Physical and functional performance assessment in pediatric oncology: a systematic review

Background Research indicates reduced physical performance from diagnosis into survivorship of pediatric cancer patients. However, there is no systematic information or guideline available on the methods to assess physical performance and function in this population. The purpose was to systematically compile and describe assessments of physical performance and function in patients and survivors of pediatric cancer, including cardiorespiratory fitness, muscle strength, speed, balance, flexibility, functional mobility, gait and motor performance test batteries. Methods We searched the databases PubMed, SPORTDiscus, and Cochrane Database and performed abstract and full-text selection of 2619 articles according to the Cochrane Handbook of Systematic Reviews. Information on patients characteristics, assessments, information on validity and reliability, and relevant references was extracted. Results In summary, 63 different assessments were found in 149 studies including 11639 participants. Most studies evaluated cardiorespiratory fitness and muscle strength with the majority conducted off treatment. Some outcomes (e.g. speed) and diagnoses (e.g. neuroblastoma) were severely underrepresented. With the exception of gait, leukemia patients represented the largest group of individuals tested. Conclusions Insufficient data and patient heterogeneity complicate uniform recommendations for assessments. Our results support researchers and practitioners in selecting appropriate assessment to meet their specific research questions or individual daily practice needs. Impact This systematic review includes 149 studies and provides a comprehensive summary of 63 assessments to evaluate cardiorespiratory fitness, muscle strength, speed, balance, flexibility, functional mobility, gait or motor performance test batteries in patients and survivors of pediatric cancer. We present the most studied fields within the pediatric cancer population, which are cardiorespiratory fitness and muscle strength, off treatment phase, and leukemia patients. We propose research priorities by identification of subgroups in terms of cancer type, phase of treatment, and outcome of interest that are underrepresented in studies currently available.


INTRODUCTION
Age-appropriate healthy physical and functional development of infants, children, and adolescents is an important prerequisite for participation in physical activity and sports representing a major determinant of a long-term active and healthy lifestyle. 1 Physical and functional performance of children and adolescents during and after cancer treatment has been the interest of a growing number of studies during past decades. Current literature presents increasing evidence that childhood cancer patients and survivors are challenged by physical performance limitations such as reduced cardiorespiratory fitness, muscle strength, balance, gait, functional mobility, and flexibility/range of motion. [2][3][4] Influencing factors for these impairments might be the cancer itself, side effects of medical therapy, and inactivity during and after treatment. 5 Study results demonstrate reduced physical performance shortly after diagnosis, 6 during acute treatment, 3 and persisting throughout survivorship. 7 This is specifically concerning as physical performance limitations are linked to an increased incidence of unemployment and low income. 8 At the same time, preliminary exercise intervention studies provide promising results in terms of efficacy to improve physical performance and fitness. [9][10][11] However, evaluation of those positive effects found in research interventions with childhood cancer populations is difficult due to the large number of different physical and functional performance assessments that have been used in pediatric oncology research. An overview of assessments could help future researchers when planning a study on exercise and fitness in child and adolescent cancer patients and survivors. Few attempts have been done to summarize and describe tests performed and used in this population. Grimshaw et al. 12 summarized subjective and objective tools to measure physical function and physical activity in the age group 0-18 years with a focus on the evaluation of measurement properties. Another group of researchers 13 listed evaluation tools used in childhood cancer physical activity/exercise studies or community-based programs that assess motor performance, physical literacy, wellbeing, quality of life (QoL), and health behavior, but assessments of physical performance and fitness were excluded. However, no review has predefined the categories of physical and functional performance relevant to health and exercise science 14 in order to systematically search and summarize them. Thus, the aim of the present systematic review is to summarize in detail all assessments used to measure cardiorespiratory fitness, muscle strength, speed, balance, flexibility, functional mobility, gait, and motor performance in interventional and non-interventional studies with childhood cancer patients and survivors. This summary is intended to support researchers and therapists in selecting the most appropriate assessments for their individual purposes and needs.

MATERIALS AND METHODS
This paper was written according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA Statement). 15 Data sources and searches We systematically searched PubMed, Cochrane Central Register of Controlled Trials and SPORTDiscus from database inception to 13 February 2020. The search strategy (see Appendix 1) included Medical Subject Headings with terms and text words to identify studies conducted with children, adolescents, or adults during or after childhood cancer treatment who underwent any assessment for either physical or functional performance. In addition, references of relevant reviews and reference lists of included studies were screened. The specific outcomes of interest were assessments for cardiorespiratory fitness, muscle strength, speed, balance, flexibility, functional mobility, gait, and motor performance assessed in test batteries. Motor performance test batteries measuring physical performance provide an important overview of performance levels and motor development and are of great importance in children and adolescents. In pediatric oncology and chronically ill children, they are usually assessing performance of general motor skills. Data on validity and reliability of the included assessments in pediatric cancer patients and survivors were extracted from the included full texts and associated references.

