Machine-assisted interpolation algorithm for semi-automated segmentation of highly deformable organs.

PURPOSE: Accurate and robust auto-segmentation of highly deformable organs (HDOs), for example, stomach or bowel, remains an outstanding problem due to these organs' frequent and large anatomical variations. Yet, time-consuming manual segmentation of these organs presents a particular challenge to time-limited modern radiotherapy techniques such as on-line adaptive radiotherapy and high-dose-rate brachytherapy. We propose a machine-assisted interpolation (MAI) that uses prior information in the form of sparse manual delineations to facilitate rapid, accurate segmentation of the stomach from low field magnetic resonance images (MRI) and the bowel from computed tomography (CT) images. METHODS: Stomach MR images from 116 patients undergoing 0.35T MRI-guided abdominal radiotherapy and bowel CT images from 120 patients undergoing high dose rate pelvic brachytherapy treatment were collected. For each patient volume, the manual delineation of the HDO was extracted from every 8th slice. These manually drawn contours were first interpolated to obtain an initial estimate of the HDO contour. A two-channel 64 × 64 pixel patch-based convolutional neural network (CNN) was trained to localize the position of the organ's boundary on each slice within a five-pixel wide road using the image and interpolated contour estimate. This boundary prediction was then input, in conjunction with the image, to an organ closing CNN which output the final organ segmentation. A Dense-UNet architecture was used for both networks. The MAI algorithm was separately trained for the stomach segmentation and the bowel segmentation. Algorithm performance was compared against linear interpolation (LI) alone and against fully automated segmentation (FAS) using a Dense-UNet trained on the same datasets. The Dice Similarity Coefficient (DSC) and mean surface distance (MSD) metrics were used to compare the predictions from the three methods. Statistically significance was tested using Student's t test. RESULTS: For the stomach segmentation, the mean DSC from MAI (0.91 ± 0.02) was 5.0% and 10.0% higher as compared to LI and FAS, respectively. The average MSD from MAI (0.77 ± 0.25 mm) was 0.54 and 3.19 mm lower compared to the two other methods. Only 7% of MAI stomach predictions resulted in a DSC < 0.8, as compared to 30% and 28% for LI and FAS, respectively. For the bowel segmentation, the mean DSC of MAI (0.90 ± 0.04) was 6% and 18% higher, and the average MSD of MAI (0.93 ± 0.48 mm) was 0.42 and 4.9 mm lower as compared to LI and FAS. Sixteen percent of the predicted contour from MAI resulted in a DSC < 0.8, as compared to 46% and 60% for FAS and LI, respectively. All comparisons between MAI and the baseline methods were found to be statistically significant (p-value < 0.001). CONCLUSIONS: The proposed MAI algorithm significantly outperformed LI in terms of accuracy and robustness for both stomach segmentation from low-field MRIs and bowel segmentation from CT images. At this time, FAS methods for HDOs still require significant manual editing. Therefore, we believe that the MAI algorithm has the potential to expedite the process of HDO delineation within the radiation therapy workflow.

Dosimetric Quality of Online Adapted Pancreatic Cancer Treatment Plans on an MRI-Guided Radiation Therapy System.

PURPOSE: Stereotactic magnetic resonance image-guided adaptive radiation therapy (SMART) is an emerging technique that shows promise in the treatment of pancreatic cancer and other abdominopelvic malignancies. However, it is unknown whether the time-limited nature of on-table adaptive planning may result in dosimetrically suboptimal plans. The purpose of this study was to quantitatively address that question through systemic retrospective replanning of treated on-table adaptive pancreatic cancer cases. METHODS AND MATERIALS: Of 74 consecutive adapted fractions, 30 were retrospectively replanned based on deficiencies in planning target volume (PTV) and gross tumor volume (GTV) coverage or doses to organs-at-risk (OARs) that exceeded ideal constraints. Retrospective plans were created by adjusting dose-volume objectives in an iterative fashion until deemed optimized. The goal of replanning was to improve PTV/GTV coverage while keeping the dose to gastrointestinal OARs the same or lower or to reduce OAR doses while keeping PTV coverage the same or higher. The global maximum dose was required to be maintained within 2% of that of the treated adaptive plan to eliminate it as a confounding factor. A threshold of 5% improvement in PTV coverage or 5% decrease in OAR dose was used to define a clinically significant improvement. RESULTS: Of the 30 replans, 7 obtained at least 5% PTV coverage improvement. The average increase in PTV coverage for these plans was 11%. No plans were clinically significantly improved in terms of OAR sparing. Changes in beam-on time did not show any correlation. Statistical analysis via a linear mixed-effects model with a nested random effect suggested that both GTV and PTV coverage were improved over SMART process plans by 0.91 cc (P = .02) and 2.03 cc (P < .001), respectively. CONCLUSIONS: Dosimetric plan quality of at least 10% of SMART fractions may be improved through more extensive replanning than is currently performed on-table. Further work is needed to develop an automated replanning workflow to streamline the in-depth replanning process to better fit into an on-table adaptive workflow.

