Radiologist-Level Performance by Using Deep Learning for Segmentation                     of Breast Cancers on MRI Scans

Hirsch, Lukas; Huang, Yu; Luo, Shaojun; Saccarelli, Carolina Rossi; Gullo, Roberto Lo; Naranjo, Isaac Daimiel; Bitencourt, Almir; Onishi, Natsuko; Ko, Eun Sook; Leithner, Doris; Avendaño, Daly; Eskreis‐Winkler, Sarah; Hughes, Mary; Martinez, Danny F.; Pinker, Katja; Juluru, Krishna; El-Rowmeim, Amin; Elnajjar, Pierre; Morris, Elizabeth A.; Makse, Hernán A.; Parra, Lucas C.; Sutton, Elizabeth

doi:10.1148/ryai.200231

Cited by 19 publications

(22 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the visual ensemble selection proves to provide acceptable segmentation results in 77% of the test cases and results are globally within inter-radiologist reproducibility. Our approach provides a 3D segmentation of lesions, while some of the most recent studies still segment in 2D [13,16,21], despite tumor volume measured by MR imaging being a strong predictor of patient survival [27]. For advanced radiomic studies or follow-up studies, 3D segmentation is also an important task to achieve [3].…”

Section: Discussionmentioning

confidence: 99%

“…The CNN models were trained using multicentric MRI, a prerequisite for a higher generalization of these models, and they were also evaluated using a multi-scanner test dataset. Compared to many studies, for which DSC was the only evaluation criterion [15,16], HD95 was added as a criterion for the maximal distance between two segmentations. Contrary to DSC, this criterion was not included in the loss function for the training of the models and was therefore more independent.…”

Section: Discussionmentioning

confidence: 99%

“…There are fewer studies using deep learning methods to segment breast tumors using DCE-MRI than using mammograms, partly due to the availability of very large mammography datasets [12]. Studies based on DCE-MRI used wellestablished CNN segmentation models [13][14][15][16] based on U-Net [5], DeepMedic [17], or SegNet [18] architectures or less common models [19,20]. Several studies [14,16,19,21] took advantage of all the information given by the DCE-MRI by using the different post-contrast or subtraction (post-contrast minus the pre-contrast acquisition) images.…”

Section: Introductionmentioning

confidence: 99%

“…Studies based on DCE-MRI used wellestablished CNN segmentation models [13][14][15][16] based on U-Net [5], DeepMedic [17], or SegNet [18] architectures or less common models [19,20]. Several studies [14,16,19,21] took advantage of all the information given by the DCE-MRI by using the different post-contrast or subtraction (post-contrast minus the pre-contrast acquisition) images. For instance, Piantadosi et al [21] used images from three different time points (pre-contrast, first and last post-contrast images).…”

Section: Introductionmentioning

confidence: 99%

“…For instance, Piantadosi et al [21] used images from three different time points (pre-contrast, first and last post-contrast images). In the same way, Hirsch et al [16] built several models taking different post-contrast images as input while Zhang et al [19] fed both post-contrast and subtraction images as input to a hierarchical CNN.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Visual ensemble selection of deep convolutional neural networks for 3D segmentation of breast tumors on dynamic contrast enhanced MRI

