Robison, Leslie L.

PURPOSE Radiation-associated cardiac disease is a major cause of morbidity/mortality among childhood cancer survivors. Radiation dose-response relationships for cardiac substructures and cardiac diseases remain unestablished. METHODS Using the 25,481 5-year survivors of childhood cancer treated from 1970 to 1999 in the Childhood Cancer Survivor Study, we evaluated coronary artery disease (CAD), heart failure (HF), valvular disease (VD), and arrhythmia. We reconstructed radiation doses for each survivor to the coronary arteries, chambers, valves, and whole heart. Excess relative rate (ERR) models and piecewise exponential models evaluated dose-response relationships. RESULTS The cumulative incidence 35 years from diagnosis was 3.9% (95% CI, 3.4 to 4.3) for CAD, 3.8% (95% CI, 3.4 to 4.2) for HF, 1.2% (95% CI, 1.0 to 1.5) for VD, and 1.4% (95% CI, 1.1 to 1.6) for arrhythmia. A total of 12,288 survivors (48.2%) were exposed to radiotherapy. Quadratic ERR models improved fit compared with linear ERR models for the dose-response relationship between mean whole heart and CAD, HF, and arrhythmia, suggesting a potential threshold dose; however, such departure from linearity was not observed for most cardiac substructure end point dose-response relationships. Mean doses of 5-9.9 Gy to the whole heart did not increase the risk of any cardiac diseases. Mean doses of 5-9.9 Gy to the right coronary artery (rate ratio [RR], 2.6 [95% CI, 1.6 to 4.1]) and left ventricle (RR, 2.2 [95% CI, 1.3 to 3.7]) increased risk of CAD, and to the tricuspid valve (RR, 5.5 [95% CI, 2.0 to 15.1]) and right ventricle (RR, 8.4 [95% CI, 3.7 to 19.0]) increased risk of VD. CONCLUSION Among children with cancer, there may be no threshold dose below which radiation to the cardiac substructures does not increase the risk of cardiac diseases. This emphasizes their importance in modern treatment planning.

10027 Background: Thoracic radiotherapy (RT) is a risk factor for cardiac disease among survivors of childhood cancer based on studies considering RT doses to the entire heart. Dose to specific cardiac substructures may provide more precise dose-response associations to guide RT planning. We report associations between RT dose to cardiac substructures and risk of specific cardiac outcomes. Methods: We determined the cumulative incidences of CTCAE grade 3 – 5 coronary artery disease (CAD), heart failure (HF), and valvular disease (VD) among 25,481 5-year childhood cancer survivors diagnosed 1970 – 1999 in the CCSS. Median age at diagnosis was 6.1 years (0 – 20 years) and at last follow-up 29.8 years (5.6 – 65.9 years). For the 12,228 individuals receiving RT, fields were reconstructed on an age scaled computational phantom. Mean doses to the entire heart, cardiac chambers, valves, and left main, anterior descending (LAD), circumflex, and right coronary (RCA) arteries were estimated. Adjusted piecewise exponential models (including cumulative anthracycline dose) evaluated associations between mean RT dose to each structure and outcomes. Results: At 30 years from diagnosis, the cumulative incidences of grades 3 – 5 CAD, HF, and VD were 2.3% (95% CI 2.0 – 2.5), 2.6% (95% CI 2.4 – 2.9), and 0.6% (95% CI 0.5 – 0.8) respectively. Low dose RT (mean 5 – 9.9 Gy) to the RCA (relative rate [RR] = 2.6, 95% CI 1.6 – 4.1, p < 0.001), LAD (RR = 1.9, 95% CI 1.1 – 3.3, p = 0.019), and left ventricle (RR = 2.2, 95% CI 1.3 – 3.7, p = 0.002) was associated with increased risk of CAD relative to those not exposed; mean dose of 5 – 9.9 Gy to the entire heart was not (RR = 1.1, 95% CI = 0.8 – 1.6, p = 0.59). Table shows the associated cumulative incidences of CAD for these structures. Mean RT doses of 5 – 9.9 Gy to the aortic valve (RR = 4.6, 95% CI 1.5 – 14.0, p = 0.008) and tricuspid valve (RR = 5.5, 95% CI 2.0 – 15.1, p = 0.001) were associated with risk of VD compared to those with no RT; mean heart dose 5 – 9.9 Gy was not (RR = 0.6, 95% CI = 0.2 – 1.3, p = 0.16). No cardiac structure was significantly associated with an increased risk of HF at a mean dose of 5 – 9.9 Gy. Mean RT doses below 5 Gy were not associated with increased risks of CAD, HF, or VD; doses ≥10 Gy to nearly all substructures and the entire heart were. Conclusions: Low dose RT to the coronary arteries and cardiac valves (5 – 9.9 Gy) from RT is associated with increased risk of CAD and VD, respectively. Sensitivity to low dose RT may vary across the heart and doses to specific structures may be more predictive of late cardiac disease than whole heart dose. These data can guide RT planning to prioritize substructures for avoidance.[Table: see text]

11570 Background: Intergroup Rhabdomyosarcoma Study Group (IRSG) protocols included treatment modifications, which may have ameliorated late health outcomes for rhabdomyosarcoma (RMS) survivors treated in more recent era. Methods: We evaluated chronic health conditions (CHCs) and late mortality ( > 5 years from diagnosis) among survivors treated 1970-1990 (IRSG I-III) and 1991-1999 (IRSG IV), and associations with specific treatments to identify treatment-related factors for adverse outcomes. Associations between treatments and CHCs and mortality were evaluated using Fine and Gray’s proportional hazards method accounting for competing risks. Results: 856 survivors treated 1970-90 (median diagnosis age 5.4 years [0- 20]) and 306 treated 1991-99 (median diagnosis age 5.5 years [0-20]) were included. Significant exposure differences between eras included higher percentage (53% vs. 17%, p < 0.01) receiving ≥ 20gm/m2 cumulative alkylators in 1991-99, but more receiving platinums (13% vs 5%, p < 0.01) and abdomen/ pelvis radiation (29% vs. 23%, p = 0.04) in 1970-90. 20-year cumulative incidence for any (40% vs. 28%, p < 0.01), ≥2 (16% vs. 7%, p < 0.01), and endocrine (8% vs. 2.5%, p < 0.01) grade 3-5 CHCs was higher in 1970-90 compared to 1991-99. The hazard ratio (HR) for any (HR 0.7, 95% Confidence Interval [CI] 0.55-0.9), ≥2 (HR 0.38, 95% CI 0.22-0.66) and endocrine (HR 0.25, 95% CI 0.09-0.67) grade 3-5 CHC was lower for 1991-99 survivors than 1970-90. The effect of era (1991-99 vs 1970-90, HR 0.73; 95% CI 0.59-0.91) on CHC was not attenuated when treatment variables were added to multivariable model. Exposures with increased risk of grade 3-5 CHC included platinums (hearing, HR 2, 95% CI 1.07-3.8), anthracycline ≥250mg/m2 (cardiovascular, HR 2.7, 95% CI 1.2-6) and abdomen/ pelvis radiation (second malignant neoplasms, HR 2.1, 95% CI 1.1-4, gastrointestinal, HR 7.4, 95% CI 3.5-16 and endocrine, HR 2.5, 95% CI 1.4-4.4). Gonadal dysfunction was the most common endocrine CHC. There was no difference in all cause or cause-specific mortality between the two cohorts. Conclusions: RMS survivors from the IRSG IV era are at reduced risk for late onset chronic health conditions compared to previous era.