A comparison of the shielding effectiveness of silicone oil vitreous substitutes when used with Palladium-103 and Iodine-125 eye plaques.

PURPOSE: Episcleral eye plaques provide excellent local control of ocular melanoma, but vision sparing remains a significant problem with 30% or more of patients experiencing significant visual acuity degradation. The use of silicone oil shielding with Iodine-125 plaques has previously been reported to improve critical structure sparing. We hypothesized that the use of Palladium-103 would improve the shielding effectiveness of silicone oil due to the strong energy dependence of the photoelectric effect. This Monte Carlo simulation study reports a comparison of the shielding effects of silicone oil when used in conjunction with Pd-103 and with I-125 plaques. MATERIALS AND METHODS: GEANT4 was used to simulate eye plaque treatments to an eye with either water-equivalent vitreous humor, or silicone oil in place of the vitreous humor. Two solid gold plaques, 15 and 23 mm, were simulated loaded with I-125 and with Pd-103 source seeds. Seed activity was normalized such that 85 Gy was delivered to the tumor apex in the water-equivalent cases. Tumor apex dose, central axis dose, and inner sclera dose reductions with silicone oil were evaluated. RESULTS: Silicone oil resulted in an underdosing to the tumor apex of 6.1% and 7.5% in the 15 mm plaque for I-125 and Pd-103, respectively, and 3.4% and 4.3% in the 23 mm plaque for I-125 and Pd-103, respectively. When renormalized to 85 Gy to the tumor apex in all scenarios, silicone oil reduced the dose to the inner sclera 90° from the plaque by 19-32% for the 15 and 23 mm plaques using I-125, and by 33-65% for the 15 and 23 mm plaques using Pd-103. CONCLUSIONS: The combination of silicone oil and Pd-103 eye plaques offers the potential for greatly improved sparing to normal structures compared to Pd-103 plaques alone or I-125 plaques with or without silicone oil.

Technical Note: Dosimetric effects of couch position variability on treatment plan quality with an MRI-guided Co-60 radiation therapy machine.

PURPOSE: Magnetic resonance imaging (MRI) guidance in radiation therapy brings real-time imaging and adaptive planning into the treatment vault where it can account for interfraction and intrafraction movement of soft tissue. The only commercially available MRI-guided radiation therapy device is a three-head (60)Co and MRI system with an integrated treatment planning system (TPS). Couch attenuation of the beam of up to 20% is well modeled in the TPS. Variations in the patient's day-to-day position introduce discrepancies in the actual couch attenuation as modeled in the treatment plan. For this reason, the authors' institution avoids plans with beams that pass through or near the couch edges. This study investigates the effects of differential beam attenuation by the couch due to couch shifts in order to determine whether couch edge avoidance restrictions can be lifted. Couch shifts were simulated using a Monte Carlo treatment planning system and ion chamber measurements performed for validation. METHODS: A total of 27 plans from 23 patients were investigated. Couch shifts of 1 and 2 cm were introduced in combinations of lateral and vertical directions to simulate patient position variations giving 16 shifted plans per reference plan. The 1 and 2 cm shifts were based on shifts recorded in 320 treatment fractions. RESULTS: Following TG176 recommendations for measurement methods, couch attenuation measurements agreed with TPS modeled attenuation to within 2.1%. Planning target volume D95 changed less than 1% for 1 and 2 cm couch shifts in only the x-direction and less than 3% for all directions. CONCLUSIONS: Dosimetry of all plans tested was robust to couch shifts up to ±2 cm. In general, couch shifts resulted in clinically insignificant dosimetric deviations. It is conceivable that in certain cases with large systematic couch shifts and plans that are particularly sensitive to shifts, dosimetric changes might rise to a clinically significant level.

Visual Acuity, Contrast Sensitivity and Color Vision Three Years After Iodine-125 Brachytherapy for Choroidal and Ciliary Body Melanoma.