Study selection
After exclusion of duplicates, three teams of two researchers each independently reviewed titles and abstracts of the identified articles. Studies were excluded for the following reasons: (i) less than 75% of the population were diagnosed with cancer <21 years, (ii) the outcome was no measure of either physical or functional performance as defined above, (iii) any non-original articles (e.g. reviews, congress abstracts, commentaries or letters without data), (iv) duplicates that were not identified as such before, (v) studies without description of assessment used (vi) studies/assessment with less than five participants, or (vii) fulltexts that were not available in English or German. We included all types of studies and had no restriction in terms of publication date. In case of disagreement between the two reviewers, articles were discussed between these two and if no consensus could be reached, a third reviewer was consulted. After final inclusion of abstracts, the respective full texts were reviewed independently as described above.
Data extraction, synthesis, and analysis Relevant data from the included full texts were extracted and organized into standardized data tables. During the data extraction process, the following information was extracted from all texts: study citation, characteristics of the study population (sample sizes, age ranges, diagnoses, stage of cancer treatment), assessments used and their measurement properties, and relevant references for further information. In terms of measurement properties, information regarding validity and reliability of the assessment was only extracted if those were evaluated in the childhood cancer study sample. Based on these tables, assessments were sorted into predefined health-related and skill-related categories as defined by the American College of Sports Medicine 14 (cardiorespiratory fitness, muscle strength, speed, balance, flexibility). In addition to these main motor domains, functional mobility, gait, and motor performance test batteries which have been identified to be of high relevance for the population of children with cancer 3,16,17 and for coping with everyday life and participation with peers 18 were included. For each single assessment (e.g. 6-minute walk test (6MWT) in the category cardiorespiratory fitness) all information about the study participants was merged from the studies using this particular assessment. Diagnoses were grouped into categories, i.e. leukemia/lymphoma (as hematological tumors), bone tumor, CNS tumors, and others. This classification was made because individuals after bone tumors or CNS tumors are known to suffer from more severe motor deficits due to the underlying disease. 19 In case of insufficient information, study authors were contacted via email. If no answer was received, information was taken from the manuscript as specific as possible. Figure 1 displays the flow of studies through the review process. After 81 duplicates were removed, 2619 records underwent abstract screening, of which 2295 were excluded and 324 articles were retained for full-text review. Additional 40 articles, identified through reference list screening of reviews and other sources, were then added to full-test screening, resulting in 364 articles. Of those, 215 articles were excluded with reasons and 149 full texts 3,6,10,20-165 (see Appendix 2 for detailed study information) were included for data extraction, representing 5.7% of screened abstracts and 40.7% of screened full text articles. Agreement between the reviewers for abstract screening ranged between 82 and 96% and for the full text screening between 84 and 93%. In n = 1575 cases (13.5%) a classification was not possible due to a missing detailed description in the full texts. While 22 studies (14.8%) took place during active cancer treatment, 9 studies (6.0%) were conducted during maintenance therapy and 99 studies (66.4%) after treatment. Nineteen studies (12.8%) included participants during different phases of medical treatment. The age of participants ranged between 1.0 and 68.3 years. Most studies analyzed a parameter of cardiorespiratory fitness, followed by strength, motor performance in test batteries, flexibility, functional mobility, gait and balance. Only five studies evaluated speed (Fig. 2). Considering the incidence of childhood cancer, 166 the number of individuals tested in the categories of physical performance and function deviates from the incidence of the tumor type. An overview of the distribution of diagnoses within each category and overall childhood cancer incidence rates are presented in Fig. 3. Some physical performance categories, like cardiorespiratory fitness and muscle strength, were tested in many different types of cancer. However, in all categories, with the exception of gait, individuals with leukemia were overrepresented. For gait and motor performance test batteries, the inclusion of bone tumor patients was far above the percentage incidence of bone tumors, whereas in the other six categories, bone tumors as well as other solid tumors were investigated less frequently.

