Coarse-to-Fine Adversarial Networks and Zone-Based Uncertainty Analysis for NK/T-Cell Lymphoma Segmentation in CT/PET Images

Hu, Xiaobin; Guo, Rui; Chen, Jieneng; Li, Hongwei; Waldmannstetter, Diana; Zhao, Yu; Li, Biao; Shi, Kuangyu; Menze, Bjoern

doi:10.1109/jbhi.2020.2972694

Cited by 39 publications

(26 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They divided 109 samples into 80 for training and 29 for test, and achieved the DSC of 66.64%. In addition, Hu et al 4 . proposed a coarse‐to‐fine adversarial network, achieving the DSC of 71.15% and the HD of 5.98 mm on natural killer/T‐cell lymphoma segmentation in 83 PET‐CT samples.…”

Section: Resultsmentioning

confidence: 99%

“…trained two subnetworks in 2D multiview and 3D, and then captured features in different‐axis observations for lymphoma segmentation in PET images. Another research proposed a coarse‐to‐fine adversarial network to achieve natural killer/T‐cell lymphoma segmentation in PET/CT images 4 . Until recently, the most of researches still focus on extracting feature representations from single modality without an in‐depth consideration on complementary information 13,14 …”

Section: Introductionmentioning

confidence: 99%

“…1), which makes lymphoma segmentation task much more challenging than other tumor segmentation tasks; (b) PET responds sensitively to the metabolic activity of tissues, 2 while its low resolution and poor signal‐to‐noise ratio require the combination of CT that specializes at structural imaging; 3 (c) Traditional single‐modality processing is inherently limited without utilization of complementary information especially when the purpose is to consider both functional and structural scopes. For example, for extra‐nodal natural killer/T‐cell lymphoma, which is a lymphoma subtype occurring predominantly in nasal and oropharyngeal regions, lesions and surrounding normal tissues can appear identical in CT imaging 4 . Similarly, high radioactivity uptake in PET imaging can be related to normal regions with physiological uptake like heart and kidneys 5,6 .…”

Section: Introductionmentioning

confidence: 99%

“…For example, for extra-nodal natural killer/T-cell lymphoma, which is a lymphoma subtype occurring predominantly in nasal and oropharyngeal regions, lesions and surrounding normal tissues can appear identical in CT imaging. 4 Similarly, high radioactivity uptake in PET imaging can be related to normal regions with physiological uptake like heart and kidneys. 5,6 Thus, synthesizing different spatially visual features derived from multiple imaging modes is necessary, while the feature-level fusion method for complementary information is currently insufficient.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Diffuse large B‐cell lymphoma segmentation in PET‐CT images via hybrid learning for feature fusion

et al. 2021

Self Cite

View full text Add to dashboard Cite

Purpose: Diffuse large B-cell lymphoma (DLBCL) is an aggressive type of lymphoma with high mortality and poor prognosis that especially has a high incidence in Asia. Accurate segmentation of DLBCL lesions is crucial for clinical radiation therapy. However, manual delineation of DLBCL lesions is tedious and time-consuming. Automatic segmentation provides an alternative solution but is difficult for diffuse lesions without the sufficient utilization of multimodality information. Our work is the first study focusing on positron emission tomography and computed tomography (PET-CT) feature fusion for the DLBCL segmentation issue. We aim to improve the fusion performance of complementary information contained in PET-CT imaging with a hybrid learning module in the supervised convolutional neural network. Methods: First, two encoder branches extract single-modality features, respectively. Next, the hybrid learning component utilizes them to generate spatial fusion maps which can quantify the contribution of complementary information. Such feature fusion maps are then concatenated with specific-modality (i.e., PET and CT) feature maps to obtain a representation of the final fused feature maps in different scales. Finally, the reconstruction part of our network creates a prediction map of DLBCL lesions by integrating and up-sampling the final fused feature maps from encoder blocks in different scales. Results: The ability of our method was evaluated to detect foreground and segment lesions in three independent body regions (nasopharynx, chest, and abdomen) of a set of 45 PET-CT scans. Extensive ablation experiments compared our method to four baseline techniques for multimodality fusion (input-level (IL) fusion, multichannel (MC) strategy, multibranch (MB) strategy, and quantitative weighting (QW) fusion). The results showed that our method achieved a high detection accuracy (99.63% in the nasopharynx, 99.51% in the chest, and 99.21% in the abdomen) and had the superiority in segmentation performance with the mean dice similarity coefficient (DSC) of 73.03% and the modified Hausdorff distance (MHD) of 4.39 mm, when compared with the baselines (DSC: IL:

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Diffuse large B‐cell lymphoma segmentation in PET‐CT images via hybrid learning for feature fusion

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Another study designed a variant 3-D U-Net network architecture to segment lung tumors from PET/CT images—each input channel (individual U-Net network) shared complementary information in the decoder section with the other channel [ 94 ]. Likewise, a coarse U-Net and adversarial network have been used to segment extranodal natural killer/T cell lymphoma nasal cancer lesions [ 95 ] and the performance of a U-Net was compared to a W-Net [ 96 ] on segmenting multiple malignant bone lesions, both studies used PET/CT images [ 97 ]. Finally, Zhou and Chellappa applied a CNN network combined with prior knowledge (roundness and relative position) to segment cervical tumors in PET images [ 38 ].…”

Section: Overview Of Deep Learning Applications In Medical Imagingmentioning

confidence: 99%

Application and Construction of Deep Learning Networks in Medical Imaging

Torres-Velázquez

Chen

et al. 2021

IEEE Trans. Radiat. Plasma Med. Sci.

View full text Add to dashboard Cite

Deep learning (DL) approaches are part of the machine learning (ML) subfield concerned with the development of computational models to train artificial intelligence systems. DL models are characterized by automatically extracting high-level features from the input data to learn the relationship between matching datasets. Thus, its implementation offers an advantage over common ML methods that often require the practitioner to have some domain knowledge of the input data to select the best latent representation. As a result of this advantage, DL has been successfully applied within the medical imaging field to address problems such as disease classification and tumor segmentation for which it is difficult or impossible to determine which image features are relevant. Therefore, taking into consideration the positive impact of DL on the medical imaging field, this paper reviews the key concepts associated with its evolution and implementation. The sections of this review summarize the milestones related to the development of the DL field, followed by a description of the elements of deep neural network and an overview of its application within the medical imaging field. Subsequently, the key steps necessary to implement a supervised DL application are defined, and associated limitations are discussed.

show abstract

A Simple Mean-Teacher UNet Model for Efficient Abdominal Organ Segmentation

Zhao¹,

Chu²

2022

Fast and Low-Resource Semi-Supervised Abdominal Organ Segmentation

View full text Add to dashboard Cite

Coarse-to-Fine Adversarial Networks and Zone-Based Uncertainty Analysis for NK/T-Cell Lymphoma Segmentation in CT/PET Images

Cited by 39 publications

References 28 publications

Diffuse large B‐cell lymphoma segmentation in PET‐CT images via hybrid learning for feature fusion

Diffuse large B‐cell lymphoma segmentation in PET‐CT images via hybrid learning for feature fusion

Application and Construction of Deep Learning Networks in Medical Imaging

A Simple Mean-Teacher UNet Model for Efficient Abdominal Organ Segmentation

Contact Info

Product

Resources

About