et al. 2022

View full text Add to dashboard Cite

Objectives To develop a visual ensemble selection of deep convolutional neural networks (CNN) for 3D segmentation of breast tumors using T1-weighted dynamic contrast-enhanced (T1-DCE) MRI. Methods Multi-center 3D T1-DCE MRI (n = 141) were acquired for a cohort of patients diagnosed with locally advanced or aggressive breast cancer. Tumor lesions of 111 scans were equally divided between two radiologists and segmented for training. The additional 30 scans were segmented independently by both radiologists for testing. Three 3D U-Net models were trained using either post-contrast images or a combination of post-contrast and subtraction images fused at either the image or the feature level. Segmentation accuracy was evaluated quantitatively using the Dice similarity coefficient (DSC) and the Hausdorff distance (HD95) and scored qualitatively by a radiologist as excellent, useful, helpful, or unacceptable. Based on this score, a visual ensemble approach selecting the best segmentation among these three models was proposed. Results The mean and standard deviation of DSC and HD95 between the two radiologists were equal to 77.8 ± 10.0% and 5.2 ± 5.9 mm. Using the visual ensemble selection, a DSC and HD95 equal to 78.1 ± 16.2% and 14.1 ± 40.8 mm was reached. The qualitative assessment was excellent (resp. excellent or useful) in 50% (resp. 77%). Conclusion Using subtraction images in addition to post-contrast images provided complementary information for 3D segmentation of breast lesions by CNN. A visual ensemble selection allowing the radiologist to select the most optimal segmentation obtained by the three 3D U-Net models achieved comparable results to inter-radiologist agreement, yielding 77% segmented volumes considered excellent or useful. Key Points • Deep convolutional neural networks were developed using T1-weighted post-contrast and subtraction MRI to perform automated 3D segmentation of breast tumors. • A visual ensemble selection allowing the radiologist to choose the best segmentation among the three 3D U-Net models outperformed each of the three models. • The visual ensemble selection provided clinically useful segmentations in 77% of cases, potentially allowing for a valuable reduction of the manual 3D segmentation workload for the radiologist and greatly facilitating quantitative studies on non-invasive biomarker in breast MRI.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Visual ensemble selection of deep convolutional neural networks for 3D segmentation of breast tumors on dynamic contrast enhanced MRI

et al. 2022

View full text Add to dashboard Cite

show abstract

Editorial for “Deep Learning‐Based Segmentation of Locally Advanced Breast Cancer on MRI in Relation to Residual Cancer Burden: A Multi‐Institutional Cohort Study”

Wyatt

Huang

2023

Magnetic Resonance Imaging

View full text Add to dashboard Cite

3D Breast Cancer Segmentation in DCE‐MRI Using Deep Learning With Weak Annotation

Park,

Kim,

Nam

et al. 2023

Magnetic Resonance Imaging

View full text Add to dashboard Cite

BackgroundDeep learning models require large‐scale training to perform confidently, but obtaining annotated datasets in medical imaging is challenging. Weak annotation has emerged as a way to save time and effort.PurposeTo develop a deep learning model for 3D breast cancer segmentation in dynamic contrast‐enhanced magnetic resonance imaging (DCE‐MRI) using weak annotation with reliable performance.Study TypeRetrospective.PopulationSeven hundred and thirty‐six women with breast cancer from a single institution, divided into the development (N = 544) and test dataset (N = 192).Field Strength/Sequence3.0‐T, 3D fat‐saturated gradient‐echo axial T1‐weighted flash 3D volumetric interpolated brain examination (VIBE) sequences.AssessmentTwo radiologists performed a weak annotation of the ground truth using bounding boxes. Based on this, the ground truth annotation was completed through autonomic and manual correction. The deep learning model using 3D U‐Net transformer (UNETR) was trained with this annotated dataset. The segmentation results of the test set were analyzed by quantitative and qualitative methods, and the regions were divided into whole breast and region of interest (ROI) within the bounding box.Statistical TestsAs a quantitative method, we used the Dice similarity coefficient to evaluate the segmentation result. The volume correlation with the ground truth was evaluated with the Spearman correlation coefficient. Qualitatively, three readers independently evaluated the visual score in four scales. A P‐value <0.05 was considered statistically significant.ResultsThe deep learning model we developed achieved a median Dice similarity score of 0.75 and 0.89 for the whole breast and ROI, respectively. The volume correlation coefficient with respect to the ground truth volume was 0.82 and 0.86 for the whole breast and ROI, respectively. The mean visual score, as evaluated by three readers, was 3.4.Data ConclusionThe proposed deep learning model with weak annotation may show good performance for 3D segmentations of breast cancer using DCE‐MRI.Level of Evidence3Technical EfficacyStage 2

show abstract

Radiologist-Level Performance by Using Deep Learning for Segmentation of Breast Cancers on MRI Scans

Abstract: This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content.

Cited by 19 publications

References 34 publications

Visual ensemble selection of deep convolutional neural networks for 3D segmentation of breast tumors on dynamic contrast enhanced MRI

Visual ensemble selection of deep convolutional neural networks for 3D segmentation of breast tumors on dynamic contrast enhanced MRI

Editorial for “Deep Learning‐Based Segmentation of Locally Advanced Breast Cancer on MRI in Relation to Residual Cancer Burden: A Multi‐Institutional Cohort Study”

3D Breast Cancer Segmentation in DCE‐MRI Using Deep Learning With Weak Annotation

Contact Info

Product

Resources

About