BrachyView: Combining LDR seed positions with transrectal ultrasound imaging in a prostate gel phantom

Self-supervised learning for accelerated 3D high-resolution ultrasound imaging

PURPOSE

Ultrasound (US) imaging has been widely used in diagnosis, image-guided intervention, and therapy, where high-quality three-dimensional (3D) images are highly desired from sparsely acquired two-dimensional (2D) images. This study aims to develop a deep learning-based algorithm to reconstruct high-resolution (HR) 3D US images only reliant on the acquired sparsely distributed 2D images.

METHODS

We propose a self-supervised learning framework using cycle-consistent generative adversarial network (cycleGAN), where two independent cycleGAN models are trained with paired original US images and two sets of low-resolution (LR) US images, respectively. The two sets of LR US images are obtained through down-sampling the original US images along the two axes, respectively. In US imaging, in-plane spatial resolution is generally much higher than through-plane resolution. By learning the mapping from down-sampled in-plane LR images to original HR US images, cycleGAN can generate through-plane HR images from original sparely distributed 2D images. Finally, HR 3D US images are reconstructed by combining the generated 2D images from the two cycleGAN models.

RESULTS

The proposed method was assessed on two different datasets. One is automatic breast ultrasound (ABUS) images from 70 breast cancer patients, the other is collected from 45 prostate cancer patients. By applying a spatial resolution enhancement factor of 3 to the breast cases, our proposed method achieved the mean absolute error (MAE) value of 0.90 ± 0.15, the peak signal-to-noise ratio (PSNR) value of 37.88 ± 0.88 dB, and the visual information fidelity (VIF) value of 0.69 ± 0.01, which significantly outperforms bicubic interpolation. Similar performances have been achieved using the enhancement factor of 5 in these breast cases and using the enhancement factors of 5 and 10 in the prostate cases.

CONCLUSIONS

We have proposed and investigated a new deep learning-based algorithm for reconstructing HR 3D US images from sparely acquired 2D images. Significant improvement on through-plane resolution has been achieved by only using the acquired 2D images without any external atlas images. Its self-supervision capability could accelerate HR US imaging.

A review of brachytherapy physical phantoms developed over the last 20 years: clinical purpose and future requirements

Within the brachytherapy community, many phantoms are constructed in-house, and less commercial development is observed as compared to the field of external beam. Computational or virtual phantom design has seen considerable growth; however, physical phantoms are beneficial for brachytherapy, in which quality is dependent on physical processes, such as accuracy of source placement. Focusing on the design of physical phantoms, this review paper presents a summary of brachytherapy specific phantoms in published journal articles over the last twenty years (January 1, 2000 - December 31, 2019). The papers were analyzed and tabulated by their primary clinical purpose, which was deduced from their associated publications. A substantial body of work has been published on phantom designs from the brachytherapy community, but a standardized method of reporting technical aspects of the phantoms is lacking. In-house phantom development demonstrates an increasing interest in magnetic resonance (MR) tissue mimicking materials, which is not yet reflected in commercial phantoms available for brachytherapy. The evaluation of phantom design provides insight into the way, in which brachytherapy practice has changed over time, and demonstrates the customised and broad nature of treatments offered.

Endorectal digital prostate tomosynthesis (endoDPT): a proof-of-concept study

In this study we present endorectal digital prostate tomosynthesis (endoDPT), a proposed method of high resolution prostate imaging using an endorectal x-ray sensor and an external x-ray source. endoDPT may be useful for visualizing the fine detail of small structures such as low dose rate brachytherapy (LDRBT) seeds that are difficult to visualize with current methods. The resolution of endoDPT was characterized through measurement of the modulation transfer function (MTF) and artifact spread function (ASF) in computational and physical phantoms. The qualitative resolution of endoDPT was assessed relative to computed tomography (CT) through imaging of LDRBT seeds implanted inex vivocanine prostates. The x-ray sensor MTF reached 10% at 11.50 mm^-1, the reconstruction algorithm MTF reached a maximum at 7.08 mm^-1, and the ASF was 2.5 mm (full-width at half-maximum). Fine structures in LDRBT seeds like the 0.05 mm thick shell were visible with endoDPT but not CT. All endoDPT images exhibited an overshoot artifact. The measured MTFs were consistent with other studies using similar x-ray sensors and demonstrated improved resolution compared to digital breast tomosynthesis; this result was due to the smaller endoDPT x-ray sensor detection element size and quantitatively demonstrates the high resolution of endoDPT. The ASF demonstrated worse depth resolution compared to in-plane resolution, due to partial angular sampling; partial angular sampling also caused the observed overshoot artifact in the endoDPT images. However, endoDPT still was able to visualize fine structures such as the LDRBT seed shell to a much higher degree than CT. This high-resolution visualization may be useful for improvements in patient specific LDRBT dosimetry. Overall, these results indicate endoDPT is capable of high in-plane spatial resolution and is thus well poised for optimization and studies assessing clinical utility.

