Survival of patients with neuroblastoma before versus after reduction of therapy due to the change in age cut-off from 12 to 18 months in Children’s Oncology Group (COG) risk stratification

10013

Background: In 2006, COG reclassified subgroups of toddlers diagnosed with neuroblastoma from high-(HR) to intermediate-risk (IR), when the age cut-off for increased risk was raised from 365 days (12 mo) to 547 days (18 mo) (London, J Clin Onc 2005). The aim of this retrospective study was to determine if excellent outcome was maintained after a reduction of therapy. Methods: Children < 3 yrs old at diagnosis, enrolled on a COG biology study from 1990-2018, were eligible (n = 9,189). Therapy was reduced for two “cohorts of interest” based on the age cut-off change: 365-546 days old with INSS stage 4, MYCN not amplified, favorable INPC, hyperdiploid tumors (12-18mo/Stage4/FavBiology); and, 365-546 days old with INSS stage 3 tumors with non-amplified MYCN and unfavorable INPC (12-18mo/Stage3/NotAmp/UnfavINPC). Log rank tests compared event-free (EFS) and overall survival (OS) curves. Results: In the cohorts of interest, patient (pt) characteristics were similar for ≤2006 vs > 2006. For 12-18mo/Stage4/FavBiology, 5-year EFS and OS (± std error) before (≤2006; n = 40) versus after (> 2006; n = 55) the reduction in therapy were similar: 89±5.1% versus 87±4.6% (p = 0.7), and 89±5.1% versus 94±3.2% (p = 0.4), respectively (Table). For 12-18mo/Stage3/NotAmp/UnfavINPC, the 5-year EFS and OS were both 100%, before (n = 6) and after (n = 4) 2006. For the combined cohorts for ≤2006 versus > 2006, EFS and OS were 91±4.4% versus 88±4.3% (p = 0.6), and 91±4.5% versus 95±2.9% (p = 0.5), respectively. Within HR pts diagnosed ≤2006, EFS/OS were 91±4.4%/91±4.5% for (12-18mo/Stage4/FavBiology plus 12-18mo/Stage3/NotAmp/UnfavINPC) vs 38±1.3%/43±1.3% for all other HR pts < 3 yrs old (p < 0.0001; Table). Within IR pts diagnosed > 2006, EFS/OS were 88±4.3%/95±2.9% for (12-18mo/Stage4/FavBiology plus 12-18mo/Stage3/NotAmp/UnfavINPC) vs 88±0.9%/95±0.6% for all other IR pts < 3 yrs old (p = 0.9). Conclusions: Our 19 year study demonstrates that excellent outcome is maintained among toddlers with Stage4/FavBiology and Stage3/Not Amp/Unfav INPC neuroblastoma with a significant reduction of therapy from high- to intermediate-risk treatment. Importantly, pts were likely spared acute toxicity and late effects known to be associated with HR therapy. Efforts to identify additional pt cohorts who may not require HR therapy to achieve long-term survival are critical to improve the long-term health of children with neuroblastoma. [Table: see text]

Late morbidity and mortality among survivors of neuroblastoma treated with contemporary therapy: A report from the Childhood Cancer Survivor Study

10044

Background: Survival rates for neuroblastoma vary widely based on risk group. Therapies have evolved over the past four decades to de-intensify treatment for individuals with low/intermediate risk disease and intensify therapy for those with high risk disease. Risk stratification is predicted to result in differential outcomes in late morbidity and mortality; the magnitude of these differences has not been well studied. Methods: We evaluated late mortality, subsequent malignant neoplasms (SMN) and chronic health conditions (CHC) graded according to CTCAE v4.03 among 491 5-year CCSS survivors of neuroblastoma diagnosed 1987-1999 at ≥1 year of age. Using age, stage at diagnosis, and treatment, survivors were classified into risk groups (low [n=182]; intermediate [n=70]; high [n=239]). Standardized mortality ratios (SMR) and standardized incidence ratios (SIR) of SMN were calculated using rates from NCHS and SEER, respectively. Cox regression models estimated hazard ratios (HR) and 95% confidence intervals (CI) for CHC compared to 1,029 CCSS siblings. Results: Among survivors (48% male; median age 22 years, range 7-42; median follow-up 19 years, range 5-29), 80.4% with low risk disease were treated with surgery alone, while 77.8% with high risk disease received surgery, radiation, chemotherapy ± transplant. The 15-year cumulative incidence of all-cause mortality was 9.2% (CI: 7.1-11.4), with a recurrence-related mortality of 7.3% (CI: 5.3-9.3) and SMN-related mortality of 0.3% (CI: 0-0.7). All-cause mortality was significantly higher in all risk groups: (low, SMR=5.8 [CI: 2.6-13.0]; intermediate, SMR=5.7 [CI: 1.4-23.5]; high, SMR=38.6 [CI: 27.9-53.5]). The risk of SMN was elevated among high risk survivors (SIR=25.1, CI: 16.7-37.6), but did not differ from the US population for survivors of low or intermediate risk disease. Table describes the HR of CHCs (grades 1-5 and 3-5) in survivors, by risk group, as compared with siblings, as well as categories of CHCs for which survivors were at increased risk. Conclusions: Long-term survivors of neuroblastoma have a high risk of late morbidity and mortality; risk is particularly pronounced among survivors of high risk disease. Vigilant lifelong medical surveillance will be required for this relatively young population as they age.[Table: see text]

Racial and ethnic disparities in risk and survival in children with neuroblastoma: An updated analysis

10036

Background: Biologic and socioeconomic factors contribute to health disparities among patients with pediatric cancer. In an analysis of Children’s Oncology Group (COG) neuroblastoma (NBL) patients (pts) diagnosed between 2001-2009, non-Hispanic Black (Black) pts were previously shown to have a higher prevalence of high-risk disease and worse event-free survival (EFS) compared to non-Hispanic White (White) pts. Here, we analyzed data in the International Neuroblastoma Risk Group Data Commons (INRGdc) to validate these findings. Methods: Three-year EFS and overall survival (OS) of COG pts diagnosed between 2001-2009 (Cohort 1; n = 4,358) and 2010-2016 (Cohort 2; n = 3,689) in the INRGdc with known race and ethnicity were estimated by the Kaplan-Meier method. Cox proportional hazards regression models were used to evaluate differences in EFS and OS between racial/ethnic groups. The association of clinical characteristics and tumor biomarkers with racial/ethnic groups were analyzed using Chi-square tests. Results: The distribution of race/ethnicity for pts in Cohort 1 and Cohort 2 was as follows, respectively: White: 72% (n = 3,136) and 70% (n = 2,575); Black: 11.4% (n = 495) and 10.7% (n = 397); Hispanic: 12.2% (n = 532) and 14.1% (n = 522); Asian and Hawaiian: 4% (n = 178) and 4.6% (n = 172); Native American: 0.4% (n = 17) and 0.6% (n = 23). In both cohorts, a higher proportion of Black pts had INSS stage 4 disease, age ≥ 18mo, and unfavorable histology tumors when compared to White pts (Cohort 1: p = 0.003; p < 0.001; p < 0.001, respectively vs Cohort 2: p = 0.014; p < 0.001; p < 0.001, respectively). No significant differences in the proportion of pts with MYCN amplified or diploid tumors were detected between Black and White pts in either cohort. Black pts had a higher prevalence of high-risk disease compared to White pts in both Cohorts 1 and 2 (p < 0.001 and p < 0.001, respectively). Among all pts in Cohort 1, EFS was 73% and OS was 83%. In Cohort 1, Black pts had worse EFS (68% vs 73%; HR = 1.31, 95%CI 1.11-1.55, p = 0.002) and OS (78% vs 84%; HR = 1.41, 95% CI 1.16-1.70, p = 0.001) compared to White pts. Among all pts in Cohort 2, EFS was 81% and OS was 88%. Black pts in Cohort 2 also had worse EFS compared to White pts (76% vs 82%; HR = 1.35, 95% CI = 1.03-1.76, p = 0.027), although no significant difference in OS was observed (p = 0.21). In analyses restricted to high-risk pts, no statistically significant difference in EFS and OS in Black vs White pts was detected in either cohort. Conclusions: In the modern treatment era, Black NBL pts continue to have a higher prevalence of high-risk disease and inferior 3-year EFS compared to White pts. The lack of significant difference in survival among high-risk NBL pts by race suggests that Black and White pts are receiving comparable treatments and responding similarly. The socioeconomic and/or genomic factors contributing to the higher proportion of Black pts with high-risk disease requires further investigation.