Literature search results
Results on methods to assess physical and functional performance In total, 63 different assessments were used to evaluate at least one of the eight categories of physical performance and/or function. Between 2 and 16 different assessments were used to evaluate one of the eight categories. The largest heterogeneity in assessment type, calculated as the number of assessment types divided by the number of studies, was in gait with 8 different assessments from 19 studies (0.42), balance with 6 different assessments from 15 studies, and speed with 2 different assessments from 5 studies (0.40). To assess motor performance with test batteries, 16 different test batteries were used in 49 studies (0.33). Strength was evaluated with 16 different measures in 57 studies (0.28) and mobility with 5 assessments in 29 studies (0.17). The greatest homogeneity in measurement techniques was for cardiorespiratory fitness (0.10) and flexibility (0.08) (7 and 3 different methods in 68 and 40 studies, respectively). The different methods are summarized in Tables 1-8 with more details in Appendices 3-10.
Cardiorespiratory fitness (also referred to as endurance, aerobic fitness, or aerobic capacity) was evaluated by a total of 68 studies , any non-original articles (e.g. reviews, congress abstracts, commentaries or letters without data) (n = 31), duplicates that were not identified as such before (n = 10), (v) studies without description of assessment used (n =9), (vi) studies/assessment with less than five participants (n = 11), or (vii) full-texts that were not available in English or German (n = 13).
Physical and functional performance assessment in pediatric oncology: a. . . R Söntgerath et al.
including n = 7936 patients/survivors using 7 lab-and field-based assessment methods (Table 1 and Appendix 3). While the most frequently used assessments (maximal cardiopulmonary exercise test (CPET) and 6MWT) were administered in all diagnostic subgroups during all phases of treatment and a very wide age range, no other assessment was applied during treatment. In terms of measurement properties, the 9MWT has shown to be both reliable and valid in the pediatric oncology population. 92,94 Muscle strength (i.e. muscular endurance or power) was evaluated in 57 studies including n = 5679 childhood cancer patients and survivors using 16 different laboratory and fieldbased assessment methods (Table 2 and Appendix 4). Muscle strength was assessed either by laboratory or field tests focusing on the upper and lower extremities as well as several assessments of core and back muscle strength. While leukemia and/or lymphoma patients and off treatment phase were included in all assessments, some researchers included other cancer diagnoses or phases of medical treatment. In addition, a wide range of age groups was assessed. Isokinetic dynamometry, hand-held dynamometry, and repetition maximum tests are the only assessments that have been shown to be reliable with pediatric cancer cohorts. 92,96 Speed (ability to perform a movement within a short period of time 14 ) was assessed in five studies using two different assessments, which comprised a total of n = 251 childhood cancer survivors aged between 6 and 30 years (Table 3 and Appendix 5). All testing took place after cessation of treatment. Only field tests, namely shuttle run tests, as the 10 × 5m shuttle run and the 4 × 10  The bar on the far right shows the incidence of childhood cancer as a reference. Note: Number of persons tested specified here (e.g. n = 8136 for cardiorespiratory fitness) differs from number of study participants (n = 7936, see section on cardiorespiratory fitness above), because some study participants were tested using several test methods.   (Table 4 and Appendix 6). The nature of assessments was based on posturography and non-posturography methods. While posturography was only performed after medical treatment, two studies conducted balance tests during treatment. 114,148 Bone tumor patients were only included in one study, 103 while CNS cancer cohorts were the population of main interest. In terms of age, a wide spectrum including very young children, as well as older adult survivors of childhood cancer (up to an age of 63 years) were analyzed. No information was available on the validity or reliability of any balance assessment in the pediatric oncology population.
Flexibility was assessed in 40 studies, applying three different test methods which included a total of n = 4309 patients/survivors (Table 5 and Appendix 7). Goniometry, measuring ankle joint range of motion, was performed in most studies including a large number of participants of all ages, with a wide range of diagnoses during all phases of medical treatment. Reliability was analyzed in children with acute lymphoblastic leukemia (ALL) in two studies. 92,158 In addition, two other flexibility tests measuring hip flexion and trunk flexibility were performed with leukemia/ lymphoma, CNS tumor, and other childhood cancer patients and survivors. However, trunk flexibility assessment was only conducted in one study including (young) adults after childhood cancer treatment 62 while the sit and reach test was applied more often with all age groups during all phases of treatment. Measurement properties were not analyzed within the childhood cancer population.
