Search

PurposeRecent deep-learning based synthetic computed tomography (sCT) generation using magnetic resonance (MR) images have shown promising results. However, generating sCT for the abdominal region poses challenges due to the patient motion, including respiration and peristalsis. To address these challenges, this study investigated an unsupervised learning approach using a transformer-based cycle-GAN with structure-preserving loss for abdominal cancer patients.MethodA total of 120 T2 MR images scanned by 1.5 T Unity MR-Linac and their corresponding CT images for abdominal cancer patient were collected. Patient data were aligned using rigid registration. The study employed a cycle-GAN architecture, incorporating the modified Swin-UNETR as a generator. Modality-independent neighborhood descriptor (MIND) loss was used for geometric consistency. Image quality was compared between sCT and planning CT, using metrics including mean absolute error (MAE), peak signal-to-noise ratio (PSNR), structure similarity index measure (SSIM) and Kullback-Leibler (KL) divergence. Dosimetric evaluation was evaluated between sCT and planning CT, using gamma analysis and relative dose volume histogram differences for each organ-at-risks, utilizing treatment plan. A comparison study was conducted between original, Swin-UNETR-only, MIND-only, and proposed cycle-GAN.ResultsThe MAE, PSNR, SSIM and KL divergence of original cycle-GAN and proposed method were 86.1 HU, 26.48 dB, 0.828, 0.448 and 79.52 HU, 27.05 dB, 0.845, 0.230, respectively. The MAE and PSNR were statistically significant. The global gamma passing rates of the proposed method at 1%/1 mm, 2%/2 mm, and 3%/3 mm were 86.1 ± 5.9%, 97.1 ± 2.7%, and 98.9 ± 1.0%, respectively.ConclusionThe proposed method significantly improves image metric of sCT for the abdomen patients than original cycle-GAN. Local gamma analysis was slightly higher for proposed method. This study showed the improvement of sCT using transformer and structure preserving loss even with the complex anatomy of the abdomen.

Purpose: This study was conducted to develop a deep learning-based automatic segmentation (DLBAS) model of head and neck organs for radiotherapy (RT) in dogs, and to evaluate the feasibility for delineating the RT planning.Materials and Methods: The segmentation indicated that there were potentially 15 organs at risk (OARs) in the head and neck of dogs. Post-contrast computed tomography (CT) was performed in 90 dogs. The training and validation sets comprised 80 CT data sets, including 20 test sets. The accuracy of the segmentation was assessed using both the Dice similarity coefficient (DSC) and the Hausdorff distance (HD), and by referencing the expert contours as the ground truth. An additional 10 clinical test sets with relatively large displacement or deformation of organs were selected for verification in cancer patients. To evaluate the applicability in cancer patients, and the impact of expert intervention, three methods–HA, DLBAS, and the readjustment of the predicted data obtained via the DLBAS of the clinical test sets (HA_DLBAS)–were compared.Results: The DLBAS model (in the 20 test sets) showed reliable DSC and HD values; it also had a short contouring time of ~3 s. The average (mean ± standard deviation) DSC (0.83 ± 0.04) and HD (2.71 ± 1.01 mm) values were similar to those of previous human studies. The DLBAS was highly accurate and had no large displacement of head and neck organs. However, the DLBAS in the 10 clinical test sets showed lower DSC (0.78 ± 0.11) and higher HD (4.30 ± 3.69 mm) values than those of the test sets. The HA_DLBAS was comparable to both the HA (DSC: 0.85 ± 0.06 and HD: 2.74 ± 1.18 mm) and DLBAS presented better comparison metrics and decreased statistical deviations (DSC: 0.94 ± 0.03 and HD: 2.30 ± 0.41 mm). In addition, the contouring time of HA_DLBAS (30 min) was less than that of HA (80 min).Conclusion: In conclusion, HA_DLBAS method and the proposed DLBAS was highly consistent and robust in its performance. Thus, DLBAS has great potential as a single or supportive tool to the key process in RT planning.

Background: While atlas segmentation (AS) has proven to be a time-saving and promising method for radiation therapy contouring, optimal methods for its use have not been well-established. Therefore, we investigated the relationship between the size of the atlas patient population and the atlas segmentation auto contouring (AC) performance.Methods: A total of 110 patients' head planning CT images were selected. The mandible and thyroid were selected for this study. The mandibles and thyroids of the patient population were carefully segmented by two skilled clinicians. Of the 110 patients, 100 random patients were registered to 5 different atlas libraries as atlas patients, in groups of 20 to 100, with increments of 20. AS was conducted for each of the remaining 10 patients, either by simultaneous atlas segmentation (SAS) or independent atlas segmentation (IAS). The AS duration of each target patient was recorded. To validate the accuracy of the generated contours, auto contours were compared to manually generated contours (MC) using a volume-overlap-dependent metric, Dice Similarity Coefficient (DSC), and a distance-dependent metric, Hausdorff Distance (HD).Results: In both organs, as the population increased from n = 20 to n = 60, the results showed better convergence. Generally, independent cases produced better performance than simultaneous cases. For the mandible, the best performance was achieved by n = 60 [DSC = 0.92 (0.01) and HD = 6.73 (1.31) mm] and the worst by n = 100 [DSC = 0.90 (0.03) and HD = 10.10 (6.52) mm] atlas libraries. Similar results were achieved with the thyroid; the best performance was achieved by n = 60 [DSC = 0.79 (0.06) and HD = 10.17 (2.89) mm] and the worst by n = 100 [DSC = 0.72 (0.13) and HD = 12.88 (3.94) mm] atlas libraries. Both IAS and SAS showed similar results. Manual contouring of the mandible and thyroid required an average of 1,044 (±170.15) seconds, while AS required an average of 46.4 (±2.8) seconds.Conclusions: The performance of AS AC generally increased as the population of the atlas library increased. However, the performance does not drastically vary in the larger atlas libraries in contrast to the logic that bigger atlas library should lead to better results. In fact, the results do not vary significantly toward the larger atlas library. It is necessary for the institutions to independently research the optimal number of subjects.