1
|
Fassia MK, Balasubramanian A, Woo S, Vargas HA, Hricak H, Konukoglu E, Becker AS. Deep Learning Prostate MRI Segmentation Accuracy and Robustness: A Systematic Review. Radiol Artif Intell 2024; 6:e230138. [PMID: 38568094 PMCID: PMC11294957 DOI: 10.1148/ryai.230138] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/26/2023] [Revised: 02/24/2024] [Accepted: 03/19/2024] [Indexed: 04/28/2024]
Abstract
Purpose To investigate the accuracy and robustness of prostate segmentation using deep learning across various training data sizes, MRI vendors, prostate zones, and testing methods relative to fellowship-trained diagnostic radiologists. Materials and Methods In this systematic review, Embase, PubMed, Scopus, and Web of Science databases were queried for English-language articles using keywords and related terms for prostate MRI segmentation and deep learning algorithms dated to July 31, 2022. A total of 691 articles from the search query were collected and subsequently filtered to 48 on the basis of predefined inclusion and exclusion criteria. Multiple characteristics were extracted from selected studies, such as deep learning algorithm performance, MRI vendor, and training dataset features. The primary outcome was comparison of mean Dice similarity coefficient (DSC) for prostate segmentation for deep learning algorithms versus diagnostic radiologists. Results Forty-eight studies were included. Most published deep learning algorithms for whole prostate gland segmentation (39 of 42 [93%]) had a DSC at or above expert level (DSC ≥ 0.86). The mean DSC was 0.79 ± 0.06 (SD) for peripheral zone, 0.87 ± 0.05 for transition zone, and 0.90 ± 0.04 for whole prostate gland segmentation. For selected studies that used one major MRI vendor, the mean DSCs of each were as follows: General Electric (three of 48 studies), 0.92 ± 0.03; Philips (four of 48 studies), 0.92 ± 0.02; and Siemens (six of 48 studies), 0.91 ± 0.03. Conclusion Deep learning algorithms for prostate MRI segmentation demonstrated accuracy similar to that of expert radiologists despite varying parameters; therefore, future research should shift toward evaluating segmentation robustness and patient outcomes across diverse clinical settings. Keywords: MRI, Genital/Reproductive, Prostate Segmentation, Deep Learning Systematic review registration link: osf.io/nxaev © RSNA, 2024.
Collapse
Affiliation(s)
- Mohammad-Kasim Fassia
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| | - Adithya Balasubramanian
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| | - Sungmin Woo
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| | - Hebert Alberto Vargas
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| | - Hedvig Hricak
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| | - Ender Konukoglu
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| | - Anton S. Becker
- From the Departments of Radiology (M.K.F.) and Urology (A.B.), New York-Presbyterian Weill Cornell Medical Center, 525 E 68th St, New York, NY 10065-4870; Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY (S.W., H.A.V., H.H., A.S.B.); and Department of Biomedical Imaging, ETH-Zurich, Zurich Switzerland (E.K.)
| |
Collapse
|
2
|
Impact of Training Data, Ground Truth and Shape Variability in the Deep Learning-Based Semantic Segmentation of HeLa Cells Observed with Electron Microscopy. J Imaging 2023; 9:jimaging9030059. [PMID: 36976110 PMCID: PMC10058680 DOI: 10.3390/jimaging9030059] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/30/2022] [Revised: 02/16/2023] [Accepted: 02/17/2023] [Indexed: 03/06/2023] Open
Abstract
This paper investigates the impact of the amount of training data and the shape variability on the segmentation provided by the deep learning architecture U-Net. Further, the correctness of ground truth (GT) was also evaluated. The input data consisted of a three-dimensional set of images of HeLa cells observed with an electron microscope with dimensions 8192×8192×517. From there, a smaller region of interest (ROI) of 2000×2000×300 was cropped and manually delineated to obtain the ground truth necessary for a quantitative evaluation. A qualitative evaluation was performed on the 8192×8192 slices due to the lack of ground truth. Pairs of patches of data and labels for the classes nucleus, nuclear envelope, cell and background were generated to train U-Net architectures from scratch. Several training strategies were followed, and the results were compared against a traditional image processing algorithm. The correctness of GT, that is, the inclusion of one or more nuclei within the region of interest was also evaluated. The impact of the extent of training data was evaluated by comparing results from 36,000 pairs of data and label patches extracted from the odd slices in the central region, to 135,000 patches obtained from every other slice in the set. Then, 135,000 patches from several cells from the 8192×8192 slices were generated automatically using the image processing algorithm. Finally, the two sets of 135,000 pairs were combined to train once more with 270,000 pairs. As would be expected, the accuracy and Jaccard similarity index improved as the number of pairs increased for the ROI. This was also observed qualitatively for the 8192×8192 slices. When the 8192×8192 slices were segmented with U-Nets trained with 135,000 pairs, the architecture trained with automatically generated pairs provided better results than the architecture trained with the pairs from the manually segmented ground truths. This suggests that the pairs that were extracted automatically from many cells provided a better representation of the four classes of the various cells in the 8192×8192 slice than those pairs that were manually segmented from a single cell. Finally, the two sets of 135,000 pairs were combined, and the U-Net trained with these provided the best results.
