Awareness of pattern of failure after surgery, as well as knowledge of the relevant anatomy and organs at risk (OAR), must be the basis for target volume definition. In clinical, reoperative, and pathologic studies, mainly four different sites of failure after surgery have been described: local recurrence, LN failure, peritoneal metastasis, or distant metastasis. Results of pattern of failure based on reoperation series in charge of the University of Minnesota and autopsy studies are summarized in Table 9.1 (Gunderson and Sosin 1982). These findings imply that mainly local recurrence of the tumor bed, in the area of anastomotic junction, and local-regional LN failure appeared after surgery. Some autopsy studies even described local-regional failure rates up to 93 % (Horn 1955). These results demonstrate the potential benefit of local-regional radiotherapy.

Pattern of lymphatic spread differs according to the localization of the primary tumor. On the one hand, tumors localized in the fundus tend to spread to splenic nodes (12–42 %), but less commonly to subpyloric lymph nodes. On the other hand, subpyloric nodes (~50 %) constitute the primary site of lymphatic spread of antral tumors, and splenic nodes are rather rarely involved (<10 %) (Smalley et al. 2002).

However, peritoneal seeding is relatively frequent in gastric cancer and has been described in the literature in up to 43 % of postgastrectomy patients (Gunderson and Sosin 1982).

Distant metastasis can be found most commonly in the liver, in approximately 30 % of cases and 10 % in the lung. Yet, with effective locoregional therapy, these secondary failures could be averted (Fig. 9.3).

Patients are usually simulated in supine position with arms raised above the head. For enhancing reproducibility, a variety of immobilization systems, for example, wing boards and vacuum cushions, are available. Legs are placed on knee fixations and patients should be advised to fast for 3 h before simulation.

CT-based radiotherapy planning for gastric cancers is obligatory. CT images should be acquired from above the diaphragm to below the kidneys at the level of vertebrae L4/L5. Application of intravenous contrast is useful to identify the tumor and lymph nodes, but administration of oral contrast medium to visualize gastric distension is only recommended for diagnostic CT, not for planning CT. Surgical clips may be used for matching, but the radiation oncologist must be aware that the anatomy changes after surgery and that clips may not represent the resection margin, but rather vessel clips. Optimally, additional imaging like MRI or PET can be used for treatment planning and available CT slices should not exceed a thickness of 3 mm.

Accurate target volume definition is based on the surgical and pathology reports and the combination of imaging, including CT, endoscopy, and endoscopic ultrasound (EUS) as compulsory and PET as recommended modalities.