Predicting response to chemotherapy in neuroblastoma using deep learning: A report from the International Neuroblastoma Risk Group

10039

Background: Metaiodobenzylguanidine (MIBG) scans are a radionucleotide imaging modality used to evaluate neuroblastoma stage at diagnosis and also determine disease response following therapy. Curie scoring is used to semi-quantitatively assess disease burden from an MIBG scan on a scale from none (0) to widespread throughout the body (30). While a Curie score ≤2 after six cycles of induction chemotherapy has been shown to be prognostic of outcome, there is no established correlation between diagnostic Curie score and outcome. Deep learning models, such as convolutional neural networks (CNN), have been shown to learn generalizable patterns within images for successful classification of metastases and detection of multiple adult cancers. We hypothesized a CNN could be developed to predict response to induction chemotherapy, a proxy for outcome, using diagnostic MIBG scans. Methods: DICOM MIBG scans and associated clinical data from a Children’s Oncology Group (COG) pilot study for children diagnosed with high-risk neuroblastoma (ANBL12P1; NCT1798004) were deidentified and linked to clinical data by the Pediatric Cancer Data Commons and obtained from the International Neuroblastoma Risk Group Data Commons. Patients were defined as having a poor response to induction chemotherapy if their Curie score after four cycles of induction chemotherapy was ≥2. An independent external validation cohort was comprised of 29 images from 26 high-risk patients treated at the University of Chicago with clinically-annotated diagnostic and post-cycle six induction DICOM MIBG scans. The CNN was trained using 2D whole body MIBG scans obtained at diagnosis. We developed the CNN using a transfer learning approach using the Xception architecture as the base layer. Hyperparameter optimization was performed using an 80%-20% train-validation strategy. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC). Results: Among 146 patients with high-risk neuroblastoma enrolled on ANBL12P1, 104 had available diagnostic and end-induction MIBG scans. There were no differences in clinical or biological characteristics between included and excluded patients. The base model CNN was able to predict which patients had a poor response to induction chemotherapy with an AUROC of 0.72 in the validation set from the ANBL12P1 cohort. Additionally, the CNN was able to predict patient response to therapy with an AUROC of 0.64 in an independent external dataset from University of Chicago. Conclusions: Our study suggests it is feasible to apply machine learning of diagnostic MIBG scans to predict response to chemotherapy for high-risk neuroblastoma patients. Given these promising results, further work to improve AUROC and performance within larger datasets is ongoing.

Outcome in patients with refractory high-risk neuroblastoma

10537

Background: Outcome for high-risk neuroblastoma (HRNBL) patients (pts) with refractory disease at end of induction (EOI) is poor. The impact of therapies such as I-131-MIBG or irinotecan/temozolomide/dinutuximab (I/T/DIN) prior to autologous stem cell transplant (ASCT) on outcome is unknown. Methods: A multi-center, retrospective study of HRNBL pts diagnosed between 2008-2018 with refractory disease at EOI was conducted. Demographics, tumor biology, treatment response, and outcomes were abstracted. 3-year (yr) EFS and OS from time of diagnosis were estimated by the Kaplan-Meier method. Results: 3-yr EFS and OS were 54% and 79% for the 136 pts analyzed. 91 pts received no additional therapy prior to ASCT (Cohort 1); 32 pts received post-induction therapy prior to ASCT (Cohort 2); and 13 pts did not undergo ASCT (Cohort 3). The prevalence of metastatic disease in Cohort 1, 2, and 3 was 65%, 97%, and 85%. 3-yr EFS and OS were not statistically different between Cohort 1 (3-yr EFS and OS; 62% and 81%) and Cohort 2 [3-yr EFS and OS; 49% (p = 0.48) and 82% (p = 0.19)]. Outcome for Cohort 3 pts was significantly worse than Cohort 1 [3-yr EFS: 15% vs. 62% (p < .001); and 3-yr OS: 48% vs. 81% (p = 0.003)] and Cohort 2 [3-yr EFS: 15% vs. 49% (p < .001); and 3-yr OS 48% vs. 82% (p = 0.035)]. For Cohort 2 pts with metastatic disease, post-induction therapy included I/T/DIN (n = 12), MIBG (n = 16), MIBG plus I/T/DIN (n = 1), and other (n = 2). Metastatic disease response was observed in 10/12 (83%) pts who received I/T/DIN and 9/16 (56%) who received MIBG. MIBG plus I/T/DIN (n = 1) or MIBG with chemotherapy (n = 1) also induced response. Among the 21 pts with metastatic disease response, 3-yr EFS and OS were 69% and 94%; significantly better than Cohort 2 patients who did not respond to post-induction therapy [3-yr EFS and OS: 11% (p = 0.016) and 66% (p = 0.2)]. 6 Cohort 2 pts achieved a complete response (CR) in metastatic sites following I/T/DIN (n = 5) or MIBG (n = 1), and all are alive without relapse with median follow-up of 3.4 years (range 2.7-8.1). The single Cohort 3 patient who achieved a metastatic CR with I/T/DIN and did not undergo ASCT remains disease-free 2.4 years from diagnosis. Conclusions: Patient characteristics differed in the 3 Cohorts, reflecting the influence of refractory disease on treatment decisions. For Cohort 2 pts, outcome was better for those with metastatic disease at EOI who responded to post-induction therapy compared to those who did not. Pts who achieved a metastatic CR of refractory disease had excellent survival. Prospective studies testing the efficacy of I/T/DIN in pts with refractory metastatic disease at EOI are warranted.

Outcomes and toxicities in patients (pts) non-randomly assigned to immunotherapy Children’s Oncology Group (COG) ANBL0032

10523

Background: Immunotherapy with the anti-GD2 antibody dinutuximab plus sargramostim (GM-CSF), aldesleukin (IL-2) and isotretinoin following consolidation therapy improved outcome for high-risk neuroblastoma (HRNBL) pts enrolled on COG ANBL0032. Randomization was halted in 2009; subsequent pts were non-randomly assigned to immunotherapy. Toxicities and survival were evaluated. Methods: HRNBL pts < 31 years old with a pre-autologous stem cell transplant (ASCT) response of ≥ partial response (PR) were eligible. Demographics, INSS stage, tumor biology, 1993 INRC pre-ASCT response and toxicities were summarized using descriptive statistics. Five-year (yr) EFS and OS from time of study enrollment were estimated. Results: From 2009-2015, 1,183 pts were non-randomly assigned to immunotherapy. 96.7% (n = 1,144) were ≥18 months old and 83.1% (n = 765/921) had stage 4 disease. 45.1% (n = 363/805) of tumors with known biology were MYCN amplified, 94.5% (n = 749/793) had unfavorable histology, and 54.9% (n = 397/723) were diploid. Pre-ASCT, 352 (29.8%) pts had complete response (CR), 418 (35.3%) had very good partial response (VGPR), and 413 (34.9%) had PR. 1,042 (88.1%) pts underwent a single and 141 (11.9%) underwent tandem ASCT. For the entire cohort, 5-yr EFS was 61.1±1.9% and 5-yr OS was 71.9±1.7%. 5-yr EFS and OS for pts ≥18 months of age with stage 4 disease (n = 746) were 58.4±2.3% and 71.0±2.1%. 5-yr EFS and OS were 82.3±4.8% and 86.7±4.2% among pts with stage 3 disease (n = 110). EFS but not OS was superior for those with a CR/VGPR pre-ASCT vs. PR (5-yr EFS: 64.2±2.2% vs. 55.4±3.2%, p = 0.0133; OS: 72.7±2.1% vs. 70.5±2.9%, p = 0.3811). There was a trend toward improved OS for those treated with tandem vs. single transplant (5-yr EFS: 65.9±4.3% vs. 60.4±2.1%, p = 0.1282; OS: 76.5±3.8% vs. 71.2±1.9%, p = 0.0704). Grade ≥3 toxicities ( > 10% of pts) during GM-CSF and IL-2-containing cycles, respectively, included pain (15.6/11.4%), fever (15.1/32.7%), anemia (18.9/21.7%), thrombocytopenia (13.9/17.4%), lymphopenia (12.3/16.0%), and hypokalemia (13.3/25.2%). Additional Grade ≥3 toxicities ( > 10% of pts) included hypoxia (10.1%) during GM-CSF-containing cycles, and anaphylaxis (12.0%), neutropenia (16.1%), hyponatremia (16.5%), and hypotension (13.8%) during IL-2-containing cycles. Conclusions: In this large cohort of HRNBL pts treated with immunotherapy, 5-yr EFS was 61.1%. Superior EFS was observed for pts with stage 3 disease and for those with CR/VGPR pre-ASCT. IL-2-containing cycles were associated with increased toxicity. Clinical trial information: NCT00026312.

Randomized phase II trial of MIBG versus MIBG/vincristine/irinotecan versus MIBG/vorinostat for relapsed/refractory neuroblastoma: A report from the New Approaches to Neuroblastoma Therapy Consortium

10500

Background: 131I-metaiodobenzylguanidine (MIBG) remains one of the most active agents for neuroblastoma. It is not clear if putative radiation sensitizers improve upon this activity. The primary aim of this trial was to identify the MIBG treatment regimen with highest response rate among: MIBG monotherapy (Arm A); MIBG/Vincristine/Irinotecan (Arm B); MIBG/Vorinostat (Arm C). The secondary aim was to compare toxicity across arms. Methods: We conducted a multicenter, randomized phase II trial. Patients 1-30 years with relapsed/refractory high-risk neuroblastoma were eligible with at least one MIBG-avid site and adequate autologous stem cells (ASCs). All patients received MIBG 18 mCi/kg on Day 1 and ASC on day 15. Patients on Arm A received only MIBG; patients on Arm B also received vincristine (2 mg/m2) IV on Day 0 and irinotecan (50 mg/m2) IV daily on Days 0-4; patients on Arm C also received vorinostat (180 mg/m2) orally once daily on days -1 to 12. The primary endpoint was response after one course according to NANT response criteria. The trial was designed as a pick-the-winner study with a maximum of 105 eligible and evaluable patients to ensure an 80% chance that the arm with highest response rate is selected, if that response rate is at least 15% higher than the other arms. Results: 114 patients enrolled. Three patients were ineligible and 6 eligible patients never received MIBG, leaving 105 eligible and evaluable patients (36 Arm A; 35 Arm B; and 34 Arm C; 55 boys; median age 6.5 years). 9 patients had received prior MIBG monotherapy, 65 prior irinotecan, and 7 prior vorinostat. After one course, the response rates (Partial Response or better) on Arms A, B, and C were 17% (95% CI 7-33%), 14% (5-31%), and 32% (18-51%). An additional 4, 4, and 7 patients met NANT Minor Response criteria [partial response in one disease category (e.g., bone marrow) and stable disease in other categories] on Arms A, B, and C, respectively. On Arms A, B, and C, rates of any grade 3+ non-hematologic toxicity were 19%, 49% and 32%; rates of grade 3+ diarrhea were 0%, 11%, 0%; and rates of grade 3+ febrile neutropenia were 6%, 11%, and 0%. Conclusions: The combination of vorinostat/MIBG had the highest response rate, with manageable toxicity. Vincristine and irinotecan do not improve the response rate to MIBG and are associated with increased toxicity. These data provide response rates for MIBG monotherapy in a contemporary patient population assessed with current response criteria. Clinical trial information: NCT02035137.

