Patients
This retrospective study was approved by the institutional research ethics board of our hospital. Informed consent for treatment and the use of 4D-CBCT-IGRT and its images for this study was obtained from all patients. Eight patients who completed 4D-CBCT-IGRT for gastric MALT lymphoma radiotherapy at our hospital between May 2017 and October 2019 were included in this study.
4D-CT imaging and structure generation
All patients were instructed to fast for at least 8 h before planning CT simulation and treatment to minimize variations in stomach volume. They underwent CT simulation in the supine position with their arms raised; a LightSpeed RT (GE Healthcare, Chicago, IL) or a Discovery RT CT scanner (GE Healthcare) was used for the CT simulation. 4D-CT scans were performed using a real-time position management system (Varian Medical Systems, Palo Alto, CA) or smart deviceless 4D (GE Healthcare) [24]. The scan parameters were set to 120 kV, 70 mA, a gantry rotation time of 0.5–1.0 s, a slice thickness of 2.5 mm, and cine mode. The cine durations were set to the respiratory cycles plus the gantry rotation time. The cine images were sorted into 10 phases using a phase-binning algorithm. The average intensity projection (AIP) CT images were generated from the projection data of all phases. In cases where AIP CT images could not be generated, slow CT images were acquired in the axial mode, with a gantry rotation time of 4 s and a slice thickness of 2.5 mm [25,26].
All CT images were exported to the treatment planning system (Pinnacle3, version 9.10; Phillips Radiation Oncology Systems, Fitchburg, WI) and were registered by the hardware arrangement. The gross tumor volume (GTV) was identified based on the endoscopic examination findings, and it was confirmed that the entire stomach appropriately covered the GTV. The CTV was defined as the entire stomach [19,27]. The CTV delineated based on the 4D-CT images was defined as the CTV-4D. A PTV was defined as the CTV-4D along with an additional margin, which accounts for intra- and interfractional variations in stomach volume, respiratory movement, and patient set-up error [10,19]. All patients underwent the treatment with an individually-defined PTV considering age, performance status, and the dose volume histogram, in terms of target coverage and OAR doses. The structure of the OARs was delineated based on the AIP or slow CT images. The PTV structures with 5, 10, 15, 20, 25, and 30 mm margins from the CTV-4D were generated for the retrospective evaluation (Fig. 1). AIP or slow CT images and all structures were exported into the Elekta X-ray volume imaging (XVI) software, version 5.0.4 (Elekta Oncology Systems, Crawley, UK) as references to be used for image guidance.
4D-CBCT imaging and target localization method
During the initiation of each actual treatment session, the patient was positioned based on body skin marks and aligned at the isocenter. Before the daily treatment fraction of radiotherapy, 4D-CBCT imaging based on skin marks (skin matching) was performed using the Elekta Symmetry 4D IGRT System (Elekta Oncology Systems, Crawley, UK). The projection data of 4D-CBCT were sorted into 10 respiratory-phase bins. The scan parameters were set to 120 kV, 20 mA, 16 ms per frame, and a slice thickness of 2 mm, with a gantry rotation speed (GRS) of 50° min-1 [17,19]. Automatic registration between planning CT and 4D-CBCT images was performed based on the bone anatomy (bone matching) using the Elekta XVI software. Subsequently, the manual registration between planning CT and 4D-CBCT images was performed based on the stomach anatomy using the axial, coronal, and sagittal images until moving images of the stomach in all 10 phases of the 4D-CBCT images were symmetrically positioned within the PTV structure in the planning CT images (4D soft-tissue matching).
Evaluation of the required PTV for target localization methods
We retrospectively evaluated the required PTV to cover the entire stomach, which was confirmed using daily 4D-CBCT images, according to the PTV structures with 5, 10, 15, 20, 25, and 30 mm margins from the CTV-4D. We acquired daily 4D-CBCT images of 10 phases with 15 fractions for each patient (a total of 150 phases per patient). We also compared the required PTV for three target localization methods of the skin, bone, and 4D soft-tissue matchings using daily 4D-CBCT images. The covering phase of the stomach was defined as the phase in which the PTV structures covered the overall stomach and was evaluated by the consensus of two radiotherapists with 4 and 18 years of experience, respectively. For each patient, we calculated the covering ratio (CR) of the stomach with PTV structures of 5–30 mm margins, based on the 4D-CBCT images of a total of 150 phases [CR (%) = (number of stomach covering phases / total of 150 phases) × 100] (Fig. 2) in three target localization methods, and defined a minimum PTV with an average CR of ≥ 95 % for all patients as a requirement [19]. A Kruskal–Wallis test was performed to compare the CRs of the three target localization methods. Subsequently, a Dunn–Bonferroni test was performed to compare the CRs of the three methods as a post hoc analysis if the Kruskal–Wallis test result was significant [28]. Statistical significance was defined as a P value < 0.05. All statistical calculations were performed using the SPSS software, version 25.0 (SPSS Inc., Chicago, IL, USA).