Collapse
|
3
|
Astono IP, Welsh JS, Rowe CW, Jobling P. Objective quantification of nerves in immunohistochemistry specimens of thyroid cancer utilising deep learning. PLoS Comput Biol 2022; 18:e1009912. [PMID: 35226665 PMCID: PMC8912900 DOI: 10.1371/journal.pcbi.1009912] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/27/2021] [Revised: 03/10/2022] [Accepted: 02/10/2022] [Indexed: 11/18/2022] Open
Abstract
Accurate quantification of nerves in cancer specimens is important to understand cancer behaviour. Typically, nerves are manually detected and counted in digitised images of thin tissue sections from excised tumours using immunohistochemistry. However the images are of a large size with nerves having substantial variation in morphology that renders accurate and objective quantification difficult using existing manual and automated counting techniques. Manual counting is precise, but time-consuming, susceptible to inconsistency and has a high rate of false negatives. Existing automated techniques using digitised tissue sections and colour filters are sensitive, however, have a high rate of false positives. In this paper we develop a new automated nerve detection approach, based on a deep learning model with an augmented classification structure. This approach involves pre-processing to extract the image patches for the deep learning model, followed by pixel-level nerve detection utilising the proposed deep learning model. Outcomes assessed were a) sensitivity of the model in detecting manually identified nerves (expert annotations), and b) the precision of additional model-detected nerves. The proposed deep learning model based approach results in a sensitivity of 89% and a precision of 75%. The code and pre-trained model are publicly available at https://github.com/IA92/Automated_Nerves_Quantification. The study of nerves as a prognostic marker for cancer is becoming increasingly important. However, accurate quantification of nerves in cancer specimens is difficult to achieve due to limitations in the existing manual and automated quantification methods. Manual quantification is time-consuming and subject to bias, whilst automated quantification, in general, has a high rate of false detections that makes it somewhat unreliable. In this paper, we propose an automated nerve quantification approach based on a novel deep learning model structure for objective nerve quantification in immunohistochemistry specimens of thyroid cancer. We evaluate the performance of the proposed approach by comparing it with existing manual and automated quantification methods. We show that our proposed approach is superior to the existing manual and automated quantification methods. The proposed approach is shown to have a high precision as well as being able to detect a significant number of nerves not detected by the experts in manual counting.
Collapse
Affiliation(s)
- Indriani P. Astono
- School of Engineering, The University of Newcastle, Newcastle, Australia
- * E-mail:
| | - James S. Welsh
- School of Engineering, The University of Newcastle, Newcastle, Australia
| | - Christopher W. Rowe
- School of Medicine and Public Health, The University of Newcastle, Newcastle, Australia
| | - Phillip Jobling
- School of Biomedical Sciences and Pharmacy, The University of Newcastle, Newcastle, Australia
| |
Collapse
|
4
|
Douglass MJJ, Keal JA. DeepWL: Robust EPID based Winston-Lutz analysis using deep learning, synthetic image generation and optical path-tracing. Phys Med 2021; 89:306-316. [PMID: 34492498 DOI: 10.1016/j.ejmp.2021.08.012] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 07/02/2021] [Revised: 08/03/2021] [Accepted: 08/27/2021] [Indexed: 12/23/2022] Open
Abstract
Radiation therapy requires clinical linear accelerators to be mechanically and dosimetrically calibrated to a high standard. One important quality assurance test is the Winston-Lutz test which localises the radiation isocentre of the linac. In the current work we demonstrate a novel method of analysing EPID based Winston-Lutz QA images using a deep learning model trained only on synthetic image data. In addition, we propose a novel method of generating the synthetic WL images and associated 'ground-truth' masks using an optical path-tracing engine to 'fake' mega-voltage EPID images. The model called DeepWL was trained on 1500 synthetic WL images using data augmentation techniques for 180 epochs. The model was built using Keras with a TensorFlow backend on an Intel Core i5-6500T CPU and trained in approximately 15 h. DeepWL was shown to produce ball bearing and multi-leaf collimator field segmentations with a mean dice coefficient of 0.964 and 0.994 respectively on previously unseen synthetic testing data. When DeepWL was applied to WL data measured on an EPID, the predicted mean displacements were shown to be statistically similar to the Canny Edge detection method. However, the DeepWL predictions for the ball bearing locations were shown to correlate better with manual annotations compared with the Canny edge detection algorithm. DeepWL was demonstrated to analyse Winston-Lutz images with an accuracy suitable for routine linac quality assurance with some statistical evidence that it may outperform Canny Edge detection methods in terms of segmentation robustness and the resultant displacement predictions.
