Personalized treatment in pediatric sarcoma: prognostic role of imaging and biomarkers

This thesis investigates the role of imaging biomarkers in pediatric rhabdomyosarcoma and osteosarcoma, focusing on early treatment response, prognostic stratification, and implications for clinical trial design.

Rhabdomyosarcoma (Part I)

The first part evaluates early tumor size changes and diffusion-weighted MRI (DW-MRI) as prognostic markers in rhabdomyosarcoma. A systematic review found that early size response, measured variably using RECIST, WHO, or 3D criteria, lacks validated cutoffs and does not predict survival in non-progressive rhabdomyosarcoma. Studies often excluded patients with progressive disease, limiting insight into progression definitions. Consequently, early tumor size change should not guide treatment adaptation in non-progressive rhabdomyosarcoma, though progression remains a negative prognostic factor.

In DW-MRI analyses from EpSSG rhabdomyosarcoma trials, an increase in apparent diffusion coefficient (ADC) was observed post-chemotherapy. However, ADC changes showed no association with survival. High inter-scanner variability and lack of standardized acquisition protocols limited conclusions. DW-MRI is currently unsuitable for guiding treatment decisions. Future studies must validate ADC metrics and incorporate prospective data with harmonized protocols and centralized review.

Furthermore, this thesis contributed to international imaging guidelines for rhabdomyosarcoma, developed within European and transatlantic collaborations. These guidelines standardize imaging protocols and response assessments, laying a foundation for ongoing trials such as FaR-RMS. Centralized imaging platforms are essential for quality control, inter-rater variability analysis, and accurate response assessment, particularly in patients with progressive disease.

Osteosarcoma (Part II)

The second part addresses diagnosis, risk stratification, and treatment development in childhood osteosarcoma. Clinical guidelines from the European Standard Clinical Practice (ESCP) project provide updated management strategies.

The prognostic value of molecular biomarkers was assessed through gene expression and sequencing data. High MYC expression and the G2 gene expression signature were independently associated with poor outcomes. MYC amplification lacked prognostic significance. Results suggest potential value of combined transcriptomic profiling, though further validation and standardization are required.

Quantitative [¹⁸F]FDG PET-CT imaging at diagnosis, including SULpeak, total lesion glycolysis (TLG), and metabolic tumor volume (MTV), showed no consistent association with survival. Previous studies lacked methodological robustness or standardization. Future prognostic PET studies must adhere to harmonized acquisition protocols and include multivariable validation in cohorts including an appropriate number of patients.

Finally, a systematic review of recent phase 2 trials in relapsed/refractory osteosarcoma highlighted increased use of targeted and immunotherapies but persistent methodological heterogeneity. Most trials were single-arm, limiting comparability. Randomized trials are recommended, especially for combination therapies. The inclusion of adolescents in adult trials and survival endpoints over tumor size response are positive developments. Consensus on trial design, endpoint definition, and biomarker integration is essential to advance treatment.