Abstract
BACKGROUND: Changes in musculoskeletal health and function are increasingly recognised as important long-term consequences of treatment for malignant bone tumours in paediatric populations. Accurate and reliable assessment methods are critical for monitoring these changes during treatment. We aimed to determine the inter- and intra-rater reliability of MRI-based measurements of muscle cross-sectional area in children undergoing treatment for paediatric musculoskeletal sarcoma. METHODS: We conducted a retrospective cohort study at a tertiary paediatric cancer centre. Eligible participants were aged 2-17 years and had undergone routine clinical MRI scans as part of their treatment for musculoskeletal sarcoma. Muscle cross-sectional area was measured, focusing on key lower limb muscle/muscle groups. Three trained raters independently performed all measurements. Inter- and intra-rater reliability were assessed using intraclass correlation coefficients (ICCs) with 95% confidence intervals (CIs). Measurement precision was evaluated using minimal detectable change (MDC) values, expressed as a percentage of mean muscle size. RESULTS: Nineteen patients (mean age 11.6 ± 2.6 years; 11 female and 8 male) were included. Diagnoses included osteosarcoma (n = 12), Ewing sarcoma (n = 5), rhabdomyosarcoma (n = 1) and synovial sarcoma (n = 1). The most common tumour location was the distal femur (n = 9), followed by proximal tibia (n = 2), pelvis (n = 2), proximal fibula (n = 1), distal fibula (n = 1) and other sites (n = 4). Metastatic disease was present in seven patients, while 12 had localised disease. Once images affected by tumour were excluded, a minimum of 60 images of each muscle/muscle group were included for analysis. Inter-rater reliability was excellent for psoas (ICC = 0.97, 95% CI: 0.95-0.98), gracilis (ICC = 0.96, CI: 0.93-0.98), medial gastrocnemius (ICC = 0.91, CI: 0.86-0.94), mid-femoral muscle circumference (ICC = 0.99, CI: 0.99-0.99) and mid-tibial muscle circumference (ICC = 0.99, CI: 0.99-0.99). Good inter-rater reliability was found for rectus femoris (ICC = 0.88, CI: 0.83-0.95) and biceps femoris (ICC = 0.83, CI: 0.75-0.89). Intra-rater reliability was excellent across all muscles assessed (ICCs: 0.92-0.99). MDC values indicated highest measurement precision for mid-femoral muscle circumference (10.62%), mid-tibial muscle circumference (11.69%) and psoas (18.26%), enabling detection of clinically meaningful changes over time. CONCLUSIONS: This study demonstrates that MRI measurement of muscle cross-sectional area in children with musculoskeletal sarcoma is a reliable tool. MDC values allow for identification of true muscle loss, supporting early intervention.