Indications and results of diagnostic biopsy in pediatric renal tumors: A retrospective analysis of 317 patients with critical review of SIOP guidelines.

OBJECTIVES: According to the Renal Tumor Study Group (RTSG) of the International Society of Paediatric Oncology (SIOP), diagnostic biopsy of renal tumors prior to neoadjuvant chemotherapy is not mandatory unless the presentation is atypical for a Wilms tumor (WT). This study addresses the relevance of this strategy as well as the accuracy and safety of image-guided needle biopsy. METHODS: Clinical, radiological, and pathological data from 317 children (141 males/176 females, mean age: 4 years, range, 0-17.6) consecutively treated in one SIOP-affiliated institution were retrospectively analyzed. RESULTS: Presumptive chemotherapy for WT was decided for 182 patients (57% of the cohort), 24 (8%) were operated upfront, and 111 (35%) were biopsied at diagnosis. A non-WT was confirmed after surgery in 5/182 (3%), 11/24 (46%), and 28/111 (25%), respectively. Age at diagnosis was the most commonly (46%) used criterion to go for biopsy but a nine-year threshold should be retrospectively considered more relevant. Tumor volumes of clear cell sarcoma of the kidney and WT were significantly higher than those of other tumors (P = 0.002). The agreement between core-needle biopsy (CNB) and final histology was 99%. No significant morbidity was associated with CNB. CONCLUSION: The use of SIOP criteria to identify patients eligible for presumptive WT neoadjuvant chemotherapy or upfront surgery avoided biopsy in 65% of children and led to a 97% rate of appropriate preoperative chemotherapy. Image-guided CNB is a safe and accurate diagnostic procedure. The relevance of SIOP biopsy criteria might be improved by using an older age threshold (9 years instead of 6 years) and by adding initial tumor volume.

From childhood to adulthood: long-term outcome of medulloblastoma patients. The Institut Curie experience (1980-2000).

Medulloblastoma patients treated at the Institute Curie between 1980 and 2000 were reviewed. Only patients whose primary treatment included craniospinal radiation were considered. Surviving patients were identified and evaluated by means of self-report questionnaires using the Health Utility Index (HUI). Psychosocial functioning, employment, and other health-related indicators were recorded. Seventy-three patients were treated during the study period. At a median follow-up from diagnosis of 14.4 years, 49 patients were alive and 45 surviving patients could be contacted. Late sequelae were frequent, particularly neurological deficits (71%) and endocrine complications (52%). Impairments of psychosocial functioning, including employment, driving capacity, independent living, and marital status, were identified in most patients. Most long-term medulloblastoma survivors suffer persistent deficits in several domains, with a significant impact on their psychosocial functioning. These findings reinforce the importance of early intervention programs for all survivors in order to reduce the psychosocial impacts of their disease.

[Genital bleeding secondary to chemotherapy in adolescent girls]. / Métrorragies de l'adolescente au cours des chimiothérapies responsables de thrombopénies sévères.

UNLABELLED: Severe genital bleeding during adolescence can occur in the event of thrombocytopenia related to chemotherapy. Preventive hormonal treatment to induce therapeutic amenorrhea is recommended by some clinicians. Nevertheless, the adverse effects of oestroprogestative treatment, such as thromboembolic risk or hepatic toxicity, could potentialize the adverse effects of some chemotherapies. AIM OF THE STUDY: To assess retrospectively the risk of genital bleeding associated with thrombocytopenia secondary to chemotherapy in a population of adolescent girls for whom therapeutic amenorrhea was induced or not. PATIENTS AND METHODS: Among 140 girls, 12 to 18 years old, who were subjected to chemotherapy between 1991 and 1998, 24 girls presented at least one thrombocytopenic event (platelet level < 20 x 10(9) l(-1)) and were included. RESULTS: Six out of 24 adolescent girls received hormonal therapy to induce amenorrhea. Only one bleeding event was observed, in a girl who had not received preventive oestroprogestative treatment. Vital prognosis was preserved. Furthermore, spontaneous amenorrhea occurring before chemotherapy was observed in 33% of the patients and permanent secondary ovarian insufficiency in 17% of the patients. CONCLUSION: The risk of genital bleeding among adolescent population at risk of chemotherapy-induced thrombocytopenia is low. The high frequency of amenorrhea, secondary to weight loss or to chemotherapy toxicity, should raise questions as to the usefulness of preventive oestroprogestative treatment.

