1 INTRODUCTION

Treatment for childhood cancer has improved considerably over the last decades, which has led to a steady increase in the long-term survival rate.^{1, 2} The present 5-year survival rate after childhood cancer is approximately 80% in Europe.^{1, 3} Most childhood cancer patients receive multimodal treatment, with varying combinations of chemotherapy-regimens, sometimes including hematopoietic stem cell transplantation (HSCT), radiotherapy, surgery and more recently, targeted therapy.^4-6 Although treatment is necessary for cure, unfortunately, all modalities potentially affect normal tissues which can result in adverse late health outcomes.^{7, 8}

To gain insight in the occurrence of and risk factors for these health outcomes, several childhood cancer survivors (CCS) cohorts in Europa, Northern-America and Australia/New Zealand have been established.^{7, 9-12} These studies have yielded a wealth of data on the health status, quality of life, and health care needs of the growing population of (young) adults with a history of pediatric cancer treatment, with implications for clinical care.^13-16 The majority of survivors included in these cohorts were diagnosed during 1940s–1990s when two dimensionally-planned radiotherapy (2D-RT), planning without Computed Tomography (CT) imaging, was the standard of care.^{9, 10} Since then, CT-based three dimensionally-planned radiation treatment (3D-RT) and multi-beam radiation delivery techniques have been implemented in pediatric oncology, sometimes referred to as ‘conformal radiotherapy’.¹⁷ Compared to 2D-RT, the dose from 3D-RT delivered to healthy tissues surrounding the tumor in the high-dose region is generally lower at the cost of a larger area of healthy tissue receiving low-dose irradiation.^{18, 19} While lower rates of harmful side effects can be expected due to reductions in high-dose volumes^{18, 19} the long-term effects of exposing more organs at risk (OARs) to the so-called low-dose bath remains unknown.²⁰ As such, the validity of extrapolating currently available evidence on late effects generated from data on 2D-RT treatments to patients treated in the 3D-era remains uncertain. Although small studies on specific childhood cancer types have reported on side effects of 3D-RT,²¹ large-scale observational studies including follow-up of patients beyond the 5-year survival mark are rare to date.²²

This study has expanded upon the resources of the Dutch Childhood Cancer Survivor Study (DCCSS) LATER Cohort of 5-year CCS^{10, 11, 23} with the purpose of filling this evidence gap. In particular, the overall aim of the Dutch Pediatric 3D-RT study is to evaluate adverse late health outcomes among children treated with radiotherapy in the era of CT-based treatment planning, and to assess dose response relationships for selected health outcomes. In addition, our sub-aims are to compare the frequency of late health outcomes in this cohort to children treated with 2D-RT and to evaluate self-reported health-related quality of life. The current report describes the study design, data collection methods, and baseline characteristics of the 3D-RT study.

2 METHODS

2.1 Study population/inclusion criteria

Eligible survivors for the Dutch pediatric 3D-RT study were identified from the DCCSS LATER cohort. The original DCCSS LATER cohort^{11, 24} included patients diagnosed in the period 1963–2001. Recently, we have expanded the cohort with patients diagnosed after this date. The base source of our study includes 4537 patients (Figure 1) diagnosed with a malignant neoplasm or selected other neoplasia²⁵ before the age of 18 years, during 2000–2012, in a Dutch pediatric oncology center, and who survived at least 5 years after diagnosis. The 3D-RT study includes a sub-cohort with all patients who received external beam radiotherapy (EBRT) with curative intent planned in the 3D-era, defined as the period since introduction of CT-planning in the respective academic radiotherapy center, excluding total body irradiation (TBI), for a primary tumor and/or a recurrence. One Dutch pediatric oncology center was excluded (Table S1), because virtually all pediatric cancer patients treated with radiotherapy in this center received TBI as conditioning regimen for HSCT. HSCT was the local focus of this center; TBI treatment was an exclusion criterion for our study.

During the data collection phase, we identified potentially eligible 5-year survivors of childhood cancer who received radiotherapy during the calendar period since introduction of CT-based treatment planning. The calendar year of introduction of CT-based treatment planning in routine care for children, and thus the start year of inclusion for this cohort study varied by center (2000, 2001 or 2002) in the six included centers (Table S1). We aimed to collect all digital radiotherapy files available in the centers. The detailed information on who did and who not did undergo 3D radiotherapy is currently being evaluated by use of the dose files (number and direction of beams) in combination with the information derived from the letter of the pediatric radiation oncologist based on radiotherapy technique used. Owing to the focus of the study on long-term health outcomes rather than acute side effects, we excluded patients who passed away within 2 years after their first episode of radiotherapy.

