Data Resource Profile: The French Childhood Cancer Observation Platform (CCOP)

Data resource basics

Rationale and objectives

Childhood cancers are rare diseases, substantially different from adult cancers. They most often develop from embryonal or undifferentiated cells and carcinomas only occur exceptionally. Almost half of childhood cancers occur before the age of 5 years. Intensive protocols used in young people are very effective, altogether resulting in 5-year overall survival to date of ∼80%.^1–3 Several dozen thousands of people in France have had a cancer in childhood, sometimes with sequelae, and with various consequences for their education, professional life, lifestyle and overall quality of life (QoL).^4–13 Furthermore, large childhood cancer survivors cohorts in North America and Europe^14–16 have shown a risk of late adverse effects.^17–21 Former patients are exposed to second cancers^22–26 and adverse chronic disease conditions (such as cardiovascular,^27–37 cerebrovascular^38–40 and pulmonary diseases,^41–43 endocrine and metabolic disorders,^44–48 and neurocognitive disorders^49,50) which may occur decades after the end of the treatment. There is thus an essential need for a systematic large-scale and comprehensive long-term surveillance of former childhood cancer cases, in the context of constantly evolving health care management. That is why we have set up the Childhood Cancer Observation Platform (CCOP), a nationwide infrastructure based on the French National Registry of Childhood Cancers (RNCE)³ and medico-administrative databases. Launched in 2013 under the auspices of the French paediatric oncology society (SFCE), the CCOP integrates a large range of standardized data from multiple sources to support observational and interventional research in paediatric oncology. Acronyms used in this paper are listed in Table 1.

Data resource area and population coverage

The CCOP includes all incident cancer or intracranial tumour cases diagnosed in those under 15 years of age in mainland France registered in the RNCE since 2000 (∼1700 new cases per year), amounting to 29 747 cases over the 2000–2016 period, of which 29 546 were primary tumours. A 2-year lag time is necessary for full registration in the RNCE, which means that 2016 is the last complete year in the tables herein. There is no mandatory reporting of childhood cancers in France. RNCE data are manually collected from the medical records by trained clinical research assistants in all French paediatric haematology and oncology departments (∼30 centres). In order to recover missing cases treated in other departments or deceased before any specialized treatment, an annual request is performed for any child diagnosed with a cancer or intracranial tumour from the Information System Medicalization Programme (ISMP) of the University Hospitals and Cancer Centres, from the National health insurance long-term diseases (LTD) files and from the National Registry of medical Causes of Death (CepiDC). The completeness of the RNCE cannot be quantified (e.g. by the capture recapture method) as all valid and reliable sources of information available contribute to RNCE and are therefore not independent sources. The number and quality of sources constitute our guarantee of completeness.

The RNCE provides the CCOP with baseline and basic follow-up data, including administrative data, data on initial diagnosis and treatment, relapses and second childhood cancers before the age of 18 years, vital status and causes of death. Four following sets of data are added in the CCOP (Figure 1).

Figure 1

Structure and progress of the Childhood Cancer Observation Platform (CCOP).

• The GEOCAP database (GEOlocation of Pediatric CAncers)

GEOCAP includes the spatial coordinates of the residences at the time of diagnosis and, since 2002, that of contemporaneous controls representative of the French paediatric population (5000 controls per year). Addresses are geocoded by the GEOCIBLE company, using the HERE^® mapping company street databases and the BD ADRESSE^® database from the National Geographic Institute (IGN). We use a geographic information system (GIS) to evaluate residential proximity to sources of environmental and residential exposures. GIS-based exposure assessment relies on models developed and validated by experts in each field. Deprivation scores are also assigned to each residence based on sociodemographic data from the National institute for statistical and economic studies (INSEE) at the smallest census unit in which the residence is located. GEOCAP is used for surveillance and aetiological research on environmental factors, and to appreciate territorial and sociodemographic disparities in incidence and survival. GEOCAP has complete GIS-based data for about 29 500 cases and 60 000 controls. Less than 1% of the addresses are missing.

• The virtual biobank BIOCAP (BIObank of Pediatric CAncers, BB-0033-00086)

BIOCAP aims to provide standardized information on tumour and non-tumour material stored in the hospital biobanks for childhood cancer cases diagnosed from 2011 onwards (since 2009 for the Ile-de-France region), with clinical annotations on diagnosis and follow-up. BIOCAP will eventually support basic and translational research on mechanisms involved in carcinogenesis and tumour spread, and contribute to identifying aetiological and prognostic markers and leads for treatment. Several dozen biobanks contribute to BIOCAP, after local authorizations at various levels (biobank, hospital, clinical research department, paediatric oncology department). Data transfer complies with the legal requirements regarding privacy and security, which are specified in local data transfer agreements. The raw files of the biobanks are securely transferred to the CCOP each year, harmonized and crosschecked with the RNCE to identify the cases, tumours and events associated with each specimen. An interface has been developed to provide investigators with tabulated information on specimens in response to requests relating to clinical and follow-up items. Biobanking data have been collected for 19 698 specimens identified as corresponding to 4944 cases diagnosed over 2009–2016, with at least one tumour specimen (42.9% of the eligible cases) stored in the hospital biobanks. Use in research projects will depend on the real availability of the materials and the consent given for the specimens. The distribution by gender, age at diagnosis and diagnostic group is similar to that of the whole RNCE for 2000–2016 (Table 2). Ile-de-France, as a pilot region between 2009 and 2011, is over-represented.