Peripheral blood stem cell collection in 24 low-weight infants: experience of a single centre.

Peripheral blood stem cells (PBSC) harvest may be difficult in young children. Extracorporeal separator line priming by red blood cells is usually required to improve haemodynamic tolerance and efficacy of collection. We present our experience with 24 children weighing less than 15 kg treated between January 1997 and September 1999, in whom we tried to avoid systematic blood priming. The median age and weight at the time of apheresis were 2.4 years and 12 kg, respectively. A total of 48 PBSC were performed. When haemoglobin was less than 12 g/dl, packed red cells were transfused before collection (40% of aphereses). The median cell yield per apheresis was 7.1 (2.2-30.6)x10(6)/kg CD34(+) cells and 16.0 (3.3-44.3)x10(5) CFU-GM/kg. Initial collection failed in three cases. Four children required an additional haematopoietic progenitor mobilization. This procedure allowed PBSC collection without transfusion in 37.5% of children, and was safe (two serious and five mild transient side effects) and effective (median CD34(+) cells collected per child: 7.1 x 10(6)/kg (4.6-30.6) and CFU-GM: 15.1 x 10(5)/kg (4.7-44.3)). Despite their low weight, insertion of a femoral catheter was avoided in 43% of children.

[Diode laser thermotherapy and chemothermotherapy in the treatment of retinoblastoma]. / Thermothérapie et thermochimiothérapie au laser diode dans le traitement du rétinoblastome.

INTRODUCTION: The use of transpupillary thermotherapy alone or associated with systemic chemotherapy is a therapeutic modality of ocular retinoblastoma that allows ocular preservation without external beam irradiation of the eye. We present our experience with thermotherapy in the treatment of selected cases of retinoblastoma. MATERIAL AND METHODS: This paper reports a retrospective case series of patients treated for retinoblastoma by thermotherapy or chemothermotherapy (carboplatin IV followed by thermotherapy) in a single institution from October 1994 to December 2000. Data collected include general characteristics of the treated children, tumor characteristics, and the results of the treatments on local tumor control. Transpupillar thermotherapy was delivered with a diode laser through an operating microscope. Each tumor was treated separately and laser intensity, spot size, and duration were adapted to the size of the tumor and the clinical response. Chemothermotherapy consisted in thermotherapy delivered shortly after an intravenous injection of carboplatin (560 mg/m(2)) at day 1, followed by thermotherapy alone at day 8 if the lesion was 6mm or more in diameter. This cycle was administered every 28 days. The choice between thermotherapy and chemothermotherapy depended on the initial size of the lesions. Thermotherapy was used when the lesion measured 3mm or less. Lesions measuring more than 15 mm, or associated with substantial vitreous seeding, retinal detachment, or optic nerve head involvement are not suitable for these techniques. RESULTS: During the study period, 239 children were treated in our institution and 109 of them (147 eyes, 372 tumors) could be treated conservatively without external beam radiation. The median tumor diameter at the moment of thermotherapy or chemothermotherapy was 2mm (range, 0.2-15.0mm). One hundred and ninety-four tumors were treated by chemothermotherapy and 18 by thermotherapy alone. In 75% of the cases, the treatment was administered after two courses of chemotherapy (etoposide and carboplatin). After a mean follow-up of 55 months (range, 16-89 months), tumor control was obtained in 87.1% of lesions after chemothermotherapy and 77.8% after thermotherapy. Salvage enucleation was necessary for seven lesions (seven eyes) but none in the cases where thermotherapy was used alone. No severe systemic side effects were noted. DISCUSSION: Diode laser delivers hyperthermia on the tumor bed and its use alone or in association with systemic administration of carboplatin makes it possible to preserve the eye without external beam irradiation, with few side effects and less cumulative doses of chemotherapy. CONCLUSION: Thermotherapy and chemothermotherapy provide excellent local tumor control and eye preservation in selected cases of retinoblastoma.