Effect of Tandem Autologous Stem Cell Transplant vs Single Transplant on Event-Free Survival in Patients With High-Risk Neuroblastoma: A Randomized Clinical Trial

IMPORTANCE

Induction chemotherapy followed by high-dose therapy with autologous stem cell transplant and subsequent antidisialoganglioside antibody immunotherapy is standard of care for patients with high-risk neuroblastoma, but survival rate among these patients remains low.

OBJECTIVE

To determine if tandem autologous transplant improves event-free survival (EFS) compared with single transplant.

DESIGN, SETTING, AND PARTICIPANTS

Patients were enrolled in this randomized clinical trial from November 2007 to February 2012 at 142 Children's Oncology Group centers in the United States, Canada, Switzerland, Australia, and New Zealand. A total of 652 eligible patients aged 30 years or younger with protocol-defined high-risk neuroblastoma were enrolled and 355 were randomized. The final date of follow-up was June 29, 2017, and the data analyses cut-off date was June 30, 2017.

INTERVENTIONS

Patients were randomized to receive tandem transplant with thiotepa/cyclophosphamide followed by dose-reduced carboplatin/etoposide/melphalan (n = 176) or single transplant with carboplatin/etoposide/melphalan (n = 179).

MAIN OUTCOMES AND MEASURES

The primary outcome was EFS from randomization to the occurrence of the first event (relapse, progression, secondary malignancy, or death from any cause). The study was designed to test the 1-sided hypothesis of superiority of tandem transplant compared with single transplant.

RESULTS

Among the 652 eligible patients enrolled, 297 did not undergo randomization because they were nonrandomly assigned (n = 27), ineligible for randomization (n = 62), had no therapy (n = 1), or because of physician/parent preference (n = 207). Among 355 patients randomized (median diagnosis age, 36.1 months; 152 [42.8%] female), 297 patients (83.7%) completed the study and 21 (5.9%) were lost to follow-up after completing protocol therapy. Three-year EFS from the time of randomization was 61.6% (95% CI, 54.3%-68.9%) in the tandem transplant group and 48.4% (95% CI, 41.0%-55.7%) in the single transplant group (1-sided log-rank P=.006). The median (range) duration of follow-up after randomization for 181 patients without an event was 5.6 (0.6-8.9) years. The most common significant toxicities following tandem vs single transplant were mucosal (11.7% vs 15.4%) and infectious (17.9% vs 18.3%).

CONCLUSIONS AND RELEVANCE

Among patients aged 30 years or younger with high-risk neuroblastoma, tandem transplant resulted in a significantly better EFS than single transplant. However, because of the low randomization rate, the findings may not be representative of all patients with high-risk neuroblastoma.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT00567567.

A revised Children's Oncology Group (COG) neuroblastoma risk classification system: Report from the COG biology study ANBL00B1

10012

Background: The COG risk classification system previously used the International Neuroblastoma Staging System (INSS). The pre-treatment INRG staging system (INRGSS) has been adopted internationally, requiring integration of INRGSS with known prognostic biological and clinical characteristics to evaluate outcomes and assess whether this incorporation will require revision to the established COG risk classifier. Methods: 4,037 newly diagnosed neuroblastoma patients were enrolled on COG ANBL00B1 between 2006-2014. Staging per the INSS and INRGSS was determined. Tumor biological and histologic features, including MYCN status [amplified (A) versus not amplified (NA)], ploidy, histology, and segmental chromosome aberrations (SCA) including 1p and 11q LOH, were assessed centrally. Survival analyses were performed to identify independent prognostic factors and to calculate event-free and overall survival (EFS, OS) for combinations of variables used to determine risk group assignments according to COG and INRG classification templates. Results: Using the original COG risk classifier 1,309 low (LR), 992 intermediate (IR) and 1,736 high-risk (HR) patients were identified with 5-year EFS of 91.4±2.1%, 84.3±2.9%, 45.2±3.1%, and OS of 98.1±1.0%, 94.0±1.9%, 54.1±3.0%, respectively. Outcomes based on combinations of clinical and biological prognostic factors were determined and compared for subsets of patients according to the COG (version1) and INRG risk classification systems to develop a revised COG risk classifier that incorporates the INRGSS (version 2, subset shown in table). Conclusions: Use of INRGSS requires a revision to the COG risk classifier. By combining INRGSS and presence of SCA together with age, MYCN status, ploidy, and histology to determine outcome of patients treated with modern era therapies, we developed a revised risk classification system to inform therapy and COG clinical trial eligibility. Clinical trial information: NCT00904241. [Table: see text]

Long term morbidity and mortality among survivors of infant neuroblastoma: A report from the Childhood Cancer Survivor Study (CCSS)

10051

Background: Infants with neuroblastoma typically have low-risk disease with excellent survival. Therapy has been de-intensified over time to minimize late effects, however the impact on survivors’ risk of late mortality, subsequent malignant neoplasms (SMN), and chronic health conditions (CHC) is unclear. Methods: We evaluated late mortality, SMNs and CHCs (graded according to CTCAE v4.03), overall and by diagnosis era, among 990 5-year neuroblastoma survivors diagnosed at < 1 year of age between 1970-1999. Cumulative mortality, standardized mortality ratios (SMR), and standardized incidence ratios (SIR) of SMNs were estimated using the National Death Index and SEER rates, respectively. Cox proportional hazards estimated hazard ratios (HR) and 95% confidence intervals (CI) for CHC, compared to 5,051 CCSS siblings. Results: Among survivors (48% female; median attained age: 24 years, range 6-46), there was increased treatment with surgery alone across the 1970s, 1980s and 1990s (21.5%, 35.3%, 41.1%, respectively), but decreased treatment with combination surgery + radiation (22.5%, 5.3%, 0.3%, respectively) and surgery + radiation + chemotherapy (28.7%, 14.7%, 9.3%, respectively). The 20-year cumulative mortality was 2.3% (95% CI, 1.4-3.8), primarily due to SMNs (SMRSMN= 10.0, 95% CI, 4.5-22.3). The 20-year cumulative incidence of SMN was 1.2% (95% CI, 0.3-3.2), 2.5% (95% CI, 1.3-4.4), and zero for those diagnosed in the 1970s, 1980s, and 1990s, respectively. SIR was highest for renal SMNs (SIR 12.5, 95% CI, 1.7-89.4). Compared to siblings, survivors were at increased risk for grade 1-5 CHC (HR 2.1, 95% CI, 1.9-2.3) with similar HR across eras (HR1970s= 1.9, 95% CI, 1.6-2.2; HR1980s= 2.2, 95% CI, 1.9-2.6; HR1990s= 2.0, 95% CI, 1.7-2.4). The HR of severe, disabling, life-threatening and fatal CHC (grades 3-5) decreased in more recent eras (HR1970s= 4.7, 95% CI, 3.4-6.6; HR1980s= 4.4, 95% CI, 3.2-6.2; HR1990s= 2.9, 95% CI, 2.0-4.3). Conclusions: Survivors of infant neuroblastoma remain at increased risk for late mortality, SMN, and CHCs many years after diagnosis. However, the risk of grade 3-5 CHCs has declined in more recent eras, likely reflecting de-intensification of therapy.

Phase 1/2 KEYNOTE-051 study of pembrolizumab (pembro) in pediatric patients (pts) with advanced melanoma or a PD-L1+ advanced, relapsed, or refractory solid tumor or lymphoma

10525

Background: In phase 1 of the KEYNOTE-051 study (NCT02332668), the 2 mg/kg Q3W approved adult dose of pembro was also determined to be the pediatric recommended phase 2 dose. In phase 2, presented herein, we further evaluated this dose in pediatric pts with advanced cancer. Methods: Children aged 6 mo to < 18 y with advanced melanoma or a PD-L1+ advanced, relapsed or refractory solid tumor or lymphoma, measurable disease per RECIST v1.1, and performance score ≥50 using Lansky Play or Karnofsky scales received pembro 2 mg/kg Q3W for 35 cycles or until confirmed disease progression per immune-related RECIST by investigator review, intolerable toxicity, or pt/investigator decision to discontinue. Tumor imaging was performed every 8 wk for the first 6 mo, then every 12 wk thereafter. AEs were graded by NCI CTCAE v4.0. Key efficacy endpoints were ORR, disease control rate (DCR), and PFS per RECIST v1.1 by investigator and OS. Results: Of 369 pts prescreened, 364 were evaluable for PD-L1 expression; of these, 121 (33.2%) were PD-L1+. 66 pts were enrolled; median follow-up was 2.5 mo (range, 0.2-18). As of the data cutoff (Nov 7, 2016), 23 (34.8%) pts were still on treatment. Median age was 13 y (range, 1-17), 77.3% had metastatic disease, and 34.8% had ≥3 prior lines of therapy for recurrent/metastatic disease. Primary diagnoses were non-CNS solid tumors (n = 45), CNS tumors (n = 16), and lymphoma (n = 5). 5 (7.6%) pts had grade 3-4 treatment-related AEs (TRAEs), most commonly neutropenia (n = 2). No treatment-related deaths occurred; 1 pt discontinued for a TRAE (grade 3 AST increased). 1 pt each with Hodgkin lymphoma, adrenocortical carcinoma, mesothelioma, and glioblastoma had partial response for an ORR of 6.1% (95% CI, 1.7-14.8); 7 (10.6%) pts had stable disease for a DCR of 16.7% (95% CI, 8.6-27.9). Median PFS and OS were 1.8 mo and 9.2 mo, respectively; 12-mo PFS was 10.2% and OS was 40.5%. Potential effects of pembro on the developing immune system (eg, T and B cells, vaccinated antibodies) will also be presented. Conclusions: Pembro showed low toxicity and warrants further study to determine activity in select pediatric tumors. Enrollment in KEYNOTE-051 is ongoing. Clinical trial information: NCT02332668.

