Osteosarcoma is a highly morbid bone-tumor with poor prognosis. Neoadjuvant-chemotherapy(NACT) is the current standard of care. The response of NACT is judged on Histopathology-examination(HPE) after surgical resection of tumor. However, a non-invasive and accurate methods for evaluation of treatment response during the course of therapy is highly desirable. In this research, a Simple-linear-iterative-clustering supervoxels(SLICs) algorithm based methodology using multiparametric MRI (T2,DWI and ADC) has been developed for identification of sub-parts of tumor (active-tumor, necrosis). The volume of active-tumor and necrosis were estimated using this novel approach in patients with OS, before NACT(baseline) and after 3 cycles of NACT(follow-up). The level of necrosis estimated using SLICs and measure with HPE showed a close match. SLICs based estimation of necrosis level is a non-invassive technique that can be useful in response evaluation of cancer imaging.