Unmet Need:The use of quantitative metrics derived from PET images to help make clinical decisions, has shown tremendous promise. Examples of such metrics include standardized uptake value (SUV), total lesion glycolysis, and metabolic tumor volume. Quantitative FDG-PET imaging has been shown to separate responders from nonresponders early in a treatment regimen. Such stratification allows clinical trials to achieve statistically significant decisions in shorter duration with fewer subjects.
Further, on an individual patient level, the stratification of a non-responding patient prevents continued treatment of the patient with a high-dose and/or high-risk regimen and allows rapidly switching the patient to alternative treatment regimens. The inventors provide a novel method for comparing methods and determining the optimal segmentation method for metabolic tumor volume.
Technical Overview:This invention is a mathematical/statistical method of identifying the optimum segmentation method for metabolic tumor volume using 18F FDG PET/CT in human solid tumors. This statistical method assumes a linear relationship between the true and measured metabolic tumor volumes for each of the tumor segmentation method. This relationship is characterized by individual slope, bias, and noise terms.
These linear relationship parameters can then be used to compute a noise-to-slope ratio, which ranks different methods on the basis of how precisely they measure the metabolic tumor volume.
Stage of Development:The algorithm has been validated with clinical subjects.
Publication: https://www.ncbi.nlm.nih.gov/pubmed/26982626https://www.ncbi.nlm.nih.gov/pubmed/22713231