BrachyView: development of an algorithm for real-time automatic LDR brachytherapy seed detection

BrachyView is a novel in-body imaging system developed to provide real-time intraoperative dosimetry for low dose rate prostate brachytherapy treatments. Seed positions can be reconstructed after in-vivo implantation using a high-resolution pinhole gamma camera inserted into the patient rectum. The obtained data is a set of 2D projections of the seeds on the image plane. The 3D reconstruction algorithm requires the identification of the seed's centre of mass. This work presents the development and techniques adopted to build an algorithm that provides the means for fully automatic seed centre of mass identification and 3D position reconstruction for real-time applications. The algorithm presented uses a local feature detector, speeded up robust features, to perform detection of brachytherapy seed 2D projections from images, allowing for robust seed identification. Initial results have been obtained with datasets of 30, 96 and 98 I-125 brachytherapy seeds implanted into a prostate gel phantom. It can detect 97% of seeds and correctly match 97% of seeds. The average overall computation time of 2.75 s per image and improved reconstruction accuracy of 22.87% for the 98 seed dataset was noted. Elimination processes for initial false positive detection removal have shown to be extremely effective, resulting in a 99.9% reduction of false positives, and when paired with automatic frame alignment and subtraction procedures allows for the effective removal of excess counts generated by previously implanted needles. The proposed algorithm will allow the BrachyView system to be used as a real-time intraoperative dosimetry tool for low dose rate prostate brachytherapy treatments.

BrachyView: Reconstruction of seed positions and volume of an LDR prostate brachytherapy patient plan using a baseline subtraction algorithm

PURPOSE

BrachyView is a novel in-body imaging system developed with the objective to provide real-time intraoperative dosimetry for low dose rate (LDR) prostate brachytherapy treatments. The BrachyView coordinates combined with conventional transrectal ultrasound (TRUS) imaging, provides the possibility to localise the effective position of the implanted seeds inside the prostate volume, providing a unique tool for intra-operative verification of the quality of the implantation. This research presents the first complete LDR brachytherapy plan reconstructed by the BrachyView system and is used to evaluate the effectiveness of an imaging algorithm with baseline subtraction.

METHODS

A plan featuring 98 I-125 brachytherapy seeds, with an average activity of 0.248 mCi, were implanted into a prostate gel phantom under TRUS guidance. Images of implanted seeds were obtained by the BrachyView after the implantation of seeds. The baseline subtraction algorithm is applied as a pixel-to-pixel counts subtraction and is applied to every second projection obtained after the implantation of each needle. Seed positions and effectiveness of the baseline reconstruction in the identification of seeds were verified by a high-resolution post-implant CT scan.

RESULTS

A complete brachytherapy plan has been reconstructed with a 100% detection rate. This is possible due to the effectiveness of the baseline subtraction, with its application an overall increase of 11.3% in position accuracy and 8.2% increase in detection rate was noted.

CONCLUSION

It has been demonstrated that the BrachyView system shows the potential to be a solution to providing clinics with the means for intraoperative dosimetry for LDR prostate brachytherapy treatments.

A simulation study of BrachyShade, a shadow-based internal source tracking system for HDR prostate brachytherapy

This paper presents a simulation study of BrachyShade, a proposed internal source-tracking system for real time quality assurance in high dose rate prostate brachytherapy. BrachyShade consists of a set of spherical tungsten occluders located above a pixellated silicon photodetector. The source location is estimated by minimising the mean squared error between a parametric model of the shadow image and acquired images of the shadows projected on the detector plane. A novel algorithm is finally employed to correct the systemic error resulting from Compton scattering in the medium. The worst-case error obtained with BrachyShade for a 13.5 ms image acquisition is less than 1.3 mm in the most distant part of the treatment volume, while for 75% of source locations an error of less than 0.42 mm was achieved.

Semiconductor real-time quality assurance dosimetry in brachytherapy

With the increase in complexity of brachytherapy treatments, there has been a demand for the development of sophisticated devices for delivery verification. The Centre for Medical Radiation Physics (CMRP), University of Wollongong, has demonstrated the applicability of semiconductor devices to provide cost-effective real-time quality assurance for a wide range of brachytherapy treatment modalities. Semiconductor devices have shown great promise to the future of pretreatment and in vivo quality assurance in a wide range of brachytherapy treatments, from high-dose-rate (HDR) prostate procedures to eye plaque treatments. The aim of this article is to give an insight into several semiconductor-based dosimetry instruments developed by the CMRP. Applications of these instruments are provided for breast and rectal wall in vivo dosimetry in HDR brachytherapy, urethral in vivo dosimetry in prostate low-dose-rate (LDR) brachytherapy, quality assurance of HDR brachytherapy afterloaders, HDR pretreatment plan verification, and real-time verification of LDR and HDR source dwell positions.