The LMCE5 unselected cohort of 25 children consecutively diagnosed with untreated stage 4 neuroblastoma over 1 year at diagnosis.

The Lyon-Marseille-Curie-Est (LMCE) of France cooperative group has previously reported successive series of unselected stage four children older than 1 year at diagnosis with metastatic neuroblastoma (LMCE 1 and 3). The goal of LMCE 5 study was to increase progression free survival rate as compared to LMCE 1 and 3. Based on improvements reported with post induction chemotherapy, the LMCE 5 used post induction for all children, but omitted total body irradiation and immunomagnetic purging in megatherapy regimen for all children. Twenty-five sequentially diagnosed children received an induction regimen which compared with previous induction included an increased dose of etoposide and cyclophosphamide, delivered similar dose of cisplatinum, and deleted doxorubicin and vincristin. After surgery treatment was stratified based on response and eligible children received etoposide carboplatin (LMCE 5A : n=10)+/-doxorubicin (LMCE 5B-C n=13) followed by megatherapy (melphalan without total body irradiation and unpurged peripheral blood stem cell rescue). The increase in drug doses during induction did not improve remission rate. The progression free survival at 6 years is 8%. It is significantly worse than LMCE 3, and equivalent to LMCE 1 study though toxic death rate has decreased with increasing experience. Failure to improve the response rate during induction and reducing the megatherapy regimen may be the main factors in this disappointing result. Modified strategies for induction, non toxic alternative to total body irradiation, and post megatherapy regimen should be developed.

[Reaching a team consensus in pediatric oncology when treatment proves ineffective]. / Recherche d'un consensus d'équipe en oncologie pédiatrique en cas de traitements éprouvés inefficaces.

Modern treatments are able to cure 3/4 of paediatric cancers, but on another hand there is still a too high number of patients which will enter a resistant phase and test the ability of the medical and nursing team to face together the new modalities of care. At diagnosis, the reasonable hope to cure the child and the exciting feeling to be endued with a repairing mission strengthen among the team both the self-esteem and a rational and emotional consensus. Conversely, at the time of relapse or resistant disease, the feeling of a failure, added to anxiety and reciprocal aggressiveness, will contradictorily affect the group. In this atmosphere of loss of control of reality and emotions, elaborating new priorities of care is mandatory: medical positions have of course to be in keeping, but this is not enough since the nursing team has, through sharing information and feeling, to give their cure project of the child's cancer as lost while retaining and expanding their ability to care the child with cancer. Thanks to an over-investment of the symptomatic, psychological, and affective help support to the child, the nursing team will succeed in regaining this self-esteem absolutely necessary to their own professional survival. Consensus in the team is ultimately linked to its empathic ability of exchanging with the families which in turn results in giving them guidelines for the last weeks and months of life and helping them to be less emotionally unprovided when bereavement time has come: a pedagogy of the continuity of care is thus taking place which, equally distant from forsaking and therapeutic relentlessness, chooses to give up the idea of curing in order to go on caring.

Carboplatin before and during radiation therapy for the treatment of malignant brain stem tumours: a study by the Société Française d'Oncologie Pédiatrique.