Table 2.

Characteristics of the cases currently included in the Childhood Cancer Observation Platform (primary tumours in those under 15 years of age, mainland France, 2000–2016)

	Data collection
	Total-baseline (2000–2016)		5-Year follow-up (2000–2012)		PediaRT (2013–2016)		BIOCAP (2009–2016)
	n	%	n	%	n	%	n	%
Total eligible	29 546	–	22 611	–	1743a	–	11 534	–
Total collected	29 546	100.0	11 511	51.0	641	36.8	4944	42.9
Year of diagnosis
2000–2004	8542	28.9	2495	21.7	−	−	−
2005–2008	6934	23.5	4701	40.8	−	−	−
2009–2012	7135	24.1	4315	37.5	−	−	2438	49.3
2013–2016	6935	23.5	–		641	100.0	2506	50.7
Region of residence at diagnosis
Nord-Ouest	3168	10.7	949	8.2	70	10.9	540	10.9
Ouest	4493	15.2	2374	20.6	147	22.9	205	4.1
Sud-Ouest	3164	10.7	1531	13.3	78	12.2	546	11.0
Ile-de-France	8570	29.0	3317	28.8	92	14.4	2238	45.3
Est	3221	10.9	1084	9.4	102	15.9	699	14.1
Rhône-Alpes-Auvergne	3740	12.7	1604	13.9	62	9.7	508	10.3
Sud-Est	3190	10.8	652	5.7	90	14.0	208	4.2
Category of diagnosis
Leukaemias	8402	28.4	3196	27.8	18	2.8	1414	28.6
Lymphomas	3177	10.8	1207	10.5	48	7.5	444	9.0
Central nervous system tumours	7356	24.9	2527	22.0	284	44.3	1344	27.2
Sympathetic nervous system tumours	2389	8.1	1285	11.2	68	10.6	369	7.5
Retinoblastomas	850	2.9	112	1.0	≤5	–	144	2.9
Renal tumours	1628	5.5	746	6.5	65	10.1	333	6.7
Liver tumours	296	1.0	119	1.0	≤5	–	68	1.4
Bone tumours	1417	4.8	558	4.8	56	8.7	282	5.7
Soft-tissue sarcomas	1857	6.3	839	7.3	70	10.9	323	6.5
Germ-cell tumours	1103	3.7	597	5.2	20	3.1	101	2.0
Carcinomas	1002	3.4	298	2.6	10	1.6	114	2.3
Other	69	0.2	27	0.2	≤5	–	8	0.2
Gender
Male	16 139	54.6	6266	54.4	368	57.4	2688	54.4
Female	13 407	45.4	5245	45.6	273	42.6	2256	45.6
Age at diagnosis
<1 year	3078	10.4	1381	12.0	9	1.4	476	9.6
1–4 years	10 136	34.3	4004	34.8	190	29.6	1751	35.4
5–9 years	7809	26.4	3016	26.2	239	37.3	1328	26.9
≥10 years	8523	28.8	3110	27.0	203	31.7	1389	28.1
Relapse or second cancer in childhood	–	–	2315	20.1	–	–	–	–
Deaths	5416	18.3	2182	19.0	197	30.7	842	17.0
Current age (surviving cases)
<18 years	13 853	46.9	5042	54.0	414	93.2	3460	84.3
18–24 years	7248	24.5	3266	35.0	30	6.8	642	15.7
≥25 years	3029	10.3	1021	10.9	–	–	–	–
Time since diagnosis (surviving cases)
<5 years	4697	19.5	−	−	361	81.3	1584	38.6
5–9 years	7424	30.8	3734	40.0	83	18.7	2518	61.4
10–14 years	7035	29.2	4231	45.4	–	–	–	–
≥15 years	4973	20.6	1364	14.6	–	–	–	–

a

Estimation based on 2000–2012 period.