A phase III randomized clinical trial (RCT) of tandem myeloablative autologous stem cell transplant (ASCT) using peripheral blood stem cell (PBSC) as consolidation therapy for high-risk neuroblastoma (HR-NB): A Children's Oncology Group (COG) study

LBA3

Background: ASCT improves event-free survival (EFS) for HR-NB. Pilot studies suggest that intensification of myeloablative therapy using tandem ASCT further improves outcome for HR-NB. We conducted a multicenter RCT comparing tandem vs. single consolidation in patients with HR-NB. Methods: Between 11/2007 and 2/2012, 652 eligible patients (pts) with newly diagnosed HR-NB received induction therapy: 6 cycles of chemotherapy including initial 2 cycles of dose-intensive cyclophosphamide/topotecan followed by PBSC collection. Randomization occurred at end induction to single ASCT with carboplatin-etoposide-melphalan (CEM) or tandem ASCT with thiotepa–cyclophosphamide ASCT followed by a modified CEM (TC:CEM). HR pts with non-MYCN amplified Stage 3 (age>18mos) or Stage 4 (age 12-18 mos) tumors were non-randomly assigned to single ASCT (CEM). EFS and overall survival (OS) were analyzed as intent-to-treat. Results: Median age at study entry was 3.1 yrs, 88% (n=574 pts) had Stage 4 disease and 38.2% (n=249 tumors) had MYCN amplification. A total of 355 pts were randomized (CEM n=179 pts; TC:CEM n=176 pts) and 27 patients were non-randomly assigned to CEM. Of randomized pts, 249 patients received post-consolidation immunotherapy on COG trials. Treatment-related mortality was 2.6% (Induction n=7 [1%]; Consolidation n=10 [2.8%; n=8 CEM, n=2 TC:CEM]). Rates of severe mucosal, infectious or liver toxicities were similar between arms. 3-year EFS and OS from diagnosis were 51.1±2.0% and 68.2±1.9%, respectively. EFS and OS for randomized cohort are shown in the Table. Conclusions: Tandem myeloablative consolidation therapy improves survival probability in patients with high-risk neuroblastoma, especially in the setting in post-consolidative immunotherapy. Clinical trial information: NCT00567567. [Table: see text]

Metastatic neuroblastoma confined to distant lymph nodes (stage 4N) to predict outcome in patients with stage 4 disease: A study from the International Neuroblastoma (NB) Risk Group (INRG) Database

10015

Background: Patients with metastatic NB typically have a poor prognosis; however case series have suggested that those with 4N disease may have improved outcomes. Methods: Retrospective analysis of data from INRG database for patients diagnosed 1990–2002. 4N patients (INSS stage 4 disease confined to distant lymph nodes) were compared to the balance of stage 4 patients (‘non-4N’), excluding those with missing metastatic site data. 5-yr estimates of overall (OS) and event-free survival (EFS) were calculated ± standard error (Kaplan-Meier method). Patient characteristics were compared by Mann-Whitney or Fisher’s exact/Chi-square tests. Results: 2,250 INSS stage 4 patients with complete data were identified, of whom 146 (6%) had 4N disease. For 4N patients, EFS and OS (5-yr: 77% ± 4%, 85% ± 3%), were significantly better than EFS and OS (5-yr: 35% ± 1%, 42% ± 1%) for non-4N stage 4 patients (p<0.0001). 4N patients were more likely to be younger (median age at diagnosis 1.2 yr vs 2.5 yr for non-4N; p<0.0001) and have tumors with favourable International Neuroblastoma Pathologic Classification (INPC) (63% vs 26%, p<0.0001), differentiating grade (21% vs 8%, p=0.006), lower MKI (p=0.0011) and non-MYCN amplified tumors (89% vs 69%, p<0.0001). Within subgroups defined by age at diagnosis and MYCN status, 4N pattern of disease remained significantly associated with improved outcomes. For patients aged ≥547 days at diagnosis and MYCN non-amplified, 5-yr EFS for 4N patients (n=42) was 63% ± 8% vs 27% ± 2% for non-4N (n=785); OS 74% ± 7% vs 38% ± 2% (both p<0.0001). Within this subgroup, favourable INPC and differentiating grade remained more frequent in the 4N vs non-4N patients (45% vs 10%, p<0.0001; 45% vs 8%, p=0.0017, respectively). Conclusions: 4N represents a subgroup of metastatic patients with better outcome than other INSS stage 4 patients. These findings indicate that the biology and response to treatment of 4N tumors differs from other stage 4 tumors, and different therapies should be considered for this cohort. Future exploration of biological factors determining pattern of metastatic spread and response to therapy is warranted.

Comparison of Taqman low density array (TLDA) five-gene assay for tumor cells in bone marrow and blood with histologic bone marrow examination and imaging for disease assessment and outcome in patients with recurrent/refractory neuroblastoma (NBL): A new approaches to neuroblastoma therapy (NANT) study

10039

Background: Accurate quantification of tumor burden in NBL patients is needed to define homogenous populations for therapy and establish response criteria that predict outcome. The 5-gene TLDA assay was developed for quantification of NBL cells in bone marrow (BM) and blood (BLD). Methods: Expression of CHGA, DCX, DDC, PHOX2B, and TH (NBL genes) was quantified with TLDA and reported as the geometric mean cycle threshold for the 5 genes (DGS=detection gene score; inversely related to tumor content, 40=negative). Sixty-three patients with recurrent/refractory NBL had TLDA performed on 107 BM and 99 BLD samples (66 paired) at 140 time points. Based on central review of reports, tumor longest diameter (LD) on CT/MRI (n=118), and number of 123I-MIBG avid sites (n=120) were recorded. Percentage of tumor cells in BM was from institutional reports of bilateral BM aspirates/biopsies (n=109). Overall response was assessed per NANT Response Criteria. Spearman rank correlation was performed. Results:TLDA detected tumor cells in 62/99(63%) BLD (average DGS=37.45) and 91/107(85%) BM samples (average DGS=33.42). 39/91(42%) with positive BM TLDA were negative by morphology. BLD and BM TLDA were correlated r = 0.6540, p< 0.0001 with stronger correlation with lower BM DGS scores. The BM and BLD DGS correlated with % BM involvement (BM r= -0.63, p<0.0001; BLD r= -0.35, p=0.0023) and number of MIBG sites (BM r=-0.34, p=0.001, BLD r=-0.51 p<0.0001) but not LD. Number of MIBG sites was also correlated with % BM involvement (r= 0.45, p<0.001) and LD (r= 0.28, p=0.0039). Analysis of 43 BM pairs demonstrated decreasing DGS correlated with overall progressive disease (r= 0.39, p=0.01). Conclusions: This TLDA assay detects NBL cells in both BM and BLD at high rates, and frequently detects tumor cells when BM morphology is negative. Quantification of tumor with DGS correlates with % BM involvement and number MIBG sites. TLDA may provide an additional parameter to delineate response in NBL. Clinical trial information: NCT01587300.

A pilot study of tandem high-dose chemotherapy with stem cell rescue as consolidation for high-risk neuroblastoma: Children's Oncology Group study ANBL00P1

Increasing treatment intensity has improved outcomes for children with neuroblastoma. We performed a pilot study in the Children’s Oncology Group (COG) to assess feasibility and toxicity of a tandem myeloablative regimen without total body irradiation (TBI) supported by autologous CD34 selected peripheral blood stem cells. Forty-one patients with high-risk neuroblastoma were enrolled; eight patients did not receive any myeloablative consolidation procedure, and seven received only one. Two patients out of 41 (4.9%) experienced transplant-related mortality. CD34 selection was discontinued after subjects were enrolled due to serious viral illness. From the time of study enrollment, the overall 3-year event-free survival (EFS) and overall survival (OS) were 44.8±9.6% and 59.2±9.2% (N=41). These results demonstrate that tandem transplantation in the cooperative group setting is feasible and support a randomized comparison of single versus tandem myeloablative consolidation with PBSC support for high-risk neuroblastoma.

Segmental chromosomal alterations have prognostic impact in neuroblastoma: a report from the INRG project

Background:

In the INRG dataset, the hypothesis that any segmental chromosomal alteration might be of prognostic impact in neuroblastoma without MYCN amplification (MNA) was tested.