Abstract Background Recent genome wide association studies (GWAS) of Hodgkin lymphoma (HL) have identified several associations at both HLA and non-HLA loci. However, much of HL heritability remains unexplained. Methods To identify novel risk loci, we performed a meta-analysis of 3 HL GWAS including a total of 1,810 cases and 7,879 controls. Results were replicated in an independent set of 1,163 cases and 2,580 controls, for a total of 3,097 and 11,097 cases and controls combined, respectively. participants in discovery and replication stages were of European descent. quality control and imputation we conducted a meta-analysis addressing 1,004,829 variants (λ= 1.10, λ1000= 1.03). Associations between SNP genotypes and HL risk were evaluated under a log-additive model of inheritance adjusting for sex, study center and significant principal components to control for population stratification. We performed an analysis with all HL cases and then conducted stratified analyses by histological subtype (classical, nodular sclerosis and mixed cellularity), age at diagnosis (nodular sclerosis among those diagnosed at 15- 35 years in all studies, and those diagnosed at 35 and older years in the European Study only) and EBV tumor status (negative and positive). We then used a bioinformatic approach (FunciSNP) to identify potential functional variants associated with HD risk correlated with risk loci of interest. We extracted publically available ENCODE data on biofeatures to identify potential functional motifs associated with the index SNP or correlated SNPs. Finally, we measured expression levels of the two alternative mRNA transcripts in lymphoblastoid cell lines (LCLs) from 49 post-therapy HL patients and 25 unaffected controls. RT-PCR was carried out in triplicate. Relative expression levels were calculated relative to TBP as housekeeping gene. Linear models were used to assess correlation between genotype and TCF3expression levels. Results The meta-analysis identified a novel susceptibility variant at chromosome 19p13.3 (rs1860661) associated with HL risk (Odds Ratio [OR]= 0.78, P=2.0*10-8, I2=0%). variant is located in intron 2 of TCF3 (also known as E2A), a regulator of B- and T-cell lineage commitment. was also significantly associated with HL (OR= 0.85, P=0.002) in the replication series of 1,281 cases and 3,218 controls. the combined analysis consisting of the discovery and replication sets, rs1860661was strongly associated with HL (OR=0.81,=3.5*10-10), with no evidence of heterogeneity between contributing studies (Phom=0.41, I2=0%). The number of G alleles defined by rs1860661 was significantly associated with a reduced risk of each HL subtype except EBV positive HL. rs1860661 and two correlated SNPs, rs10413888 (r2=0.90) and rs8103453 (r2=0.89) identified by FunciSNP analysis map in or near marks of open chromatin and in DNAse hypersensitivity sites in TCF3 in CD20+ B cell lines., the protective minor alleles of these SNPs as defined by the G-G-G haplotype map to the binding sites for ZBTB7a (rs10413888 and rs1860661) and (rs8103453) transcription factors, likely improving the binding efficiency to the sites which may result in increased transcription rates of TCF3. TCF3 is encoded by two alternative transcripts (E12 and E47). Higher expression levels of TCF3-E47, whose transcription start site is located close to rs1860661, was associated with the rs1860661-G allele in controls (P=0.02), but not in HL patients (P=0.22). Conclusion/Discussion TCF3 is essential for the commitment of lymphoid progenitors to both B-cell and T-cell lineage development. A molecular and phenotypic hallmark of classical HL is the loss of the B-cell phenotype in HRS cells, including lack of demonstrable B-cell receptor and most B-cell specific markers such as CD19 or CD20. HRS cells have a low level of TCF3, particularly homodimers of the isoform E47, due to expression of the ABF-1 and ID2 inhibitors that bind to TCF3. Thus, higher TCF3 levels in HRS precursor cells may lead to enhanced retention of the B cell phenotype, thereby conferring a protective effect. These data suggest a link between the 19p13.3 locus including TCF3 and HL risk, indicating that TCF3 could be relevant to HL etiology and pathogenesis. Disclosures: Link: Genentech: Consultancy; Millenium: Consultancy; Pharmacyclics: Consultancy; Spectrum: Consultancy.

