Effect on Heart and Lung Doses Reduction of Abdominal and Thoracic Deep Inspiratory Breath-hold Assuming Involved-field Radiation Therapy in Patients with Simulated Esophageal Cancer.

PURPOSE: The purpose of this study was to evaluate the lung and heart doses in volumetric-modulated arc therapy (VMAT) using involved-field irradiation in patients with middle-to-lower thoracic esophageal cancer during free breathing (FB), abdominal deep inspiratory breath-hold (A-DIBH), and thoracic DIBH (T-DIBH) images. METHODS: Computed tomography images of A-DIBH, T-DIBH, and FB from 25 patients with breast cancer were used to simulate patients with esophageal cancer. The irradiation field was set at an involved-field, and target and risk organs were outlined according to uniform criteria. VMAT optimization was performed, and lung and heart doses were evaluated. RESULTS: A-DIBH had a lower lung V20 Gy than FB and a lower lung V40 Gy, V30 Gy, V20 Gy than T-DIBH. The heart all dose indices were lower in T-DIBH than FB, and the heart V10 Gy was lower in A-DIBH than FB. However, the heart Dmean was comparable with A-DIBH and T-DIBH. CONCLUSIONS: A-DIBH had significant dose advantages for lungs compared to FB and T-DIBH, and the heart Dmean was comparable to T-DIBH. Therefore, when performing DIBH, A-DIBH is suggested for radiotherapy in patients with middle-to-lower thoracic esophageal cancer, excluding irradiation of the prophylactic area.

Dose and organ displacement comparisons with breast conservative radiotherapy using abdominal and thoracic deep-inspiration breath-holds: A comparative dosimetric study.

Deep-inspiration breath-hold (DIBH) reduces the radiation dose to the heart and lungs during breast radiotherapy in cancer. However, there is not enough discussion about suitable breathing methods for DIBH. Therefore, we investigated the radiation doses and organ and body surface displacement in abdominal DIBH (A-DIBH) and thoracic DIBH (T-DIBH). Free-breathing, A-DIBH, and T-DIBH computed tomography images of 100 patients were used. After contouring the targets, heart, and lungs, radiotherapy plans were created. We investigated the heart and lung doses, the associations between the heart and left lung displacements, and the thorax and abdominal surface displacements. No significant differences were observed in the target dose indices. However, the heart and lung doses were significantly lower in A-DIBH than in T-DIBH for all the indices; the mean heart and lung doses were 1.69 and 3.48 Gy, and 1.91 and 3.55 Gy in A-DIBH and T-DIBH, respectively. The inferior displacement of the heart and the left lung was more significant in A-DIBH. Therefore, inferior expansion of the heart and lungs may be responsible for the respective dose reductions. The abdominal surface displaced more than the thoracic surface in both A-DIBH and T-DIBH, and thoracic surface displacement was greater in T-DIBH than in A-DIBH. Moreover, A-DIBH can be identified because abdominal surface displacement was greater in A-DIBH than in T-DIBH. In conclusion, A-DIBH and T-DIBH could be distinguished by comparing the abdominal and thoracic surfaces of A-DIBH and T-DIBH, thereby ensuring the implementation of A-DIBH and reducing the heart and lung doses.

Examination of the dose distribution of volumetric modulated arc radiotherapy using a high-definition multi-leaf collimator for breast cancer patients with irradiated regional lymph nodes.

Background: A high-definition multi-leaf collimator (HD-MLC) with 5- and 10-mm fine MLCs is useful for radiotherapy. However, it is difficult to irradiate the mammary gland and supraclavicular region using a HD-MLC because of the narrow field of volumetric modulated arc radiotherapy (VMAT). Therefore, we aimed to evaluate the dose distribution of the VMAT dose using a HD-MLC in 15 patients with left breast cancer undergoing postoperative irradiation of breast and regional lymph nodes, including the internal mammary node. Materials and methods: The following four plans were generated: three-arc VMAT using HD-MLC (HD-VMAT), two tangential arcs and one-arc VMAT using HD-MLC (tHD-VMAT), three-dimensional conformal radiotherapy (3DCRT) using HD-MLC, and two-arc VMAT using the Millennium 120-leaf MLC (M-VMAT). We assessed the doses to the target volume and organs at risk. Results: The target dose distributions were higher for HD-VMAT than 3DCRT. There were no significant differences in the heart mean dose (Dmean) or lung volume receiving 20 Gy (V20 Gy) between HD-VMAT and 3DCRT. The heart Dmean and lung V20 Gy of tHD-VMAT were higher than those of HD-VMAT, and the heart Dmean of M-VMAT was higher than that of HD-VMAT. However, the target doses of tHD-VMAT, M-VMAT, and HD-VMAT were equivalent. Conclusions: In cases of the mammary gland and regional lymph node irradiation, including the internal mammary node in patients with left breast cancer, HD-VMAT was not inferior to M-VMAT and provided a better dose distribution to the target volume and organs at risk compared with 3DCRT and tHD-VMAT.

Comparing the toxicity and disease control rate of radiotherapy for prostate cancer between salvage settings after high-intensity focused ultrasound therapy and initial settings.