Childhood malignant brain stem tumours have a very poor prognosis with a median survival of 9 months despite radiotherapy. No chemotherapy has improved survival. However, carboplatin has been reported to have activity in glial tumours as well as antitumour synergy with radiation. Our aims were to test the response rate of these tumours to carboplatin alone and to evaluate the efficacy on survival of carboplatin alone followed by concurrent carboplatin and radiotherapy. Patients younger than 16 years with typical clinical and radiological presentation of infiltrating brain stem tumour, as well as histologically-documented cases in the atypical forms, were eligible. Two courses of carboplatin (1050 mg/m2 over 3 days) were administered initially. This treatment was followed by a chemoradiotherapy phase including five weekly carboplatin courses (200 mg/m2) and conventional radiotherapy. 38 eligible patients were included. No tumour response was observed after the initial phase. This schedule of first-line carboplatin followed by concurrent carboplatin and radiotherapy did not improve survival.

Non-metastatic Ewing's sarcoma of the ribs: the French Society of Pediatric Oncology Experience.

From 1984 to 1997, 57 consecutive patients with non-metastatic Ewing's sarcoma of the ribs were treated according to multimodal French Society of Pediatric Oncology (SFOP) protocols EW 84, EW 88 and EW 93. The results of treatment were reviewed and analysed. Median age was 12 years. 34 patients had large tumours (greatest tumour dimension > or = 8 cm); pleural effusion was noted in 26. A tumour-positive margin after surgery was noted in 15 patients. Histological response after chemotherapy was assessed in 34 patients. 34 patients received radiation therapy. With a median follow-up of 5 years, the projected overall and relapse-free survival rates were 69 and 62%, respectively. The major site of relapse was local. None of the following was significant in predicting relapse: tumour size, gender, age at diagnosis, existence of pleural effusion, level of rib tumour, rib component, type of local control, surgical margin (positive or negative). Response to chemotherapy was the sole significant prognostic factor (P=0.004). Patients with pleural effusion had a higher percentage of relapse if they were treated without local radiation therapy. Our study confirms the prognostic significance of response to initial chemotherapy. Radiation therapy may be withheld in selected cases, but seems necessary in patients with pleural effusion.

[Metastatic osteosarcoma: prognosis factors and treatment]. / Les ostéosarcomes métastatiques; facteurs pronostiques et traitements.

Long term outcome and prognosis factors of patients with metastatic osteosarcoma were evaluated on 29 observations from 3 centres reviewed retrospectively. Twenty-nine patients less than 18 years old were treated from 1990 to 1998. Only 29 of these patients had received similar treatments associating chemotherapy and surgery, adapted according to histological and clinical response to treatment, as recommended by the SFOP. Overall survival at five years was 26%, and disease free survival 14%. Eight patients are alive, four in first complete remission (CR) and four in second CR. Three of the four patients alive in first CR had bone metastases at diagnosis. On univariate analysis, factors predicting survival are: the numbers of organs affected by metastatic lesions, the number of lung nodules and the type of surgery. This is, to our knowledge, the first report of long term survivors with bone metastases at diagnosis. Metastatic osteosarcoma prognosis remain poor. A randomised study would help to define the best possible treatment for this disease.

GH deficiency caused by cranial irradiation during childhood: factors and markers in young adults.