PURPOSE: To report visual acuity, contrast sensitivity and color vision prior to, 1 year after, 2 years after and 3 years after iodine-125 brachytherapy for choroidal and ciliary body melanoma (CCM). DESIGN: Prospective interventional case series. PARTICIPANTS: Thirty-seven patients (37 eyes) with CCM. METHODS: Patients had best-corrected Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity, Pelli-Robson contrast sensitivity and Hardy-Rand-Rittler color vision measurement; comprehensive ophthalmology examination; optical coherence tomography; and ultrasonography at baseline prior to, 1 year after, 2 years after and 3 years after I-125 brachytherapy. MAIN OUTCOME MEASURES: Visual acuity, contrast sensitivity and color vision prior to, 1 year after, 2 years after and 3 years after brachytherapy. RESULTS: Nineteen (19) men and 18 women with mean age of 58 years (SD 13, range 30-78) prior to, 1 year after, 2 years after and 3 years after brachytherapy had mean best-corrected visual acuity of 77 letters (20/32), 65 letters (20/50), 56 letters (20/80) and 47 letters (20/125); contrast sensitivity of 30, 26, 22 and 19 letters; color vision of 26, 20, 17 and 14 test figures, respectively. Decrease in visual acuity, contrast sensitivity and color vision was statistically significant from baseline at 1 year, 2 years, and 3 years after brachytherapy. Decreased acuity at 3 years was associated with mid-choroid and macula melanoma location, ≥ 4.1 mm melanoma height, radiation maculopathy and radiation optic neuropathy. CONCLUSION: 1, 2 and 3 years after brachytherapy, eyes with CCM had significantly decreased visual acuity, contrast sensitivity and color vision.

Image-guided radiotherapy using a mobile kilovoltage x-ray device.

A mobile isocentric C-arm kilovoltage imager has been evaluated as a potential tool for image-guided radiotherapy. The C-arm is equipped with an amorphous silicon flat panel for high-quality image acquisition. Additionally, the device is capable of cone beam computed tomography (CT) and volumetric reconstruction. This is achieved through the application of a modified Feldkamp algorithm with acquisition over a 180 degrees scan arc. The number of projections can be varied from 100 to 1000, resulting in a reconstructed volume 20 cm in diameter by 15-cm long. While acquisition time depends upon number of projections, acceptable quality images can be obtained in less than 60 seconds. Image resolution and contrast of cone-beam phantom images have been compared with images from a conventional CT scanner. The system has a spatial resolution of > or = 10 lp/cm and resolution is approximately equal in all 3 dimensions. Conversely, subject contrast is poorer than conventional CT, compromised by the increased scatter and underlying noise inherent in cone beam reconstruction, as well as the absence of filtering prior to reconstruction. The mobility of the C-arm makes it necessary to determine the C-arm position relative to the linear accelerator isocenter. Two solutions have been investigated: (1) the use of fiducial markers, embedded in the linac couch, that can subsequently be registered in the image sets; and (2), a navigation approach for infrared tracking of the C-arm relative to the linac isocenter. Observed accuracy in phantom positioning ranged from 1.0 to 1.5 mm using the navigation approach and 1.5 to 2.5 mm using the fiducial-based approach. As part of this work, the impact of respiratory motion on cone-beam image quality was evaluated, and a scheme for retrospective gating was devised. Results demonstrated that kilovoltage cone beam CT provides spatial integrity and resolution comparable to conventional CT. Cone-beam CT studies of patients undergoing radiotherapy have demonstrated acceptable soft tissue contrast, allowing assessment of daily changes in target anatomy. Of the 2 approaches developed to register images to the linac isocenter, the navigation method demonstrated superior accuracy for daily patient positioning relative to the fiducial-based method. Finally, significant image degradation due to respiratory motion was observed. It was demonstrated that this could be improved by correlating the acquisition of individual 2D projections with respiration for retrospective reconstruction of phase-based volumetric datasets.

Respiratory correlated cone-beam computed tomography on an isocentric C-arm.

A methodology for 3D image reconstruction from retrospectively gated cone-beam CT projection data has been developed. A mobile x-ray cone-beam device consisting of an isocentric C-arm equipped with a flat panel detector was used to image a moving phantom. Frames for reconstruction were retrospectively selected from complete datasets based on the known rotation of the C-arm and a signal from a respiratory monitor. Different sizes of gating windows were tested. A numerical criterion for blur on the reconstructed image was suggested. The criterion is based on minimization of an Ising energy function, similar to approaches used in image segmentation or restoration. It is shown that this criterion can be used for the determination of the optimal gating window size. Images reconstructed from the retrospectively gated projection sequences using the optimal gating window data showed a significant improvement compared to images reconstructed from the complete projection datasets.