A limited number of relatively contradictory studies have suggested that the development of serious ototoxicity in children treated with cisplatin or, more rarely, carboplatin could be partly related to genetic risk factors affecting detoxification enzymes and membrane transporters of platinum derivatives. The objective of this study is therefore to identify genetic variants associated with the development of platinum ototoxicity in patients treated with cisplatin or carboplatin (minimum follow-up of 3 years) for one of the following diseases: neuroblastoma, hepatoblastoma, retinoblastoma, malignant germ cell tumour, osteosarcoma, high-risk or recurrent Wilms' tumour, non-parameningealrhabdomyosarcoma. A total of 180 patients, corresponding to 60 cases with grade 3 or 4 ototoxicity and 120 controls with no signs of ototoxicity (separate complete audiograms for each ear) will be included. A saliva sample will be used to obtain DNA for pharmacogenetic studies. The value of this study will be to define a population at high risk of developing ototoxicity in order to adapt treatment, or even develop preventive treatment of ototoxicity based on antioxidant medications

• Platinum derivatives and their toxicity Cisplatin, carboplatin, and oxaliplatin are major cytotoxic agents used in the treatment of many types of cancer. They exert their cytotoxic action by forming DNA adducts. However, their structural differences confer different pharmacokinetic and pharmacodynamic properties, and especially different toxicity profiles. Cisplatin induces a risk of nephrotoxicity at all ages, peripheral neuropathy of the extremities in young adults and a risk of ototoxicity in children. Carboplatin is essentially haematotoxic and sometimes ototoxic (especially at high doses), while oxaliplatin is essentially neurotoxic. Ototoxicity is measured by pure-tone audiometry and is scored from grade 0 to grade 4, where grades 3 and 4 correspond to severe ototoxicity, according to the Brock grading system, the reference system for measuring ototoxicity at the time of treatment of these children and up until the present time (Brock et al., 1991) (Peters et al., 2000). Ototoxicity limits the cumulative dose of platinum derivatives in children. Cisplatin remains a major drug in paediatric oncology (osteosarcoma, high-risk neuroblastoma, malignant germ cell tumours, hepatoblastoma, medulloblastoma, other brain tumours). Cisplatin ototoxicity is fairly frequent and can be very severe (more than 50% of grade 1 or higher ototoxicity; 15% of grade 3 or 4 ototoxicity) (Bertolini et al., 2004). Carboplatin, an important component of the treatment of many embryonal tumours (neuroblastoma, medulloblastoma, retinoblastoma) is less frequently responsible for ototoxicity (between 1 and 5% of grade 1 or higher ototoxicity), but can nevertheless occasionally cause serious ototoxicity (1% of grade 4) (Jehanne et al., 2009). Ototoxicity can also appear and deteriorate for several years after treatment (median time to discovery of carboplatin ototoxicity: 3.7 years), while no cases of recovery of high-grade hearing loss have been reported (Bertolini et al., 2004; Jehanne et al., 2009). These ototoxicity phenomena can have very harmful consequences for the child's social development and schooling. Oxaliplatin will be excluded from this study, as few children are currently treated with this molecule, which is generally not ototoxic.
• Detoxification of platinum derivatives The glutathione (GSH) conjugation pathway is involved in detoxification of platinum derivatives and decreases their cytotoxicity by inhibiting the formation of DNA adducts. Conjugation to GSH is ensured by glutathione-S-transferases (GSTs), which constitute a multigenic family of enzymes involved in phase II metabolism that inactivate electrophilic molecules by conjugating them to glutathione, thereby promoting their elimination. GSTP1 is particularly involved in the conjugation of platinum derivatives. Elimination of glutathione conjugates then involves membrane transporters. These conjugates can be expelled from target tumour cells and then eliminated, essentially in the urine, by efflux membrane transporters, particularly MRP2 (Sun et al., 2009). Two other enzymes, thiopurine S-methyltransferase (TPMT) and catechol O-methyltransferase, have been reported to be associated with increased toxicity of cisplatin when their activity is decreased (Ross et al., 2009).
• Genetic factors predisposing to or protecting against the toxicity of platinum derivatives Several GST isoenzymes present polymorphisms modifying their activity. This is the case for the GSTM1 and GSTT1 isoenzymes, which can undergo deletions responsible for loss of activity of the enzyme, and GSTM3, for which the GSTM3*B allele has been associated with induction of enzyme activity by cisplatin (Peters et al., 2000). An A/G polymorphism situated in the substrate binding domain of the GSTP1 isoenzyme in position +313 of exon 5 (A313G) induces substitution of a isoleucine by a valine (Ile105Val). This polymorphism is relatively frequent (about 10% of homozygous mutants in Caucasian populations). It influences the activity of the enzyme and has been associated with better survival of patients receiving platinum-based chemotherapy (Goekkurt et al., 2006; Stoehlmacher et al., 2002; Sun et al., 2009). This polymorphism has also been shown to influence the toxicity of platinum-based treatments. For example, one study showed that the neurotoxicity of oxaliplatin was more frequent in subjects homozygous for GSTP1 105Ile allele than in heterozygous subjects or subjects homozygous for the GSTP1 105Val allele (odds ratio = 5.75). (Lecomte et al., 2006). This protective effect of the mutant allele has also been demonstrated for docetaxel peripheral neuropathies (odds ratio = 6.11) and cisplatin ototoxicity in testicular cancers in young men (Mir et al., 2009; Oldenburg et al., 2007).

