Late toxicities, especially second malignancies, with GWAS analysis in hepatoblastoma cases treated in the Japanese study group for pediatric liver tumor protocol-2.

Abstract

10043 Background: This study focused on late toxicities, especially second malignancies (SMNs), experienced by patients with hepatoblastoma (HB) treated with CITA (cisplatin and pirarubicin) in the Japanese Study Group for Pediatric Liver Tumor protocol-2 (JPLT-2). Methods: In the JPLT-2, 361 children with HB were eligible for inclusion. We evaluated late effects including ototoxicity, cardiotoxicity and SMN in these patients. PRETEXT I/II patients were typically treated by initial resection at diagnosis followed by CITA. Otherwise, patients received preoperative CITA, followed by surgery and postoperative chemotherapy. Non-responders to CITA received ITEC (ifosfamide, pirarubicin, etoposide, and carboplatin) or intensified chemotherapeutic regimens. In 267 patients whose germline DNA samples were available, genotyping analysis was performed using the Illumina Human Omni Express Exome-8 v.1.3 BeadChip. Associations between single nucleotide polymorphisms (SNPs) and toxicities were assessed using logistic regression analysis. Results: Among the patients (n = 300) who survived more than 5 years, ototoxicity was detected in 61 and cardiotoxicity in 18. The presence of ototoxicity was associated with a significantly higher dose of cisplatin, and the Brock grade of ototoxicity was correlated with the total dose of cisplatin (P < 0.01). SMNs occurred in 13 patients (9 leukemia, 1 myelodysplastic disease, 1 lymphoma, 1 Ewing sarcoma and 1 thyroid tumor). The total dose of pirarubicin in SMN cases was significantly higher than those without SMNs (P = 0.035). The mean age at onset of HB was 13 months in the patients with SMNs and 23.3 months in the other patient (P = 0.12). However, two-way analysis of variance showed that SMN occurrence was significantly correlated with the age of HB onset and the total dose of pirarubicin (P < 0.0001). We identified ten inherited SNPs, including six SNPs associated with ototoxicity such as ACYP2, two SNPs associated with cardiotoxicity and two candidate SNPs associated with SMNs. Conclusions: The total dose of cisplatin was significantly associated with ototoxicity development. A low incidence of cardiotoxicity but high rates of SMTs, which were correlated with a high dose of pirarubicin and younger age at diagnosis, were found in this study. We also identified SNPs associated with SMNs in HB. Our study provides strong evidence that germline mutation may modify the effect of chemotherapy for childhood cancers on late toxicities, especially during early childhood. Clinical trial information: UMIN000001116.