Publications

@article{Boone2023,

title = {ROOD-MRI: Benchmarking the robustness of deep learning segmentation models to out-of-distribution and corrupted data in MRI},

author = {Lyndon Boone and Mahdi Biparva and Parisa Mojiri Forooshani and Joel Ramirez and Mario Masellis and Robert Bartha and Sean Symons and Stephen Strother and Sandra E. Black and Chris Heyn and Anne L. Martel and Richard H. Swartz and Maged Goubran},

url = {https://doi.org/10.1016/j.neuroimage.2023.120289 https://linkinghub.elsevier.com/retrieve/pii/S1053811923004408},

doi = {10.1016/j.neuroimage.2023.120289},

issn = {10538119},

year  = {2023},

date = {2023-09-01},

journal = {NeuroImage},

volume = {278},

pages = {120289},

abstract = {Deep artificial neural networks (DNNs) have moved to the forefront of medical image analysis due to their success in classification, segmentation, and detection challenges. A principal challenge in large-scale deployment of DNNs in neuroimage analysis is the potential for shifts in signal-to-noise ratio, contrast, resolution, and presence of artifacts from site to site due to variances in scanners and acquisition protocols. DNNs are famously susceptible to these distribution shifts in computer vision. Currently, there are no benchmarking platforms or frameworks to assess the robustness of new and existing models to specific distribution shifts in MRI, and accessible multi-site benchmarking datasets are still scarce or task-specific. To address these limitations, we propose ROOD-MRI: a novel platform for benchmarking the Robustness of DNNs to Out-Of-Distribution (OOD) data, corruptions, and artifacts in MRI. This flexible platform provides modules for generating benchmarking datasets using transforms that model distribution shifts in MRI, implementations of newly derived benchmarking metrics for image segmentation, and examples for using the methodology with new models and tasks. We apply our methodology to hippocampus, ventricle, and white matter hyperintensity segmentation in several large studies, providing the hippocampus dataset as a publicly available benchmark. By evaluating modern DNNs on these datasets, we demonstrate that they are highly susceptible to distribution shifts and corruptions in MRI. We show that while data augmentation strategies can substantially improve robustness to OOD data for anatomical segmentation tasks, modern DNNs using augmentation still lack robustness in more challenging lesion-based segmentation tasks. We finally benchmark U-Nets and vision transformers, finding robustness susceptibility to particular classes of transforms across architectures. The presented open-source platform enables generating new benchmarking datasets and comparing across models to study model design that results in improved robustness to OOD data and corruptions in MRI.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Bahlmann2023,

title = {Targeting tumour-associated macrophages in hodgkin lymphoma using engineered extracellular matrix-mimicking cryogels},

author = {Laura C. Bahlmann and Chang Xue and Allysia A. Chin and Arianna Skirzynska and Joy Lu and Brigitte Thériault and David Uehling and Yulia Yerofeyeva and Rachel Peters and Kela Liu and Jianan Chen and Anne L. Martel and Martin Yaffe and Rima Al-awar and Rashmi S. Goswami and Jarkko Ylanko and David W. Andrews and John Kuruvilla and Rob C. Laister and Molly S. Shoichet},

url = {https://doi.org/10.1016/j.biomaterials.2023.122121 https://linkinghub.elsevier.com/retrieve/pii/S0142961223001291},

doi = {10.1016/j.biomaterials.2023.122121},

issn = {01429612},

year  = {2023},

date = {2023-06-01},

journal = {Biomaterials},

volume = {297},

pages = {122121},

abstract = {Tumour-associated macrophages are linked with poor prognosis and resistance to therapy in Hodgkin lymphoma; however, there are no suitable preclinical models to identify macrophage-targeting therapeutics. We used primary human tumours to guide the development of a mimetic cryogel, wherein Hodgkin (but not Non-Hodgkin) lymphoma cells promoted primary human macrophage invasion. In an invasion inhibitor screen, we identified five drug hits that significantly reduced tumour-associated macrophage invasion: marimastat, batimastat, AS1517499, ruxolitinib, and PD-169316. Importantly, ruxolitinib has demonstrated recent success in Hodgkin lymphoma clinical trials. Both ruxolitinib and PD-169316 (a p38 mitogen-activated protein kinase (p38 MAPK) inhibitor) decreased the percent of M2-like macrophages; however, only PD-169316 enhanced the percentage of M1-like macrophages. We validated p38 MAPK as an anti-invasion drug target with five additional drugs using a high-content imaging platform. With our biomimetic cryogel, we modeled macrophage invasion in Hodgkin lymphoma and then used it for target discovery and drug screening, ultimately identifying potential future therapeutics.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Moncayo2023,

title = {Removing non-nuclei information from histopathological images: A preprocessing step towards improving nuclei segmentation methods},

author = {Ricardo Moncayo and Anne L. Martel and Eduardo Romero},

url = {https://linkinghub.elsevier.com/retrieve/pii/S2153353923001293},

doi = {10.1016/j.jpi.2023.100315},

issn = {21533539},

year  = {2023},

date = {2023-04-01},

journal = {Journal of Pathology Informatics},

pages = {100315},

abstract = {Disease interpretation by computer-aided diagnosis systems in digital pathology depends on reliable detection and segmentation of nuclei in hematoxylin and eosin (HE) images. These 2 tasks are challenging since appearance of both cell nuclei and background structures are very variable. This paper presents a method to improve nuclei detection and segmentation in HE images by removing tiles that only contain background information. The method divides each image into smaller patches and uses their projection to the noiselet space to capture different spatial features from non-nuclei background and nuclei structures. The noiselet features are clustered by a K-means algorithm and the resultant partition, defined by the cluster centroids, is herein named the noiselet code-book. A part of an image, a tile, is divided into patches and represented by the histogram of occurrences of the projected patches in the noiselet code-book. Finally, with these histograms, a classifier learns to differentiate between nuclei and non-nuclei tiles. By applying a conventional watershed-marked method to detect and segment nuclei, evaluation consisted in comparing pure watershed method against denoising-plus-watershed in an open database with 8 different types of tissues. The averaged F-score of nuclei detection improved from 0.830 to 0.86 and the dice score after segmentation increased from 0.701 to 0.723.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Chen2023,

title = {A radiomic biomarker for prognosis of resected colorectal cancer liver metastases generalizes across MRI contrast agents},

author = {Jianan Chen and Helen M C Cheung and Paul J Karanicolas and Natalie G Coburn and Guillaume Martel and Albert Lee and Chirag Patel and Laurent Milot and Anne L Martel},