Cranial irradiation alters hypothalamic-pituitary function. We reevaluated 90 patients with GH deficiency caused by fractionated cranial irradiation performed at age 4.9 +/- 0.4 (SE) yr when they were 15.7 +/- 0.2 yr old. Group 1 received 18 Grays (Gy) (7 cases) or 24 Gy (21 cases) for acute lymphoblastic leukemia; group 2, 30-40 Gy for medulloblastoma (22 cases); group 3, 45-60 Gy for optic glioma and various tumors (30 cases); and group 4, 40-50 Gy for retinoblastoma (10 cases). The mean GH peaks after an arginine insulin test in group 3 (1.9 +/- 0.4 microg/liter) was lower than in groups 1 (4.8 +/- 0.5 microg/liter, P < 0.001) and 2 (3.4 +/- 0.5 microg/liter, P < 0.03). The mean plasma IGF-I concentrations in group 3 [-3.8 +/- 0.2 z score (zs)] was lower than in groups 1 (-2.4 +/- 0.3 zs, P < 0.001) and 2 (-3.1 +/- 0.2 zs, P < 0.02), as was the mean in group 4 (-3.9 +/- 0.3 zs, P < 0.01 compared with group 1 and P < 0.05 compared with group 2). GH peaks and IGF-I were correlated positively (P = 0.0001) and negatively with dose (P < 0.001 for GH and P = 0.0001 for IGF-I), but not with age at irradiation. Among the 43 patients with GH peaks below 3 microg/liter, 41 (95%) had plasma IGF-I less than -2 zs. The body mass index (BMI), plasma insulin, and leptin were similar in the four groups. They were positively correlated with each other (P < 0.001 for BMI compared with insulin and with leptin, respectively, and P < 0.01 for insulin compared with leptin), but not with age or dose of irradiation, or with markers of GH secretion. In conclusion, in patients with GH deficiency caused by cranial irradiation, the residual GH secretion and plasma IGF-I depend on the dose. Almost all the patients with severe GH deficiency had low plasma IGF-I. BMI, leptin, and insulin seem to be independent of GH status.

Optimal duration of preoperative therapy in unilateral and nonmetastatic Wilms' tumor in children older than 6 months: results of the Ninth International Society of Pediatric Oncology Wilms' Tumor Trial and Study.

PURPOSE: To determine the optimal duration of preoperative chemotherapy to further increase the proportion of stage I tumors by comparison of two regimens in the treatment of patients older than 6 months who have unilateral Wilms' tumor. PATIENTS AND METHODS: Eligible patients (n = 382) initially received four weekly doses of vincristine (VCR) and two courses of actinomycin D (AMD) and were randomized either to be operated on (4-week group [n = 193]) or to receive 4 more weeks of the same chemotherapy regimen (8-week group [n = 189]). The assessment criterion was the observed percentage of stage I tumors. After surgery, patients were assigned according to tumor stage and histology to four different treatment groups: stage I and favorable histology (n = 5) were to have no further treatment (NFT); stage I and standard histology or anaplasia (n = 244), VCR and AMD for 17 weeks (AV); stages II and III and favorable or standard histology, VCR, AMD, and an anthracycline for 27 weeks (AVE) with no abdominal radiotherapy for stage II N0 disease (n = 75) or with a 15-Gy dose of abdominal irradiation (RTH) in case of stages IIN1 and III (n = 56). Anaplastic tumors staged higher than I or clear-cell sarcoma of the kidney (14), AMD, VCR, an anthracycline, and ifosfamide for 36 weeks (DEVI). RESULTS: No advantage was found in favor of prolonged preoperative treatment. The percentages obtained for the 4-week and the 8-week groups, respectively, were as follows: stage I, 64% versus 62%; intraoperative tumor rupture rate, 1% versus 3%; 2-year EFS, 84% versus 83%; and 5-year OS, 92% versus 87%. Two-year EFS and 5-year OS rates, respectively, of the different treatment groups were as follows: NFT, 100% for both EFS and OS; AV, 88% and 93%; AVE, 84% and 88%; AVE RTH, 71% and 85%; and DEVI, 71% and 71%. The rate of abdominal recurrences in stage II N0 nonirradiated patients was 6.6%. CONCLUSION: The 4-week schedule pre-nephrectomy chemotherapy regimen should be considered the standard treatment. Clinical trials should continue to improve the cure rate of high-risk patients and the quality of life of children with a more favorable prognosis.

Hepatic regenerating nodules: a mimic of recurrent cancer in children.

