Paediatric Cancer Predisposition Documentation Tool – Standardized Reporting Form for Children and Adolescents With Suspected Cancer Predisposition Syndrome

More comprehensive genetic diagnostics in children with cancer, enabled by modern sequencing techniques have shown that germline variants causing genetic cancer predisposition can be detected in an increasing proportion of patients. Many individuals carrying a predisposing germline variant exhibit distinct characteristics regarding family history, tumor type, age at manifestation and therapy toxicity. However, comprehensive phenotypic characterization and automated electronic documentation in searchable databases are essential to fully integrate genetic and clinical features. Therefore, we have developed a structured Paediatric Cancer Predisposition Tool – PERCEPT to facilitate more accurate documentation of even subtle clinical features of patients with or with suspected germline cancer predisposition or suspected germline cancer predisposition. It improves the comparability in multicentre studies and the automated recognition of phenotypic patterns in international searchable databases.


Introduction
Genetic testing in children and adolescents with malignant diseases for suspected genetic cancer predisposition syndromes, have revealed a previously clinically underestimated high proportion of individuals with constitutional genetic variants associated with an increased cancer risk [1][2][3].
Pathogenic germline variants leading to cancer susceptibility are reported in at least 8-10% of childhood cancer patients [4]. Patients with secondary malignancies, a family history of cancer, the occurrence of malignancies at an unusually young age, or rare or adult-type tumors are more likely to have an underlying genetic cancer predisposition syndrome (CPS) [5].
Up to now the approach to identify patients with CPS was based on the awareness of phenotypic ndings and family history. In clinical routine, the work-up of newly diagnosed cancer patients focuses on imminent medical challenges and prompt initiation of therapy, deferring from comprehensive phenotypic evaluation. In addition, clinical oncologists often do not have the training and expertise to recognize subtle morphological ndings and standard phenotypic traits [6,7]. Moreover, in contrast to the presentation of hereditary cancer predisposition syndromes with characteristic signs and inheritance [8], malignant diseases caused by germline variants also occur in a substantial number of individuals lacking a remarkable family history due to de-novo variants, incomplete penetrance, or variable expression of cancer susceptibility genes [9].
So far the bottle neck towards comprehensively uncovering individuals affected by a CPS was the lack of systematic genetic germline screening of all children and adolescents with malignancies. Meanwhile technological advances have signi cantly reduced the cost and turnover time of gene panel screening, whole exome and whole genome analysis enabling the identi cation of pathogenic germline variants in an increasing number of patients. There fore systematic clinical assessment and phenotypical description of individuals with suspected or con rmed pathogenic germline variants becomes more and more important.
The "Paediatric Cancer Predisposition Documentation Tool" (PERCEPT) presented here has been developed as an additional tool for paediatric oncologists and medical geneticists. PERCEPT is a secondline instrument to be applied after identi cation of childhood cancer patients with suspected or con rmed cancer predisposing germline variants. It aims to serve as a tool for the systematic assessment and documentation of clinical characteristics. It will raise awareness in clinicians regarding syndromic features and would enable harmonized future studies in the eld. This is the joint work of members of the CPS WG of the GPOH, SIOPE's Host Genome Working Group, the I-BFM Host Genetic Variation Working Group, the Working Group Tumor Genetics of the German Society of Human Genetics and the COST Action CA16223 "LEukaemia GENe Discovery by data sharing, mining and collaboration (LEGEND)".

Methods And Results
Questionnaire development Documentation forms and questionnaires previously used for genetic counseling at participating institutions were collected and reviewed. The existing questionnaires mainly aimed to record and describe growth disorders, intellectual disability, or other congenital anomalies depending on the clinical and scienti c focus of the respective institutions. The structures of the existing questionnaires were largely retained, but speci c contents regarding predisposition syndromes for paediatric cancers were systematically included. To cover all the facets relevant to cancer predisposition syndromes, additional features and questions were identi ed by two approaches. First, features of currently known and well described cancer predisposition syndromes were added in the PERCEPT form. Second, characteristics of malignancies commonly associated with predisposing syndromes were also systematically integrated using Human Phenotype Ontology (HPO; http://human-phenotype-ontology.github.io/) terms exclusively throughout the form.
As described above, the PERCEPT form is not intended for patient identi cation and description in rstline clinical care, but is to be applied in a second step for a more comprehensive characterisation of individuals with CPS. A rst-line selection tool was primarly described by Jongmans et al. [10] and later updated by Ripperger et al. [11]. The latter became an essential screening tool for GPOH trials and is a mandatory prerequisite for certi cation by the German Cancer Foundation [11]. The screening tool is intended to aid paediatric oncologists in identifying children with malignancies who may have a possible underlying cancer predisposition syndrome. In contrast, our Pediatric Cancer Predisposition Documentation Tool is designed for characterisation of children and adolescents with a malignant disease within or after diagnostic workup.