Methods:

The presence of any segmental chromosomal alteration (chromosome 1p deletion, 11q deletion and/or chromosome 17q gain) defined a segmental genomic profile. Only tumours with a confirmed unaltered status for all three chromosome arms were considered as having no segmental chromosomal alterations.

Results:

Among the 8800 patients in the INRG database, a genomic type could be attributed for 505 patients without MNA: 397 cases had a segmental genomic type, whereas 108 cases had an absence of any segmental alteration. A segmental genomic type was more frequent in patients >18 months and in stage 4 disease (P<0.0001). In univariate analysis, 11q deletion, 17q gain and a segmental genomic type were associated with a poorer event-free survival (EFS) (P<0.0001, P=0.0002 and P<0.0001, respectively). In multivariate analysis modelling EFS, the parameters age, stage and a segmental genomic type were retained in the model, whereas the individual genetic markers were not (P<0.0001 and RR=2.56; P=0.0002 and RR=1.8; P=0.01 and RR=1.7, respectively).

Conclusion:

A segmental genomic profile, rather than the single genetic markers, adds prognostic information to the clinical markers age and stage in neuroblastoma patients without MNA, underlining the importance of pangenomic studies.

Genome-based outcome prediction in MYCN nonamplified high-risk neuroblastoma

9534

Background: Less than 40% of children with high-risk neuroblastoma achieve long-term survival, and at diagnosis, it is not possible to identify patients who will be cured. Microarray studies have proposed expression signatures associated with outcome within high-risk cohorts. However, integrating this technology as a clinical test has been difficult, in part due to the lack of available frozen tissue and high quality RNA. The nCounter overcomes this obstacle, using formalin-fixed paraffin embedded tissue (FFPE). Our objective is to test the correlation of a previously published “ultra high-risk” microarray gene signature developed in the MYCN nonamplified high-risk population (Asgharzadeh et al, J Natl Cancer Inst, 2006) with the gene expression signature obtained with the nCounter (NanoString Technologies) using RNA isolated from FFPE MYCN nonamplified high-risk neuroblastoma samples. Methods: FFPE tumor samples linked to clinical outcome data were obtained from 6 collaborative institutions. Tumor content of each sample was assessed morphologically. RNA was isolated using the RNeasy FFPE-kit. Customized probes corresponding to the candidate genes were designed by NanoString. Hybridization reactions were performed in duplicate using 100 ng of RNA. Positive and negative control probes and housekeeping probes were included in every reaction and were used to normalize data for differences in purification, hybridization, and capture efficiencies Results: Forty-two MYCN nonamplified high-risk neuroblastoma samples were analyzed by the nCounter. The cohort 5-year event-free survival and overall survival were 44.6 +/- 8.0% and 53.8 +/- 8.4% respectively. Highly degraded RNA (RIN ~ 1.5-2.8) was obtained. Unsupervised clustering and principle components analysis on normalized expression data showed differential expression of most of the genes with clustering of cases depending upon outcome (FDR = 0.05). Conclusions: Our results demonstrate that the nCounter can yield gene expression profiles that are similar to microarray gene signatures. Further investigation of the clinical utility the nCounter technology to prognosticate outcome in patients with high-risk neuroblastoma is warranted.

Development of an open-source, flexible framework for interinstitutional data sharing and collaboration

9583

Background: Clinical information, “-omic” datasets, and tissue samples are becoming more difficult to harmonize and manage for advanced data mining. We believe that clinical research data can be centralized and provide direct access to sample availability and associated data from a variety of information stores. Methods: We obtained a standardized set of anonymized patient data from the International Neuroblastoma Risk Group. The cohort consists of more than 11,000 children diagnosed worldwide between 1974 and 2002. The data consist of 34 metrics, such as age at diagnosis, stage of tumor, and other clinical and biological markers. We instantiated the dataset into a Postgres database, and using the Django web framework, created a data model for rapid development of tools and views and built a front-end interface for generating complex queries. To test the feasibility of accessing information on disparate and geographically distinct data samples, we have a formal agreement with the Children's Oncology Group Tumor Bank at The Research Institute at Nationwide Children's Hospital. Based on query results, we consume the Tumor Bank tissue inventory data through a web-facing application programming interface. The end-user is presented only with the number of patients who match their query search terms and for whom tissue samples are available. Results: We have completed our initial implementation and have agreements for collaboration with other international consortium groups. We have created a paradigm for statisticians to securely update and add data, and a verification system checks for internal validity and provides a report of the transaction. Our system can initiate queries and accept results in a variety of standards-compliant formats, and will be available in demonstration form by May 2012. Conclusions: Querying patient data while interrogating external sources allows researchers to observe which ancillary data and samples are available and to quickly download data or request any samples. While designed around a neuroblastoma dataset, our system can be applied to a variety of clinical scenarios and will be made available through an open-source license.

Outcome analysis of non-high-risk neuroblastoma patients enrolled on Children’s Oncology Group trials P9641 and A3961

9533

Background: Patients with non-high-risk neuroblastoma (low- and intermediate-risk) generally have an excellent event free (EFS) and overall (OS) survival with current therapy. However, within this heterogeneous patient population, there are patients that may benefit from further therapy reduction and patients that may benefit from augmented therapy. Methods: Survival tree regression analysis was performed in patients enrolled on P9641 and A3961 with OS the primary endpoint. Univariate Cox proportional hazards models determined statistically significant prognostic factors. Secondary analysis of cases with MYCN non-amplified, non-high-risk neuroblastoma classified patients by age, INSS stage, Shimada histology and genomic features to determine 5-year EFS and OS. Favorable genomics were defined as hyperdiploid tumors without 1p or 11q loss of heterozygosity (LOH). Those with LOH at 1p or 11q or with a diploid DNA index were considered unfavorable. Patients without genomic data were excluded. Results: In the survival tree analysis, ploidy and genomic features were found to be statistically significant. In the secondary analysis, patients <18 months of age with stage 2 or 3, favorable histology tumors with favorable genomic profile had 5-year EFS and OS rates of 92.9±3.9% and 100% respectively. Patients with stage 2 and 3 tumors, unfavorable histology and unfavorable genomic features had EFS of 68.4±8.0% and OS of 78.4±7.0%. Patients <12 months of age with stage 4 disease and either unfavorable histologic or genomic features had EFS of 69.1±5.4% and OS of 88.5±3.8%. Conclusions: Excellent outcomes in non-high-risk neuroblastoma patients with favorable histologic and genomic features suggest further reduction in therapy is possible in this cohort while maintaining survival and decreasing side effects. Select patients with unfavorable features have suboptimal OS and would benefit from modifications in therapy. Histologic and genomic criteria can be used to identify patients with non-high-risk neuroblastoma that may benefit from modifications in therapy. These data support the conduct of a cooperative group clinical trial to refine therapy for patients with non-high-risk disease.

Relationship of divergent ancestral genetic variation on chromosome 6p22 and racial disparities in survival in neuroblastoma

9516

Background: An increased prevalence of high-risk disease and worse outcome are observed in children with neuroblastoma who self-report as black versus white. We sought to determine whether genetic variation would explain this racial disparity. Methods: After quality control, we analyzed 511,836 germline genetic variants in 2,709 ethnically diverse children with neuroblastoma enrolled on Children’s Oncology Group study ANBL00B1 from 2001 to 2009. Genetic variation was summarized by conducting principal components analysis. The first principal component (PC1) separated patients with African ancestry from all others. PC1 was used as a continuous variable for ordinal regression with risk group and a Cox proportional hazard model of EFS. To identify genetic mechanisms for the observed disparities, we developed a method using genome-wide variation data applied to high-risk versus non-high risk samples. We identified a comprehensive list of loci with significant divergence between the ancestral populations. We then tested each such locus for association with high-risk phenotype using logistic regression with the proportion of African ancestry estimated by ADMIXTURE as covariate. Finally, top SNPs were added to multivariate models of both risk and event-free survival to determine if any top associations could abrogate observed disparities. Results: PC1 was associated with both risk (p = 0.007) and EFS (p = 0.037). We identified 72 population-divergent SNPs nominally associated with high-risk disease (p < 0.001). The risk allele for one of the top SNPs: rs9295536 (p = 9.2x10-8) was more common in the African ancestral population, was associated with high-risk phenotype and poor outcome in all patients and validated in a Caucasian-only subanalysis. In multivariate testing, this SNP abrogated the PC1 association with EFS (p = 0.18). Conclusions: A SNP with high divergence between ancestral populations on chromosome 6p22 accounts for the observed racial disparity in survival and is also a common genetic variant associated with survival in patients derived from either European or African ancestry (Bonferroni adjusted p < 0.05). Studies to elucidate the function of this SNP are underway.

Criteria for evaluation of disease extent by (123)I-metaiodobenzylguanidine scans in neuroblastoma: a report for the International Neuroblastoma Risk Group (INRG) Task Force

BACKGROUND

Neuroblastoma is an embryonic tumour of the sympathetic nervous system, metastatic in half of the patients at diagnosis, with a high preponderance of osteomedullary disease, making accurate evaluation of metastatic sites and response to therapy challenging. Metaiodobenzylguanidine (mIBG), taken into cells via the norepinephrine transporter, provides a sensitive and specific method of assessing tumour in both soft tissue and bone sites. The goal of this report was to develop consensus guidelines for the use of mIBG scans in staging, response assessment and surveillance in neuroblastoma.