BACKGROUND: Pseudometastatic lesions of the liver may be discovered incidentally in children previously treated for malignant tumour. OBJECTIVE: To describe the radiological pattern of these lesions and to analyse their pathogenesis. MATERIALS AND METHODS: Nine children, 2-12 years' old at the time of diagnosis, are described in this retrospective multicentre report. The primary tumours were: nephroblastoma (n = 2), neuroblastoma (n = 2), Ewing's tumour/PNET (n = 2), non-Hodgkin's lymphoma (n = 1), and osteosarcoma (n = 2), treated by surgery (8/9), chemotherapy (9/9), intensive chemotherapy and bone-marrow transplantation (5/9), and radiotherapy (7/9). Three children suffered veno-occlusive disease (VOD) during treatment. The hepatic assessment was performed by sonography (8/9), Doppler (7/9), multiphase spiral CT (8/9) and MRI (7/9). RESULTS: Lesions were discovered 15 months to 16 years after completing treatment. CT was the most sensitive modality for diagnosis. Lesions were multiple in eight cases, measured 2-50 mm, and appeared hypervascular on the arterial phase of CT and/or MRI in every case. Metastases were excluded on the basis of histological verification (n = 2) and clinical and radiological follow-up. CONCLUSION: Pseudometastatic hypervascular hepatic nodules can appear after treatment of a malignant tumour in children. The hypothesis of benign regenerative lesions secondary to treatment and/or VOD is considered.

[Current treatment of retinoblastoma. 153 children treated between 1995 and 1998]. / Le traitement actuel du rétinoblastome. 153 enfants traités entre 1995 et 1998.

PURPOSE: Treatment of retinoblastoma has changed significantly over the past few years. There are fewer indications for external beam radiation and a new treatment modality, chemotherapy, has appeared. MATERIAL: and methods: We reviewed a series of 153 children treated for retinoblastoma between 1995 and 1998. There were 67 boys and 86 girls: 76 unilateral and 77 bilateral retinoblastomas. Indications for treatments and outcome were reviewed for 230 eyes and for each tumor. Age at diagnosis varied from 0 to 94 months with a median age of 12 months. A family history of retinoblastoma was found in 24 cases. Three children were seen for treatment of recurrence. RESULTS: Among the 76 cases of unilateral retinoblastoma, 56 were enucleated and 20 were treated conservatively (5 with external beam radiation). Among the 154 eyes with bilateral retinoblastoma, 48 were enucleated and 106 were treated conservatively (49 by external beam radiation). Local treatments included chemothermotherapy, laser alone, cryotherapy, and (125)I plaques. Conservative management other than external beam radiation was used for 81 eyes and was successful in 72 (89%). CONCLUSION: We discuss the indications and results of primary chemotherapy and local treatments. External beam radiation is still often indicated in bilaterally advanced cases. In other forms of retinoblastoma, chemothermotherapy is a very reliable and useful treatment.

LMCE3 treatment strategy: results in 99 consecutively diagnosed stage 4 neuroblastomas in children older than 1 year at diagnosis.

PURPOSE: To tailor postinduction therapy for stage 4 neuroblastoma in children who are older than 1 year at diagnosis according to status after induction. PATIENTS AND METHODS: From March 1987 to December 1992, 99 patients who were consecutively admitted were included in the Lyon-Marseille-Curie East of France (LMCE)3 strategy. After induction with the French Society of Pediatric Oncology NB87 regimen and surgery, patients who were in complete remission immediately proceeded to consolidation therapy with vincristine, melphalan, and fractionated total-body irradiation (VMT). All other patients underwent a postinduction strategy before VMT, either an additional megatherapy regimen or further chemotherapy with etoposide/carboplatin. RESULTS: The progression-free survival (PFS) is 29% at 7 years from diagnosis, which compares favorably with that of a similar cohort of 72 patients previously reported by our group (LMCE1; PFS of 20% at 5 years and 8% at 14 years, P =.004). In the multivariate analysis, only age younger than 3 years at diagnosis (P =.0085) and achievement of complete or very good partial remission after NB87 and surgery (P =.00024) remained significant. The PFS of the 87 patients who were included in the postinduction strategy was significantly better than that of the comparable 62 patients on the LMCE1 study (32% v 11% at 7 years; P =.005). CONCLUSION: The progressive improvements in the LMCE results over the last 10 years suggest that improvements in supportive care measures and increases in each component of this strategy (induction, postinduction, consolidation) may all contribute to increased survival rates.