Abstract Abstract 664 Background: Survivors of childhood acute lymphoblastic leukemia (ALL) who received prophylactic cranial radiation therapy (CRT) are at increased risk for deficits in neurocognitive skills, including attention, working memory and processing speed. As survivors age, global brain injury from CRT may reduce cognitive reserve, placing them at risk for early onset dementia or long-term memory deficits. The prevalence of dementia and memory abnormalities in adult survivors of childhood ALL has not previously been established. Methods: Analyses were conducted on 265 of 445 (60%) eligible survivors of childhood ALL (median age 36 years, range 26–54 years; 52% female) treated with 18Gy (n=127) or 24Gy (n=138) CRT with a median time from CRT of 30 years (range 15–46 years). Participants completed the Wechsler Memory Scale IV, including the Brief Cognitive Status Exam (BSCE), and the Wechsler Abbreviated Scale of Intelligence (WASI). Age-adjusted standard scores were calculated and the BSCE was also adjusted for education level. Prevalence of memory impairment (<1 SD below age-expected mean), stratified by RT dose exposure, is reported and logistic regression used to identify risk factors for impairment. A subset of survivors (n=85) completed brain magnetic resonance imaging (MRI), including assessment of hippocampal volume, cortical thickness, white matter volume, diffusion tensor imaging, and functional MRI (fMRI) during a cued-recall memory task. Results: Survivors who received 24 Gy CRT had increased impairment on global measures of both short-term recall (33%; p<0.001) and long-term recall (30%; p<0.001), while no increase in impairment was seen after 18Gy. Impaired short-term recall was associated with smaller right (p=0.02) and left (p<0.01) temporal lobe volumes, while impaired long-term recall was associated with thinner parietal and frontal cortices. On subtests evaluating narrative memory (i.e. story recall) and design memory, increasing RT dose (24 vs. 18Gy) was associated with an increased prevalence of long-term memory impairment (narrative: 28% vs. 12%, p=0.001; designs: 13% vs. 3.2%, p=0.003). However, no CRT dose response was identified for short term narrative and design memory. Survivors with impaired long-term memory for designs demonstrated a compensatory increase in left hippocampal fMRI activation (p=0.005), and the effect was greater in the higher dose group (p = 0.04). The mean score for long-term narrative memory among survivors who received 24Gy was equivalent to the mean score of a 70–74 year old adult population. Neither young age (0–4 years) at CRT (Odds Ratio [OR] 1.4, 95% confidence interval [CI] 0.8–2.7), time from CRT (OR 1.0, 95% CI 0.9–1.1) nor intrathecal methotrexate exposure (OR 3.9, 95% CI 0.4–36.1) were significantly associated with long-term memory deficits. Reduced cognitive status (by BSCE) was identified after 18Gy (9%, p=0.11) and 24Gy (18%, p<0.001), suggesting a CRT dose-response effect. On diffusion tensor imaging, increased radial diffusivity in the frontal, parietal and temporal regions, an inverse measure of white matter integrity, was associated with reduced BSCE. Current employment rates were equivalent (63%) in both CRT dose groups, suggesting no difference in functional status. Conclusions: Aging adult survivors of ALL who received 24Gy CRT have reduced cognitive status and significant impairment in short-term and long-term memory. There appears to be a dose response effect selective for long-term narrative and design memory, but not for short term narrative and design memory. These patterns are consistent with early onset of age-related (long-term) memory loss, and early stage dementia, yet at a median age of only 36 years. After 24Gy, survivors have the narrative memory equivalent to a 70 year-old in the general population. Survivors with memory impairment demonstrated reduced integrity on structural and functional neuroimaging in anatomical regions established as essential for memory formation and long-term recall. However, these memory impairments do not seem to affect functional status (employment rates) suggesting that, rather than frank dementia, deficits in middle adulthood are consistent with mild cognitive impairment (MCI). Longitudinal evaluation of this population is needed as MCI often progresses into early onset dementia with age. Disclosures: No relevant conflicts of interest to declare.