Functional mobility was measured in 29 studies including a total of n = 4421 patients using five different assessment methods ( Table 6 and Appendix 8). Of these, the Timed Up and Go Test (TUG) was administered in two ways: covering either a 3 m or a 10 m distance. While the TUG 3 m and the Timed Up and Down Stairs Test (TUDS) were applied within several studies, including various childhood cancer diagnoses during all phases of medical treatment, the TUG 10 m was only used with ALL patients during maintenance and/or off treatment. 128,130,131 Two additional functional tests (stand up from bed rest exam and floor to stand performance test) were both administered within one study each during treatment for childhood cancer. 79,148 Only the TUG 3 m was performed with older (up to age 64 years) adult survivors of childhood cancer, while all other assessments were conducted with children, adolescents, and young adults. The TUG 3 m demonstrated high validity and reliability 96 while the TUDS and TUG 10 m both have shown to be reliable. 131 Gait analyses was carried out in 19 studies, using eight different methods including a total of n = 545 patients/survivors (Table 7 and Appendix 9). A wide variety of systems were used to assess gait in childhood cancer populations. While few studies used video-recording, partly in combination with force platforms and sometimes electromyography (EMG) measurements, single studies used specific systems, visual observation, or a timed walking test. Except for the EMG analysis of gait and visual observation, all systems assessed gait within various groups of childhood cancer diagnoses. However, only two methods (GAITRite and visual observation) were performed during treatment. 50,148,150 No information is available on validity and/or reliability of any gait analysis system in the pediatric oncology population.
Motor performance test batteries were assessed in 49 studies using 16 different motor test batteries and included a total of 1955 participants (Table 8 and Appendix 10). Most tests were applied in leukemia/lymphoma cohorts after medical treatment. Except for the Functional Mobility Assessment (FMA), which was used in survivors up to age 42 years, all motor test batteries are designed for children and adolescents. Considering all three Bruininks-Oseretsky Test (BOT) versions (BOTMP, BOT-2, BOT-2 SF), the BOT and MOON-Test (Motor performance in pediatric oncology) are the only motor performance test batteries evaluated for feasibility in all diagnosis groups and during all phases of cancer treatment and with young adults. The Gross Motor Function Measure (GFMF) and GFMF-ALL Test Battery are the only assessments that have been evaluated in terms of measurement properties in pediatric oncology populations, 160 while the University of Québec in Chicoutimi-University of Québec in Montréal (UQAC-UQAM) has been validated using the Jackknife method. 85

DISCUSSION
This systematic review summarizes the available studies assessing physical performance and function in pediatric cancer patients and survivors. Based on the included 149 studies with 11,639 participants and 63 different assessment tools, we found important characteristics of the distribution and characteristics of the assessments ( Table 9). The majority of studies (45.6%)         assessing physical or functional performance evaluated cardiorespiratory fitness as an outcome. The 68 studies testing for cardiorespiratory fitness using seven different assessment tools highlight a high homogeneity in the choice of methods. Flexibility was also frequently examined with very uniform assessments. In contrast, muscle strength tests and motor performance batteries have also been evaluated in a high number of studies (57 resp. 49), although with enormous variation in assessment tools. Therefore, the idea of harmonizing physical and functional performance assessments arises to improve comparability of study results. However, harmonization does not seem appropriate nor reasonable across all pediatric cancer types, age groups, treatment phases, and research questions. Speed as a physical performance measure has rarely been evaluated. It can be hypothesized that speed, assessed via shuttle run or other running tests, is difficult to assess during cancer treatment, because children are in a reduced overall condition during cancer treatment. In addition, the health benefits of speed for children and young people appear to be less prominent in the literature than cardiorespiratory fitness and muscle strength 167 and are therefore less focused in children with and after cancer.
In terms of treatment phase, most studies (66%) have been conducted after cessation of cancer therapy with childhood cancer survivors. The evaluation of persistent physical limitations is of great importance, as they may be limiting to working ability and participation. 168 Nevertheless, a continuous monitoring of physical performance should be carried out from the time of diagnosis in order to detect physical limitations at an early stage and prevent further deterioration in a sense of early rehabilitation. At the same time, assessment of physical performance from diagnosis onward is important to determine the need for structured exercise. However, since physical fitness, medical side effects, and motivation vary considerably over the course of the therapy, and are dependent on age, diagnoses, and cancer stage, assessment tools evaluating physical performance and function in children with cancer have to fulfill many requirements. To be feasible and safe, different assessments might be chosen according to different groups of patients.
