Non-invasive biological information about residual neuroblastoma tumour tissue could allow treatment monitoring without the need for repeated biopsies. Magnetic resonance spectroscopy (MRS) can be performed with standard MR-scanners, providing specific biochemical information from selected tumour regions. By proton 1H-MRS, lipids, certain amino acids and lactate can be detected and their relative concentrations estimated in vivo. Using experimental models of neuroblastoma, we have described the potential of 1H-MRS for the prediction of tumour tissue viability and treatment response. Whereas viable neuroblastoma tissue is dominated by the choline 1H-MRS resonance, cell death as a consequence of spontaneous necrosis or successful treatment with chemotherapy, angiogenesis inhibitors, or NSAIDs is associated with decreased choline content. Therapy-induced neuroblastoma cell death is also associated with enhanced 1H-MRS resonances from mobile lipids and polyunsaturated fatty acids. The mobile lipid/choline ratio correlates significantly with cell death and based on the dynamics of this ratio tumour regression or continued growth (drug resistance) after chemotherapy can be predicted in vivo. The implications of these findings are discussed with focus on the potentials and limitations of introducing 1H-MRS for clinical assessment of treatment response in children with neuroblastoma. Biochemical monitoring of neuroblastoma with 1H-MRS could enable tailoring of individual therapy as well as provide early pharmacodynamic evaluation of novel therapeutic modalities.