Automated Segmentation of Tissues Using CT and MRI: A Systematic Review

Lenchik, Leon; Heacock, Laura; Weaver, Ashley A.; Boutin, Robert D.; Cook, Tessa S.; Itri, Jason N.; Filippi, Christopher G.; Gullapalli, Rao P.; Lee, James; Zagurovskaya, Marianna; Retson, Tara; Godwin, Kendra; Nicholson, Joey; Narayana, Ponnada A.

doi:10.1016/j.acra.2019.07.006

Cited by 94 publications

(79 citation statements)

References 460 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Although skeletal muscles are most commonly evaluated by CT at the L3 level, measurements at T10‐L5 have been validated 19,22 . Importantly, paravertebral muscle density and size can now be measured automatically with machine learning algorithms in large cohorts, 23 such as in the National Lung Screening Trial 24 …”

Section: Discussionmentioning

confidence: 99%

Opportunistic muscle measurements on staging chest CT for extremity and truncal soft tissue sarcoma are associated with survival

Phan

Thorpe

Wong

et al. 2020

Journal of Surgical Oncology

Self Cite

View full text Add to dashboard Cite

Background and Objectives Computed tomography (CT) measurements of sarcopenia have been proposed as biomarkers associated with outcomes in various cancers and have typically been evaluated at the L3 vertebral level. However, staging imaging for patients with extremity and truncal soft tissue sarcoma (STS) often only includes chest CT imaging which precludes evaluation at L3. Therefore, we sought to evaluate muscle metrics at T12 on standard staging chest CT scans and evaluate for correlation with overall and event‐free survival in patients with STS. Methods CT chest imaging for 89 patients with intermediate and high‐grade STS (53 male, 36 female; 58.5 ± 19.0 years old, follow‐up 37.4 ± 27.1 months) was reviewed on PACS at T12 for skeletal muscle density (SMD) and skeletal muscle index (SMI). Results Overall survival increased with increased SMD on univariate (hazard ratio [HR] = 0.61 [0.43, 0.86]) and age‐adjusted analysis (HR = 0.65 [0.42, 0.89]. Event‐free survival also increased with increased SMD in univariate analyses (HR = 0.68 [0.49, 0.95]) but did not maintain significance after adjusting for age (HR = 0.68 [0.43, 1.07]). SMI was not a predictor of overall or event‐free survival. Conclusions Higher SMD measured on routinely obtained staging chest CTs in STS patients is associated with improved survival.

show abstract

Section: Discussionmentioning

confidence: 99%

Opportunistic muscle measurements on staging chest CT for extremity and truncal soft tissue sarcoma are associated with survival

Phan

Thorpe

Wong

et al. 2020

Journal of Surgical Oncology

Self Cite

View full text Add to dashboard Cite

show abstract

“…[4][5][6] Computer-aided automated multiorgan segmentation would be a compelling approach to the roadblock. 7 Early research on automated segmentation algorithms focused on mathematical modeling of the morphological information of organs. For instance, level-set, 8 SNAKE, 9 and graph cut 10 focus on attracting descriptors to organ boundaries, driven by intensity gradient and neighborhood structures.…”

Section: Introductionmentioning

confidence: 99%

“…Deep networks incorporate representation as part of the learning, in contrast with hand crafted features in conventional regression methods. 7 In medical segmentation problems, it is common to adopt a supervised learning setting, where images and the corresponding clinical manual labels are used during training, and the resultant network is used to infer the labels automatically on new images. Their superior ability to model the complexity in multiorgan shapes, context information, and the intersubject varieties has been demonstrated on several benchmark datasets.…”

Section: Introductionmentioning

confidence: 99%

Fully automated multiorgan segmentation in abdominal magnetic resonance imaging with deep neural networks

