Interfractional Variations Of Tumor Volume And Centroid Position During Stereotactic Ablative Radiotherapy For Lung Tumor

Background and Objectives Stereotactic body radiation therapy(SABR) has been adopted to treat early stage non-small-cell lung cancer(NSCLC) and lung oligometastases in recent years. SABR is characterized by delivering high doses in a few fractions(typically between 1 and 5 fractions). SABR technique can minimize the exposure dose to adjacent normal tissues and reach a higher biological equivalent dose(BED) which is twice the BED of conventional radiation therapy. However, intrafractional and interfractional variations of tumor position during treatment may compromise the accuracy of radiation delivery. Furthermore, variation of tumor volume in response to radiation treatment may impact on accurate delivery. During conventional radiotherapy treatment for lung cancer, though a notable heterogeneity of tumor volume change was observed, a trend of decrease in average tumor volume was also detected after full course of treatment. Unlike conventional radiotherapy, SABR involves high fractionated doses delivered within short time, leading to possible different patterns of variation in tumor position and volume. The aim of this study was to evaluate the interfractional variations of tumor volume and centroid position over time during SABR for primary or metastatic lung tumor. It will further provide objective basis for whether it is necessary to modify the treatment plan during SABR treatment or to perform adaptive radiation therapy.Materials and Methods From October 2011 to October 2014, a total of 66 patients with 71 primary oroligo-metastatic lung tumors received SABR treatment in the department of radiation oncology of the Affiliated Cancer Hospital of Zhengzhou University. Treatment simulation was conducted on a 16-slice CT scanner(Brilliance CT; Big Bore; Philips). Of all tumors, 34 tumors received three dimensional-CT(3D-CT) scan and received SABR delivery in a CT-vision accelerator(Siemens oncor 5710). For this part of patients who underwent 3D-CT scanning only, gross target volume(GTV) was delineated based on CT images. For the margin of planning target volume(PTV), 10 milimeters was added to craniocadual direction and 5 milimeters was added to both anterior-posterior and left-right directions. Before each fraction, 3D-CT scan performed by CT on-rail system was used to calibrate the set-up error. 37 tumors received four dimensional-CT(4D-CT) scan and received SABR delivery in a Varian accelerator(Varian true beam 1407). For this part of patients, no clinical target volume(CTV) margin was added to internal target volume(ITV) and an isotropic margin of 5 milimeters was added in superior-inferior, anterior-posterior and left-right directions for PTV. Before each fraction, CBCT scan performed by Varian system was used to calibrate the set-up error. Median age was 66 years. 34 tumors were treated with three dimensional SABR(3D-SABR) and 37 tumors were treated with 4D-SABR. After SABR treatment, the CT scans acquired for verification(including 3D-CT and CBCT scans) were registered with their corresponding simulation CT scans based on bone structure. Then, The GTV or internal target volume ITV was contoured on all verification CT scans and compared to the initial GTV in treatment plan system. Then centroid position of each GTV or ITV was obtained in treatmentplan system. Measurement of interfractional variation of tumor volume was based on the GTV or ITV in the first verification CT scan. General estimate equation was used to analyze the trend of tumor volume change and the potential impact factors. Measurement of interfractional variation of tumor centroid position was based on the GTV or ITV in the simulation CT scan. The interfractional variation of tumor centroid position was calculated in three dimensions: superior-inferior(SI), anterior-posterior(AP) and left-right(LR). Finally, the 3D vector shift was also calculated.Results The volumetric shrinkage of lung tumors was not statistically significant during both 3D-SABR and 4D-SABR course(P=0.069 vs. 0.453). Diameter is a significant predictor of the tumor volumetric decrease as result shows that the smaller tumors(≤3cm) shrank faster than bigger tumors(>3cm)(P=0.000). No significant difference was observed between 3D-SABR and 4D-SABR methods(P=0.246). For the 34 tumors who adopted 3D-SABR, the overall 3D vector shift was 5.2±3.1 mm(range:0.2-15.9mm). The interfractional variation of GTV centroid position were-0.7±4.5 mm in AP direction, 0.2±3.1 mm in SI direction and 0.4±2.4 mm in LR direction, respectively. For the 37 tumors who adopted 4D-SABR, the overall 3D vector shift was 4.4 ± 2.5 mm(range: 0.4-13.8 mm). The interfractional variation of GTV centroid position were-0.7 ± 2.7 mm in AP direction,-1.4 ± 3.4 mm in SI direction and-0.5 ± 2.2 mm in LR direction, respectively.Conclusions Small but insignificant tumor volume regression was observed during lung SABR. It should be not necessary to perform another field shrink during SABR treatment as conventional radiotherapy. However, a transient increase in tumor volume occurred after the first fraction of SABR, so the margin added to GTV or ITV should consider these variations. While the mean interfractional variations of tumor centroid position were minimal in three directions, variations more than 5 mm account for approximate a third of all, indicating additional margin for PTV, especially in AP direction. Ideally, individualized PTV for lung SABR should be determined by combination of interfractional variations of tumor centroid position and tumor volume as well as daily set-up error. Especially when tumors are adjacent to the vital structures, such as main bronchus and chest wall, patients may benefit from monitoring the volumetric changes, modifying the target and treatment planning individually. Regardless, the final check should be performed with image-guidance for the verification of treatment target volume to prevent target from underdosing and normal tissue from overdosing.