 months was 16.6, with a strong negative adverse prognostic effect of MYCN-A in young patients. This finding is supported by previous work demonstrating patients <12 months with MYCN-A have a 2-year OS of ~30%.22 Initial studies into this age-dependent MYCN-A effect, in which age was dichotomized at 12 months, hypothesized a transient differentiation period for neuroblasts.12 More recent data demonstrate a nonlinear relationship, as MYCN-A incidence peaks between 15 and 20 months of age.6 The available data from this and previous studies highlight the complex interplay between age, overall prognosis, prognostic impact of MYCN-A, and incidence of MYCN-A all taking place in this transition period between infancy and toddlerhood.
Our classification tree examined the impact of MYCN status in subgroups, identifying specific subgroups for which MYCN-A is most impactful. As MYCN-A tumors are inherently higher risk tumors, most patients with MYCN-A are subject to the intense and prolonged standard of care treatment for high-risk neuroblastoma. For a subset of patients with otherwise lower risk features, our study yields evidence of the disproportionately large negative prognostic impact of MYCN-A in this setting, reinforcing the need for MYCN testing in nearly all neuroblastoma cases. Furthermore, of those patients who relapse, MYCN-A has been identified as an independent risk factor for worse postrelapse survival,23 underscoring the need to abrogate the effects of MYCN-A upfront. As MYCN directed therapies enter clinical trials, our novel description of the relative impact of MYCN-A in specific subgroups may nominate ideal patient candidates for these trials.
All of the HRs for MYCN were attenuated in the high-risk population compared with the overall cohort, yet MYCN still remained prognostic (Table 3). These findings demonstrate that MYCN-A can still have a negative prognostic impact even among patients already deemed to be high risk. Among patients with high-risk disease, the prognostic impact of MYCN status varied depending on whether patients received high-dose chemotherapy with stem cell rescue, with a greater negative impact of MYCN-A for patients who did not receive this therapy. While we are unable to fully explain this result, we postulate two potential explanations. First, it is possible that the group of patients who did not undergo transplant may have had otherwise more favorable features, consistent with our findings that MYCN has an outsized prognostic impact in more favorable groups. Second, it is possible that the diminished impact of MYCN-A in patients treated with high-dose chemotherapy may reflect a role for increased chemotherapy dose intensity in overcoming the adverse prognostic impact of MYCN-A.
Moving forward, as newer agents intended to target Myc family proteins are developed, our results suggest potential subgroups of particular interest for these approaches. Examples of potential strategies to target MYCN and downstream dependencies include bromodomain inhibition, dual HDAC/PI3K inhibition, MDM2 inhibition, and aurora A kinase inhibition.24–28 Whether patients with MYCN-A as their sole unfavorable prognostic factor will be more likely to benefit from these strategies will require further study.
Although this study was undertaken using the largest cohort available, the strength of our cohort size may be mitigated by different INRG member groups using divergent, though validated, techniques to identify MYCN status and also different treatment strategies.10,29,30 Given the size of the cohort, we were not able to centrally confirm MYCN status. We note as well that MYCN status was not missing at random, with patients missing MYCN status more likely to have unfavorable features and to be diagnosed prior to 1999 (Table 1). In completing our classification tree, given the novelty of this method, we made an a priori (but arbitrary) decision to not proceed with a branch if subgroup sample size was <5 subjects. As such, a larger sample size may have been able to provide evidence for additional splits. Our analysis does not account for different treatment approaches, including observation for some patients. However, our subanalysis focused on patients with high-risk disease mitigates this limitation as nearly all such patients would have been treated with intensive therapy.
In summary, this study represents a novel approach in demonstrating the context dependence of the prognostic impact of MYCN-A. Though it remains significant in all settings, we demonstrate the degree of impact is not uniform, but greatly dependent on the context of clinical and biological features. The differences in the degree to which MYCN-A adversely impacts prognosis can enable providers to more accurately weigh MYCN status when assessing an individual patient’s risk of treatment failure. Moreover, we have identified subgroups for which MYCN status has little impact, which may have implications in settings in which MYCN status cannot be assessed. Finally, as new agents targeting Myc family proteins mature, our results indicate specific subgroups of patients in which it would be most crucial to abrogate the effects of MYCN-A, thus identifying the best patient candidates for this class of new therapies.