et al. 2020

View full text Add to dashboard Cite

Purpose: Segmentation of multiple organs-at-risk (OARs) is essential for magnetic resonance (MR)only radiation therapy treatment planning and MR-guided adaptive radiotherapy of abdominal cancers. Current practice requires manual delineation that is labor-intensive, time-consuming, and prone to intra-and interobserver variations. We developed a deep learning (DL) technique for fully automated segmentation of multiple OARs on clinical abdominal MR images with high accuracy, reliability, and efficiency. Methods: We developed Automated deep Learning-based abdominal multiorgan segmentation (ALAMO) technique based on two-dimensional U-net and a densely connected network structure with tailored design in data augmentation and training procedures such as deep connection, auxiliary supervision, and multiview. The model takes in multislice MR images and generates the output of segmentation results. 3.0-Tesla T1 VIBE (Volumetric Interpolated Breath-hold Examination) images of 102 subjects were used in our study and split into 66 for training, 16 for validation, and 20 for testing. Ten OARs were studied, including the liver, spleen, pancreas, left/right kidneys, stomach, duodenum, small intestine, spinal cord, and vertebral bodies. An experienced radiologist manually labeled each OAR, followed by reediting, if necessary, by a senior radiologist, to create the ground-truth. The performance was measured using volume overlapping and surface distance. Results: The ALAMO technique generated segmentation labels in good agreement with the manual results. Specifically, among the ten OARs, nine achieved high dice similarity coefficients (DSCs) in the range of 0.87-0.96, except for the duodenum with a DSC of 0.80. The inference completed within 1 min for a three-dimensional volume of 320 9 288 9 180. Overall, the ALAMO model matched the state-of-the-art techniques in performance. Conclusion: The proposed ALAMO technique allows for fully automated abdominal MR segmentation with high accuracy and practical memory and computation time demands.

show abstract

“…The lung is a frequently targeted organ for automated processing, but most applications focus on pathology or nodule detection on CT data, which is generally more widely used than MRI. Pulmonary MRI and the application of neural networks for automatic post‐processing remains still infrequent but rapidly growing 46 . Recently Guo et al proposed an interesting automated processing for lung segmentation, but it was limited to proton‐based ventilation MRI and still needs manually placed seeding points 47 .…”

Section: Discussionmentioning

confidence: 99%

The impact of segmentation on whole‐lung functional MRI quantification: Repeatability and reproducibility from multiple human observers and an artificial neural network

Willers

Bauman

Andermatt

et al. 2020

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose To investigate the repeatability and reproducibility of lung segmentation and their impact on the quantitative outcomes from functional pulmonary MRI. Additionally, to validate an artificial neural network (ANN) to accelerate whole‐lung quantification. Method Ten healthy children and 25 children with cystic fibrosis underwent matrix pencil decomposition MRI (MP‐MRI). Impaired relative fractional ventilation (RFV) and relative perfusion (RQ) from MP‐MRI were compared using whole‐lung segmentation performed by a physician at two time‐points (At1 and At2), by an MRI technician (B), and by an ANN (C). Repeatability and reproducibility were assess with Dice similarity coefficient (DSC), paired t‐test and Intraclass‐correlation coefficient (ICC). Results The repeatability within an observer (At1 vs At2) resulted in a DSC of 0.94 ± 0.01 (mean ± SD) and an unsystematic difference of −0.01% for RFV (P = .92) and +0.1% for RQ (P = .21). The reproducibility between human observers (At1 vs B) resulted in a DSC of 0.88 ± 0.02, and a systematic absolute difference of −0.81% (P < .001) for RFV and −0.38% (P = .037) for RQ. The reproducibility between human and the ANN (At1 vs C) resulted in a DSC of 0.89 ± 0.03 and a systematic absolute difference of −0.36% for RFV (P = .017) and −0.35% for RQ (P = .002). The ICC was >0.98 for all variables and comparisons. Conclusions Despite high overall agreement, there were systematic differences in lung segmentation between observers. This needs to be considered for longitudinal studies and could be overcome by using an ANN, which performs as good as human observers and fully automatizes MP‐MRI post‐processing.

show abstract

Automated Segmentation of Tissues Using CT and MRI: A Systematic Review

Cited by 94 publications

References 460 publications

Opportunistic muscle measurements on staging chest CT for extremity and truncal soft tissue sarcoma are associated with survival

Opportunistic muscle measurements on staging chest CT for extremity and truncal soft tissue sarcoma are associated with survival

Fully automated multiorgan segmentation in abdominal magnetic resonance imaging with deep neural networks

The impact of segmentation on whole‐lung functional MRI quantification: Repeatability and reproducibility from multiple human observers and an artificial neural network

Contact Info

Product

Resources

About