In terms of sample size, eight tests should be highlighted as they were performed by more than 1000 children each, namely grip and isokinetic dynamometry (muscle strength), 6MWT and maximum CPET (cardiorespiratory fitness), goniometry and sit and reach (flexibility), TUG 3 (functional mobility), and SOT (balance). Of those, the 6MWT (n = 6180), grip strength (n = 4451), and TUG 3 m (n = 4283) were the tests with the greatest number of participants. This fact suggests that those outcomes are of specific interest in pediatric oncology as scientists and clinicians seem specifically concerned about their patients' ability to perform everyday activities, since functional mobility as well as walking capacity measured with the 6MWT are considered important prerequisites to perform physically activities of everyday life. 46,169,170 Concerning the motor test batteries, geographical differences are noticeable. It can be assumed that countries use tests for which reference values of healthy kindergarten and schoolchildren are available. Especially for younger children motor performance test batteries seem to be appropriate to generate an overview of age-related motor development in comparison with age-related reference values. However, generating a database with reference values for children during and after cancer Floor to stand performance Stand up from bed rest exam Physical and functional performance assessment in pediatric oncology: a. . . R Söntgerath et al. treatment could be helpful to evaluate skills in the context of cancer treatment. Measurement properties such as validity or reliability of assessments are rarely available for the population of pediatric cancer patients or survivors. With reference to the period covered by this search, only the 9MWT, TUG 3 m, GMFM, and GMFM-ALL have been tested for validity and reliability, while others (hand held dynamometry, isokinetic dynamometry, repetition maximum test, shuttle run, gonimetry, TUDS, TUG 10 m, and UQAC-UQAM test battery) have been tested for reliability within this population. In addition, it might be useful to examine the quality criteria for children with chronic conditions in general since various chronic health conditions are present during/after pediatric cancer treatment ranging from endocrinological to orthopedic and psychosocial problems. 171 Moreover, information on quality criteria assessed with healthy children may also be helpful while choosing appropriate assessments. Cooperation of interdisciplinary professional societies and scientists could contribute to a joint evaluation of the quality criteria for chronically ill and disabled children.
Overall, due to the variety of assessments used and the small cohorts often found in childhood cancer studies, the significance of studies on physical and functional performance is very limited. Furthermore, it is hypothesized that children with severe physical limitations have been excluded from many tests that have originally been developed for healthy children.

Limitations and strength
The strength of this article lies in the comprehensive research, systematic elaboration, and overview of all methods used to test physical performance in the context of pediatric oncology. The inclusion of interventional as well as observational studies allows a complete listing and a clear focus on the assessment tools. Although the lack of a rating and recommendation of the assessments seems to be a weakness, this evaluation was deliberately avoided. Recommendations based on single studies with large sample sizes, personal experiences, or geographical preferences appear inappropriate, as the objective presentation of the assessments were the primary aim. Instead, the results enable researchers and practitioners to select methods from this paper that correspond to their individual research questions or everyday practice and inform researchers about urgent research questions. The elaboration of recommendations and contraindications for the individual therapy phases, cancer types, and study settings as well as the evaluation of the quality criteria in the target group of children with cancer should be the subject of further research in the context of an international consensus. Methodological limitations include that no distinction has been made between diagnosis subgroups (i.e. myeloid and lymphoblastic leukemia) due to lack of detailed descriptions within studies included. The reader should be aware that the frequency with which an assessment is used does not allow any direct conclusions to be drawn about the quality and suitability of the assessment. Rather, the publication date of the assessment, national availability, and translations, as well as material and costs may also have an influence.
Clinical implication This systematic review identified 149 studies assessing any category of physical or functional performance in childhood cancer patients or survivors. However, the evidence for the effectiveness of interventions to improve aspects of physical performance is very limited. 9,172 This might be related to the methodological quality of intervention studies. However, using standardized tools to assess physical and functional performance in defined subgroups of pediatric cancer patients and survivors (i.e. children with ALL during treatment) would enable metaanalyses of single cohorts and overall improve the significance of    Including other cancer diagnoses and diagnoses that were not clearly specified.
c If evaluated in a childhood cancer population. Table 9. Summary of the main findings.
Main findings are… 1. Physical function and performance were mostly evaluated after medical treatment.
2. Leukemia patients formed the most examined group while solid tumors were less studied. 5. Most assessments have not been evaluated for validity and reliability in pediatric cancer populations.
6MWT 6-minute walk test, TUG timed up and go test.
Physical and functional performance assessment in pediatric oncology: a. . . R Söntgerath et al. interest to prevent physical performance deconditioning. And finally, since survivors of childhood cancer can experience very heterogeneous late sequelae, a transferability of the test applications to children with heart or lung diseases, metabolic diseases, or other chronic conditions is conceivable and should be verified in future research projects.