METHODS

The International Neuroblastoma Risk Group (INRG) Task Force, including a multidisciplinary group in paediatric oncology of North and South America, Europe, Oceania and Asia, formed a subcommittee on metastatic disease evaluation, including expert nuclear medicine physicians and oncologists, who developed these guidelines based on their experience and the medical literature, with approval by the larger INRG Task Force.

RESULTS

Guidelines for patient preparation, radiotracer administration, techniques of scanning including timing, energy, specific views, and use of single photon emission computed tomography are included. Optimal timing of scans in relation to therapy and for surveillance is reviewed. Validated semi-quantitative scoring methods in current use are reviewed, with recommendations for use in prognosis and response evaluation.

CONCLUSIONS

Metaiodobenzylguanidine scans are the most sensitive and specific method of staging and response evaluation in neuroblastoma, particularly when used with a semi-quantitative scoring method. Use of the optimal techniques for mIBG in staging and response, including a semi-quantitative score, is essential for evaluation of the efficacy of new therapy.

Prophylaxis of venous thromboembolism in the US: improving hospital performance

Venous thromboembolism (VTE) prevention has been recognized as the most important practice for improving patient safety in hospitals. To be effective, VTE prophylaxis must be appropriately prescribed with respect to type, dose and duration. Large-scale studies of medical discharge records have highlighted low rates of appropriate thromboprophylaxis in hospitalized medical patients, especially those with cancer or severe lung disease. Lack of prophylaxis and an insufficient duration are the most common forms of inappropriate prophylaxis. Multifaceted, active, quality improvement initiatives have been developed and shown to successfully increase the appropriate prescribing of VTE prophylaxis in patients at risk. By increasing the use of appropriate VTE prophylaxis in at-risk patients, the disease burden of hospital-acquired VTE and its resulting complications can be reduced.

Consensus criteria for sensitive detection of minimal neuroblastoma cells in bone marrow, blood and stem cell preparations by immunocytology and QRT-PCR: recommendations by the International Neuroblastoma Risk Group Task Force

Disseminating disease is a predictive and prognostic indicator of poor outcome in children with neuroblastoma. Its accurate and sensitive assessment can facilitate optimal treatment decisions. The International Neuroblastoma Risk Group (INRG) Task Force has defined standardised methods for the determination of minimal disease (MD) by immunocytology (IC) and quantitative reverse transcriptase-polymerase chain reaction (QRT-PCR) using disialoganglioside G(D2) and tyrosine hydroxylase mRNA respectively. The INRG standard operating procedures (SOPs) define methods for collecting, processing and evaluating bone marrow (BM), peripheral blood (PB) and peripheral blood stem cell harvest by IC and QRT-PCR. Sampling PB and BM is recommended at diagnosis, before and after myeloablative therapy and at the end of treatment. Peripheral blood stem cell products should be analysed at the time of harvest. Performing MD detection according to INRG SOPs will enable laboratories throughout the world to compare their results and thus facilitate quality-controlled multi-centre prospective trials to assess the clinical significance of MD and minimal residual disease in heterogeneous patient groups.

SPARC enhances tumor stroma formation and prevents fibroblast activation

Tumor growth is influenced by interactions between malignant cells and the tumor stroma. Although the normal host microenvironment is nonpermissive for neoplastic progression, tumor-reactive stroma, characterized by the presence of activated fibroblasts, promotes neoplastic growth and metastasis. Secreted protein, acidic and rich in cysteine (SPARC) is a matricellular glycoprotein that is capable of inhibiting the growth of several different types of cancer. Recently, we reported that SPARC also impairs the growth of xenografts comprised of 293 cells. In this study, we show that in addition to enhancing stroma formation, SPARC prevents fibroblast activation in 293 xenografts, suggesting that the anti-cancer effects of SPARC may be due, at least in part, to the formation of tumor stroma that is not supportive of tumor growth. In vitro, 3T3 fibroblasts cocultured with SPARC-transfected 293 cells remain negative for alpha-smooth muscle actin, whereas wild-type 293 cells induce fibroblast activation. Moreover, activation of 3T3 cells and primary fibroblasts by transforming growth factor beta is blocked by SPARC treatment. We also demonstrate that SPARC significantly increases basic fibroblast growth factor-induced fibroblast migration in vitro, indicating that it may recruit host fibroblasts to the tumor stroma. Taken together, our results suggest that in addition to blocking angiogenesis, SPARC may inhibit tumor growth by promoting the assembly of stroma that is non-permissive for tumor progression.

The role of the multidrug resistance-associated protein 1 gene in neuroblastoma biology and clinical outcome

Multidrug resistance is a major obstacle to cancer treatment and leads to poor prognosis for the patient. Multidrug resistance-associated protein 1 (MRP1) can confer drug resistance in vitro and MRP1 may play a role in the development of drug resistance in several cancers including acute myeloid leukaemia, small cell lung cancer, T-cell leukaemia and neuroblastoma. The majority of patients with neuroblastoma present with widely disseminated disease at diagnosis and despite intensive treatment, the prognosis for such patients is dismal. There is increasing evidence for the involvement of the MYCN oncogene, and its down-stream target, MRP1, in the development of multidrug resistance in neuroblastoma. Given the importance of MRP1 overexpression in neuroblastoma, MRP1 inhibition may be a clinically relevant approach to improving patient outcome in this disease.

Evidence for an age cutoff greater than 365 days for neuroblastoma risk group stratification in the Children's Oncology Group

PURPOSE

In the Children's Oncology Group, risk group assignment for neuroblastoma is critical for therapeutic decisions, and patients are stratified by International Neuroblastoma Staging System stage, MYCN status, ploidy, Shimada histopathology, and diagnosis age. Age less than 365 days has been associated with favorable outcome, but recent studies suggest that older age cutoff may improve prognostic precision.

METHODS

To identify the optimal age cutoff, we retrospectively analyzed data from the Pediatric Oncology Group biology study 9047 and Children's Cancer Group studies 321p1-p4, 3881, 3891, and B973 on 3,666 patients (1986 to 2001) with documented ages and follow-up data. Twenty-seven separate analyses, one for each different age cutoff (adjusting for MYCN and stage), tested age influence on outcome. The cutoff that maximized outcome difference between younger and older patients was selected.

RESULTS

Thirty-seven percent of patients were younger than 365 days, and 64% were > or = 365 days old (4-year event-free survival [EFS] rate +/- SE: 83% +/- 1% [n = 1,339] and 45% +/- 1% [n = 2,327], respectively; P < .0001). Graphical analyses revealed the continuous nature of the prognostic contribution of age to outcome. The optimal 460-day cutoff we selected maximized the outcome difference between younger and older patients. Forty-three percent were younger than 460 days, and 57% were > or = 460 days old (4-year EFS rate +/- SE: 82% +/- 1% [n = 1,589] and 42% +/- 1% [n = 2,077], respectively; P < .0001). Using a 460-day cutoff (assuming stage 4, MYCN-amplified patients remain high-risk), 5% of patients (365 to 460 days: 4-year EFS 92% +/- 3%; n = 135) fell into a lower risk group.

CONCLUSION

The prognostic contribution of age to outcome is continuous in nature. Within clinically relevant risk stratification, statistical support exists for an age cutoff of 460 days.

p16/p14(ARF) cell cycle regulatory pathways in primary neuroblastoma: p16 expression is associated with advanced stage disease

p16 regulates the G(1)-S cell cycle transition by inhibiting the cyclin D-cyclin-dependent kinase (CDK)4/CDK6-mediated phosphorylation of retinoblastoma protein (pRb). We examined the possible derangement of the p16-CDK/cyclin D-pRb pathway in 40 primary neuroblastomas including 18 samples in the unfavorable stages (C and D) and 22 in the favorable stages (A, B, and Ds) by PCR, reverse transcription-PCR, Western blot, and immunohistochemistry and correlated the results with clinical outcome. No samples harbored alterations of the p16 gene. Interestingly, the samples in the unfavorable stages exhibited expression of p16 mRNA and protein more frequently than those in the favorable stages [mRNA, 9 of 18 (50%) versus 2 of 22 (9%), P = 0.006; protein, 5 of 16 (31%) versus 0 of 18 (0%), P = 0.013]. Alterations of the downstream components of the pathway were infrequent. pRb was deregulated in the majority of samples investigated [27 of 33 (82%), 24 with hyperphosphorylated pRb and 3 with no pRb protein]. The phosphorylation status of pRb did not correlate with p16 protein expression, suggesting that the elevated p16 protein may not be functioning properly to regulate the pathway. Among patients of all stages, p16 expression was significantly associated with a lower overall survival. There was no overexpression of MDM2, and loss of p14(ARF) expression and p53 mutation were infrequent events. Taken together, these findings suggest that up-regulated p16 expression may represent a unique feature of aggressive neuroblastoma.

Antiangiogenic therapy inhibits human neuroblastoma growth

BACKGROUND

Angiogenesis plays a crucial role in the progression and metastasis of malignant solid tumors. In many types of cancer, including neuroblastoma, high tumor vascularity correlates with poor outcome. Recently, a number of angiogenic inhibitors have been identified that had antitumor activity in preclinical studies.

PROCEDURE

To investigate the effect of the antiangiogenic agent TNP-470 on neuroblastoma tumor growth, we treated nude mice with TNP-470 after they were inoculated subcutaneously with 5 x 10(6) cells from the MYCN-amplified, human neuroblastoma cell line NBL-W-N.