url = {http://www.ncbi.nlm.nih.gov/pubmed/36816920 http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC9932499 https://www.frontiersin.org/articles/10.3389/fonc.2023.898854/full},

doi = {10.3389/fonc.2023.898854},

issn = {2234-943X},

year  = {2023},

date = {2023-02-01},

journal = {Frontiers in Oncology},

volume = {13},

pages = {898854},

abstract = {INTRODUCTION Contrast-enhanced MRI is routinely performed as part of preoperative work-up for patients with Colorectal Cancer Liver Metastases (CRLM). Radiomic biomarkers depicting the characteristics of CRLMs in MRI have been associated with overall survival (OS) of patients, but the reproducibility and clinical applicability of these biomarkers are limited due to the variations in MRI protocols between hospitals. METHODS In this work, we propose a generalizable radiomic model for predicting OS of CRLM patients who received preoperative chemotherapy and delayed-phase contrast enhanced (DPCE) MRIs prior to hepatic resection. This retrospective two-center study included three DPCE MRI cohorts (n=221) collected between January 2006 and December 2012. A 10-minute delayed Gd-DO3A-butrol enhanced MRI discovery cohort was used to select features based on robustness across contrast agents, correlation with OS and pairwise Pearson correlation, and to train a logistic regression model that predicts 3-year OS. RESULTS The model was evaluated on a 10-minute delayed Gd-DO3A-butrol enhanced MRI validation cohort (n=121), a 20-minute delayed Gd-EOB-DTPA (n=72) cohort from the same institute, and a 5-minute delayed Gd-DTPA cohort (n=28) from an independent institute. Two features were selected: minor axis length and dependence variance. The radiomic signature model stratified high-risk and low-risk CRLM groups in the Gd-DO3Abutrol (HR = 6.29},

keywords = {colorectal cancer, contrast agents, liver, metastasis, MRI, radiomics},

pubstate = {published},

tppubtype = {article}

}

@article{Han2022,

title = {Cell segmentation for immunofluorescence multiplexed images using two-stage domain adaptation and weakly labeled data for pre-training},

author = {Wenchao Han and Alison M. Cheung and Martin J. Yaffe and Anne L. Martel},

url = {https://www.nature.com/articles/s41598-022-08355-1},

doi = {10.1038/s41598-022-08355-1},

issn = {2045-2322},

year  = {2022},

date = {2022-12-01},

journal = {Scientific Reports},

volume = {12},

number = {1},

pages = {4399},

abstract = {Cellular profiling with multiplexed immunofluorescence (MxIF) images can contribute to a more accurate patient stratification for immunotherapy. Accurate cell segmentation of the MxIF images is an essential step. We propose a deep learning pipeline to train a Mask R-CNN model (deep network) for cell segmentation using nuclear (DAPI) and membrane (Na + K + ATPase) stained images. We used two-stage domain adaptation by first using a weakly labeled dataset followed by fine-tuning with a manually annotated dataset. We validated our method against manual annotations on three different datasets. Our method yields comparable results to the multi-observer agreement on an ovarian cancer dataset and improves on state-of-the-art performance on a publicly available dataset of mouse pancreatic tissues. Our proposed method, using a weakly labeled dataset for pre-training, showed superior performance in all of our experiments. When using smaller training sample sizes for fine-tuning, the proposed method provided comparable performance to that obtained using much larger training sample sizes. Our results demonstrate that using two-stage domain adaptation with a weakly labeled dataset can effectively boost system performance, especially when using a small training sample size. We deployed the model as a plug-in to CellProfiler, a widely used software platform for cellular image analysis.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Kuling_nlp_2022,

title = {BI-RADS BERT and Using Section Segmentation to Understand Radiology Reports},

author = {Grey Kuling and Belinda Curpen and Anne L. Martel},

url = {https://www.mdpi.com/2313-433X/8/5/131},

doi = {10.3390/jimaging8050131},

issn = {2313-433X},

year  = {2022},

date = {2022-05-01},

journal = {Journal of Imaging},

volume = {8},

number = {5},

pages = {131},

abstract = {Radiology reports are one of the main forms of communication between radiologists and other clinicians, and contain important information for patient care. In order to use this information for research and automated patient care programs, it is necessary to convert the raw text into structured data suitable for analysis. State-of-the-art natural language processing (NLP) domain-specific contextual word embeddings have been shown to achieve impressive accuracy for these tasks in medicine, but have yet to be utilized for section structure segmentation. In this work, we pre-trained a contextual embedding BERT model using breast radiology reports and developed a classifier that incorporated the embedding with auxiliary global textual features in order to perform section segmentation. This model achieved 98% accuracy in segregating free-text reports, sentence by sentence, into sections of information outlined in the Breast Imaging Reporting and Data System (BI-RADS) lexicon, which is a significant improvement over the classic BERT model without auxiliary information. We then evaluated whether using section segmentation improved the downstream extraction of clinically relevant information such as modality/procedure, previous cancer, menopausal status, purpose of exam, breast density, and breast MRI background parenchymal enhancement. Using the BERT model pre-trained on breast radiology reports, combined with section segmentation, resulted in an overall accuracy of 95.9% in the field extraction tasks. This is a 17% improvement, compared to an overall accuracy of 78.9% for field extraction with models using classic BERT embeddings and not using section segmentation. Our work shows the strength of using BERT in the analysis of radiology reports and the advantages of section segmentation by identifying the key features of patient factors recorded in breast radiology reports.},

keywords = {NLP},

pubstate = {published},

tppubtype = {article}

}

@article{Ciga2022,

title = {Self supervised contrastive learning for digital histopathology},

author = {Ozan Ciga and Tony Xu and Anne L Martel},

url = {https://linkinghub.elsevier.com/retrieve/pii/S2666827021000992},

doi = {10.1016/j.mlwa.2021.100198},

issn = {26668270},

year  = {2022},

date = {2022-03-01},

urldate = {2022-03-01},

journal = {Machine Learning with Applications},

volume = {7},

pages = {100198},

abstract = {Unsupervised learning has been a long-standing goal of machine learning and is especially important for medical image analysis, where the learning can compensate for the scarcity of labeled datasets. A promising subclass of unsupervised learning is self-supervised learning, which aims to learn salient features using the raw input as the learning signal. In this paper, we use a contrastive self-supervised learning method Chen et al. (2020a) that achieved state-of-the-art results on natural-scene images, and apply this method to digital histopathology by collecting and training on 60 histopathology datasets without any labels. We find that combining multiple multi-organ datasets with different types of staining and resolution properties improves the quality of the learned features. Furthermore, we find drastically subsampling a dataset (e.g., using ? 1% of the available image patches) does not negatively impact the learned representations, unlike training on natural-scene images. Linear classifiers trained on top of the learned features show that networks pretrained on digital histopathology datasets perform better than ImageNet pretrained networks, boosting task performances up to 7.5% in accuracy and 8.9% in F1. These findings may also be useful when applying newer contrastive techniques to histopathology data. Pretrained PyTorch models are made publicly available at https://github.com/ozanciga/self-supervised-histopathology.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Bilocq-Lacoste2022,

