Role of MRI in osteosarcoma for chemotherapy response evaluation and prediction: Correlation with histological necrosis
10540 Background:Histologic necrosis (HN), the current gold standard for response evaluation in osteosarcoma (OS), is accessible only after neoadjuvant chemotherapy (NACT) and may get affected by confounding factors. Thus, it would be useful to have surrogate markers of response evaluation and prognostication using magnetic resonance imaging (MRI) to individualize therapy. Method:Thirty-one treatment naïve OS patients received 3 cycles of NACT followed by surgery during 2006–2008. All patients underwent baseline and post-NACT conventional(C), diffusion-weighted (DW) and dynamic contrast enhanced (DCE) MRI. Taking ‘Huvos grading’ (good response >/= 90% HN) as reference standard, various parameters of MRI were compared with it. Tumor considered as ellipsoidal; volume (V), average tumor plane (ATP) and relative(r)-ATP (ATP/body surface area) were calculated using standard formula for ellipse. Receiver operating characteristic curves were generated to assess the best threshold and predictability. Sensitivity and specificity were calculated along with 95% confidence limits. After deriving thresholds for each parameter in univariable analysis, multivariable analysis was carried out. Results: Both pre-and post NACT, absolute and relative size parameters were well correlated with HN, though post NACT change in parameters did not. Apparent diffusion coefficient (ADC), either pre-and post NACT measurements or change following chemotherapy were not correlating well. However, on adjusting for volume, significant correlation was found. Thus, we could derive a new parameter diffusion per unit volume (DV= ADC /V). Change in shape of time intensity curve did not show significant correlation. Conclusions: In OS, NACT response can be assessed and predicted by conventional and DW- MRI early in the disease course which correlates well with HN. DV seems to be a sensitive substitute for response evaluation. For DCE-MRI, more sophisticated models in future studies might be useful. [Table: see text] No significant financial relationships to disclose.