Subclinical Lesions of the Primary Clinical Target Volume Margin in Esophageal Squamous Cell Carcinoma and Association With FDG PET/CT

Background and Objectives: An accurate delineation of the primary clinical target volume (CTVp) in esophageal squamous cell carcinoma (ESCC) significantly affects the outcomes of radiotherapy. However, when basing the CTVp on the primary gross tumor volume, there are no consistent guidelines for the size of the margin. We compared preoperative (18)F-fluorodeoxyglucose (FDG) PET/CT images and large slices of resected pathological ESCC specimens for evidence and prediction of subclinical lesions. We also investigated associations between the maximum standardized uptake value (SUVmax), metabolic tumor volumes (MTVs), and lesions to improve estimates of the CTVp. Methods:55 patients underwent FDG PET/CT before surgery, and the SUVmax and MTVs were determined. To ensure that the in situ distances between the primary and secondary tumors were preserved, the esophageal specimens collected during radical surgery were processed to minimize shrinkage, and subclinical lesions were characterized by pathological examination. A 2-dimensional logistic regression model was used to assess the associations between clinicopathological features and microscopic spread of the lesions. Results: Subclinical lesions in pathological specimens were characterized as direct invasion, multicentric occurrence lesions, intra-mural metastasis, vascular invasion, and perineural invasion in 56.4, 40.0, 30.9, 21.8, and 18.2% of patients, respectively. The mean distances of the subclinical lesions from the primary tumor were 0.79 ± 1.28 cm and 0.87 ± 1.00 cm in the cranial and caudal directions, respectively. Together the SUVmax and MTV values could predict the presence of subclinical lesions that were not detectable in PET/CT images. Conclusions: To cover 94.5% of ESCC subclinical lesions in the CTVp, a 3-cm margin along the cranial-caudal axis should be added to the primary gross tumor volume as defined by FDG-PET/CT, as well as a cutoff SUVmax value of 2.5. Although preoperative FDG PET/CT images may not reveal lesions directly, the SUVmax and MTV measurements together could predict their presence.