title = {Missed Breast Cancers on MRI in High-Risk Patients: A Retrospective Case–Control Study},

author = {Julie Bilocq-Lacoste and Romuald Ferre and Grey Kuling and Anne L. Martel and Pascal N. Tyrrell and Siying Li and Guan Wang and Belinda Curpen},

url = {https://www.mdpi.com/2379-139X/8/1/27},

doi = {10.3390/tomography8010027},

issn = {2379-139X},

year  = {2022},

date = {2022-02-01},

journal = {Tomography},

volume = {8},

number = {1},

pages = {329–340},

abstract = {Purpose: To determine if MRI features and molecular subtype influence the detectability of breast cancers on MRI in high-risk patients. Methods and Materials: Breast cancers in a high-risk population of 104 patients were diagnosed following MRI describing a BI-RADS 4–5 lesion. MRI characteristics at the time of diagnosis were compared with previous MRI, where a BI-RADS 1–2–3 lesion was described. Results: There were 77 false-negative MRIs. A total of 51 cancers were overlooked and 26 were misinterpreted. There was no association found between MRI characteristics, the receptor type and the frequency of missed cancers. The main factors for misinterpreted lesions were multiple breast lesions, prior biopsy/surgery and long-term stability. Lesions were mostly overlooked because of their small size and high background parenchymal enhancement. Among missed lesions, 50% of those with plateau kinetics on initial MRI changed for washout kinetics, and 65% of initially progressively enhancing lesions then showed plateau or washout kinetics. There were more basal-like tumours in BRCA1 carriers (50%) than in non-carriers (13%)},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Reinke2022,

title = {Metrics Reloaded - A new recommendation framework for biomedical image analysis validation},

author = {Annika Reinke and Lena Maier-Hein and Evangelia Christodoulou and Ben Glocker and Patrick Scholz and Fabian Isensee and Jens Kleesiek and Michal Kozubek and Mauricio Reyes and Michael Alexander Riegler and Manuel Wiesenfarth and Michael Baumgartner and Matthias Eisenmann and Doreen Heckmann-Nötzel and Ali Emre Kavur and Tim Rädsch and Minu D. Tizabi and Laura Acion and Michela Antonelli and Tal Arbel and Spyridon Bakas and Peter Bankhead and Arriel Benis and M. Jorge Cardoso and Veronika Cheplygina and Beth A Cimini and Gary S. Collins and Keyvan Farahani and Bram Ginneken and Fred A Hamprecht and Daniel A. Hashimoto and Michael M. Hoffman and Merel Huisman and Pierre Jannin and Charles Kahn and Alexandros Karargyris and Alan Karthikesalingam and Hannes Kenngott and Annette Kopp-Schneider and Anna Kreshuk and Tahsin Kurc and Bennett A. Landman and Geert Litjens and Amin Madani and Klaus Maier-Hein and Anne Martel and Peter Mattson and Erik Meijering and Bjoern Menze and David Moher and Karel G. M. Moons and Henning Müller and Brennan Nichyporuk and Felix Nickel and Jens Petersen and Nasir Rajpoot and Nicola Rieke and Julio Saez-Rodriguez and Clara I. Sánchez and Shravya Shetty and Maarten Smeden and Carole H. Sudre and Ronald M. Summers and Abdel A. Taha and Sotirios A. Tsaftaris and Ben Van Calster and Gael Varoquaux and Paul F Jaeger},

year  = {2022},

date = {2022-01-01},

journal = {Medical Imaging with Deep Learning},

abstract = {Meaningful performance assessment of biomedical image analysis algorithms depends on objective and appropriate performance metrics. There are major shortcomings in the current state of the art. Yet, so far limited attention has been paid to practical pitfalls associated when using particular metrics for image analysis tasks. Therefore, a number of international initiatives have collaborated to offer researchers with guidance and tools for selecting performance metrics in a problem-aware manner. In our proposed framework, the characteristics of the given biomedical problem are first captured in a problem fingerprint, which identifies properties related to domain interests, the target structure(s), the input datasets, and algorithm output. A problem category-specific mapping is applied in the second step to match fingerprints to metrics that reflect domain requirements. Based on input from experts from more than 60 institutions worldwide, we believe our metric recommendation framework to be useful to the MIDL community and to enhance the quality of biomedical image analysis algorithm validation.},

keywords = {Classification, Instance Segmentation, Medical Imaging, Metrics, Object Detection, Validation},

pubstate = {published},

tppubtype = {article}

}

@article{Cheung2021,

title = {Quantitative single-cell analysis of immunofluorescence protein multiplex images illustrates biomarker spatial heterogeneity within breast cancer subtypes},

author = {Alison Cheung and Dan Wang and Kela Liu and Tyna Hope and Mayan Murray and Fiona Ginty and Sharon Nofech-Mozes and Anne L Martel and Martin Joel Yaffe},

url = {https://breast-cancer-research.biomedcentral.com/articles/10.1186/s13058-021-01475-y},

doi = {10.1186/s13058-021-01475-y},

issn = {1465-542X},

year  = {2021},

date = {2021-12-01},

urldate = {2021-12-01},

journal = {Breast Cancer Research},

volume = {23},

number = {1},

pages = {114},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Ciga2021a,

title = {Overcoming the limitations of patch-based learning to detect cancer in whole slide images},

author = {Ozan Ciga and Tony Xu and Sharon Nofech-Mozes and Shawna Noy and Fang-i Lu and Anne L Martel},

url = {http://arxiv.org/abs/2012.00617 http://www.nature.com/articles/s41598-021-88494-z},