I. Genetic medical counseling
Genetic consultation and speci c diagnosis should be conducted by specially trained physicians. In most institutions, including ours, this task is mainly performed by physician geneticists. However, in some countries genetic counselling is performed by physicians with alternative specialization. To allow the comparison of results from different countries in multicentre data collections, documentation of the quali cation of the person responsible for PERCEPT assessment has been added.

II. Current malignancy
The current malignancy records include age at onset, occurrence of metastases, multiple primary tumors, unusual therapy toxicity, and presence of any molecular aberration indicative of a cancer predisposition syndrome. An unusually young age may be as important as the speci c type of malignancy observed. The occurrence of adult-type tumors like gastrointestinal carcinomas, melanomas, malignancies of head and neck, lung cancer, and breast cancers is extremely rare in children. The aetiology is very likely to be different in children compared to adults (e.g. long-term exposure to toxins can be largely excluded).
Information regarding the type of therapy and unusual therapy toxicity potentially associated with the germline aberration, are also included in PERCEPT. For instance, excessive toxicities have been observed in Fanconi anaemia patients after regular conditioning regimen before haematopoietic stem cell transplantation [12]. Also, patients with other DNA repair disorders or immune de ciencies may suffer from severe treatment complications [13].

III. Previous malignancies
Currently, nearly one out of ve cancer cases occurs in individuals with a history of cancer. Within the cohort of individuals who suffered from childhood cancer, 9.3% develop a secondary cancer within the following 30 years [14]. Carriers of germline variants in a certain cancer-predisposing gene are at higher risk of encountering another cancer. However, the appearance of secondary malignancies may also be attributed to the mutagenic effects of chemotherapy (mainly seen in leukaemia and myelodysplastic syndrome) and/or radiotherapy (mainly thyroid cancers, sarcomas, and brain tumors) [15]. The latency between the end of therapy and diagnosis of a secondary solid tumor is typically over 10 years whereas secondary leukaemia often occurs within few years after the primary disease. However, the effect of previous therapy on the development of a second malignancy is di cult to determine [16]. Another pitfall includes the observation that a wide range of therapy-induced cancers belongs to the spectrum of hereditary cancer predisposing syndromes. Moreover, it is well established that individuals with certain hereditary disorders have increased sensitivity to the carcinogenic effects of therapies, such as DNA damage repair, oxidative stress, and cell cycle control, which likely contributes to the development of radiation-and chemotherapy-related cancers.

IV. Pregnancy / Birth
Pregnancy, birth, and neonatal history are captured in PERCEPT to document growth retardation or overgrowth. To assess if the patient´s phenotypic signs belong to a certain cancer predisposition syndrome or can be partially attributed to other factors, additional points, including infections or exposure to toxic agents during pregnancy, are recorded.
V. Medical history / Clinical examination PERCEPT also allows recording the medical history of the patient, including early childhood development and results of clinical examinations. Cancer predisposing syndromes can be associated with developmental anomalies and skin lesions, such as café-au-lait-spots in neuro bromatosis type I, Constitutional Mismatch Repair De ciancy (CMMRD) or Fanconi anemia. However, in most cases, symptoms are subtle and vague, and only all full description of abnormalities might shed light on the complete picture of the signs and symptoms associated with a certain CPS.
While several cancer predisposition syndromes, such as Down syndrome and Beckwith-Wiedemann syndrome, are already well described and associated with characteristic dysmorphological signs, others are associated only with minor morphological signs (8). The relation between minor anomalies and childhood cancers, especially solid tumors is well documented. Moreover, it has been shown that genes entangled in organogenesis may cause minor congenital anomalies and may be involved in the development of cancer [18]. Thus, it becomes important to identify the complete spectrum of physical ndings and possible congenital and developmental abnormalities while also accounting for the possibility of altered prevalence with age. Congenital anomalies and physical ndings not explained by an underlying cancer predisposition syndrome might be due to an unknown syndrome not discovered till now or could be due to another unrelated condition which does not predispose to cancer. In addition, PERCEPT speci cally documents external factors (infection, teratogen and other) to capture differential diagnoses with acquired aetiology.