AbstractFemale Hodgkin lymphoma (HL) patients treated with chest radiotherapy (RT) have a very high risk of breast cancer. The contribution of genetic factors to this risk is unclear. We therefore examined 211 155 germline single-nucleotide polymorphisms (SNPs) for gene-radiation interaction on breast cancer risk in a case-only analysis including 327 breast cancer patients after chest RT for HL and 4671 first primary breast cancer patients. Nine SNPs showed statistically significant interaction with RT on breast cancer risk (false discovery rate, &lt;20%), of which 1 SNP in the PVT1 oncogene attained the Bonferroni threshold for statistical significance. A polygenic risk score (PRS) composed of these SNPs (RT-interaction-PRS) and a previously published breast cancer PRS (BC-PRS) derived in the general population were evaluated in a case-control analysis comprising the 327 chest-irradiated HL patients with breast cancer and 491 chest-irradiated HL patients without breast cancer. Patients in the highest tertile of the RT-interaction-PRS had a 1.6-fold higher breast cancer risk than those in the lowest tertile. Remarkably, we observed a fourfold increased RT-induced breast cancer risk in the highest compared with the lowest decile of the BC-PRS. On a continuous scale, breast cancer risk increased 1.4-fold per standard deviation of the BC-PRS, similar to the effect size found in the general population. This study demonstrates that genetic factors influence breast cancer risk after chest RT for HL. Given the high absolute breast cancer risk in radiation-exposed women, these results can have important implications for the management of current HL survivors and future patients.

PURPOSE: Survivors of childhood and adolescent cancer are at risk for long-term effects of disease and treatment. The Childhood Cancer Survivor Study assessed overall and cause-specific mortality in a retrospective cohort of 20,227 5-year survivors. PATIENTS AND METHODS: Eligible subjects were individuals diagnosed with cancer (from 1970 to 1986) before the age of 21 who had survived 5 years from diagnosis. Underlying cause of death was obtained from death certificates and other sources and coded and categorized as recurrent disease, sequelae of cancer treatment, or non–cancer-related. Age and sex standardized mortality ratios (SMRs) were calculated using United States population mortality data. RESULTS: The cohort, including 208,947 person-years of follow-up, demonstrated a 10.8-fold excess in overall mortality (95% confidence interval, 10.3 to 11.3). Risk of death was statistically significantly higher in females (SMR = 18.2), individuals diagnosed with cancer before the age of 5 years (SMR = 14.0), and those with an initial diagnosis of leukemia (SMR = 15.5) or CNS tumor (SMR = 15.7). Recurrence of the original cancer was the leading cause of death among 5-year survivors, accounting for 67% of deaths. Statistically significant excess mortality rates were seen due to subsequent malignancies (SMR = 19.4), along with cardiac (SMR = 8.2), pulmonary (SMR = 9.2), and other causes (SMR = 3.3). Treatment-related associations were present for subsequent cancer mortality (radiation, alkylating agents, epipodophyllotoxins), cardiac mortality (chest irradiation, bleomycin), and other deaths (radiation, anthracyclines). No excess mortality was observed for external causes (SMR = 0.8). CONCLUSION: While recurrent disease remains a major contributor to late mortality in 5-year survivors of childhood cancer, significant excesses in mortality risk associated with treatment-related complications exist up to 25 years after the initial cancer diagnosis.

AbstractBackgroundSurvivors of childhood non‐Hodgkin lymphoma (NHL) are at increased risk for chronic health conditions. The objective of this study was to characterize health conditions, neurocognitive function, and physical performance among a clinically evaluated cohort of 200 childhood NHL survivors.MethodChronic health and neurocognitive conditions were graded as per a modified version of the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) and impaired physical function defined as performance < 10th percentile of normative data. Multivariable regression was used to investigate associations between sociodemographic characteristics, therapeutic exposures, and outcomes.ResultsSurvivors were a median age of 10 years (range 1–19) at diagnosis and 34 years (range 20–58) at evaluation. Eighty‐eight (44%) received radiation, 46 (23%) cranial radiation, and 69 (35%) high‐dose methotrexate. Most prevalent CTCAE Grades 3–4 (severe life‐threatening) conditions were obesity (35%), hypertension (9%), and impairment of executive function (13%), attention (9%), and memory (4%). Many had impaired strength (48%), flexibility (39%), muscular endurance (36%), and mobility (36%). Demographic and treatment‐related factors were associated with the development of individual chronic diseases and functional deficits.ConclusionsClinical evaluation identified a high prevalence of chronic health conditions, neurocognitive deficits, and performance limitations in childhood NHL survivors.