doi = {10.1038/s41598-021-88494-z},

issn = {2045-2322},

year  = {2021},

date = {2021-12-01},

journal = {Scientific Reports},

volume = {11},

number = {1},

pages = {8894},

abstract = {Whole slide images (WSIs) pose unique challenges when training deep learning models. They are very large which makes it necessary to break each image down into smaller patches for analysis, image features have to be extracted at multiple scales in order to capture both detail and context, and extreme class imbalances may exist. Significant progress has been made in the analysis of these images, thanks largely due to the availability of public annotated datasets. We postulate, however, that even if a method scores well on a challenge task, this success may not translate to good performance in a more clinically relevant workflow. Many datasets consist of image patches which may suffer from data curation bias; other datasets are only labelled at the whole slide level and the lack of annotations across an image may mask erroneous local predictions so long as the final decision is correct. In this paper, we outline the differences between patch or slide-level classification versus methods that need to localize or segment cancer accurately across the whole slide, and we experimentally verify that best practices differ in both cases. We apply a binary cancer detection network on post neoadjuvant therapy breast cancer WSIs to find the tumor bed outlining the extent of cancer, a task which requires sensitivity and precision across the whole slide. We extensively study multiple design choices and their effects on the outcome, including architectures and augmentations. We propose a negative data sampling strategy, which drastically reduces the false positive rate (25% of false positives versus 62.5%) and improves each metric pertinent to our problem, with a 53% reduction in the error of tumor extent. Our results indicate classification performances of image patches versus WSIs are inversely related when the same negative data sampling strategy is used. Specifically, injection of negatives into training data for image patch classification degrades the performance, whereas the performance is improved for slide and pixel-level WSI classification tasks. Furthermore, we find applying extensive augmentations helps more in WSI-based tasks compared to patch-level image classification.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Srinidhi2021,

title = {Improving Self-supervised Learning with Hardness-aware Dynamic Curriculum Learning: An Application to Digital Pathology},

author = {Chetan L Srinidhi and Anne L Martel},

url = {http://arxiv.org/abs/2108.07183},

year  = {2021},

date = {2021-08-01},

journal = {ICCV 2021 CDpath workshop},

abstract = {Self-supervised learning (SSL) has recently shown tremendous potential to learn generic visual representations useful for many image analysis tasks. Despite their notable success, the existing SSL methods fail to generalize to downstream tasks when the number of labeled training instances is small or if the domain shift between the transfer domains is significant. In this paper, we attempt to improve self-supervised pretrained representations through the lens of curriculum learning by proposing a hardness-aware dynamic curriculum learning (HaDCL) approach. To improve the robustness and generalizability of SSL, we dynamically leverage progressive harder examples via easy-to-hard and hard-to-very-hard samples during mini-batch downstream fine-tuning. We discover that by progressive stage-wise curriculum learning, the pretrained representations are significantly enhanced and adaptable to both in-domain and out-of-domain distribution data. We performed extensive validation on three histology benchmark datasets on both patch-wise and slide-level classification problems. Our curriculum based fine-tuning yields a significant improvement over standard fine-tuning, with a minimum improvement in area-under-the-curve (AUC) score of 1.7% and 2.2% on in-domain and out-of-domain distribution data, respectively. Further, we empirically show that our approach is more generic and adaptable to any SSL methods and does not impose any additional overhead complexity. Besides, we also outline the role of patch-based versus slide-based curriculum learning in histopathology to provide practical insights into the success of curriculum based fine-tuning of SSL methods. Code is released at https://github.com/srinidhiPY/ICCV-CDPATH2021-ID-8},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Ma2021,

title = {Loss odyssey in medical image segmentation},

author = {Jun Ma and Jianan Chen and Matthew Ng and Rui Huang and Yu Li and Chen Li and Xiaoping Yang and Anne L Martel},

url = {https://linkinghub.elsevier.com/retrieve/pii/S1361841521000815},

doi = {10.1016/j.media.2021.102035},

issn = {13618415},

year  = {2021},

date = {2021-07-01},

journal = {Medical Image Analysis},

volume = {71},

pages = {102035},

keywords = {convolutional neural networks, loss function, segmentation},

pubstate = {published},

tppubtype = {article}

}

@article{Petrick2021,

title = {SPIE-AAPM-NCI BreastPathQ challenge: an image analysis challenge for quantitative tumor cellularity assessment in breast cancer histology images following neoadjuvant treatment},

author = {Nicholas Petrick and Shazia Akbar and Kenny H. Cha and Sharon Nofech-Mozes and Berkman Sahiner and Marios A. Gavrielides and Jayashree Kalpathy-Cramer and Karen Drukker and Anne L. Martel and BreastPathQ Challenge Group},

url = {https://www.spiedigitallibrary.org/journals/journal-of-medical-imaging/volume-8/issue-03/034501/SPIE-AAPM-NCI-BreastPathQ-challenge–an-image-analysis-challenge/10.1117/1.JMI.8.3.034501.full},

doi = {10.1117/1.JMI.8.3.034501},

issn = {2329-4302},

year  = {2021},

date = {2021-05-01},

journal = {Journal of Medical Imaging},

volume = {8},

number = {03},

pages = {034501},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Hesse2020,

title = {Intensity Augmentation to Improve Generalizability of Breast Segmentation Across Different MRI Scan Protocols},

author = {Linde S Hesse and Grey Kuling and Mitko Veta and Anne L. Martel},

url = {https://ieeexplore.ieee.org/document/9166708/},

doi = {10.1109/TBME.2020.3016602},

issn = {0018-9294},

year  = {2021},

date = {2021-03-01},

urldate = {2021-03-01},

journal = {IEEE Transactions on Biomedical Engineering},

volume = {68},

number = {3},

pages = {759–770},

keywords = {_breast_segmentation},

pubstate = {published},

tppubtype = {article}

}

@article{Ciga2021b,

title = {Learning to segment images with classification labels},

author = {Ozan Ciga and Anne L. Martel},

url = {https://linkinghub.elsevier.com/retrieve/pii/S1361841520302760},

doi = {10.1016/j.media.2020.101912},

issn = {13618415},

year  = {2021},

date = {2021-02-01},

journal = {Medical Image Analysis},

volume = {68},

pages = {101912},

abstract = {Two of the most common tasks in medical imaging are classification and segmentation. Either task requires labeled data annotated by experts, which is scarce and expensive to collect. Annotating data for segmentation is generally considered to be more laborious as the annotator has to draw around the boundaries of regions of interest, as opposed to assigning image patches a class label. Furthermore, in tasks such as breast cancer histopathology, any realistic clinical application often includes working with whole slide images, whereas most publicly available training data are in the form of image patches, which are given a class label. We propose an architecture that can alleviate the requirements for segmentation-level ground truth by making use of image-level labels to reduce the amount of time spent on data curation. In addition, this architecture can help unlock the potential of previously acquired image-level datasets on segmentation tasks by annotating a small number of regions of interest. In our experiments, we show using only one segmentation-level annotation per class, we can achieve performance comparable to a fully annotated dataset.},