VII. Previous examinations
The chapter "previous examinations" refers to previous genetic counselling and testing conducted. Until recently, genetic analyses were focused on speci c genes based on syndromic features and family history, as well as chromosome analyses and Sanger Sequencing. However, the focus has now shifted to next generation sequencing methods that made comprehensive genetic testing feasible.
This chapter also includes genetic analysis of tumors, primarily looking for somatic variants, with possible impacts on diagnosis, prognosis and therapy. These analyses may also reveal a suspicion for a germline pathogenic variant, e.g. in TP53 or IKZF1 [15,19].
Imaging performed previously can add to syndromic features found in clinical examinations as some symptoms and signs are only detectable by instrument-based investigations (e.g. radiography). Examples include malformations of bones, cysts of organs, benign tumors, malrotation of organs.
Family history and guidelines to generate a family tree are included in PERCEPT. While collecting information regarding cancer history, information on all childhood-and adult-onset malignancies, age at onset, and type and site of each malignancy up to third-degree relatives should be included. Inherited pathogenic variants can be accompanied by different degrees of penetrance and can cause cancer in childhood and/or adulthood leading to different disease presentations. Any cases of death with unknown cause in the family, along with age at the time of death, should be included. Furthermore, diseases like immunode ciency and other congenital disorders, as well as family members with mental retardation or syndromic features, should be annotated and clinical reports, pathology reports, and any pre-existing genetic analyses for family members should be collected.
While assessing information on families with certain tumor syndromes, an unusually young age at cancer diagnosis may be as important as the speci c type of malignancy observed. The occurrence of adult-type tumors, such as gastrointestinal carcinomas, melanomas, malignancies of head and neck, lung cancers, and breast cancers is extremely rare in children.

IX. Family tree
A common pedigree nomenclature is mandatory to allow for a standard inheritance assessment. We followed the nomenclature published by the National Cancer Institute (https://visualsonline.cancer.gov/details.cfm?imageid=10346). Because cancer predisposition syndromes mainly follow an autosomal dominant pattern of inheritance, the occurrence of several malignancies in a family is particularly indicative of a cancer predisposition syndrome running in a family. However, the presence is not excluded if the index patient is the only known case as there are other modes of inheritance as well, which include rare autosomal recessive, X-linked recessive or more complex inheritance modes (e.g. methylation dependent disorders).
In contrast, no or low frequency of malignancies in a family can be indicative of de novo pathogenic variants, incomplete penetrance, or related to small family size. Moreover, in cases of autosomal recessive cancer predisposition syndromes, the parents are carriers and the family tree is most likely unremarkable. In such cases, a possible consanguinity between parents or in the family would be of relevance.

Discussion
Due to the rapid technical advances in sequencing methods in the recent years, the number of patients diagnosed with malignant diseases in childhood and adolescence who carry a disease causing germline variant has increased signi cantly. This has also increased the clinical attention for these individuals because these conditions are now perceived as a more frequent diagnosis with sometimes subtle manifestations.
This in turn highlights the demand for a suitable routine procedure to characterize and follow-up patients with malignant diseases that have a high likelihood of having a CPS with consequences for cancer treatment and long-term surveillance. In the past, easy-to-use selection tools based on family history, type of malignancy, number of malignancies in the same patient, some basic congenital anomalies, and other speci c features and excessive toxicity were used in the process of reliable identi cation of candidates for further genetic assessment in daily routine [10,11].
The intention of PERCEPT was to extend this structured procedure by generating a tool for standardized documentation of targeted anamnesis and comprehensive phenotypic description. This information may prove useful in two directions. First, to ensure detailed clinical characterisation and documentation of individuals with de ned genetic variants in searchable databases; and second, to improve the development of automated tools for the detection of cancer predisposition syndromes based on clinical features. Thus, it is important to identify the complete spectrum of physical ndings and developmental abnormalities and relate them to patient age as there might be an increased or decreased prevalence with age. Therefore the systematic recording of clinical features at the time of tumor diagnosis is needed, when the full picture of the underlying syndrome might not be completely obvious and re-evaluation needs to be scheduled. This proceeding may provide a relevant source for expanding the phenotype of CPS and contribute novel ndings to the underlying development of cancer.
Moreover, identifying and understanding links between congenital abnormalities, childhood development and the risk of developing cancer will have impact for risk estimation, surveillance, prognosis, families and potential personalized therapies.
International co-operation within SIOPE (https://siope.eu/) and COST (https://www.cost.eu/) has contributed signi cantly to the formulation of a series of surveillance recommendations and the discovery of novel gene variants in familial and syndromic malignancies. The different structural and organizational requirements in the participating countries have been considered while designing PERCEPT. The format of the PERCEPT form is easy to implement into electronic data capture and supports the analysis of genetic information in conjunction with clinical features in multicentre, international studies.
Especially because the future perspective might be genetic testing of all pediatric patients diagnosed with malignant diseases in childhood and adolescence [3,20], a comprehensive and standardized phenotypic description of individuals with childhood malignancies is mandatory for the comprehensive interpretation of genetic data and comparison between different study cohorts. We hope that PERCEPT might serve as a useful tool to improve standardized documentation and educate and train doctors who are engaged in the eld of cancer predisposition syndromes. In view of the constant growth of knowledge in this eld, it is obvious that the contents of PERCEPT must be constantly amended.

Declarations
Funding COSTAction CA16223 LEukaemia GENe Discovery by datasharing, mining, and collaboration (LEGEND) is supported by the EU Framework Programme Horizon 2020.