Supplementary Material
ACKNOWLEDGMENTS
The International Neuroblastoma Risk Group database is supported in part by the William Guy Forbeck Research Foundation, the Little Heroes Pediatric Cancer Research Fund, the Children’s Neuroblastoma Cancer Foundation, the Neuroblastoma Children’s Cancer Foundation, and the Super Jake Foundation. Data included in the International Neuroblastoma Risk Group database were provided by the Children’s Oncology Group, the Pediatric Oncology Group, the Children’s Cancer Study Group, the German Gesellschaft fuer Paediatrische Onkologie und Haematologie, the European Neuroblastoma Study Group, the International Society of Paediatric Oncology Europe Neuroblastoma Group, the Japanese Advanced Neuroblastoma Study Group, the Japanese Infantile Neuroblastoma Cooperative Study Group, the Spanish Neuroblastoma Group, and the Italian Neuroblastoma Group. This study was supported in part by Alex’s Lemonade Stand Foundation (KKM, SGD, and WBL), Little Heroes Pediatric Cancer Research Foundation (WBL), the Audrey Evans Endowed Chair (to Garrett M. Brodeur), and the Mildred V. Strouss Chair (to Katherine K. Matthay). Matthias Fischer reports grants from the German Ministry of Science and Education as part of the e:Med initiative (01ZX1603A and 01ZX1607D) and from Foerdergesellschaft Kinderkrebs-Neuroblastom-Forschung eV during the conduct of the study. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the funding sources listed here.
Funding information
Alex’s Lemonade Stand Foundation; Little Heroes Pediatric Cancer Research Foundation; Audrey Evans Endowed Chair; Mildred V. Strouss Chair; German Ministry of Science and Education, Grant Number: 01ZX1603A and 01ZX1607D; Foerdergesellschaft Kinderkrebs-Neuroblastom-Forschung eV.
Abbreviations:
- HR
hazard ratio
- INRG
International Neuroblastoma Risk Group
- INSS
International Neuroblastoma Staging System
- LDH
lactate dehydrogenase
- LOH
loss of heterozygosity
- MKI
mitosis karrhyohexis index
- MYCN-A
MYCN amplification
- OS
overall survival
- SCA
segmental chromosomal aberration
Footnotes
CONFLICT OF INTEREST
The authors declare that there is no conflict of interest.
DATA AVAILABILITY
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.
REFERENCES
- 1.Brodeur GM. Neuroblastoma: biological insights into a clinical enigma. Nat Rev Cancer. 2003;3(3):203–216. [DOI] [PubMed] [Google Scholar]
- 2.Schwab M, Alitalo K, Klempnauer KH, et al. Amplified DNA with limited homology to myc cellular oncogene is shared by human neuroblastoma cell lines and a neuroblastoma tumour. Nature. 1983;305(5931):245–248. [DOI] [PubMed] [Google Scholar]
- 3.Brodeur GM, Seeger RC, Schwab M, et al. Amplification of N-myc in untreated human neuroblastomas correlates with advanced disease stage. Science. 1984;224(4653):1121–1124. [DOI] [PubMed] [Google Scholar]
- 4.Seeger RC, Brodeur GM, Sather H, et al. Association of multiple copies of the N-myc oncogene with rapid progression of neuroblastomas. N Engl J Med. 1985;313(18):1111–1116. [DOI] [PubMed] [Google Scholar]
- 5.Cohn SL, Pearson AD, London WB, et al. The International Neuroblastoma Risk Group (INRG) classification system: an INRG Task Force report. J Clin Oncol. 2009;27(2):289–297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Thompson D, Vo KT, London WB, et al. Identification of patient subgroups with markedly disparate rates of MYCN amplification in neuroblastoma: a report from the International Neuroblastoma Risk Group project. Cancer. 2016;122(6):935–945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cohn SL, Tweddle DA. MYCN amplification remains prognostically strong 20 years after its “clinical debut.” Eur J Cancer. 2004;40(18):2639–2642. [DOI] [PubMed] [Google Scholar]
- 8.Tonini GP, Boni L, Pession A, et al. MYCN oncogene amplification in neuroblastoma is associated with worse prognosis, except in stage 4s: the Italian experience with 295 children. J Clin Oncol. 1997;15(1):85–93. [DOI] [PubMed] [Google Scholar]
- 9.Goto S, Umehara S, Gerbing RB, et al. Histopathology (International Neuroblastoma Pathology Classification) and MYCN status in patients with peripheral neuroblastic tumors: a report from the Children’s Cancer Group. Cancer. 2001;92(10):2699–2708. [DOI] [PubMed] [Google Scholar]
