Summary
Research in artificial intelligence for radiology and radiotherapy has recently become increasingly reliant on the use of deep learning‐based algorithms. While the performance of the models which these algorithms produce can significantly outperform more traditional machine learning methods, they do rely on larger datasets being available for training. To address this issue, data augmentation has become a popular method for increasing the size of a training dataset, particularly in fields where large datasets aren’t typically available, which is often the case when working with medical images. Data augmentation aims to generate additional data which is used to train the model and has been shown to improve performance when validated on a separate unseen dataset. This approach has become commonplace so to help understand the types of data augmentation techniques used in state‐of‐the‐art deep learning models, we conducted a systematic review of the literature where data augmentation was utilised on medical images (limited to CT and MRI) to train a deep learning model. Articles were categorised into basic, deformable, deep learning or other data augmentation techniques. As artificial intelligence models trained using augmented data make their way into the clinic, this review aims to give an insight to these techniques and confidence in the validity of the models produced.
The authors have developed a registration framework for mapping in vivo MRI data of the prostate with histology by implementing a number of processing steps and ex vivo MRI of the prostate specimen. Validation of DIR was challenging, particularly in prostates with few or mostly linear rather than spherical shaped features. Refinement of their MR imaging protocols to improve the data quality is currently underway which may improve registration accuracy. Additional mpMRI sequences will be registered within this framework to quantify prostate tumor location and biology.
Objectives
To test the hypothesis that observation with early salvage radiotherapy (SRT) is not inferior to ‘standard’ treatment with adjuvant RT (ART) with respect to biochemical failure in patients with pT3 disease and/or positive surgical margins (SMs) after radical prostatectomy (RP).
To compare the following secondary endpoints between the two arms: patient‐reported outcomes, adverse events, biochemical failure‐free survival, overall survival, disease‐specific survival, time to distant failure, time to local failure, cost utility analysis, quality adjusted life years and time to androgen deprivation.
Patients and Methods
The Radiotherapy – Adjuvant Versus Early Salvage (RAVES) trial is a phase III multicentre randomised controlled trial led by the Trans Tasman Radiation Oncology Group (TROG), in collaboration with the Urological Society of Australia and New Zealand (USANZ), and the Australian and New Zealand Urogenital and Prostate Cancer Trials Group (ANZUP).
In all, 470 patients are planned to be randomised 1:1 to either ART commenced at ≤4 months of RP (standard of care) or close observation with early SRT triggered by a PSA level of >0.20 ng/mL (experimental arm).
Eligible patients have had a RP for adenocarcinoma of the prostate with at least one of the following risk factors: positive SMs ± extraprostatic extension ± seminal vesicle involvement. The postoperative PSA level must be ≤0.10 ng/mL.
Rigorous investigator credentialing and a quality assurance programme are designed to promote consistent RT delivery among patients.
Results
Trial is currently underway, with 258 patients randomised as of 31 October 2013.
International collaborations have developed, including a planned meta‐analysis to be undertaken with the UK Medical Research Council/National Cancer Institute of Canada Clinical Trials Group RADICALS (Radiotherapy and Androgen Deprivation In Combination with Local Surgery) trial and an innovative psycho‐oncology sub‐study to investigate a patient decision aid resource.
Conclusion
On the current evidence available, it remains unclear if ART is equivalent or superior to observation with early SRT.
This study examined the variation of dose-volume histogram (DVH) data sourced from multiple radiotherapy treatment planning systems (TPSs). Treatment plan exports were obtained from 33 Australian and New Zealand centres during a dosimetry study. Plan information, including DVH data, was exported from the TPS at each centre and reviewed in a digital review system (SWAN). The review system was then used to produce an independent calculation of DVH information for each delineated structure. The relationships between DVHs extracted from each TPS and independently calculated were examined, particularly in terms of the influence of CT scan slice and pixel widths, the resolution of dose calculation grids and the TPS manufacturer. Calculation of total volume and DVH data was consistent between SWAN and each TPS, with the small discrepancies found tending to increase with decreasing structure size. This was significantly influenced by the TPS model used to derive the data. For target structures covered with relatively uniform dose distributions, there was a significant difference between the minimum dose in each TPS-exported DVH and that calculated independently.
A hypothetical, generic HDR (192)Ir source was designed and implemented in two commercially available TPSs employing different MBDCAs. Reference dose distributions for this source were benchmarked and used for the evaluation of MBDCA calculations employing a virtual, cubic water phantom in the form of a CT DICOM image series. The implementation of a generic source of identical design in all TPSs using MBDCAs is an important step toward supporting univocal commissioning procedures and direct comparisons between TPSs.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.