Adult height after growth hormone (GH) treatment for GH deficiency due to cranial irradiation.

BACKGROUND: The indications and factors affecting the growth in response to treatment with growth hormone (GH) of patients with cranial irradiation-induced GH deficiency remain unclear. PROCEDURE: The adult heights of 56 patients treated with GH (0.4-0.6 U/kg/week) as daily sc injections were analysed. They had been given 18 or 24 Grays (Gy) cranial irradiation for leukemia (group 1, 26 cases), 50 +/- 1 Gy for various tumors (group 2, 13 cases), 46 +/- 1 Gy for retinoblastoma (group 3, 8 cases), or 34 +/- 2 Gy with spinal irradiation for medulloblastoma (group 4, 9 cases). Twenty- five of these 56 patients had early puberty and were also treated with gonadotropin-releasing hormone (GnRH) analog. RESULTS: The standing (-1.0 +/- 0.2 in group 1, -0.7 +/- 0.3 in group 2, -1.1 +/- 0.3 in group 3, and -2.0 +/- 0.4 SD in group 4) and sitting (-1.8 +/- 0.2 in group 1, -0.4 +/- 0.4 in group 2, -1.2 +/- 0.4 in group 3, and -3. 4 +/-0.4 SD in group 4) adult heights were shor ter (P < 0.05 for standing and P < 0.001 for sitting heights) for group 4 than for each of the other groups. Of the 47 patients given cranial (and not craniospinal) irradiation, sitting adult height was shorter (P = 0. 02) and the difference between standing adult and target heights greater (P = 0.03) in those patients in whom puberty occurred at a normal age than in those treated with GnRH analog. Conclusion. The incomplete catch-up of growth seems to be mainly due to the reduction in sitting height of patients given spinal irradiation and in whom puberty occurred at a normal age. This suggests that GnRH analog treatment should be more widely used to treat children with early and/or rapidly progressing puberty after cranial irradiation.

Thyroid carcinomas after irradiation for a first cancer during childhood.

BACKGROUND: The thyroid gland is among the most radiosensitive organs. However, little is known about the long-term risk of developing a thyroid tumor after fractionated external radiotherapy for cancer during childhood. OBJECTIVE: To study the long-term risk of developing a thyroid tumor in 4096 three-year survivors of childhood cancer treated between May 1942 and December 1985 in 8 centers in France and the United Kingdom, 2827 of whom had received external radiotherapy. METHODS: A wide range of radiation doses were given to the thyroid: 1164 children received less than 0.5 Gy and 812 received more than 5.0 Gy, the average dose being 7.0 Gy. RESULTS: After mean follow-up of 15 years (range, 3-45 years), 14 patients-all of whom had received radiotherapy-developed a clinical thyroid carcinoma. Within the cohort, the relation between radiation dose to the thyroid and risk of thyroid carcinoma and adenoma was similar to that observed in patients who received radiotherapy during childhood for other reasons, such as an excess relative risk per gray of 4 to 8, up to a few gray. In contrast, compared with thyroid cancer incidence in the general population, the standardized incidence of thyroid carcinoma was much higher than expected from the dose-response relationship estimated within the cohort and from patients who received radiotherapy during childhood for other reasons: a dose of 0.5 Gy was associated with a standardized incidence ratio of 35 (90% confidence interval, 10-87) and a dose of 3.6 Gy with a standardized incidence ratio of 73 (90% confidence interval, 28-153). We did not show a reduction in excess relative risk per gray with use of an increasing number of fractions. CONCLUSION: Although we cannot estimate the exact proportion, it is probable that some or all children who are treated for cancer are predisposed to developing a thyroid carcinoma.