keywords = {digital histopathology, image segmentation, weakly supervised learning, whole slide image},

pubstate = {published},

tppubtype = {article}

}

@article{Maier-Hein2020,

title = {BIAS: Transparent reporting of biomedical image analysis challenges},

author = {Lena Maier-Hein and Annika Reinke and Michal Kozubek and Anne L. Martel and Tal Arbel and Matthias Eisenmann and Allan Hanbury and Pierre Jannin and Henning Müller and Sinan Onogur and Julio Saez-Rodriguez and Bram Ginneken and Annette Kopp-Schneider and Bennett A Landman},

url = {https://doi.org/10.1016/j.media.2020.101796 https://linkinghub.elsevier.com/retrieve/pii/S1361841520301602},

doi = {10.1016/j.media.2020.101796},

issn = {13618423},

year  = {2020},

date = {2020-12-01},

urldate = {2020-12-01},

journal = {Medical Image Analysis},

volume = {66},

pages = {101796},

abstract = {The number of biomedical image analysis challenges organized per year is steadily increasing. These international competitions have the purpose of benchmarking algorithms on common data sets, typically to identify the best method for a given problem. Recent research, however, revealed that common practice related to challenge reporting does not allow for adequate interpretation and reproducibility of results. To address the discrepancy between the impact of challenges and the quality (control), the Biomedical Image Analysis ChallengeS (BIAS) initiative developed a set of recommendations for the reporting of challenges. The BIAS statement aims to improve the transparency of the reporting of a biomedical image analysis challenge regardless of field of application, image modality or task category assessed. This article describes how the BIAS statement was developed and presents a checklist which authors of biomedical image analysis challenges are encouraged to include in their submission when giving a paper on a challenge into review. The purpose of the checklist is to standardize and facilitate the review process and raise interpretability and reproducibility of challenge results by making relevant information explicit.},

keywords = {Biomedical challenges},

pubstate = {published},

tppubtype = {article}

}

@article{Mouraviev2020,

title = {Use of Radiomics for the Prediction of Local Control of Brain Metastases After Stereotactic Radiosurgery},

author = {Mouraviev, Andrei and Detsky, Jay and Sahgal, Arjun and Ruschin, Mark and Lee, Young K and Karam, Irene and Heyn, Chris and Stanisz, Greg J and Martel, Anne L},

year  = {2020},

date = {2020-01-20},

journal = {Neuro-Oncology},

abstract = {Background

Local response prediction for brain metastases (BM) after stereotactic radiosurgery (SRS) is challenging, particularly for smaller BM, as existing criteria are based solely on unidimensional measurements. This investigation sought to determine whether radiomic features provide additional value to routinely available clinical and dosimetric variables to predict local recurrence following SRS.



Methods

408 BM in 87 patients treated with SRS were analyzed. A total of 440 radiomic features were extracted from the tumor core, and the peritumoral regions, using the baseline pre-treatment volumetric post-contrast T1 (T1c) and volumetric T2 fluid-attenuated inversion recovery (FLAIR) MRI sequences. Local tumor progression was determined based on RANO-BM criteria, with a maximum axial diameter growth of >20% on the follow-up T1c indicating local failure. The top radiomic features were determined based on resampled Random Forest (RF) feature importance. An RF classifier was trained using each set of features and evaluated using the area under the receiver operating characteristic curve (AUC).



Results

The addition of any one of the top ten radiomic features to the set of clinical features resulted in a statistically significant (p<0.001) increase in the AUC. An optimized combination of radiomic and clinical features resulted in a 19% higher resampled AUC (mean = 0.793, 95% C.I. = 0.792-0.795) than clinical features alone (0.669, 0.668-0.671).



Conclusions

The increase in AUC of the RF classifier, after incorporating radiomic features, suggests that quantitative characterization of tumor appearance on pretreatment T1c and FLAIR adds value to known clinical and dosimetric variables for predicting local failure.},

keywords = {Brain},

pubstate = {published},

tppubtype = {article}

}

@article{Ciga2019b,

title = {Learning to segment images with classification labels},

author = {Ozan Ciga, Anne L Martel},

year  = {2019},

date = {2019-12-28},

journal = {arXiv preprint arXiv:1912.12533},

abstract = {Two of the most common tasks in medical imaging are classification and segmentation.

Either task requires labeled data annotated by experts, which is scarce and expensive to

collect. Annotating data for segmentation is generally considered to be more laborious

as the annotator has to draw around the boundaries of regions of interest, as opposed

to assigning image patches a class label. Furthermore, in tasks such as breast cancer

histopathology, any realistic clinical application often includes working with whole

slide images, whereas most publicly available training data are in the form of image

patches, which are given a class label. We propose an architecture that can alleviate the

requirements for segmentation-level ground truth by making use of image-level labels

to reduce the amount of time spent on data curation. In addition, this architecture can

help unlock the potential of previously acquired image-level datasets on segmentation

tasks by annotating a small number of regions of interest. In our experiments, we show

using only one segmentation-level annotation per class, we can achieve performance

comparable to a fully annotated dataset.},

keywords = {digital pathology},

pubstate = {published},

tppubtype = {article}

}

@article{Srinidhi2019,

title = {Deep neural network models for computational histopathology: A survey},

author = {Chetan L Srinidhi, Ozan Ciga, Anne L Martel},

year  = {2019},

date = {2019-12-28},

journal = {arXiv preprint arXiv:1912.12378},

abstract = {Histopathological images contain rich phenotypic information that can be used to monitor underlying mechanisms contributing to diseases progression and patient survival outcomes. Recently, deep learning has become the mainstream methodological choice for analyzing and interpreting cancer histology images. In this paper, we present a comprehensive review of state-of-the-art deep learning approaches that have been used in the context of histopathological image analysis. From the survey of over 130 papers, we review the fields progress based on the methodological aspect of different machine learning strategies such as supervised, weakly supervised, unsupervised, transfer learning and various other sub-variants of these methods. We also provide an overview of deep learning based survival models that are applicable for disease-specific prognosis tasks. Finally, we summarize several existing open datasets and highlight critical challenges and limitations with current deep learning approaches, along with possible avenues for future research.},

