Overlap Volume Analysis Using 4DCT: Implications for Gating in SBRT of the Lung and Pancreas

Dose escalation in stereotactic body radiation therapies (SBRT) results in improved clinical outcomes, yet is directly limited by dose to normal tissue. Respiratory-gated radiation therapy has been used to treat moving tumors and reduce dose to normal tissue. The end-of-expiration (EOE) position is typically chosen due to its longer duty cycle without considering the changing tissue geometry shown by 4DCT that commonly limits achievable prescription doses for pancreas and centrally located lung SBRT. The purpose of this study is to analyze the relative geometrical relationships between tumor and normal tissues and determine the optimal gating window. 6 patients diagnosed with centrally located lung cancer as defined by RTOG 0813 (within or touching 2 cm expansion of bronchial tree, or immediately adjacent to mediastinal or pericardial pleura) were included to the study. Of the 6 patients, 4 had tumor in the upper lobes, 2 left-sided and 2 right-sided. 2 patients had tumor in the lower lobes, 1 left-sided and 1 right-sided. All patients had 4DCT. Gross tumor volume (GTV) and 2cm expansion of GTV (GTV2cm) were drawn on all 10 phases. Overlap volumes (OLVs) were calculated: OLV2OARall (overlap volume of GTV2cm and union of all organs at risk [OAR]); and OLV2OAR (overlap volume of GTV2cm and individual OAR including heart, aorta, esophagus, bronchial tree and trachea). OLVs were plotted against breathing phases and the changes in OLV during breathing were calculated. One patient diagnosed with pancreatic cancer was also included to the study. Similar OLV definition was used with different OARs including stomach, duodenum and bowel. The OLV2OARall changes during breathing were small (average of 1.9(1.4-2.8) cc) for the 4 lung patients with upper-lobe tumors. However, for the 2 patients with lower-lobe tumors, the average change in OLV2OARall during breathing showed substantial changes of 8.9 (7.1-10.6) cc, both with smallest OLV at the end-of-inhalation (EOI) phase. OLV2heart changes showed similar patterns with smallest OLV at the EOI phase (7.9cc smaller than the EOE phase) but OLV2DA showed the nearly opposite pattern for one patient and insignificant changes for the other. The pancreas patients also showed a significant change 12.2cc in OLV2OARall, but contrary to the lung patients showed the smallest OLV at the EOE phase. Similar motion patterns for individual organs were also observed with generally smaller OLV in the EOE phase. The maximal changes in OLV between breathing phases for the stomach, duodenum and small bowel were 3, 8.6 and 4.9 cc, respectively. Overlapping volume analysis shows that gating windows can be optimized for respiratory driven motion. Optimal gating window varies by tumor and normal organ geometry. For patients with lower central lung tumors, this window can be the opposite of common clinical practice.