Can combined 18F-FDG-PET and dynamic contrast-enhanced MRI predict behavior of desmoid tumors in patients with familial adenomatous polyposis?

BACKGROUND Desmoid tumors associated with familial adenomatous polyposis show variable behavior; about 10% grow relentlessly, resulting in severe morbidity or mortality. Investigations that could identify the minority of desmoid tumors that behave aggressively would allow these tumors to be treated early and spare the majority of patients who have more benign disease from unnecessary intervention. OBJECTIVE The aim of this study was to investigate whether imaging the tumor metabolic-vascular phenotype by modern methods predicts growth. DESIGN This is a prospective case series study. SETTINGS The study was conducted at a tertiary center specializing in familial adenomatous polyposis and desmoid disease. PATIENTS Nine patients with familial adenomatous polyposis (4 male, mean age 39 years) with desmoid tumor underwent 18F-FDG-PET and dynamic contrast-enhanced MRI. Standard MRI was repeated a year later to assess tumor growth. MAIN OUTCOME MEASURES The primary outcome measured was the correlation between 18F-FDG-PET and dynamic contrast-enhanced MRI parameters and subsequent desmoid growth. RESULTS Failed intravenous access precluded dynamic contrast-enhanced MRI in 1 female patient. Thirteen desmoid tumors (4 intra-abdominal, 2 extra-abdominal, 7 abdominal wall; mean area, 68 cm) were analyzed in the remaining 8 patients. Two patients died before follow-up MRI. Five tumors decreased in size, 3 increased in size, and 3 remained stable after a year. Significant correlation (Spearman rank correlation, significance at 5%) existed between maximum standardized uptake value and k(ep) (r = -0.56, p = 0.04), but not with other vascular parameters (K(trans) (r = -0.47, p = 0.09); v(e) (r = -0.11, p = 0.72); integrated area under the gadolinium-time curve at 60 seconds (r = -0.47, p = 0.10)). There was no significant difference in the maximum standardized uptake value or dynamic contrast-enhanced MRI parameters (K(trans), v(e), k(ep), integrated area under the gadolinium-time curve at 60 seconds) between the tumors that grew or decreased in size or between the tumor sites. However, vascular metabolic ratio (maximum standardized uptake value/K(trans)) was significantly different for tumor site (p = 0.001) and size (p = 0.001, 1-way ANOVA). LIMITATIONS This investigation is limited because of its exploratory nature and small patient numbers. CONCLUSIONS Although not predictive for tumor behavior, some correlations existed between dynamic contrast-enhanced MRI and 18F-FDG-PET parameters. Vascular metabolic ratio may provide further information on tumor behavior; however, this needs to be evaluated with further larger studies.