Search

In this paper, we study the reliability of a novel deep learning framework for internal gross target volume (IGTV) delineation from 4-D computed tomography (4DCT), which is applied to patients with lung cancer treated by stereotactic body radiation therapy (SBRT). Seventy seven patients who underwent SBRT followed by 4DCT scans were incorporated in this retrospective study. The IGTV_DL was delineated using a novel deep machine learning algorithm with a linear exhaustive optimal combination framework. For the purpose of comparison, three other IGTVs based on common methods was also delineated. We compared the relative volume difference (RVI), matching index (MI), and encompassment index (EI) for the above IGTVs. Then, multiple parameter regression analysis was performed to assess the tumor volume and motion range as clinical influencing factors in the MI variation. The results demonstrated that the deep learning algorithm with linear exhaustive optimal combination framework has a higher probability of achieving optimal MI compared with other currently widely used methods. For patients after simple breathing training by keeping the respiratory frequency in 10 breath per minute (BPM), the four phase combinations of 0%, 30%, 50% and 90% can be considered as a potential solution for an optimal combination to synthesize IGTV in all respiration amplitudes.

This paper expanded two-dimensional optical flow registration method to three- dimensional optical flow method (3D-OFM) to deform the four-dimensional computed tomography (4D CT) images. 4D CT can investigate the effect of breathing motion to the liver tumors and track the “trajectory” of the tumors. This algorithm was applied to the fully automated registration of 4D CT data for hepatocellular carcinoma patients. According to the respiratory phase, 4D CT data was segmented into 10-series CT images which were named CT0, CT10…CT90, and CT0 is at end inspiration and CT50 is at end expiration. In particular, The iodipin was used to help define the gross target volume of HCC to improve the image contrast. In addition, The iodipin also used to improve the registration method and verify the performance of registration method as the markers. We chose the end-exhale CT as the reference image. To verify the registration performance of this method, we qualitatively compared the subtractions of floating images and reference image and the iodipin deposition regions before and after registered, and we quantificationally computed the multi-information between floating images and reference image of 8 patients. The improving proportion of multi-information between end-exhale CT and in-exhale CT is from 1.71% to 4.17%.

Objective: To investigate the feasibility of delineating individual gross target volume (GTV) for tumors moving with respiration by using AIP/MIP CT instead of 4D-CT. Methods: CT data of three patients were acquired in Cine mode. The Cine duration time for each couch position is 4~6s, which equals to one breathing cycle plus 1s. The acquired CT data were input to a 4D-CT reconstruction system made in house to reconstruct AIP/MIP and 4D-CT. Then, the characteristics of motion tumor in AIP/MIP and 4D-CT were analyzed. The influences of window width/level on the appearance of periodical motion target in AIP/MIP CT were investigated by using a periodical motion phantom. Results: The GTV in AIP (GTVAIP) and MIP CT (GTVMIP) was bigger than that in extra-inhale phase 4D-CT (GTV4D-in) or extra-exhale phase 4D-CT (GTV4D-ex). The ratios of GTVAIP to GTV4D-in for two patients were 1.30 and 1.07, and the ratios of GTVMIP to GTV4D-ex were 1.57 and 1.21, respectively. AIP/MIP CT with appropriate window width/level can be used to derive the individual target motion envelope. Conclusions: AIP/MIP CT could be used to quickly delineate individual GTV for moving tumor, which may significantly reduce the target delineation laboring. It can be used to replace 4D-CT to derive the individual GTVs for tumors moving with respiratory.