The purpose of this retrospective study was to compare the toxicity and disease control rate of radiotherapy for prostate cancer in salvage settings after high-intensity focused ultrasound (HIFU) therapy (HIFU cohort) with those in radical settings (non-HIFU cohort). From 2012 to 2020, 215 patients were identified for this study and 17 were treated in the salvage settings after HIFU. The median follow-up time was 34.5 months (range: 7-102 months, inter-quartile range [IQR]: 16-64 months). Genitourinary (GU) and gastrointestinal (GI) adverse events were evaluated in acute and late periods with Common Terminology Criteria for Adverse Events version 5, and the rates of biochemical-clinical failure free survival (BCFS) and overall survival (OS) were estimated. The cumulative incidence of late GU Grade 2 or greater toxicity after five years was significantly different between the non-HIFU and HIFU cohorts with rates of 7.3% and 26.2%, respectively (P = 0.03). Regarding GI Grade 2 or greater toxicity, there was no significant difference between the two cohorts. The 5y-BCFS was 84.2% in the non-HIFU cohort and 69.5% in the HIFU cohort with no significant difference (P = 0.10) and the 5y-OS was 95.9% and 92.3%, respectively (P = 0.47). We concluded that the possibility of increased late GU Grade 2 or greater should be considered when applying salvage radiotherapy for local recurrence after HIFU.

Effects of sample size and data augmentation on U-Net-based automatic segmentation of various organs.

Deep learning has demonstrated high efficacy for automatic segmentation in contour delineation, which is crucial in radiation therapy planning. However, the collection, labeling, and management of medical imaging data can be challenging. This study aims to elucidate the effects of sample size and data augmentation on the automatic segmentation of computed tomography images using U-Net, a deep learning method. For the chest and pelvic regions, 232 and 556 cases are evaluated, respectively. We investigate multiple conditions by changing the sum of the training and validation datasets across a broad range of values: 10-200 and 10-500 cases for the chest and pelvic regions, respectively. A U-Net is constructed, and horizontal-flip data augmentation, which produces left and right inverse images resulting in twice the number of images, is compared with no augmentation for each training session. All lung cases and more than 100 prostate, bladder, and rectum cases indicate that adding horizontal-flip data augmentation is almost as effective as doubling the number of cases. The slope of the Dice similarity coefficient (DSC) in all organs decreases rapidly until approximately 100 cases, stabilizes after 200 cases, and shows minimal changes as the number of cases is increased further. The DSCs stabilize at a smaller sample size with the incorporation of data augmentation in all organs except the heart. This finding is applicable to the automation of radiation therapy for rare cancers, where large datasets may be difficult to obtain.

Simple low-cost approaches to semantic segmentation in radiation therapy planning for prostate cancer using deep learning with non-contrast planning CT images.

PURPOSE: Deep learning has shown great efficacy for semantic segmentation. However, there are difficulties in the collection, labeling and management of medical imaging data, because of ethical complications and the limited number of imaging studies available at a single facility. This study aimed to find a simple and low-cost method to increase the accuracy of deep learning semantic segmentation for radiation therapy of prostate cancer. METHODS: In total, 556 cases with non-contrast CT images for prostate cancer radiation therapy were examined using a two-dimensional U-Net. Initially, all slices were used for the input data. Then, we removed slices of the cranial portions, which were beyond the margins of the bladder and rectum. Finally, the ground truth labels for the bladder and rectum were added as channels to the input for the prostate training dataset. RESULTS: The highest mean dice similarity coefficients (DSCs) for each organ in the test dataset of 56 cases were 0.85 ± 0.05, 0.94 ± 0.04 and 0.85 ± 0.07 for the prostate, bladder and rectum, respectively. Removal of the cranial slices from the original images significantly increased the DSC of the rectum from 0.83 ± 0.09 to 0.85 ± 0.07 (p < 0.05). Adding bladder and rectum information to prostate training without removing the slices significantly increased the DSC of the prostate from 0.79 ± 0.05 to 0.85 ± 0.05 (p < 0.05). CONCLUSIONS: These cost-free approaches may be useful for new applications, which may include updated models and datasets. They may be applicable to other organs at risk (OARs) and clinical targets such as elective nodal irradiation.

Efficacy evaluation of 2D, 3D U-Net semantic segmentation and atlas-based segmentation of normal lungs excluding the trachea and main bronchi.

This study aimed to examine the efficacy of semantic segmentation implemented by deep learning and to confirm whether this method is more effective than a commercially dominant auto-segmentation tool with regards to delineating normal lung excluding the trachea and main bronchi. A total of 232 non-small-cell lung cancer cases were examined. The computed tomography (CT) images of these cases were converted from Digital Imaging and Communications in Medicine (DICOM) Radiation Therapy (RT) formats to arrays of 32 × 128 × 128 voxels and input into both 2D and 3D U-Net, which are deep learning networks for semantic segmentation. The number of training, validation and test sets were 160, 40 and 32, respectively. Dice similarity coefficients (DSCs) of the test set were evaluated employing Smart SegmentationⓇ Knowledge Based Contouring (Smart segmentation is an atlas-based segmentation tool), as well as the 2D and 3D U-Net. The mean DSCs of the test set were 0.964 [95% confidence interval (CI), 0.960-0.968], 0.990 (95% CI, 0.989-0.992) and 0.990 (95% CI, 0.989-0.991) with Smart segmentation, 2D and 3D U-Net, respectively. Compared with Smart segmentation, both U-Nets presented significantly higher DSCs by the Wilcoxon signed-rank test (P < 0.01). There was no difference in mean DSC between the 2D and 3D U-Net systems. The newly-devised 2D and 3D U-Net approaches were found to be more effective than a commercial auto-segmentation tool. Even the relatively shallow 2D U-Net which does not require high-performance computational resources was effective enough for the lung segmentation. Semantic segmentation using deep learning was useful in radiation treatment planning for lung cancers.