• Treatments performed and treatment responses

A systematic collection of detailed treatments and responses to treatments is ongoing for all cases. These data are collected 5 years after diagnosis by active search in the medical records of departments where the children were treated. They are used for second validation, complementation and enhanced specification of the relapses and second cancers reported in the RNCE. The systematic 5-year follow-up began in 2018 for the cases diagnosed since 2013. For the cases diagnosed before 2013, data are collected retroactively, first for the cases diagnosed in 2005–2012 and then for the cases diagnosed in 2000–2004. So far, the data are complete for 64.1% of the 2005–2012 cases and 29.2% of the 2000–2004 cases, giving a total of 51.0% completion for 2000–2012 (11 511 of 22 611 cases). Table 2 summarizes the main characteristics of the 2000–2012 cases for which data collection is complete, as well as those of all 2000–2016 RNCE cases. The distribution by age at diagnosis, period of diagnosis and diagnostic group reflects the current priorities in retroactive data collection; sympathetic nervous system tumours and germ-cell tumours are slightly over-represented. Figure 2 shows the progress of the retroactive data collection by diagnostic group and period of diagnosis. In all, 2315 cases (20.1%) have had a relapse or second cancer and 2182 (19.0%) had died by 31 December 2018. Additional information on radiotherapies for cases diagnosed since 2013 are provided by a specific database compiled by the French network of paediatric radiotherapy units (PediaRT). The cumulative doses for 641 cases (estimated as 36.8% of the 2013–2016 cases who received radiotherapy) have been provided by PediaRT to date. The cases are described in Table 2. Collection is more advanced in some regions. As expected, central nervous system tumours, bone tumours and soft-tissue sarcomas are more represented and deaths are more frequent in this subgroup of cases who received radiotherapy.

Figure 2

Current inclusions in the Childhood Cancer Observation Platform—data collection in medical records 5 years after diagnosis, by diagnosis and period of diagnosis (primary tumours in those under 15 years of age, mainland France, 2000–2012).

• The COHOPER (COHOrt of the PEdiatric cancer Registry) cohort

COHOPER is the key component of the CCOP. Its aims are (i) to evaluate the risk of late complications and (ii) to evaluate lifestyle and individual perceptions of health, with a view to formulating preventive and follow-up recommendations. The cohort ensures systematic follow-up of cases diagnosed since 2000. Relapses and second cancers before the age of 18 years are provided by the RNCE (see above). Health events in adulthood (relapses, second cancers, disabilities and diseases possibly related to treatments) are detected through an annual individual linkage with the National health insurance information system (SNDS) database. The SNDS comprises comprehensive individual data concerning all detailed reimbursements for outpatient and hospital care to the main health insurance schemes beneficiaries (96% of the French population).⁵¹ A mapping of the diseases of beneficiaries and their related health care expenditure⁵² will help compile algorithms for the identification of health events. In addition, adult survivors are eligible to complete online questionnaires on sociodemographic, occupational, lifestyle, health and QoL items at least 5 years after diagnosis. They will be contacted for the first time in fall 2020. Common survey modules will allow comparison of the respondents to the same-aged contemporaneous participants in the French general cohort CONSTANCES.⁵³ To date, 24 130 cases are alive and eligible for long-term follow-up; 10 277 are adults and 10 054 of them were diagnosed at least 5 years ago and are therefore eligible for the administration of questionnaires. Non-respondents and respondents will be compared with regard to baseline data and systematic follow-up information.

Methods used to recruit participants

Most baseline and follow-up data are obtained from medical records and preexisting medico-administrative databases, and active individual participation is not required, except for the questionnaire part of COHOPER. All cases are eligible, provided they have not expressed their opposition to registration in RNCE and CCOP data sets. Information is given to parents in the hospital at diagnosis; parents may opt out of inclusion of their child in RNCE and CCOP (only 10 did so thus far). Once former patients have become adults, they are informed by letter that their data have been included in the RNCE and CCOP databases for research and surveillance purposes, and that their medico-administrative data will be used to ensure follow-up. In compliance with the European General Data Protection Regulation (GDPR 2016/679),⁵⁴ former patients are informed of our privacy policy relating to personal data protection, and their right to access their personal data, rectify them, and object to their collection or use. This letter is about to be sent and it is not yet possible to assess the number of adult survivors who oppose the collection or use of their data.

Frequency of data collection

Vital status is updated every 2 years from the National Registry for the Identification of Individuals (RNIPP) and causes of death are collected from the National Registry of medical Causes of Death (CepiDC). Linkages with PediaRT and BIOCAP databases are annual. Systematic follow-up from medical records is performed continuously, with relapses and second cancers registered at the time of their occurrence and 5 years after the diagnosis. Linkage with the SNDS is annual from 2019 onwards. Online-questionnaires are planned every 2 years depending on financial resources.

Data collected

The catalogue of data is summarized in Table 3.

Table 3.

Summary of the catalogue of variables in the Childhood Cancer Observation Platform

a

ADICAP, Association for development of information in cytology and pathology; DNA, deoxyribonucleic acid; FISH, fluorescence in situ hybridization; RNA, ribonucleic acid.

Baseline RNCE data include administrative data (identification number, date and place of birth, gender, address at diagnosis); data on initial diagnosis [date, tumour characteristics, diagnostic codes using the WHO (World Health Organisation) international classifications (International Classification of Diseases for Oncology, 3^rd edition (ICDO-3), International Classification of Childhood Cancer, 3rd edition (ICCC-3),⁵⁵ International Classification of Diseases, 10th edition (ICD-10)), stage of extension (following Toronto recommendations⁵⁶)] and data on initial treatment (treatment protocols, episodes of surgery, chemotherapy, radiotherapy, haematopoietic stem cell transplantation and care pathway). Routine information on family and personal medical history and genetic predisposition syndromes is systematically collected.

