Long Term Neurotoxic Effects of Chemotherapy in Survivors of Bone and Soft Tissue Sarcomas. A Retrospective Study

The aim of the proposed project is to study the long-term impact of adjuvant systemic multi- agent chemotherapy (cisplatin, anthracyclines, vincristine, methotrexate, alkylating agents) in survivors (treated between 1992 and 2014 in UZ Leuven) of paediatric bone or soft tissue sarcomas on neurocognitive functioning using neuropsychological testing in combination with advanced Magnetic Resonance (MR) imaging techniques in a cross-sectional design. The majority of studies focus on either neuropsychological outcomes or structural imaging. In this project we want to combine both methodologies. MR will mainly focus on microstructural changes in white matter and differences in brain connectivity, while with our neuropsychological tests we will assess a broad range of cognitive functioning. From the obtained MR parameters and neuropsychological test results, the proposed study wants to answer the following research questions:

1. Are there subtle differences in neurocognitive function and behavior related to chemotherapy exposure for survivors of paediatric bone or soft tissue sarcomas?
2. Are there differences in morphological/functional MR parameters that could relate to subtle cognitive differences in cognitive functioning after chemotherapy exposure? Can we detect structural treatment-induced WM injury? Can we see differences in brain neural activity/connectivity during rest between the groups?
3. Is there a correlation between the obtained morphological/functional MR parameters and neuropsychological test results?
4. Is there a correlation between the possible difference in white matter microstructure and gender, age at diagnosis and the duration of treatment?

To answer these questions, we will implement the following measurements:

• From the computerized Amsterdam Neuropsychological Tasks (ANT) system we will test simple motor reaction time, sustained, focused and divided attention, inhibitory control and cognitive flexibility and motor coordination.
• Specific verbal-auditory memory and nonverbal, visual-spatial memory functioning will be registered by the Children's Memory Scale (CMS) for subjects younger than 16 years and with the auditory verbal learning test and the Rey visual design learning test for the subjects older than 16 years. Age-adjusted standardized norms are available for each test.
• The BRIEF (Behavior Rating Inventory of Executive Function) explores executive functioning in school (5-17 years), work (+18 years) and home (5-17 years and +18 years) environments.
• The assessment of intellectual functioning consists of the Wechsler Intelligence Scale for Children (WISC-III) for patients between 6 and 16 years and the Wechsler Adult Intelligence (WAIS-III) for adolescents aged 17 years or older.
• Quality of life will be investigated by the PEDSQL generic version for young adults .
• The Achenbach questionnaires (Child Behavior Checklist, Teacher Report Form, Youth Self Report Form, Adult Behavior Checklist, Adult Self Report Form) items for internalizing, externalizing, and total problems will be filled out by the parents, teacher or relatives and patients themselves.
• All participants will complete the Spielberger State-Trait Anxiety Inventory and the Beck Depression Inventory-II.

Currently, the neurocognitive follow-up of paediatric bone and brain tumour patients (NKP21/22c, National Cancer Plan Onkelinx) as well as children with congenital heart disease (IWT-TBM, 3M110068) or treated in a paediatric intensive care unit is also based on this test battery. All these tests and questionnaires have been extensively described and used to quantify neurocognitive development and Quality of life in various paediatric populations. The testing will be performed at least one year after the end of treatment.

MRI imaging to study the neural substrate of possible neurotoxic effects of chemotherapy-treatment Advanced MR neuroimaging techniques in combination with the detailed cognitive and behavioral assessment will be used. All subjects will be imaged on the same 3T Philips scanner with a 32-channel phased-array head coil, located in the radiology department at UZ Leuven. Neuroimaging consists of 3D T1-weighted FLAIR and advanced diffusion weighted- imaging (DWI), to examine white matter (WM) microstructure, and resting state functional MRI (rfMRI) to visualize neuronal activity of the brain at rest (examination with eyes closed).

1. DWI - a technique enabling the visualization and characterization of the WM architecture via the self-diffusion of water molecules - allows us to study potential chemotherapy-induced changes in the WM microstructure. Damage to WM structures - for example axonal loss or demyelination - may change quantitative DWI parameters such as fractional anisotropy (FA), mean diffusivity (MD) and diffusion kurtosis.
2. rfMRI uses the vascular response nearby electrically active neurons to indirectly visualize brain activity and allows us to determine the connectivity between the different brain regions at rest.

The duration of the proposed scan protocol is maximum 45 minutes.

Statistical analyses Neuropsychological performance Performance on neurocognitive measures of the survivors will be compared with an age-matched control group using ANOVA and simple contrasts with controls as reference and STAI and BDI scores as covariates. Partial eta squared (hp2) will be computed to estimate effect sizes (weak effect: ηp²≈0.03; moderate: ηp²≈0.06; large: ηp²≥0.14). Cohen's d will be used for simple contrasts (small effect: 0.2 < d < 0.3; medium: 0.3 < d < 0.8; large: d> 0.8). Subsequent analyses are conducted only on those neurocognitive measures where survivors demonstrated significantly lower values than the age-matched control group. The independent effects of the predictors (treatment group, gender, age at treatment, duration of treatment and socioeconomic status) on these neuropsychological score will be analyzed with a repeated measures mixed effects regression model.

Imaging State of the art image analysis approaches are applied to process and analyze the acquired brain images and to document differences in structural white matter or resting state brain activity between the study groups. DTI images will be processed and analyzed using explore DTI, in-house developed software and voxel-based statistical analysis. Furthermore, performance on neurocognitive measures and differences in FA and MD maps of the survivors will be compared with an age and sex-matched control group using ANOVA and simple contrasts with controls as reference and STAI and BDI scores as covariates. Seed-based analysis and independent component analysis (ICA) will be applied to analyze the resting state fMRI data. The association between the obtained MRI parameters and the performance on neurocognitive tests will be investigated using (voxel-based) correlation analysis. The independent effects of the predictors (treatment group, gender, age at treatment, duration of treatment and socioeconomic status) on these neuropsychological and imaging scores will be analyzed with a repeated measures mixed effects regression model.

Sample size and power Study subjects are recruited from the paediatric haemato-oncology center in UZLeuven. We expect to recruit 30 patients within the time frame of this project. A healthy control group (n=30) matched for age and gender will be included. Demographic characteristics such as social economic, school and work status are assessed.

It is difficult to perform a fundamental power calculation, because little is known about the impact of the chemotherapy in bone or soft tissue sarcomas on cognitive functioning so far. We expect to recruit 30 survivors of bone or soft tissue sarcoma in childhood. If we take these numbers into account and based on a previous studies reporting white matter FA changes and cognitive changes in childhood ALL patients who received chemotherapy, we would have a power of 98.54% to detect a difference of 1 standard deviation on cognitive tasks (1-tailed, α=0.05). For white matter FA changes, we would have a power of 98%.