- 10.Ambros PF, Ambros IM, Brodeur GM, et al. International consensus for neuroblastoma molecular diagnostics: report from the International Neuroblastoma Risk Group (INRG) Biology Committee. Br J Cancer. 2009;100(9):1471–1482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Di Cataldo A, Dau D, Conte M, et al. Diagnostic and prognostic markers in infants with disseminated neuroblastoma: a retrospective analysis from the Italian Cooperative Group for Neuroblastoma. Med Sci Monit. 2009;15(1):MT11–MT18. [PubMed] [Google Scholar]
- 12.Sansone R, Strigini P, Badiali M, et al. Age-dependent prognostic significance of N-myc amplification in neuroblastoma. The Italian experience. Cancer Genet Cytogenet. 1991;54(2):253–257. [DOI] [PubMed] [Google Scholar]
- 13.Mora J, Gerald WL, Quin J, Cheung N-K. Evolving significance of prognostic markers associated with treatment improvement in patients with stage 4 neuroblastoma. Cancer. 2002;94:2756–2765. [DOI] [PubMed] [Google Scholar]
- 14.Morgenstern DA, Potschger U, Moreno L, et al. Risk stratification of high-risk metastatic neuroblastoma: a report from the HR-NBL-1/SIOPEN study. Pediatr Blood Cancer. 2018;65(11):e27363. [DOI] [PubMed] [Google Scholar]
- 15.Ladenstein R, Potschger U, Pearson ADJ, et al. Busulfan and melphalan versus carboplatin, etoposide, and melphalan as high-dose chemotherapy for high-risk neuroblastoma (HR-NBL1/SIOPEN): an international, randomised, multi-arm, open-label, phase 3 trial. Lancet Oncol. 2017;18(4):500–514. [DOI] [PubMed] [Google Scholar]
- 16.Brodeur GM, Pritchard J, Berthold F, et al. Revisions of the international criteria for neuroblastoma diagnosis, staging, and response to treatment. J Clin Oncol. 1993;11(8):1466–1477. [DOI] [PubMed] [Google Scholar]
- 17.Lucas Moreno DG, Morgenstern Danie, Pasqualini Claudi, et al. Predicting “early” relapse/progression/death in children with INRGSS Stage M neuroblastoma using clinical and biologica factors: an INRG database analysis [abstract 119]. Adv Neuroblastoma Research Congress 2016;2016. [Google Scholar]
- 18.Attiyeh EF, London WB, Mosse YP, et al. Chromosome 1p and 11q deletions and outcome in neuroblastoma. N Engl J Med. 2005;353(21):2243–2253. [DOI] [PubMed] [Google Scholar]
- 19.Shimada H, Ambros IM, Dehner LP, et al. The International neuroblastoma pathology classification (the Shimada system). Cancer. 1999;86(2):364–372. [PubMed] [Google Scholar]
- 20.Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):25. [Google Scholar]
- 21.Greenwood M. The Errors of Sampling of the Survivorship Table. London: His Majesty’s Stationery Office; 1926. [Google Scholar]
- 22.Canete A, Gerrard M, Rubie H, et al. Poor survival for infants with MYCN-amplified metastatic neuroblastoma despite intensified treatment: the International Society of Paediatric Oncology European Neuroblastoma Experience. J Clin Oncol. 2009;27(7):1014–1019. [DOI] [PubMed] [Google Scholar]
- 23.Basta NO, Halliday GC, Makin G, et al. Factors associated with recurrence and survival length following relapse in patients with neuroblastoma. Br J Cancer. 2016;115(9):1048–1057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Puissant A, Frumm SM, Alexe G, et al. Targeting MYCN in neuroblastoma by BET bromodomain inhibition. Cancer Discov. 2013;3(3):308–323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sun K, Atoyan R, Borek MA, et al. Dual HDAC and PI3K inhibitor CUDC-907 downregulates MYC and suppresses growth of MYC-dependent cancers. Mol Cancer Ther. 2017;16(2):285–299. [DOI] [PubMed] [Google Scholar]
- 26.Petroni M, Veschi V, Gulino A, et al. Molecular mechanisms of MYCN-dependent apoptosis and the MDM2-p53 pathway: an Achille’s heel to be exploited for the therapy of MYCN-amplified neuroblastoma. Front Oncol. 2012;2:141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Richards MW, Burgess SG, Poon E, et al. Structural basis of N-Myc binding by Aurora-A and its destabilization by kinase inhibitors. Proc Natl Acad Sci US A. 2016;113(48):13726–13731. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gustafson WC, Meyerowitz JG, Nekritz EA, et al. Drugging MYCN through an allosteric transition in Aurora kinase A. Cancer Cell. 2014;26(3):414–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shapiro DN, Valentine MB, Rowe ST, et al. Detection of N-myc gene amplification by fluorescence in situ hybridization. Diagnostic utility for neuroblastoma. Am J Pathol. 1993;142(5):1339–1346. [PMC free article] [PubMed] [Google Scholar]
- 30.Crabbe DC, Peters J, Seeger RC. Rapid detection of MYCN gene amplification in neuroblastomas using the polymerase chain reaction. Diagn Mol Pathol. 1992;1(4):229–234. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.