The GEOCAP database provides the geocoded residential addresses of controls and cases at diagnosis, with allocation of contextual census data (deprivation, degree of urbanization) on the scale of the smallest census units, and indicators of some environmental exposures (roads, power lines, agricultural activities, industrial plants, air pollutants or background ionizing radiation).

BIOCAP data include for each material registered: date and sampling method, origin of sample, type of sample, method of storage, existence of derivatives, and existence of signed consent forms appended to the specimens. Table 4 reports the sampling methods by diagnosis for the 4944 cases of 2009–2016 with at least one tumour specimen registered in BIOCAP. For the 1858 haematopoietic malignancies, cytology was reported for 1398 cases (75.2%), biopsy for 279 cases (15.0%) and surgical sample for 228 cases (12.3%). For the 3086 solid tumours, surgical sample was reported for 2442 cases (79.1%), biopsy for 802 cases (26.0%) and cytology for 341 cases (11.0%).

Table 4.

Treatments and samples of the Childhood Cancer Observation Platform participants

			Haematopoietic malignancies				Solid tumours
	Total		Leukaemias		Lymphomas		CNS tumours\!\!a		SNS tumours\!\!a		Retinoblastomas		Renal tumours		Liver tumours		Bone tumours		Soft-tissue sarcomas		Germ-cell tumours		Carcinomas		Other
	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%
Detailed treatments
Chemotherapy	8807	76.5	3147	98.5	1151	95.4	1108	43.8	867	67.5	76	67.9	710	95.2	110	92.4	520	93.2	708	84.4	340	57.0	49	16.4	21	77.8
Surgery	5991	52.0	12	0.4	222	18.4	1789	70.8	1092	85.0	83	74.1	729	97.7	113	95.0	498	89.2	664	79.1	508	85.1	255	85.6	26	96.3
Radiotherapy	2864	24.9	256	8.0	353	29.2	814	32.2	350	27.2	≤5	–	260	34.9	6	5.0	173	31.0	444	52.9	158	26.5	46	15.4	≤ 5	–
Allogenic SCTa	451	3.9	387	12.1	18	1.5	≤5	–	16	1.2	≤5	–	≤ 5	–	≤5	–	12	2.2	≤ 5	–	≤ 5	–	≤ 5	–	≤ 5	–
No treatment reported	470	4.1	41	1.3	14	1.2	304	12.0	60	4.7	≤5	–	7	0.9	≤5	–	14	2.5	8	1.0	11	1.8	7	2.3	≤ 5	–
Totalb	11511		3196		1207		2527		1285		112		746		119		558		839		597		298		27
Type of radiotherapy procedure (PediaRT)
IMRTa	192	30.0	≤5	–	8	16.7	96	33.8	16	23.5	≤5	–	15	23.1	≤5	–	15	26.8	25	35.7	11	55.0	≤5	–	≤5	–
Conformal 3D radiotherapy	180	28.1	14	77.8	16	33.3	58	20.4	29	42.6	≤5	-	35	53.8	≤5	–	13	23.2	13	18.6	≤5	–	≤5	–	≤5	–
VMATa	123	19.2	≤5	–	10	20.8	60	21.1	11	16.2	≤5	–	≤5	–	≤5	–	18	32.1	15	21.4	≤5	–	≤5	–	≤5	–
Tomotherapy	110	17.2	≤5	–	14	29.2	53	18.7	9	13.2	≤5	–	9	13.8	≤5	–	6	10.7	11	15.7	≤5	–	≤5	–	≤5	–
Other radiation therapy	36	5.6	≤5	–	≤5	–	17	6.0	≤ 5	–	≤5	–	≤5	–	≤5	–	≤5	–	6	8.6	≤5	–	≤5	–	≤5	–
Totalb	641		18		48		284		68		≤5		65		≤5		56		70		20		10		≤5
Tumour sample at diagnosis (BIOCAP)
Puncture/bone marrow aspiration	1739	35.2	1334	94.3	64	14.4	46	3.4	133	36.0	≤5	–	37	11.1	24	35.3	40	14.2	50	15.5	≤5	–	≤5	–	≤5	–
Biopsy	1081	21.9	61	4.3	218	49.1	240	17.9	141	38.2	≤5	–	46	13.8	32	47.1	140	49.7	146	45.2	23	22.8	33	29.0	≤5	–
Surgical sample	2670	54.0	24	1.7	204	46.0	1216	90.5	250	67.8	11	7.6	307	92.2	62	91.2	188	66.7	230	71.2	83	82.2	88	77.2	7	87.5
Other type of sample	434	8.8	46	3.3	63	14.2	23	1.7	43	11.7	134	93.1	34	10.2	13	19.1	16	5.7	50	15.5	≤5	–	8	7.0	≤5	–
Totalb	4944		1414		444		1344		369		144		333		68		282		323		101		114		8

a

CNS, Central nervous system; IMRT, intensity modulated radiation therapy; SCT, stem cell transplantation; SNS, sympathetic nervous system; VMAT, volumetric modulated arc therapy.

b

Cases can contribute to more than one line.