In summary, the 3D-RT study source cohort includes 5-year survivors of childhood cancer in the Netherlands who fulfill the eligibility criteria for the DCSS LATER cohort, who received a form of EBRT with curative intent for a primary tumor or recurrence in the 3D-era, defined as the period since introduction of CT-planning in the respective academic radiotherapy department.

3 RESULTS

The national 3D-RT study includes 711 5-years CCS (male (n= 395; 56%)) treated with EBRT with curative intent (Figure 1). The median age at first radiotherapy was 10.6 years (interquartile range [IQR]: 5.7–14.5). The median age at follow-up was 20.5 years (IQR: 16.7–24.9), and the median follow-up time was 10.9 years (IQR: 7.9–14.3) after childhood cancer diagnosis. The median follow-up time after first radiotherapy was 10.1 years (IQR: 7.4–13.2). The most common cancer diagnoses were Hodgkin lymphoma (n = 120; 17%), medulloblastoma (n = 90; 13%) and nephroblastoma (n = 62; 9%) (Tables 1,2).

In 600 survivors (84%), EBRT was administered during the first line treatment. Of them, 64 (9%) survivors also received radiotherapy for one or more recurrence(s). The remaining 111 survivors (16%) entered the cohort after receiving EBRT for a recurrence. During the study period, cumulative treatment exposures included chemotherapy for 82% of survivors while 71% had a surgical procedure and 12% had HSCT (Table 1).

The median total prescribed dose to the primary tumor (dose to primary tumor plus boost dose, if applicable) was 45.6 Gy (IQR: 21.6–54.0 Gy). About one third of the survivors received radiotherapy to the head/cranium only (n = 240, 34%), 126 survivors to the craniospinal axis (18%) and 109 survivors received abdominopelvic radiotherapy (15%).

Table 2 represents the prescribed dose stratified by body regions and childhood cancer type, respectively. In all, 351 (49%), 166 (23%) and 75 (11%) of the survivors received a dose exceeding 50 Gy, 20–30 Gy, or less than 20 Gy, respectively. When stratified by childhood cancer type, doses exceeding 50 Gy were mainly observed among survivors of central nervous system tumors and sarcomas, while doses up to 30 Gy were mainly seen in survivors of lymphomas and renal tumors. Table 3 represents an overview of the radiated body part per childhood cancer type.

Information on prescribed doses on an individual-patient level has been collected for the cohort and is presented in this report. This characteristic enables comparisons across cohorts of childhood cancer survivors who had radiotherapy as part of their treatment regimen. Moreover, we are in the process of estimating the first series of absorbed doses as well as the volumetric aspects of radiation exposure, to specific organs and tissues in the head- and neck area. The resulting future library of radiation exposure to specific organs and tissues will be used in planned and future new late effects analyses.

4 DISCUSSION

In this paper, we described the study design, data collection, and baseline characteristics of the Dutch pediatric 3D-RT study. The aim of our work is to refine insight in the risks of specific late adverse health outcomes occurring in children treated with radiotherapy in the era of CT-based radiotherapy techniques and to investigate dose–response relationships for late health effects. In order to achieve this, we designed a multi-center resource to collect and harmonize digital radiotherapy files that allows for future organ-specific radiation dose (re)calculation among CCSs treated in the 3D era. With the future results of this project, we aim to provide empiric knowledge to enhance both treatment planning guidance in pediatrics, as well as surveillance guidelines for the follow-up of childhood cancer survivors treated in the 3D era.