keywords = {digital pathology},

pubstate = {published},

tppubtype = {article}

}

@article{Balki2019b,

title = {Sample-Size Determination Methodologies for Machine Learning in Medical Imaging Research: A Systematic Review},

author = {Indranil Balki and Afsaneh Amirabadi and Jacob Levman and Anne L Martel and Ziga Emersic and Blaz Meden and Angel Garcia-Pedrero and Saul C Ramirez and Dehan Kong and Alan R Moody and Pascal N Tyrrell},

url = {https://doi.org/10.1016/j.carj.2019.06.002 http://journals.sagepub.com/doi/10.1016/j.carj.2019.06.002},

doi = {10.1016/j.carj.2019.06.002},

issn = {0846-5371},

year  = {2019},

date = {2019-11-01},

journal = {Canadian Association of Radiologists Journal},

volume = {70},

number = {4},

pages = {344–353},

publisher = {Elsevier Inc.},

abstract = {Purpose: The required training sample size for a particular machine learning (ML) model applied to medical imaging data is often unknown. The purpose of this study was to provide a descriptive review of current sample-size determination methodologies in ML applied to medical imaging and to propose recommendations for future work in the field. Methods: We conducted a systematic literature search of articles using Medline and Embase with keywords including “machine learning,” “image,” and “sample size.” The search included articles published between 1946 and 2018. Data regarding the ML task, sample size, and train-test pipeline were collected. Results: A total of 167 articles were identified, of which 22 were included for qualitative analysis. There were only 4 studies that discussed sample-size determination methodologies, and 18 that tested the effect of sample size on model performance as part of an exploratory analysis. The observed methods could be categorized as pre hoc model-based approaches, which relied on features of the algorithm, or post hoc curve-fitting approaches requiring empirical testing to model and extrapolate algorithm performance as a function of sample size. Between studies, we observed great variability in performance testing procedures used for curve-fitting, model assessment methods, and reporting of confidence in sample sizes. Conclusions: Our study highlights the scarcity of research in training set size determination methodologies applied to ML in medical imaging, emphasizes the need to standardize current reporting practices, and guides future work in development and streamlining of pre hoc and post hoc sample size approaches.},

keywords = {Machine learning, Medical Imaging, Radiology, Sample size},

pubstate = {published},

tppubtype = {article}

}

@article{Goubran2019,

title = {Hippocampal segmentation for brains with extensive atrophy using three‐dimensional convolutional neural networks},

author = {Maged Goubran, Emmanuel Edward Ntiri, Hassan Akhavein, Melissa Holmes, Sean Nestor, Joel Ramirez, Sabrina Adamo, Miracle Ozzoude, Christopher Scott, Fuqiang Gao and Anne L Martel and Walter Swardfager, Mario Masellis, Richard Swartz, Bradley MacIntosh, Sandra E Black},

year  = {2019},

date = {2019-10-14},

urldate = {2019-10-14},

journal = {Human brain mapping},

abstract = {Hippocampal volumetry is a critical biomarker of aging and dementia, and it is widely used as a predictor of cognitive performance; however, automated hippocampal segmentation methods are limited because the algorithms are (a) not publicly available, (b) subject to error with significant brain atrophy, cerebrovascular disease and lesions, and/or (c) computationally expensive or require parameter tuning. In this study, we trained a 3D convolutional neural network using 259 bilateral manually delineated segmentations collected from three studies, acquired at multiple sites on different scanners with variable protocols. Our training dataset consisted of elderly cases difficult to segment due to extensive atrophy, vascular disease, and lesions. Our algorithm, (HippMapp3r), was validated against four other publicly available state‐of‐the‐art techniques (HippoDeep, FreeSurfer, SBHV, volBrain, and FIRST). HippMapp3r outperformed the other techniques on all three metrics, generating an average Dice of 0.89 and a correlation coefficient of 0.95. It was two orders of magnitude faster than some of the tested techniques. Further validation was performed on 200 subjects from two other disease populations (frontotemporal dementia and vascular cognitive impairment), highlighting our method's low outlier rate. We finally tested the methods on real and simulated “clinical adversarial” cases to study their robustness to corrupt, low‐quality scans. The pipeline and models are available at: https://hippmapp3r.readthedocs.ioto facilitate the study of the hippocampus in large multisite studies.},

keywords = {Brain},

pubstate = {published},

tppubtype = {article}

}

@article{Maier-Hein2019,

title = {BIAS: Transparent reporting of biomedical image analysis challenges},

author = {Lena Maier-Hein, Annika Reinke, Michal Kozubek and  Anne L. Martel and  Tal Arbel, Matthias Eisenmann, Allan Hanbuary, Pierre Jannin, Henning Müller, Sinan Onogur, Julio Saez-Rodriguez, Bram van Ginneken, Annette Kopp-Schneider, Bennett Landman},

year  = {2019},

date = {2019-10-09},

urldate = {2019-10-09},

journal = {arXiv preprint arXiv:1910.04071},

abstract = {The number of biomedical image analysis challenges organized per year is steadily increasing. These international competitions have the purpose of benchmarking algorithms on common data sets, typically to identify the best method for a given problem. Recent research, however, revealed that common practice related to challenge reporting does not allow for adequate interpretation and reproducibility of results. To address the discrepancy between the impact of challenges and the quality (control), the Biomedical I mage Analysis ChallengeS (BIAS) initiative developed a set of recommendations for the reporting of challenges. The BIAS statement aims to improve the transparency of the reporting of a biomedical image analysis challenge regardless of field of application, image modality or task category assessed. This article describes how the BIAS statement was developed and presents a checklist which authors of biomedical image analysis challenges are encouraged to include in their submission when giving a paper on a challenge into review. The purpose of the checklist is to standardize and facilitate the review process and raise interpretability and reproducibility of challenge results by making relevant information explicit.},

keywords = {Medical Imaging},

pubstate = {published},

tppubtype = {article}

}

@article{Akbar2019b,

title = {Automated and Manual Quantification of Tumour Cellularity in Digital Slides for Tumour Burden Assessment},

author = {Shazia Akbar and  Mohammad Peikari and  Sherine Salama and  Azadeh Yazdan Panah and Sharon Nofech-Mozes and  Anne L Martel},

year  = {2019},

date = {2019-10-01},

urldate = {2019-10-01},

journal = {Scientific Reports},

volume = {9},

pages = {14099},

abstract = {Aims:

The residual cancer burden index is an important quantitative measure used for assessing treatment response following neoadjuvant therapy for breast cancer. It has shown to be predictive of overall survival and is composed of two key metrics: qualitative assessment of lymph nodes and the percentage of invasive or in-situ tumour cellularity (TC) in the tumour bed (TB). Currently, TC is assessed through eye-balling of routine histopathology slides estimating the proportion of tumour cells within the TB. With the advances in production of digitized slides and increasing availability of slide scanners in pathology laboratories, there is potential to measure TC using automated algorithms with greater precision and accuracy.