A study conducted in children demonstrated an association between GSTM1 and GSTT1 polymorphisms (presence of at least one null genotype) and the development of grade 3 or higher toxicity related to treatment of medulloblastoma. However, this study, based on a small sample of children (42 children, 19 of whom experienced severe ototoxicity) and comprising patients receiving very different treatments, was unable to determine the proportion of toxicity specifically related to platinum derivatives, as some patients also received brain irradiation. Furthermore, this association between GST polymorphisms and toxicity was not observed when the analysis was specifically confined to grade 3 or higher ototoxicity. Finally, the role of GSTP1 Ile105Val polymorphism has not been evaluated in children (Barahmani et al., 2009). To our knowledge, only one study has specifically evaluated the potential impact of GST polymorphism on cisplatin ototoxicity in children. This study showed a protective effect of the GSTM3*B variant, but did not observe a protective effect of the GSTP1 105Val allele with respect to ototoxicity, as demonstrated in adults for neurotoxicity. However, these findings were based on small sample sizes (20 patients with ototoxicity versus 19 patients without ototoxicity) and the ototoxicity experienced by these patients was grade 2 (8 patients) or 3 (12 patients) with no cases of grade 4 ototoxicity (Peters et al., 2000). The results of this study therefore need to be confirmed by a study comprising a greater number of patients, cases and controls, including patients with more severe ototoxicity. Ross et al., studying drug metabolism enzymes, demonstrated a highly significant association between cisplatin ototoxicity in children and genetic variants of TPMT (rs12201199, p = 0.00022, OR = 17.0) and COMT (rs9332377, p = 0.00018, OR = 5.5). However, as the role of these enzymes in detoxification of platinum derivatives has not been previously reported, these results need to be confirmed by other studies with, in particular, matching for age, exclusion of cases of grade 2 ototoxicity from the group of severe ototoxicity, and a greater number of control patients (Ross et al., 2009). Finally, no study has investigated the genetic factors involved in carboplatin ototoxicity in children. One study concerning membrane transporters showed that a common polymorphism situated in the promoter region of the MRP2 (-24C>T) transporter gene was associated with a better response to platinum-based chemotherapy (allelic frequency: 18% of Caucasian) (Cascorbi, 2006; Sun et al., 2009). It would be interesting to determine whether polymorphisms affecting this gene can also modulate the toxicity of these treatments. A polymorphism of the gene coding for the influx transporter OCT2 (c.808G>T; p.270Ala>Ser) has also been shown to be associated in vivo and in vitro in mice with a protective effect in relation to cisplatin nephrotoxicity and ototoxicity (Ciarimboli et al.; Filipski et al., 2009), but the potential influence of this polymorphism on platinum ototoxicity has never been demonstrated in man. It therefore appears very interesting to study the potential impact of this polymorphism in the context of our study.

In summary, all of these studies clearly confirm the hypothesis that genetic factors are involved in the predisposition of some children to experience harmful ototoxic effects of platinum derivatives, particularly cisplatin and carboplatin. These studies also suggest the existence of good candidate gene variants likely to be associated with the development of this ototoxicity. However, the results of most these studies have either not been reproduced or remain controversial, indicating the need for confirmation studies based on larger sample sizes and more homogeneous patient groups in terms of treatment, as in the Otoplat study.

Inclusion Criteria:

• previous treatment with cis-or carboplatin
• Age equal to or greater than 4 years
• the presence of a separate complete audiogram for each ear in order to define the Brock grade
• informed and written consent of both parents or legal guardian
• affiliation of the child to a social security scheme

To patients:

• previous treatment with platinum compounds for neuroblastoma, hepatoblastoma, pancreatoblastoma, retinoblastoma, germ cell tumor malignant osteosarcoma, Wilms high risk or relapsed non para-meningeal rhabdomyosarcoma with ototoxicity grade 3 or 4 Brock (Grade 3 - 40dB to 2000 Hz; Grade 4-40 dB at 1000 Hz) To patients controls
• previous treatment with platinum compounds for neuroblastoma, hepatoblastoma, pancreatoblastoma, retinoblastoma, germ cell tumor malignant osteosarcoma, Wilms high risk or recurrence rhabdomyosarcoma non parameningeal without ototoxicity (grade 0 Brock)
• minimum follow-up of 3 years after stopping treatment

Exclusion Criteria:

• patients with brain tumor
• patients with tumor "paraméningée" (such as rhabdomyosarcoma of the rock)
• patients with renal toxicity (≥ Grade 2) at the time of treatment with platinum-based
• patients with pathological pre-treatment assessment of hearing
• chemotherapy during
• patients with cerebral or total body irradiation (Total Body Irradiation)

180 patients will be included : 60 patients with grade 3 or 4 Brockototoxicityand 120 patients controls with no signs of ototoxicity (separate complete audiograms for each ear)