Although 3D-RT represents the standard of care for pediatric cancer, the current evidence on health outcomes among CCSs is based on historic cohorts of mainly 2D-RT treated patients, with a different radiation dose distribution. Because of this difference in dose distribution, insight in late adverse events following 3D-RT is important. Our recent systematic literature review of observational studies on adverse health outcomes after 3D-RT included 13 studies, most of which describe small patient cohorts (range of N: 5–246) with a median follow-up of 8.8 years,²² and showed that evidence on this topic is scarce. Most available studies to date reported on selected late health outcomes. Our study is designed to address the full spectrum of adverse late somatic health outcomes, and, as such, can address both the total burden of disease as well as potential patterns of related late health effects.²⁸ Also, our cohort is about three times larger (N > 700) than the largest study included in the review (N = 246). Finally, the studies that were eligible for the systematic review typically report on the prescribed dose rather than including organ doses and dose/volume aspects of radiotherapy exposure.

At present time, the only comprehensive prospective assessment of health outcomes after 3D-RT, is the Münster-based “Register of Treatment-Associated Late Effects After Radiotherapy of Malignant Diseases in Childhood and Adolescence” (RiSK).³⁶ The RiSK collaborative group collects data based on voluntary reporting from radiotherapy departments across Germany to the study center, concerning radiation dose parameters for clinically delineated OARs and health outcomes registered during regular follow-up visits at the radiotherapy department in the initial post-treatment period.^{37, 38}

The study cohort described here will be suitable to describe the overall burden of disease and the incidence of specific health outcomes among children treated with radiotherapy in the 3D-RT era in the Netherlands, but will also serve as source cohort to select subgroups of patients of specific interest, based on childhood cancer type, type of treatment planning or radiation delivery technique (e.g., intensity-modulated radiotherapy), or area of radiation exposure. In addition, a series of nested case control studies will evaluate the effect of received dose on the surrounding OARs on occurrence of health outcomes. Of note, full characterization of the treatment technique and radiation beams among the children treated in the 3D-RT era to date awaits completion of the full evaluation of the digital radiotherapy files, because relevant technical details are not always described in medical files or physician letters. Therefore, the source cohort of 711 survivors represents a group of children who received EBRT in the era when it was possible to use CT-planning, although not all will necessarily have received 3D-RT. For example, for patients with nephroblastoma³⁹ neuroblastoma or lymphoma,^{40, 41} classic AP-PA beam set-ups have remained in use, in compliance with international trial protocols in effect during the calendar period covered by our cohort study, although the treatment planning did involve a CT-scan in the study period. Similar considerations hold for craniospinal axis irradiation techniques.

A few limitations bear mentioning, although all are related to intrinsic treatment and patient characteristics that determine eligibility for research studies aimed at delivering novel empirical evidence required to tailor clinical practice.^{42, 43} Foremost, the median follow-up period is comparatively short, in view of the long latency for many known side effects of radiotherapy in children, including subsequent malignancies and normal tissue effects. This fact cannot be changed anno 2022, since 3D-RT was introduced very gradually since the late 1990s. Also, some of the required digital files were documented and stored based on long outdated treatment planning software. For most eligible patients, the radiation physicist managed to convert (parts of) the relevant information to a common DICOM standard. For files considered irretrievable, other sources of treatment details can be used for crude assessment of dose to certain OARs,^44-46 so that valuable observation time for these patients treated early, can be included for selected analyses.

The study derives its main strength from the combination of multiple unique features. Radiotherapy files form the basis to derive valid and precise estimates of doses-volume parameters to relevant organs and we have the opportunity to ascertain health outcomes from LATER outpatient clinics operating under a coordinated common umbrella and harmonized care plans, supplemented with record linkage to national disease registries. Outcome assessment is based on a rigorous data manual, in line with definitions proposed for harmonization of survivorship research²⁸ and according to standards advocated for radiotherapy outcomes research.⁴⁷ Once established, the study platform serves as basis to supplement the data with self-reported information that cannot easily be captured from medical files or disease registries, pending funding. This may include data from the KLIK platform (Dutch: Kwaliteit van Leven In Kaart [Quality of Life in Clinical Practice]), a platform for Health Related Quality of Life questionnaires which can be used in daily clinical practice to facilitate the communication between physician and patient during consultation, empowers patients, and promotes shared decision making^{48, 49} and/or questionnaire surveys to the entire eligible cohort or those visiting the LATER outpatient clinic, as shown in previous research from the DCCSS LATER Consortium.^{28, 50} As such, this resource is available for times to come to address new research questions as they emerge in clinical practice. Also, as with all survivorship efforts, it is important to keep following this cohort to allow for evaluation of potential health problems among 3D-RT survivors as they enter later phases of life.

