Abstract
178
Objectives Intra-tumoral heterogeneity, measured by the deviation of the 3-D PET-FDG uptake pattern from ellipsoidal contours (EC), pre-therapy, has been shown to be a strong independent predictor of survival in sarcoma patients (Eary et al., 2008). Another heteroegeneity measure based on the threshold-volume characteristic has been shown to be an effective prognostic variable with in cervical cancer (Kidd and Grigsby 2008). Here the TV technique is applied to PET-FDG sarcoma data and prognostic performance relative to the EC heterogeneity evaluated.
Methods The data are from a clinical sarcoma cohort (N=224, Events=102) imaged with FDG-PET prior to neo-adjuvant chemotherapy. Tumor volumes were evaluated for EC and TV heterogeneities. Cox regression analyses were used to assess the predictive value of these heterogeneity variables after controlling for standard prognostic factors including grade and tumor type.
Results Both heterogeneity measures were found to have univariate association with outcome - the p-values were 0.06 and 0.001 for TV and EC heterogeneities, respectively. When both heterogeneity variables are included in a multivariate survival analysis with standard prognostic variables, the significance of EC heterogeneity is maintained (p<.001) but the TV heterogeneity is no longer significant (p>.70).
Conclusions The spatially invariant nature of the threshold-volume measures of heterogeneity reduce their utility in sarcoma. The spatially sensitive aspect of the EC heterogeneity measure can extract key prognostic information from PET-FDG studies of sarcoma.
Research Support NIH ROI-CA-65537 and P01-CA42045 and SFI MI-2007
- © 2009 by Society of Nuclear Medicine