BACKGROUNDHealth complications related to childhood cancer may be influenced by risky health behaviors (RHBs), particularly when RHBs co‐occur. To the authors' knowledge, only limited information is available describing how RHBs cluster among survivors of childhood cancer and their siblings and the risk factors for co‐occurring RHBs.METHODSLatent class analysis was used to identify RHB clusters using longitudinal survey data regarding smoking, alcohol use, and physical activity from adult survivors (4184 survivors) and siblings (1598 siblings) in the Childhood Cancer Survivor Study. Generalized logistic regression was used to evaluate associations between demographic characteristics, treatment exposures, psychological distress, health conditions, and cluster membership.RESULTSThree RHB clusters were identified: a low‐risk cluster, an insufficiently active cluster, and a high‐risk cluster (tobacco and risky alcohol use and insufficient activity). Compared with siblings, survivors were more likely to be in the insufficiently active cluster (adjusted odds ratio [ORadj], 1.17; 95% confidence interval [95% CI], 1.06‐1.27) and were less likely to be in the high‐risk cluster (ORadj, 0.79; 95% CI, 0.69‐0.88). Risk factors for membership in the high‐risk cluster included psychological distress (ORadj, 2.76; 95% CI, 1.98‐3.86), low educational attainment (ORadj, 7.49; 95% CI, 5.15‐10.88), income <$20,000 (ORadj, 2.62; 95% CI, 1.93‐3.57), being divorced/separated or widowed (ORadj, 1.36; 95% CI, 1.03‐1.79), and limb amputation (ORadj, 1.52; 95% CI, 1.03‐2.24). Risk factors for the insufficiently active cluster included chronic health conditions, psychological distress, low education or income, being obese or overweight, female sex, nonwhite race/ethnicity, single marital status, cranial radiation, and cisplatin exposure.CONCLUSIONSRHBs co‐occur in survivors of childhood cancer and their siblings. Economic and educational disadvantages and psychological distress should be considered in screening and interventions to reduce RHBs. Cancer 2016. © 2016 American Cancer Society. Cancer 2016;122:2747–2756. © 2016 American Cancer Society.

Nodular sclerosing Hodgkin lymphoma (NSHL) is a distinct, highly heritable Hodgkin lymphoma subtype. We undertook a genome-wide meta-analysis of 393 European-origin adolescent/young adult NSHL patients and 3315 controls using the Illumina Human610-Quad Beadchip and Affymetrix Genome-Wide Human SNP Array 6.0. We identified 3 single nucleotide polymorphisms (SNPs) on chromosome 6p21.32 that were significantly associated with NSHL risk: rs9268542 (P = 5.35 × 10−10), rs204999 (P = 1.44 × 10−9), and rs2858870 (P = 1.69 × 10−8). We also confirmed a previously reported association in the same region, rs6903608 (P = 3.52 × 10−10). rs204999 and rs2858870 were weakly correlated (r2 = 0.257), and the remaining pairs of SNPs were not correlated (r2 < 0.1). In an independent set of 113 NSHL cases and 214 controls, 2 SNPs were significantly associated with NSHL and a third showed a comparable odds ratio (OR). These SNPs are found on 2 haplotypes associated with NSHL risk (rs204999-rs9268528-rs9268542-rs6903608-rs2858870; AGGCT, OR = 1.7, P = 1.71 × 10−6; GAATC, OR = 0.4, P = 1.16 × 10−4). All individuals with the GAATC haplotype also carried the HLA class II DRB1*0701 allele. In a separate analysis, the DRB1*0701 allele was associated with a decreased risk of NSHL (OR = 0.5, 95% confidence interval = 0.4, 0.7). These data support the importance of the HLA class II region in NSHL etiology.