Collapse
Affiliation(s)
- Michael John James Douglass
- School of Physical Sciences, University of Adelaide, Adelaide 5005, South Australia, Australia; Department of Medical Physics, Royal Adelaide Hospital, Adelaide 5000, South Australia, Australia.
| | - James Alan Keal
- School of Physical Sciences, University of Adelaide, Adelaide 5005, South Australia, Australia
| |
Collapse
|
5
|
Magnetic Resonance Imaging Based Radiomic Models of Prostate Cancer: A Narrative Review. Cancers (Basel) 2021; 13:cancers13030552. [PMID: 33535569 PMCID: PMC7867056 DOI: 10.3390/cancers13030552] [Citation(s) in RCA: 20] [Impact Index Per Article: 6.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/06/2020] [Revised: 01/18/2021] [Accepted: 01/27/2021] [Indexed: 12/11/2022] Open
Abstract
Simple Summary The increasing interest in implementing artificial intelligence in radiomic models has occurred alongside advancement in the tools used for computer-aided diagnosis. Such tools typically apply both statistical and machine learning methodologies to assess the various modalities used in medical image analysis. Specific to prostate cancer, the radiomics pipeline has multiple facets that are amenable to improvement. This review discusses the steps of a magnetic resonance imaging based radiomics pipeline. Present successes, existing opportunities for refinement, and the most pertinent pending steps leading to clinical validation are highlighted. Abstract The management of prostate cancer (PCa) is dependent on biomarkers of biological aggression. This includes an invasive biopsy to facilitate a histopathological assessment of the tumor’s grade. This review explores the technical processes of applying magnetic resonance imaging based radiomic models to the evaluation of PCa. By exploring how a deep radiomics approach further optimizes the prediction of a PCa’s grade group, it will be clear how this integration of artificial intelligence mitigates existing major technological challenges faced by a traditional radiomic model: image acquisition, small data sets, image processing, labeling/segmentation, informative features, predicting molecular features and incorporating predictive models. Other potential impacts of artificial intelligence on the personalized treatment of PCa will also be discussed. The role of deep radiomics analysis-a deep texture analysis, which extracts features from convolutional neural networks layers, will be highlighted. Existing clinical work and upcoming clinical trials will be reviewed, directing investigators to pertinent future directions in the field. For future progress to result in clinical translation, the field will likely require multi-institutional collaboration in producing prospectively populated and expertly labeled imaging libraries.
Collapse
|
6
|
Abstract
Automatic and accurate prostate segmentation is an essential prerequisite for assisting diagnosis and treatment, such as guiding biopsy procedures and radiation therapy. Therefore, this paper proposes a cascaded dual attention network (CDA-Net) for automatic prostate segmentation in MRI scans. The network includes two stages of RAS-FasterRCNN and RAU-Net. Firstly, RAS-FasterRCNN uses improved FasterRCNN and sequence correlation processing to extract regions of interest (ROI) of organs. This ROI extraction serves as a hard attention mechanism to focus the segmentation of the subsequent network on a certain area. Secondly, the addition of residual convolution block and self-attention mechanism in RAU-Net enables the network to gradually focus on the area where the organ exists while making full use of multiscale features. The algorithm was evaluated on the PROMISE12 and ASPS13 datasets and presents the dice similarity coefficient of 92.88% and 92.65%, respectively, surpassing the state-of-the-art algorithms. In a variety of complex slice images, especially for the base and apex of slice sequences, the algorithm also achieved credible segmentation performance.
Collapse
|