RESULTS

TNP-470 did not significantly affect tumor growth when it was administered to mice with large tumors (>600 mm3). However, when TNP-470 was administered in the setting of microscopic disease 12 hr following tumor cell inoculation, treated mice had a significantly improved tumor-free survival at 12 weeks (P < 0.001), and overall survival at 45 weeks (P < 0.001), compared to control animals.

CONCLUSIONS

Our studies suggest that TNP-470 treatment may be most effective if it is administered in the setting of microscopic disease. We speculate that TNP-470 may inhibit neuroblastoma growth in children if treatment is initiated following intensive multimodality therapy, when residual disease is minimal.

Natural killer cell lymphoma: report of two pediatric cases, therapeutic options, and review of the literature

BACKGROUND

Natural killer (NK) cell lymphomas are rapidly fatal malignancies that to the authors' knowledge are rare in children. In the current study, the authors report the cases of two boys with NK cell lymphomas with refractory disease who both were salvaged with high dose chemotherapy and stem cell transplantation and compare these patients with those in the published experience.

METHODS

A comprehensive literature review was performed to identify other cases of pediatric patients with NK cell lymphomas, their treatment, and outcome.

RESULTS

One of the patients in the current study developed two recurrences and the other patient experienced early disease progression during front-line treatment. Both then were treated with high dose chemotherapy followed by stem cell rescue. At last follow-up, the patients remained free of disease at 15 months and 16 months, respectively, after transplantation (48 months and 22 months, respectively, from the time of diagnosis). In addition to the 2 patients in the current study, the authors found 13 pediatric patients reported in the literature to date. Of the 7 patients with localized (Stage I-II) disease, 5 patients (71%) were reported to be alive 1-107 months after diagnosis. Of the 6 patients with Stage IV disease, only the 2 patients who received high dose chemotherapy and stem cell rescue (33%) were alive at the time of last follow-up (at 30 months and 12 months, respectively). Including the patients reported in the current study, 9 of 15 children with NK cell lymphoma (all stages) (60%) were reported to be alive at the time of last follow-up.

CONCLUSIONS

Although pediatric NK cell lymphomas rapidly can become fatal, it appears that high dose chemotherapy followed by stem cell transplantation is effective therapy, especially in patients with advanced or resistant disease. Further follow-up is needed to determine whether this treatment approach will be curative.

Detection of MYCN gene amplification in neuroblastoma by fluorescence in situ hybridization: a pediatric oncology group study

To assess the utility of fluorescence in situ hybridization (FISH) for analysis of MYCN gene amplification in neuroblastoma, we compared this assay with Southern blot analysis using tumor specimens collected from 232 patients with presenting characteristics typical of this disease. The FISH technique identified MYCN amplification in 47 cases, compared with 39 by Southern blotting, thus increasing the total number of positive cases by 21%. The major cause of discordancy was a low fraction of tumor cells (< or =30% replacement) in clinical specimens, which prevented an accurate estimate of MYCN copy number by Southern blotting. With FISH, by contrast, it was possible to analyze multiple interphase nuclei of tumor cells, regardless of the proportion of normal peripheral blood, bone marrow, or stromal cells in clinical samples. Thus, FISH could be performed accurately with very small numbers of tumor cells from touch preparations of needle biopsies. Moreover, this procedure allowed us to discern the heterogeneous pattern of MYCN amplification that is characteristic of neuroblastoma. We conclude that FISH improves the detection of MYCN gene amplification in childhood neuroblastomas in a clinical setting, thus facilitating therapeutic decisions based on the presence or absence of this prognostically important biologic marker.

Treatment and outcome of 83 children with intraspinal neuroblastoma: the Pediatric Oncology Group experience

PURPOSE

To investigate whether the rate of neurologic recovery or the incidence of long-term sequelae differed for children with neuroblastoma (NB) initially treated with chemotherapy versus surgical decompression with laminectomy, we reviewed the Pediatric Oncology Group (POG) experience.

PATIENTS AND METHODS

A retrospective review of children diagnosed with intraspinal NB registered on POG NB Biology Protocol 9047 was performed. Survival, neurologic outcome, and orthopedic sequelae were evaluated according to age of the patient at diagnosis, stage of disease, duration and severity of neurologic symptoms, and therapeutic intervention.

RESULTS

Between May 1990 and January 1998, 83 children with intraspinal NB were entered onto the study. Five-year survival for this cohort of patients was 71% +/- 9%. Forty-three (52%) of the patients had neurologic symptoms at diagnosis. After treatment, six of 15 severely affected patients, who presented with paralysis, completely recovered neurologic function. Two of five patients with moderate deficits, consisting of paresis and bowel/bladder dysfunction, completely recovered neurologic function. Seventeen of 22 assessable children, who had mild symptoms comprised of paresis alone, fully recovered. Seven of 24 assessable patients who had undergone laminectomy developed scoliosis, whereas spinal deformities were only detected in one of 49 assessable patients managed without laminectomy (P =.001).

CONCLUSION

The frequency of complete neurologic recovery in children with intraspinal NB inversely correlated with the severity of the presenting neurologic deficits. The rate of neurologic recovery was similar for patients treated with chemotherapy compared to those managed with laminectomy. Fewer orthopedic sequelae were observed in the children managed with chemotherapy than were seen in children managed with laminectomy.

Natural killer cell lymphoma: report of two pediatric cases, therapeutic options, and review of the literature

BACKGROUND

Natural killer (NK) cell lymphomas are rapidly fatal malignancies that to the authors' knowledge are rare in children. In the current study, the authors report the cases of two boys with NK cell lymphomas with refractory disease who both were salvaged with high dose chemotherapy and stem cell transplantation and compare these patients with those in the published experience.

METHODS

A comprehensive literature review was performed to identify other cases of pediatric patients with NK cell lymphomas, their treatment, and outcome.

RESULTS

One of the patients in the current study developed two recurrences and the other patient experienced early disease progression during front-line treatment. Both then were treated with high dose chemotherapy followed by stem cell rescue. At last follow-up, the patients remained free of disease at 15 months and 16 months, respectively, after transplantation (48 months and 22 months, respectively, from the time of diagnosis). In addition to the 2 patients in the current study, the authors found 13 pediatric patients reported in the literature to date. Of the 7 patients with localized (Stage I-II) disease, 5 patients (71%) were reported to be alive 1-107 months after diagnosis. Of the 6 patients with Stage IV disease, only the 2 patients who received high dose chemotherapy and stem cell rescue (33%) were alive at the time of last follow-up (at 30 months and 12 months, respectively). Including the patients reported in the current study, 9 of 15 children with NK cell lymphoma (all stages) (60%) were reported to be alive at the time of last follow-up.

CONCLUSIONS

Although pediatric NK cell lymphomas rapidly can become fatal, it appears that high dose chemotherapy followed by stem cell transplantation is effective therapy, especially in patients with advanced or resistant disease. Further follow-up is needed to determine whether this treatment approach will be curative.

Expression of the putative tumour suppressor gene, p73, in neuroblastoma and other childhood tumours

BACKGROUND

The recently characterised p53 homologue, p73, has been mapped to chromosome 1p36, a region often found deleted in neuroblastoma. Although p73 has been implicated as a candidate tumour suppressor for neuroblastoma, mutations in the gene appear to be rare, thus suggesting other mechanisms for its aberrant behaviour.

PROCEDURE

We analysed p73 gene expression in a panel of primary neuroblastoma tumours and cell lines, as well as other tumours of childhood, using a reverse transcriptase-polymerase chain reaction assay.

RESULTS

Although low-level p73 expression was detected in 44/45 primary neuroblastoma tumours, no association with prognostic markers could be discerned. In no case was the level of p73 expression in the primary tumours as high as that observed in p73-expressing neuroblastoma cell lines. Expression of the p73 gene was also detected in 24/34 other childhood tumours.

CONCLUSION

Collectively, the data raise doubts over the role of this gene as a tumour suppressor in neuroblastoma.

Antiangiogenic therapy inhibits human neuroblastoma growth

BACKGROUND

Angiogenesis plays a crucial role in the progression and metastasis of malignant solid tumors. In many types of cancer, including neuroblastoma, high tumor vascularity correlates with poor outcome. Recently, a number of angiogenic inhibitors have been identified that had antitumor activity in preclinical studies.

PROCEDURE

To investigate the effect of the antiangiogenic agent TNP-470 on neuroblastoma tumor growth, we treated nude mice with TNP-470 after they were inoculated subcutaneously with 5 x 10(6) cells from the MYCN-amplified, human neuroblastoma cell line NBL-W-N.

RESULTS

TNP-470 did not significantly affect tumor growth when it was administered to mice with large tumors (>600 mm3). However, when TNP-470 was administered in the setting of microscopic disease 12 hr following tumor cell inoculation, treated mice had a significantly improved tumor-free survival at 12 weeks (P < 0.001), and overall survival at 45 weeks (P < 0.001), compared to control animals.

CONCLUSIONS

Our studies suggest that TNP-470 treatment may be most effective if it is administered in the setting of microscopic disease. We speculate that TNP-470 may inhibit neuroblastoma growth in children if treatment is initiated following intensive multimodality therapy, when residual disease is minimal.