Overall, this effort embraces and builds upon valuable collaborative work in clinical research^{51, 52} and late effects care and research^{11, 50} in academic centers in the past decades in the Netherlands coordinated by the former Dutch Childhood Oncology Group (DCOG) and continued in the Princess Máxima Center for Pediatric Oncology. It is important for the multidisciplinary treatment teams caring for survivors of childhood cancers and for survivors themselves to have accurate information about the potential late effects following radiation treatment. This information can update surveillance guidelines for the follow-up of CCSs, such as those published by the International Guideline Harmonization Group (IGHG) or ‘Pediatric Normal Tissue Effects in the Clinic’ research collaboration.^53-56

In conclusion, this study described the study design, baseline characteristics and data collection of the 3D-RT study, in which extensive data on radiotherapy and health outcomes are being collected for a nationally cohort of 711 5-year CCS treated in the era of pediatric 3D radiotherapy techniques in the Netherlands. This study can provide new insights into risks of adverse late health outcomes in CCS treated with 3D-RT and specific relations with radiation dose and volume.

AUTHOR CONTRIBUTIONS

Josien G. M. Hazewinkel-Beijer: Data curation (lead); formal analysis (lead); investigation (lead); methodology (equal); supervision (equal); visualization (lead); writing – original draft (equal); writing – review and editing (lead). Judith L. Kok: Data curation (supporting); investigation (supporting); methodology (equal); supervision (equal); writing – original draft (equal); writing – review and editing (lead). Geert O. Janssens: Data curation (supporting); funding acquisition (supporting); methodology (equal); resources (equal); writing – review and editing (supporting). Nina Streefkerk: Methodology (equal); resources (equal); writing – review and editing (supporting). Andrica C. H. de Vries: Resources (equal); writing – review and editing (supporting). Cleo Slagter: Resources (equal); writing – review and editing (supporting). John H. Maduro: Conceptualization (supporting); resources (equal); writing – review and editing (supporting). Petra S. Kroon: Resources (equal); writing – review and editing (supporting). Martha A. Grootenhuis: Conceptualization (supporting); writing – review and editing (supporting). Eline van Dulmen-den Broeder: Resources (equal); writing – review and editing (supporting). Jacqueline J. Loonen: Resources (equal); writing – review and editing (supporting). Markus Wendling: Resources (equal); writing – review and editing (supporting). Wim J. E. Tissing: Resources (equal); writing – review and editing (supporting). Helena J. van der Pal: Methodology (equal); resources (equal); writing – review and editing (supporting). Marloes Louwerens: Resources (equal); writing – review and editing (supporting). Arjan Bel: Resources (equal); writing – review and editing (supporting). Jaap den Hartogh: Conceptualization (supporting); writing – review and editing (supporting). Margriet van der Heiden-van der Loo: Resources (equal); writing – review and editing (supporting). Leontien C. M. Kremer: Conceptualization (supporting); funding acquisition (supporting); investigation (supporting); methodology (equal); project administration (equal); supervision (equal); writing – original draft (equal); writing – review and editing (supporting). Jop C. Teepen: Data curation (supporting); formal analysis (supporting); investigation (supporting); methodology (equal); project administration (equal); supervision (equal); visualization (supporting); writing – original draft (equal); writing – review and editing (lead). Cécile M. Ronckers: Conceptualization (lead); data curation (supporting); funding acquisition (lead); investigation (supporting); methodology (equal); project administration (equal); supervision (equal); writing – original draft (equal); writing – review and editing (lead).

CONFLICT OF INTEREST

C. Ronckers, Ph.D., reported a personal grant from the Dutch Cancer Society (grant no: UVA2012-5517). J. Beijer reported a grant from the Dutch Cancer Society (grant no: UVA2015-7655). The other authors reported no conflicts of interest and did not receive funding.

ETHICS STATEMENT

The Princess Máxima Center institutional review board evaluated and approved the 3D-RT study protocol, and the Medical Research Ethics Committee (MREC) Utrecht (protocol number: 20-138/C) qualified the protocol as exempt from formal review according to the ‘Medical Research Involving Human Subjects Act’ (WMO), based on the retrospective nature and type of data collection. Similarly, the pre-existing DCCSS LATER registry was declared exempt from review by the institutional review boards of the participating centers.