Methods:

We describe two methods for automated TC scoring: 1) a traditional approach to image analysis development whereby we mimic the pathologists9 workflow, and 2) a recent development in artificial intelligence in which features are learned automatically in deep neural networks using image data alone.

Results:

We show strong agreements between automated and manual analysis of digital slides. Agreements between our trained deep neural networks and experts in this study (0.82) approach the inter-rater agreements between pathologists (0.89). We also reveal properties that are captured when we apply deep neural network to whole slide images, and discuss the potential of using such visualisations to improve upon TC assessment in the future.

Conclusions:

TC scoring can be successfully automated by leveraging recent advancements in artificial intelligence, thereby alleviating the burden of manual analysis.},

keywords = {digital pathology},

pubstate = {published},

tppubtype = {article}

}

@article{Balki2019,

title = {Sample-Size Determination Methodologies for Machine Learning in Medical Imaging Research: A Systematic Review},

author = {Indranil Balki and  Afsaneh Amirabadi and  Jacob Levman and Anne L Martel and  Ziga Emersic and Blaz Meden and  Angel Garcia-Pedrero and Saul C Ramirez and  Dehan Kong and  Alan R Moody and  Pascal N Tyrrell},

year  = {2019},

date = {2019-09-12},

urldate = {2019-09-12},

journal = {Canadian Association of Radiologists Journal},

volume = {70},

number = {4},

pages = {344-353},

abstract = {Purpose

The required training sample size for a particular machine learning (ML) model applied to medical imaging data is often unknown. The purpose of this study was to provide a descriptive review of current sample-size determination methodologies in ML applied to medical imaging and to propose recommendations for future work in the field.



Methods

We conducted a systematic literature search of articles using Medline and Embase with keywords including “machine learning,” “image,” and “sample size.” The search included articles published between 1946 and 2018. Data regarding the ML task, sample size, and train-test pipeline were collected.



Results

A total of 167 articles were identified, of which 22 were included for qualitative analysis. There were only 4 studies that discussed sample-size determination methodologies, and 18 that tested the effect of sample size on model performance as part of an exploratory analysis. The observed methods could be categorized as pre hoc model-based approaches, which relied on features of the algorithm, or post hoc curve-fitting approaches requiring empirical testing to model and extrapolate algorithm performance as a function of sample size. Between studies, we observed great variability in performance testing procedures used for curve-fitting, model assessment methods, and reporting of confidence in sample sizes.



Conclusions

Our study highlights the scarcity of research in training set size determination methodologies applied to ML in medical imaging, emphasizes the need to standardize current reporting practices, and guides future work in development and streamlining of pre hoc and post hoc sample size approaches.},

keywords = {Medical Imaging},

pubstate = {published},

tppubtype = {article}

}

@article{Hesse2019,

title = {Intensity augmentation for domain transfer of whole breast segmentation in MRI},

author = {Linde S Hesse and  Grey Kuling and  Mitko Veta and  Anne L Martel},

year  = {2019},

date = {2019-09-05},

urldate = {2019-09-05},

journal = {arXiv preprint arXiv:1909.02642},

abstract = {The segmentation of the breast from the chest wall is an important first step in the analysis of breast magnetic resonance images. 3D U-nets have been shown to obtain high segmentation accuracy and appear to generalize well when trained on one scanner type and tested on another scanner, provided that a very similar T1-weighted MR protocol is used. There has, however, been little work addressing the problem of domain adaptation when image intensities or patient orientation differ markedly between the training set and an unseen test set. To overcome the domain shift we propose to apply extensive intensity augmentation in addition to geometric augmentation during training. We explored both style transfer and a novel intensity remapping approach as intensity augmentation strategies. For our experiments, we trained a 3D U-net on T1-weighted scans and tested on T2-weighted scans. By applying intensity augmentation we increased segmentation performance from a DSC of 0.71 to 0.90. This performance is very close to the baseline performance of training and testing on T2-weighted scans (0.92). Furthermore, we applied our network to an independent test set made up of publicly available scans acquired using a T1-weighted TWIST sequence and a different coil configuration. On this dataset we obtained a performance of 0.89, close to the inter-observer variability of the ground truth segmentations (0.92). Our results show that using intensity augmentation in addition to geometric augmentation is a suitable method to overcome the intensity domain shift and we expect it to be useful for a wide range of segmentation tasks.},

keywords = {Breast MRI},

pubstate = {published},

tppubtype = {article}

}

@article{Mouraviev2019,

title = {Use of Radiomics in the Prediction of Brain Metastases Local Control Post-Stereotactic Radiosurgery},

author = {A Mouraviev and J Detsky and ME Ruschin and A Sahgal and Y Lee and C Heyn and I Karam and Anne L Martel},

year  = {2019},

date = {2019-09-01},

urldate = {2019-09-01},

journal = {International Journal of Radiation Oncology• Biology• Physics},

volume = {105},

number = {1},

pages = {E82-E83},

abstract = {Results

Top 10 highest ranked radiomic features were combined, one at a time, with the clinical features, resulting in a very significant (p< 0.001) increase in AUC from the addition of the top 9 radiomic features, and no increase from adding the 10 th best radiomic feature. An optimized combination of radiomic and clinical features resulted in a dramatically higher performance (resampled AUC: mean= 0.792, 95% CI of mean= 0.790-0.793) compared to clinical features alone (0.676, 0.674-0.678).