Treatment data include dates and locations of health care delivery units for each therapeutic procedure conducted within 5 years of diagnosis, cumulative doses of each chemotherapy drug, surgical site and result of post-surgical histopathology, type of haematopoietic stem cell transplant, site of radiotherapy and response to each line of treatment. The treatments reported for the 11 511 cases diagnosed in 2000–2012 and retrospectively followed-up at 5 years are shown by diagnosis in Table 4. Almost all 4403 haematopoietic malignancies were treated by chemotherapy (4298 cases, 97.6%); allogenic stem cell transplantation was reported for 405 cases (9.2%) and radiotherapy for 609 cases (13.8%); no treatment was reported for 55 cases (1.2%). Of the 7108 solid tumours, surgery was reported for 5757 cases (81.0%), chemotherapy for 4509 cases (63.4%) and radiotherapy for 2255 cases (31.7%); no treatment was reported for 415 cases (5.8%), mainly central nervous system tumours. For the cases diagnosed since 2013, the PediaRT database provides cumulative radiation doses, minimum, maximum and median doses delivered to the target organs and the organs at risk. The types of radiotherapy administered are reported by diagnosis in Table 4.

COHOPER follow-up data include status of the disease 3 years after diagnosis; relapses and second cancers before the age of 18 are described in the same manner as initial tumours in terms of diagnosis, treatment and response to treatment. The RNIPP and the CepiDC provide vital status and causes of death. The SNDS provides indicators of precariousness (e.g. full reimbursement for low income earners); medical information (full reimbursement for certain LTDs, allowances for sick leave, disability pensions, work accident compensation, occupational disease compensation); reimbursed outpatient health care consumption (nature and date of all reimbursed medical or paramedical services with the corresponding codes in line with the French specific nomenclatures); and reimbursed hospital health care consumption (dates of hospitalization, discharge diagnoses and diagnosis-related groups, certain medical procedures).⁵¹ Causes of death, LTD and hospital diagnoses are coded using ICD-10. A self-administered online questionnaire provides information on sociodemographic and occupational characteristics (social position, educational level, disruptions of educational trajectory, employment status, marital status, household composition, income level, difficulties getting loans or suitable insurance cover), health-related behaviours (physical activity, eating habits, risky consumptions), self-reported health-scales [perceived health, QoL (the Medical Outcomes Study Short Form 12-item questionnaire (SF-12)^57,58), mental health (the 12-item General Health Questionnaire (GHQ-12)^59,60) (the latter two QoL instruments have been validated, translated and cross-culturally adapted to French-speaking patients)], self-reported impairments, limitations and disabilities, sex and reproductive life, personal disease history and screening examinations conducted since the end of cancer treatment, health-care usage and use of alternative medicines.

How is it funded?

The French health authorities are strongly committed to endowing the community with a comprehensive observational infrastructure for childhood cancers, as stated in the 2014–2019 third National Cancer Plan. The CCOP has been funded by the French national institute of cancer (INCa) and the National agency of research (ANR) (governmental ‘investment for the future’ program, ANR-10-COHO-0009 grant) under the name HOPE-EPI for 10 years. The SFCE has obtained a grant from the charitable organization ‘Enfants, cancers et santé’ to support the CCOP virtual biobank component BIOCAP. Data related to environmental exposures are collected with the support of the French National Public Health Agency [Santé Publique France (SPF)] and the French Agency for Food, Environmental and Occupational Health Safety (ANSES). Lastly, the National Registry of Childhood Hematopoietic Malignancies and the National Registry of Childhood Solid Tumours, that constitute the RNCE, benefit from perennial specific financing by INCa and SPF. The CCOP team reports to the French Institute of Health and Medical Research (INSERM).

Data resource use

Since the CCOP started in 2013, with retrospective inclusions from 2000 onwards, it has yet to generate publications relating to therapies and follow-up. However, the CCOP has already provided health agencies, health authorities and the pharmaceutical industry with estimates of the real-world risk of relapse and death. The CCOP has been used to identify and quantify specific refractory subpopulations which may be targeted by new drugs coming onto the market (e.g. car-T cells in childhood B-cell precursor lymphoblastic leukaemia).

