Although dose prediction for intensity modulated radiation therapy (IMRT) has been accomplished by a deep learning approach, delineation of some structures is needed for the prediction. We sought to develop a fully automated dose-generation framework for IMRT of prostate cancer by entering the patient CT datasets without the contour information into a generative adversarial network (GAN) and to compare its prediction performance to a conventional prediction model trained from patient contours. We propose a synthetic approach to translate patient CT datasets into a dose distribution for IMRT. The framework requires only paired-images, i.e., patient CT images and corresponding RT-doses. The model was trained from 81 IMRT plans of prostate cancer patients, and then produced the dose distribution for 9 test cases. To compare its prediction performance to that of another trained model, we created a model trained from structure images. Dosimetric parameters for the planning target volume (PTV) and organs at risk (OARs) were calculated from the generated and original dose distributions, and mean differences of dosimetric parameters were compared between the CT-based model and the structure-based model. The mean differences of all dosimetric parameters except for D98% and D95% for PTV were within approximately 2% and 3% of the prescription dose for OARs in the CT-based model, while the differences in the structure-based model were within approximately 1% for PTV and approximately 2% for OARs, with a mean prediction time of 5 seconds per patient. Accurate and rapid dose prediction was achieved by the learning of patient CT datasets by a GAN-based framework. The CT-based dose prediction could reduce the time required for both the iterative optimization process and the structure contouring, allowing physicians and dosimetrists to focus their expertise on more challenging cases.
BACKGROUND AND PURPOSE: Optimal tube voltage and tube current settings are not well established. The purpose of our study was to investigate the image quality on 3D CT angiograms of the brain at various kilovoltage settings by evaluating the depiction of simulated intracranial lesions by using a vascular phantom. MATERIALS AND METHODS: An anthropomorphic vascular phantom with superimposed bone skull structures was designed to simulate various intracranial aneurysms with aneurysmal blebs. We performed CT angiography by using a 64-detector row CT scanner for various effective tube currents with 4 tube voltages of 80, 100, 120, and 135 kV(p). Simulated aneurysm enhancement and image noise were quantified; SNR and CNR were calculated. The depiction of the simulated aneurysms and blebs on 3D CT angiograms obtained with the volume-rendering technique was subjectively assessed. The effective dose was calculated on the basis of a CTDIw. The results of several protocols were compared by using the Student t test. RESULTS: At identical doses levels (CTDIw), the mean SNR and CNR at 100 kV(p) were significantly higher than those at 80, 120, and 135 kV(p); and the mean qualitative image score at 100 kV(p) was significantly superior to those at 80 and 135 kV(p). CONCLUSIONS: Our phantom study suggests that the tube voltage of 100 kV(p) is desirable for cerebral 3D CT angiograms, and the higher or lower kilovoltage settings may result in the degradation in diagnostic image quality.
The present study compares dosimetric parameters between volumetric-modulated arc therapy (VMAT) and 3D conformal radiation therapy (3D-CRT) in lung tumors adjacent to the chest wall treated with stereotactic body radiation therapy (SBRT). The study focused on the radiation dose to the chest wall of 16 patients who had developed radiation-induced rib fractures (RIRF) after SBRT using 3D-CRT. The targets in all patients were partially overlapping with the fractured ribs, and the median overlapping rib–PTV distance was 0.4 cm. Stereotactic body radiation therapy was re-planned for all patients. The prescribed dose was 48 Gy in four fractions to cover at least 95% of the planning target volume (PTV). Evaluated dosimetric factors included D98% and the conformation number (CN) of the PTV, the D2cm3, V40 and V30 of the fractured ribs, the V30 of the chest wall, and the Dmean, V20 and V5 of the lung. A comparison of 3D-CRT with the VMAT plan for PTV revealed that CN was significantly improved in the VMAT plan, whereas D98% did not significantly differ between the two plans. Regarding organs at risk (OARs), the D2cm3, V40 and V30 of fractured ribs, the V30 of the chest wall, and the Dmean, V20 and V5 of the lung, were significantly decreased in the VMAT plan. We concluded that the dose to OARs such as ribs and chest wall could be reduced with improved target conformity using VMAT instead of 3D-CRT for SBRT to treat peripheral lung tumors.