Conclusion

The increase in performance from incorporating radiomic features suggests that quantitative characterization of tumor structure through the use of radiomics adds complementary information to what is typically available in the clinical workflow.},

keywords = {Brain},

pubstate = {published},

tppubtype = {article}

}

@article{Ciga2019c,

title = {Multi-layer Domain Adaptation for Deep Convolutional Networks},

author = {Ozan Ciga and Jianan Chen and Anne Martel},

url = {http://arxiv.org/abs/1909.02620},

doi = {10.1007/978-3-030-33391-1_3},

year  = {2019},

date = {2019-09-01},

journal = {DART workshop, MICCAI2019},

address = {Shenzhen, China},

abstract = {Despite their success in many computer vision tasks, convolutional networks tend to require large amounts of labeled data to achieve generalization. Furthermore, the performance is not guaranteed on a sample from an unseen domain at test time, if the network was not exposed to similar samples from that domain at training time. This hinders the adoption of these techniques in clinical setting where the imaging data is scarce, and where the intra- and inter-domain variance of the data can be substantial. We propose a domain adaptation technique that is especially suitable for deep networks to alleviate this requirement of labeled data. Our method utilizes gradient reversal layers and Squeezeand-Excite modules to stabilize the training in deep networks. The proposed method was applied to publicly available histopathology and chest X-ray databases and achieved superior performance to existing state-of-the-art networks with and without domain adaptation. Depending on the application, our method can improve multi-class classification accuracy by 5-20% compared to DANN introduced in (Ganin, 2014).},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{Akbar2019,

title = {The transition module: a method for preventing overfitting in convolutional neural networks},

author = {S Akbar, M Peikari, S Salama, S Nofech-Mozes, AL Martel},

year  = {2019},

date = {2019-05-04},

journal = {Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization},

volume = {7},

number = {3},

pages = {260-265},

abstract = {Abstract

Digital pathology has advanced substantially over the last decade with the adoption of slide scanners in pathology labs. The use of digital slides to analyse diseases at the microscopic level is both cost-effective and efficient. Identifying complex tumour patterns in digital slides is a challenging problem but holds significant importance for tumour burden assessment, grading and many other pathological assessments in cancer research. The use of convolutional neural networks (CNNs) to analyse such complex images has been well adopted in digital pathology. However, in recent years, the architecture of CNNs has altered with the introduction of inception modules which have shown great promise for classification tasks. In this paper, we propose a modified ‘transition’ module which encourages generalisation in a deep learning framework with few training samples. In the transition module, filters of varying sizes are used to encourage class-specific filters at multiple spatial resolutions followed by global average pooling. We demonstrate the performance of the transition module in AlexNet and ZFNet, for classifying breast tumours in two independent data-sets of scanned histology sections; the inclusion of the transition module in these CNNs improved performance.},

keywords = {digital pathology},

pubstate = {published},

tppubtype = {article}

}

@article{Fashandi2019,

title = {An investigation of the effect of fat suppression and dimensionality on the accuracy of breast MRI segmentation using U-nets},

author = {Homa Fashandi and Grey Kuling and YingLi Lu and Hongbo Wu and Anne L. Martel},

url = {http://hdl.handle.net/1807/93313},

doi = {10.1002/mp.13375},

year  = {2019},

date = {2019-01-04},

urldate = {2019-01-04},

journal = {Medical Physics},

abstract = {Purpose

Accurate segmentation of the breast is required for breast density estimation and the assessment of background parenchymal enhancement, both of which have been shown to be related to breast cancer risk. The MRI breast segmentation task is challenging, and recent work has demonstrated that convolutional neural networks perform well for this task. In this study, we have investigated the performance of several 2D U‐Net and 3D U‐Net configurations using both fat‐suppressed and nonfat suppressed images. We have also assessed the effect of changing the number and quality of the ground truth segmentations.



Materials and methods

We designed 8 studies to investigate the effect of input types and the dimensionality of the U‐Net operations for the breast MRI segmentation. Our training data contained 70 whole breast volumes of T1‐weighted sequences without fat suppression(WOFS) and with fat suppression(FS). For each subject, we registered the WOFS and FS volumes together before manually segmenting the breast to generate ground truth. We compared 4 different input types to the U‐nets: WOFS, FS, MIXED(WOFS and FS images treated as separate samples) and MULTI(WOFS and FS images combined into a single multi‐channel image). We trained 2D U‐Nets and 3D U‐Nets with this data, which resulted in our 8 studies (2D‐WOFS, 3D‐WOFS,2D‐FS,3D‐FS,2D‐MIXED,3D‐MIXED,2D‐MULTI, and 3D‐MULT). For each of these studies, we performed a systematic grid search to tune the hyperparameters of the U‐Nets. A separate validation set with 15 whole breast volumes was used for hyperparameter tuning. We performed Kruskal‐Walis test on the results of our hyperparameter tuning and did not find a statistically significant difference in the 10 top models of each study. For this reason, we chose the best model as the model with the highest mean Dice Similarity Coefficient(DSC) value on the validation set. The reported test results are the results of the top model of each study on our test set which contained 19 whole breast volumes annotated by 3 readers fused with the STAPLE algorithm. We also investigated the effect of the quality of the training annotations and the number of training samples for this task.



Results

The study with the highest average DSC result was 3D‐MULTI with 0.96 ± 0.02. The second highest average is 2D WOFS (0.96 ± 0.03), and the third is 2D MULTI (0.96 ± 0.03). We performed the Kruskal‐Wallis 1‐way ANOVA test with Dunn's multiple comparison tests using Bonferroni p‐value correction on the results of the selected model of each study and found that 3D‐MULTI, 2D‐MULTI, 3D‐WOFS, 2D‐WOFS, 2D‐FS, and 3D‐FS were not statistically different in their distributions, which indicates that comparable results could be obtained in fat‐suppressed and nonfat suppressed volumes and that there is no significant difference between the 3D and 2D approach. Our results also suggested that the networks trained on single sequence images or multiple sequence images organized in multi‐channel images perform better than the models trained on a mixture of volumes from different sequences. Our investigation of the size of the training set revealed that training a U‐Net in this domain only requires a modest amount of training data and results obtained with 49 and 70 training datasets were not significantly different.



Conclusions

To summarize, we investigated the use of 2D U‐Nets and 3D U‐Nets for breast volume segmentation in T1 fat suppressed and without fat suppressed volumes. Although our highest score was obtained in the 3D MULTI study, when we took advantage of information in both fat suppressed and non fat suppressed volumes and their 3D structure, all of the methods we explored gave accurate segmentations with an average DSC on > 94% demonstrating that the U‐Net is a robust segmentation method for breast MRI volumes.},

note = {This is the pre-peer reviewed version. The definitive version is available at: https://aapm.onlinelibrary.wiley.com/doi/abs/10.1002/mp.13375},

keywords = {_breast_segmentation, Breast MRI, deep learning, segmentation},

pubstate = {published},

tppubtype = {article}

}