The GEOCAP component of the CCOP started earlier and has published results. These include the absence of an association with exposure to nuclear plant gaseous discharges,⁶¹ a higher risk of childhood leukaemia close to high-voltage power lines⁶² and, for myeloid leukaemias, major roads,⁶³ a higher risk of childhood leukaemia (pre-B cell) with high residential exposure to natural UV radiation, the absence of an association between leukaemia and background ionizing radiation,⁶⁴ and a lower risk of pre-B-cell leukaemia in most deprived municipalities.⁶⁵

The CCOP is currently contributing to research projects in the fields of aetiological research (environmental factors, perinatal factors, genetic predisposing factors), observational clinical research (relapse-free survival in the real world, prognostic factors in the real world, evaluation of specific practices in leukaemia management, reduction of late effects of radiotherapy, QoL and disabilities after bone surgery), intervention research on secondary prevention [use of MOOCs (Massive Open Online Courses) to educate patients on late effects of treatments]. The CCOP is involved in an ENCCA (European Network for Cancer Research in Children and Adolescents) project on relapse-free survival after infant neuroblastoma, linked to the European Union JARC (Joint Action on Rare Cancers).

What are the main strengths and weaknesses?

A major asset of the CCOP is its permanent systematic data collection, shared with the RNCE, enabling low-cost timely reactive mobilization for research and surveillance. The CCOP’s background work involves homogenous continuous quality control practices. The regular linkage with the SNDS ensures prospective follow-up of medico-administrative data with optimal statistical power. The comprehensive set of standardized data collected for each individual is formatted for multidisciplinary research. Being registry-based ensures the precision and quality of diagnosis coding, completeness, and the ability to compare respondents and non-respondents. The CCOP thus constitutes a resource for health agencies and health authorities, as has already been demonstrated in the contexts of compiling regulatory submissions and addressing environmental alerts. The CCOP generates reliable information on chemotherapy and radiotherapy for observational clinical research and provides a baseline for the design of clinical interventions. The CCOP has invested time and resources to ensure compliance with the European GDPR.⁵⁴ The SNDS medico-administrative databases constitute an exceptional resource for the identification of diseases. However, judicious exploitation of the considerable quantity of data necessitates the cautious design of sophisticated algorithms. Pertinent skills are increasingly being developed by the multidisciplinary collaborative SNDS users’ network (REDSIAM) and the French health agencies, which will be an asset for CCOP users.

The CCOP also has weaknesses. First, the cohort is young, and optimal power for long-term follow-up has yet to be reached. At present, the follow-up is more than10 years for half of the cohort, but more than15 years for only a quarter of the cohort. Some internal and external projects have, however, already started in observational clinical research and intervention research (on prognostic factors in the real world and influence of socio-economic factors, on QoL and disabilities after treatments, and on secondary prevention after treatments and reduction of late effects after treatments). Second, another current limitation is the absence of clinical consultation data and ad hoc biological data; in addition, research on sociodemographic, occupational and perceived health fields has yet to be developed. A better knowledge of all the consequences of childhood cancers, not only clinical consequences, is a necessary step towards improved care and secondary prevention. Former patients are indeed vulnerable to a wide range of difficulties affecting their later QoL, making it essential to evaluate lifestyle and individual perceptions of health in this population. These issues for the most part cannot be covered by medico-administrative databases and require specific surveys. COHOPER questionnaire surveys of adult survivors will thus contribute to study the long-term physical, psychosocial and economic consequences of childhood cancer well after its treatment and cure. Like all longitudinal cohorts, the survey will be based on the active participation of volunteers and therefore subject to selection and attrition bias due to low participation rates. Survey participants and non-participants will be compared for their baseline characteristics, their health and their healthcare usage, at inclusion and during the follow-up, using RNCE and SNDS data. Integration of different sources of data will allow precise analysis of health events in the real world, and their relationships with personal, social, occupational and environmental factors.

In conclusion, the CCOP constitutes a powerful infrastructure for the support of research, health surveillance and public health decision making in the field of childhood cancers.

Data resource access

Only projects collaborative with French paediatric oncologists and the CCOP team can use the clinical data. To access the data, researchers must express their interest through the CCOP website (instruction at http://rnce.inserm.fr). Data will be opened in October 2020.

Profile in a nutshell

• The Childhood Cancer Observation Platform (CCOP), based on the French National Registry of Childhood Cancers (RNCE), was launched in 2013 to permanently follow up all primary childhood cancers diagnosed in those under 15 years of age in mainland France since 2000. Its objective is to support clinical and epidemiological research, health surveillance and public health decisions.

• To date, 29 546 cases are included and eligible for passive systematic follow-up through a linkage with the National health insurance information system (SNDS). Follow-up from medical records 5 years after diagnosis is being completed (11 511 cases collected), and online questionnaires will be proposed to adult five-year survivors (10 054 eligible cases).

• Besides baseline and basic follow-up RNCE data, the CCOP includes spatial coordinates of the residences at diagnosis, allowing allocation of deprivation and environmental exposures indicators (GEOCAP database, for cases diagnosed since 2000 and contemporaneous control since 2002), information on specimens stored in the hospital biobanks (BIOCAP database, for cases diagnosed since 2009), data on treatments, including doses of radiotherapy (PediaRT database, for cases diagnosed since 2013), responses to treatment, information on health care consumption (data from SNDS) and social, lifestyle and QoL issues (data from questionnaires).

• Calls to projects will be released annually.

