A prospective evaluation of staging and target volume definition of lymph nodes by 18FDG PET/CT in patients with squamous cell carcinoma of thoracic esophagus.
Wen Yu,Xiao-Long Fu,Ying-Jian Zhang,Jia-Qing Xiang,Lei Shen,Joe Y Chang
DOI: https://doi.org/10.1016/j.ijrobp.2010.10.065
IF: 8.013
2011-01-01
International Journal of Radiation Oncology*Biology*Physics
Abstract:To determine an optimal standardized uptake value (SUV) threshold for detecting lymph node (LN) metastases in esophageal cancer using (18)F-Fluorodeoxyglucose positron emission tomography/computer tomography (18FDG PET/CT) and to define the resulting nodal target volume, using histopathology as a "gold standard."Sixteen patients with esophageal squamous cell carcinoma who underwent radical esophagectomy and three-field LN dissection after 18FDG PET/CT and CT scans were enrolled into this study. Locations of LN groups were recorded according to a uniform LN map. Diagnostic performance of different SUV thresholds was assessed by receiver operating characteristic analysis. The optimal cutoff SUV was determined by plotting the false-negative rate (FNR) and false-positive rate (FPR), the sum of both error rates (FNR+FPR), and accuracy against a hypothetical SUV threshold. For each patient, nodal gross tumor volumes (GTVNs) were generated with CT alone (GTVNCT), PET/CT (GTVNPET), and pathologic data (GTVNpath). GTVNCT or GTVNPET was compared with GTVNpath by means of a conformity index (CI), which is the intersection of the two GTVNs divided by the sum of them minus the intersection, e.g., CICT&path=GTVNCT&path/(GTVNCT+GTVNpath-GTVNCT&path).LN metastases occurred in 21 LN groups among the 144 specimens taken from the 16 patients. The area under the receiver operating characteristic curve was 0.9017±0.0410. The plot of error rates showed a minimum of FNR+FPR for an SUV of 2.36, at which the sensitivity, specificity, and accuracy were 76.19%, 95.93%, and 93.06%, respectively, whereas those of CT were 33.33%, 94.31%, and 85.42% (p values: 0.0117, 0.7539, and 0.0266). Mean GTVNCT, GTVNPET, and GTVNpath were 1.52±2.38, 2.82±4.51, and 2.68±4.16 cm3, respectively. Mean CICT&path and CIPET&path were 0.31 and 0.65 (p value=0.0352).Diagnostic superiority of PET/CT at an SUV threshold of 2.36 over CT has potential value in nodal target volume definition, but whether this can contribute to better treatment outcomes needs prospective analyses of recurrences in a larger cohort of patients.