Clinical evaluation of deep learning and atlas-based auto-segmentation for critical organs at risk in radiation therapy.

INTRODUCTION: Contouring organs at risk (OARs) is a time-intensive task that is a critical part of radiation therapy. Atlas-based automatic segmentation has shown some success at reducing this time burden on practitioners; however, this method often requires significant manual editing to reach a clinically accurate standard. Deep learning (DL) auto-segmentation has recently emerged as a promising solution. This study compares the accuracy of DL and atlas-based auto-segmentation in relation to clinical 'gold standard' reference contours. METHODS: Ninety CT datasets (30 head and neck, 30 thoracic, 30 pelvic) were automatically contoured using both atlas and DL segmentation techniques. Sixteen critical OARs were then quantitatively measured for accuracy using the Dice similarity coefficient (DSC) and Hausdorff distance (HD). Qualitative analysis was performed to visually classify the accuracy of each structure into one of four explicitly defined categories. Additionally, the time to edit atlas and DL contours to a clinically acceptable level was recorded for a subset of 9 OARs. RESULTS: Of the 16 OARs analysed, DL delivered statistically significant improvements over atlas segmentation in 13 OARs measured with DSC, 12 OARs measured with HD, and 12 OARs measured qualitatively. The mean editing time for the subset of DL contours was 50%, 23% and 61% faster (all P < 0.05) than that of atlas segmentation for the head and neck, thorax, and pelvis respectively. CONCLUSIONS: Deep learning segmentation comprehensively outperformed atlas-based contouring for the majority of evaluated OARs. Improvements were observed in geometric accuracy and visual acceptability, while editing time was reduced leading to increased workflow efficiency.

Software-based evaluation of a class solution for prostate IMRT planning.

AIM: To use plan analysis software to evaluate a class solution for prostate intensity modulated radiotherapy (IMRT) planning. BACKGROUND: Class solutions for radiotherapy planning are increasingly being considered for streamlining planning. Plan analysis software provides an objective approach to evaluating radiotherapy plans. MATERIALS AND METHODS: Three iterations of a class solution for prostate IMRT planning (T1, T2 and Tfinal) were compared to the clinical plan of 74 prostate patients using radiotherapy plan analysis software (Plan IQ™, Sun Nuclear Corporation). A set of institution-specific plan quality metrics (scores) were established, based on best practice guidelines. RESULTS: For CTV coverage, Tfinal was not significantly different to the clinical plan. With the exception of 95% PTV coverage, Tfinal metrics were significantly better than the clinical plan for PTV coverage. In the scoring analysis, mean dose, 95% and 107% isodose coverage scores were similar for all the templates and clinical plan. 100% coverage of the CTV clinical plan was similar to Tfinal but scored higher than T1 and T2. There were no significant differences between Tfinal and the clinical plan for the metrics and scores associated with organs at risk. The total plan score was similar for Tfinal and the clinical plan, although the scores for volume receiving total dose outside the PTV were higher for Tfinal than for the clinical plan (P < 0.0001). CONCLUSIONS: The radiotherapy plan analysis software was useful for evaluating a class solution for prostate IMRT planning and provided evidence that the class solution produced clinically acceptable plans for these patients.

Effects of contrast materials in IMRT and VMAT of prostate using a commercial Monte Carlo algorithm.

Contrast materials help in contouring in radiotherapy. The primary aim of this study is to investigate the effects of contrast materials in bladder on the dosimetry during prostate intensity modulated radiation therapy and volumetric modulated arc therapy. The study also investigates the difference of the two dose calculation options namely 'dose to medium (Dm)' and 'dose to water (Dw)' in a commercial Monte Carlo based treatment planning system. Eight IMRT treatment plans were retrospectively studied which were used to treat high risk prostate cancer patients. The treatment plans generated in Monaco treatment planning system use seven coplanar beams and calculated 'Dm' as the clinical option. These plans were recalculated, keeping the segments, beam angle and monitor units the same, with different relative electron densities assigned to the structure 'bladder' to mimic the presence of contrast material. The same plans were recalculated using the 'Dw' option. Further, keeping the IMRT constraints and plan calculation properties the same, these plans were re-optimised with the delivery method changed to volumetric modulated arc therapy and calculated using both 'Dm' and 'Dw' options. For all the four scenarios, it was found that for the target volumes CTV and PTV, 'minimum dose' is the only endpoint studied having a significant difference with the presence of contrast material. For bladder, the endpoint V40 Gy is affected. Any significant dosimetric effect is found only when the relative electron density of the contrast material is 1.2 or more. Also, the dosimetric difference is greater when 'Dm' option is used for calculation. For rectum, the dosimetry remains unaffected. Hence, contrast materials should be contoured and assigned appropriate relative electron densities during IMRT and VMAT treatment planning of prostate. Also, the difference in dose reported with the two dose calculation options (Dm and Dw) in the presence of contrast materials is significant.

Enhanced degree of temporal coherence through temporal and spatial phase coupling within a focused supercontinuum.

In the diffraction of a supercontinuum source, a redistribution of amplitude and phase at the focal region is incurred by the coupling between the supercontinuum and the spatial phase caused by the lens diffraction, making it extremely difficult to predict the focal behaviour. We show that the coupling between the temporal phase of a SC source and the spatial phase from the diffraction by a low numerical aperture (NA) lens causes dramatic alterations in the spectra and the temporal coherence near the focal region, and that this effect is maximized in points of singularity. Furthermore, we show that such an enhancement in temporal coherence can be controlled by the pulse evolution through the photonic crystal fiber, in which nonlinear and disperive effects such as the soliton fission process provides the key phase evolution necessary for dramatically changing the coherence time of the focused electromagnetic wave.

Polarization effects in a highly birefringent nonlinear photonic crystal fiber with two-zero dispersion wavelengths.

A theoretical and experimental study is presented on polarized pulsed propagation from a highly birefringent nonlinear photonic crystal fiber with two-zero dispersion wavelengths. Experimental observations show that the input polarization state can maintain its linearity and that the fiber birefringence creates different spectral properties dependent on the input polarization orientation. The most extensive spectra are obtained for a coupling polarization angles aligned with the fast and slow axis, which is created by the high-order dispersion and Kerr nonlinearity.