Acknowledgements

We are particularly grateful to the French paediatric oncology society (SFCE) and Prof. François Doz and Prof. Yves Perel, who were the SFCE chairs when CCOP started, for their support of the CCOP protocol and for obtaining a grant for BIOCAP.

We would like to thank all the clinical research assistants who collected CCOP data from medical files, and we warmly thank the heads and staff of the paediatric hematology and oncology departments for having hosted them: Dr Catherine Devoldere (Centre Hospitalier Universitaire, Amiens), Prof. Alain Martinot and Dr Brigitte Nelken (Centre Hospitalier Régional Universitaire, Lille), Dr Anne-Sophie Defachelles (Centre Oscar Lambret, Lille), Prof. Pascale Schneider (Centre Hospitalier Universitaire, Rouen), Prof. Isabelle Pellier (Centre Hospitalier Universitaire, Angers), Dr Loïc De Parscau and Dr Liana Carausu (Centre Hospitalier Régional Universitaire, Brest), Dr Damien Bodet and Dr Odile Minckes (Centre Hospitalier Universitaire, Caen), Dr Caroline Thomas (Centre Hospitalier Universitaire, Nantes), Prof. Frédéric Millot (Centre Hospitalier Universitaire, Poitiers), Prof. Virginie Gandemer (Centre Hospitalier Universitaire, Rennes), Prof. Philippe Colombat (Centre Hospitalier Régional Universitaire, Tours), Dr Laurence Joly and Dr Anne Notz-Carrere (Centre Hospitalier Universitaire, Bordeaux), Prof. Anne Lienhardt-Roussie and Dr Christophe Piguet (Centre Hospitalier Universitaire, Limoges), Dr Marie-Pierre Castex (Centre Hospitalier Universitaire, Toulouse), Prof. André Baruchel (Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpital Saint-Louis and Hôpital Robert Debré, Paris), Prof. Guy Leverger (AP-HP, Hôpital Armand Trousseau, Paris), Prof. Michel Zerah (AP-HP, Hôpital Necker-Enfants Malades, Paris), Dr Daniel Orbach (Institut Curie, Paris), Dr Dominique Valteau-Couanet (Institut Gustave Roussy, Villejuif), Dr Daniel Amsallem and Dr Véronique Laithier (Centre Hospitalier Universitaire, Besançon), Prof. Frédéric Huet and Dr Elodie Bottolier-Colomb (Centre Hospitalier Universitaire, Dijon), Prof. Pascal Chastagner (Centre Hospitalier Régional Universitaire, Nancy), Prof. Christine Pietrement and Dr Claire Pluchart (Centre Hospitalier Universitaire, Reims), Prof. Catherine Paillard (Hôpitaux Universitaires de Strasbourg), Prof. Justyna Kanold (Centre Hospitalier Universitaire, Clermont-Ferrand), Prof. Dominique Plantaz (Centre Hospitalier Universitaire, Grenoble), Prof. Yves Bertrand, Dr Perrine Marec-Berard and Dr Christophe Bergeron [IHOPe (Institut of paediatric hemato-oncology), Hospices Civils de Lyon and Centre Léon Bérard, Lyon], Prof. Jean-Louis Stephan (Centre Hospitalier Universitaire, Saint-Etienne), Prof. Gérard Michel (Assistance Publique-Hôpitaux de Marseille, Hôpital de La Timone, Marseille), Prof. Nicolas Sirvent (Centre Hospitalier Universitaire, Montpellier), Prof. Pierre-Simon Rohrlich (Centre Hospitalier Universitaire-Fondation Lenval, Nice).

We would like to thank all the members of the French Group of paediatric radiotherapists (GFRP) who contributed to PediaRT data collection: Dr Alexandre Escande (Centre Oscar Lambret, Lille), Dr Charlotte Demoor Goldschmidt (Centre Hospitalier Universitaire, Angers), Dr Jean-Louis Habrand and Dr Dinu Stefan (Centre François Baclesse, Caen), Dr Julie Leseur (Centre Eugène Marquis, Rennes), Dr Emmanuel Jouglar and Dr Stéphane Supiot (Institut de cancérologie de l’Ouest, Saint-Herblain), Dr Sophie Chapet (Centre Hospitalier Régional Universitaire, Tours), Dr Aymeri Huchet (Centre Hospitalier Universitaire, Bordeaux), Prof. Anne Laprie and Dr Anne Ducassou (Institut universitaire du cancer de Toulouse), Dr Claire Alapetite, Dr Fatima Meniai-Merzouki and Dr Sylvie Helfre (Institut Curie, Paris), Dr Stéphanie Bolle and Dr Valentine Martin (Institut Gustave Roussy, Villejuif), Dr Gilles Truc (Centre Georges François Leclerc, Dijon), Dr Valérie Bernier-Chastagner (Institut de Cancérologie de Lorraine, Nancy), Dr Céline Vigneron (Centre Paul Strauss, Strasbourg), Dr Caroline Pasteris (Centre Hospitalier Universitaire, Grenoble), Dr Christian Carrie, Dr Line Claude and Dr Ronan Tanguy (Centre Léon Bérard, Lyon), Dr Xavier Muracciole and Dr Laetitia Padovani (Assistance Publique-Hôpitaux de Marseille, Hôpital de La Timone, Marseille), Dr Kerr Christine and Dr Druet Xavier (Institut du Cancer de Montpellier), Dr Pierre-Yves Bondiau and Dr Audrey Claren (Centre Antoine Lacassagne, Nice).