Residential pesticide exposure and neuroblastoma

Neuroblastoma is the most common neoplasm in children under 1 year of age. We examined the relation between residential exposure to pesticides and neuroblastoma, using data from a case-control study of risk factors for neuroblastoma. Incident cases of neuroblastoma (N = 538) were identified through the Pediatric Oncology Group and the Children's Cancer Group. One age-matched control was identified for each case by random digit dialing. Telephone interviews with each parent collected information on residential exposure to pesticides. Pesticide use in both the home and garden were modestly associated with neuroblastoma [odds ratio (OR) = 1.6 (95% confidence interval [95% CI] = 1.0-2.3, and OR = 1.7 (95% CI = 0.9-2.1), respectively]. Compared with infants [OR = 1.0 (95% CI = 0.6-2.0)], stronger associations were found for garden pesticides in children diagnosed after 1 year of age [OR = 2.2 (95% CI = 1.3-3.6)], which suggests that pesticides may act through a mechanism more common for neuroblastomas in older children. There was no evidence of differential pesticide effects in subgroups of neuroblastoma defined by MYCN oncogene amplification or tumor stage.

Prognostic significance of EPHB6, EFNB2, and EFNB3 expressions in neuroblastoma

BACKGROUND

EPH family receptor tyrosine kinases and their ligand ephrins play pivotal roles in development. High-level expression of transcripts encoding EPHB6 receptors (EPHB6), its ligands ephrin-B2 and ephrin-B3 (EFNB2, EFNB3) is predictive of favorable disease outcome of neuroblastoma (NB). When combined with TrkA expression, the expression of EPHB6, EFNB2, or EFNB3 predicts more accurately the disease outcome than each of the four variables alone.

PROCEDURE

Cox regression and Kaplan-Meier analyses were used to assess the prognostic significance of EPHB6, EFNB2, EFNB3, and TrkA expressions in NB without MYCN amplification.

RESULTS

High-level expression of EFNB3 or TrkA predicted favorable NB outcome of NB without MYCN amplification (p < 0.03). As found in the general NB population, EPHB6, EFNB2, or EFNB3 expression in combination with TrkA expression was significantly predictive of the disease outcome of normal MYCN NB (p < 0.01).

CONCLUSIONS

EPHB6, EFNB2, and EFNB3 expressions may permit further refinement of the prognostic stratification of NB into favorable and unfavorable groups.

Schwann cell-conditioned medium inhibits angiogenesis in vitro and in vivo

BACKGROUND

Neuroblastomas are biologically heterogeneous tumors that consist of two main cell populations: neuroblastic/ganglionic cells and Schwann cells. The amount of Schwannian stroma strongly impacts prognosis. Low tumor vascularity, localized stage, and favorable outcome are associated with tumors that are Schwannian stroma-rich/stroma-dominant.

PROCEDURE

To investigate if Schwann cells play a role in inhibiting angiogenesis in neuroblastoma tumors, we examined the ability of human Schwann cell-conditioned medium to affect bFGF- and VEGF-induced endothelial cell proliferation and migration, and in vivo angiogenesis.

RESULTS

Schwann cell-conditioned medium significantly inhibited bFGF- and VEGF-induced endothelial cell proliferation and migration. This effect appears to be specific for endothelial cells as smooth muscle cell and fibroblast proliferation were not inhibited by this medium. Schwann cell-conditioned medium also inhibited in vivo angiogenesis in rat corneal assays.

CONCLUSIONS

Schwann cells produce a potent inhibitor(s) of angiogenesis that may be responsible for the low level of vascularity and more benign clinical behavior of Schwannian stroma-rich/stroma-dominant neuroblastoma tumors. Studies to identify the inhibitor(s) are ongoing.

MYCN expression is not prognostic of adverse outcome in advanced-stage neuroblastoma with nonamplified MYCN

PURPOSE

The clinical significance of MYCN expression in children with neuroblastoma (NB) remains controversial. To determine the prognostic significance of MYCN expression in the absence of MYCN amplification, we analyzed MYCN mRNA and protein expression in tumors from 69 patients.

PATIENTS AND METHODS

Sixty-nine NB tumor samples with nonamplified MYCN from patients with stage C or D disease were obtained from the Pediatric Oncology Group Neuroblastoma Tumor Bank. MYCN mRNA was analyzed using a real-time reverse transcriptase polymerase chain reaction assay, and MYCN protein was examined by Western blot analyses.

RESULTS

The estimated 5-year event-free survival (EFS) and survival (S) rates plus SE for the cohort were 57% +/- 17% and 60% +/- 16%, respectively. Infants younger than 1 year had significantly higher rates of EFS and S than children >/= 1 year of age (P =.003 and P <.001, respectively); patients with stage C disease had better outcome than those with stage D NB (P <.001); and patients with hyperdiploid tumors had better outcome than those with diploid NB (P <.001). Surprisingly, outcome was slightly better for patients with high versus low levels of MYCN mRNA expression (4-year S, 70% +/- 13% v 50% +/- 16%; P =.290), and for patients with tumors that expressed MYCN protein (4-year S, 73% +/- 19% v 53% +/- 15%, respectively; P =.171).

CONCLUSION

High levels of MYCN expression are not prognostic of adverse outcome in patients with advanced-stage NB with nonamplified MYCN. A trend associating high levels of MYCN expression with improved outcome was observed.

Schwann cell-conditioned medium inhibits angiogenesis

Neuroblastomas are biologically heterogeneous tumors that consist of two main cell populations: neuroblastic/ganglionic cells and Schwann cells. The amount of Schwannian stroma strongly impacts prognosis, and favorable outcome is associated with tumors that are Schwannian stroma rich/stroma dominant. At the present time, there is controversy regarding the origin of Schwann cells in neuroblastoma tumors. However, recent studies have suggested that the Schwann cells in mature neuroblastoma tumors may be normal cells that produce soluble substances that enhance the survival and differentiation of neuroblastoma cell lines. Previously, we reported that in neuroblastoma, high vascular index correlated with clinically aggressive disease. In contrast, tumors with favorable histology and abundant Schwannian stroma had low tumor vascularity. As a first step toward investigating whether Schwann cells also play a role in inhibiting angiogenesis in neuroblastoma tumors, we examined the ability of conditioned medium collected from normal human Schwann cells to affect basic fibroblast growth factor- and vascular endothelial growth factor-induced endothelial cell proliferation and migration and in vivo angiogenesis. In vitro angiogenesis assays were also performed with conditioned medium collected from Schwann cells derived from a Schwannian stroma-dominant neuroblastoma tumor. Our results indicate that Schwann cells derived from either adult nerve or tumor tissue produce a potent inhibitor(s) of angiogenesis. Expression studies revealed tissue inhibitor of metalloproteinase (TIMP)-2 in conditioned medium collected from both normal and tumor-derived Schwann cells. In addition, TIMP-2 was detected in the cytoplasm of Schwann cells and ganglion cells in stroma-rich/stroma-dominant neuroblastoma tumors by immunohistochemistry studies. We postulate that the low level of vascularity and more benign clinical behavior of Schwannian stroma-rich/stroma-dominant neuroblastoma tumors result from the Schwann cell production of TIMP-2 and/or other inhibitors of angiogenesis.

Synergy between 5' and 3' flanking regions of the human tyrosine hydroxylase gene ensures specific, high-level expression in neuroblastoma cells

Factors regulating tyrosine hydroxylase (TH) gene transcription are of major importance in the studies of malignant and degenerative diseases of catecholamine-synthesizing tissues. In this study, we used transient transfection of a reporter gene to show that high-level, tissue-specific TH expression was only achieved when the reporter gene was cloned between a 5' TH promoter sequence (-513-+1), and, a 3' TH gene flanking sequence (end of exon 14-+976). We also show that TH mRNA expression level is closely linked to the expression level of the proto-oncogene, MYCN in neuroblastoma tumor cell lines. Taken together our data indicate that MYCN may regulate TH expression in neuroblastoma cells, but not through binding to the 5' or 3' TH gene flanking sequences used in our experiments.

Implications of EPHB6, EFNB2, and EFNB3 expressions in human neuroblastoma

Neuroblastoma (NB) is a common pediatric tumor that exhibits a wide range of biological and clinical heterogeneity. EPH (erythropoietin-producing hepatoma amplified sequence) family receptor tyrosine kinases and ligand ephrins play pivotal roles in neural and cardiovascular development. High-level expression of transcripts encoding EPHB6 receptors (EPHB6) and its ligands ephrin-B2 and ephrin-B3 (EFNB2, EFNB3) is associated with low-stage NB (stages 1, 2, and 4S) and high TrkA expression. In this study, we showed that EFNB2 and TrkA expressions were associated with both tumor stage and age, whereas EPHB6 and EFNB3 expressions were solely associated with tumor stage, suggesting that these genes were expressed in distinct subsets of NB. Kaplan-Meier and Cox regression analyses revealed that high-level expression of EPHB6, EFNB2, and EFNB3 predicted favorable NB outcome (P<0.005), and their expression combined with TrkA expression predicted the disease outcome more accurately than each variable alone (P<0.00005). Interestingly, if any one of the four genes (EPHB6, EFNB2, EFNB3, or TrkA) was expressed at high levels in NB, the patient survival was excellent (>90%). To address whether a good disease outcome of NB was a consequence of high-level expression of a "favorable NB gene," we examined the effect of EPHB6 on NB cell lines. Transfection of EPHB6 cDNA into IMR5 and SY5Y expressing little endogenous EPHB6 resulted in inhibition of their clonogenicity in culture. Furthermore, transfection of EPHB6 suppressed the tumorigenicity of SY5Y in a mouse xenograft model, demonstrating that high-level expressions of favorable NB genes, such as EPHB6, can in fact suppress malignant phenotype of unfavorable NB.