We are grateful to Prof. Hélène Cavé and Prof. Claude Preudhomme for their strong support for the BIOCAP project. We would like to thank Camille Boin for her valuable work in supervising BIOCAP for 3 years, and the pathology and biology departments and biobanks who provided the CCOP with biological resources information and devoted time to explaining their data: Centre de Ressources Biologiques (CRB), Institut Curie (BB-0033–000048, Paris); CRB, Institut Gustave Roussy (BB-0033–000074, Villejuif); CRB, Paris Sud (BB-0033–000089, Le Kremlin-Bicêtre); Plateforme de Ressources Biologiques (PRB), Hôpital Necker-Enfants Malades (BB-0033–000065, Paris); Tumorothèque, Hôpital Saint-Louis (BB-0033–000067, Paris); Biobanque, Hôpital Sainte-Anne (BB-0033–000026, Paris); Services d’anatomo-cytopathologie et hématologie, Hôpital Robert Debré (Paris); Services d’anatomo-cytopathologie et hématologie, Hôpital Armand-Trousseau (Paris); CRB, Centre Hospitalier Universitaire de Clermont-Ferrand (BB-0033–000039); CRB cancérologie, Centre Hospitalier Universitaire de Grenoble (BB-0033–000069); CRB, Hospices Civils de Lyon (BB-0033–000046); CRB institutionnel, Centre Léon Bérard (BB-0033–000050, Lyon); CRB, Centre Hospitalier Universitaire de Saint-Etienne (BB-0033–000041); CRB Cancer, Centre Hospitalier Universitaire de Bordeaux (BB-0033–000036); Tumorothèque, Centre Hospitalier Universitaire de Toulouse (BB-0033–000014); HIMIP (Cytothèque hémopathies malignes, BB-0033–000060, Toulouse); Biobanque, Picardie (BB-0033–000017, Amiens); Service d’anatomo-cytopathologie, Centre Hospitalier Universitaire d’Amiens-Picardie; CRB, Centre Hospitalier Universitaire Charles Nicolle (Rouen); Service d’anatomo-cytopathologie, Centre Hospitalier Régional Universitaire de Lille; TRFC (Tumorothèque Régionale de Franche-Comté, BB-0033–000024, Besançon); Service d’anatomo-cytopathologie, Centre Hospitalier Régional de Besançon; CRB Ferdinand Cabanne (BB-0033–000044, Dijon); Services d’anatomo-cytopathologie et hématologie, Centre Hospitalier Régional Universitaire de Nancy; Services d’anatomo-cytopathologie et hématologie, Centre Hospitalier Universitaire de Reims; CRB, Hôpitaux Universitaires de Strasbourg; Services d’anatomo-cytopathologie, Hôpitaux Universitaires de Strasbourg; Services d’anatomo-cytopathologie, Centre Hospitalier Universitaire de Nice.

Conflict of interest

None declared.

Table 1.

List of acronyms used

ANR	French National agency of research
ANSES	French Agency for Food, Environmental and Occupational Health Safety
BIOCAP	virtual BIObank of Pediatric CAncers
CCOP	Childhood Cancer Observation Platform
CepiDC	French National Registry of medical Causes of Death
COHOPER	COHOrt of the PEdiatric cancer Registry
ENCCA	European Network for Cancer Research in Children and Adolescents
GDPR	European General Data Protection Regulation
GEOCAP	GEOlocation of Pediatric Cancers database
GHQ-12	The 12-item General Health Questionnaire
GIS	Geographic information system
ICCC-3	International Classification of Childhood Cancer, 3rd edition
ICD-10	International Classification of Diseases, 10th edition
ICDO-3	International Classification of Diseases for Oncology, 3rd edition
IGN	French National Geographic Institute
INCa	French National institute of Cancer
INSEE	French National Institute for Statistical and Economic Studies
INSERM	French Institute of Health and Medical Research
ISMP	Information System Medicalization Programme
JARC	European Union Joint Action on Rare Cancers
LTD	Long-term diseases
MOOC	Massive Open Online Course
PediaRT	Pediatric cancers RadioTherapy database
QoL	Quality of Life
REDSIAM	multidisciplinary collaborative SNDS users’ network
RNCE	French National Registry of Childhood Cancers
RNIPP	French National Registry for the Identification of Individuals
SFCE	French paediatric oncology society
SF-12	The Medical Outcomes Study Short Form 12-item questionnaire
SNDS	French National health insurance information system database
SPF	Santé Publique France, French National Public Health Agency
WHO	World Health Organisation