Coarse-to-fine Nasopharyngeal Carcinoma Segmentation in MRI via Multi-stage Rendering

Li, Yang; Peng, Hong; Dan, Tingting; Hu, Yu; Tao, Guihua; Cai, Hongmin

doi:10.1109/bibm49941.2020.9313574

Cited by 9 publications

(7 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These data were captured using a Siemens MAGNETOM Verio 3T device, with Gadobenate Dimeglumine Injection as the contrast agent. The second dataset consists of NPC MRIs from 277 patients at The First People's Hospital of Foshan, China 20 . These data were captured using a GE Discovery MR750w 3.0T and Philips Achieva 1.5T devices, with Gadoteric Acid Meglumine Salt Injection as the contrast agent.…”

Section: Resultsmentioning

confidence: 99%

MFNet: Meta‐learning based on frequency‐space mix for MRI segmentation in nasopharyngeal carcinoma

Li,

Chen,

et al. 2024

J Cellular Molecular Medi

View full text Add to dashboard Cite

Deep learning techniques have been applied to medical image segmentation and demonstrated expert‐level performance. Due to the poor generalization abilities of the models in the deployment in different centres, common solutions, such as transfer learning and domain adaptation techniques, have been proposed to mitigate this issue. However, these solutions necessitate retraining the models with target domain data and annotations, which limits their deployment in clinical settings in unseen domains. We evaluated the performance of domain generalization methods on the task of MRI segmentation of nasopharyngeal carcinoma (NPC) by collecting a new dataset of 321 patients with manually annotated MRIs from two hospitals. We transformed the modalities of MRI, including T1WI, T2WI and CE‐T1WI, from the spatial domain to the frequency domain using Fourier transform. To address the bottleneck of domain generalization in MRI segmentation of NPC, we propose a meta‐learning approach based on frequency domain feature mixing. We evaluated the performance of MFNet against existing techniques for generalizing NPC segmentation in terms of Dice and MIoU. Our method evidently outperforms the baseline in handling the generalization of NPC segmentation. The MF‐Net clearly demonstrates its effectiveness for generalizing NPC MRI segmentation to unseen domains (Dice = 67.59%, MIoU = 75.74% T1W1). MFNet enhances the model's generalization capabilities by incorporating mixed‐feature meta‐learning. Our approach offers a novel perspective to tackle the domain generalization problem in the field of medical imaging by effectively exploiting the unique characteristics of medical images.

show abstract

Section: Resultsmentioning

confidence: 99%

MFNet: Meta‐learning based on frequency‐space mix for MRI segmentation in nasopharyngeal carcinoma

Li,

Chen,

et al. 2024

J Cellular Molecular Medi

View full text Add to dashboard Cite

show abstract

“…We compare the accuracy of our model with four state-of-the-art models and carried out the Kruskal–Wallis test. Five of them are conventional medical image segmentation models (Att-UNet [ 39 ], FCN [ 36 ], DeeplabV3 [ 54 ], TransNet [ 55 ], FastTransNet [ 56 ]), and the other is a nasopharyngeal cancer image segmentation model (RendUNet [ 57 ]). Figure 3 shows the segmentation results for each model.…”

Section: Resultsmentioning

confidence: 99%

“…For FCN [ 36 ], our model’s parameters are reduced by 314.64% and the FLOPs are reduced by 167.24%. For RendUNet [ 57 ], the parameters of our model are reduced by 1190.14% and the FLOPs are reduced by 533.55%. For TransNet, the parameters of our model are reduced by 2865.33% and the FLOPs are reduced by 228.09%.…”

Section: Resultsmentioning

confidence: 99%

“…However, LW-UNet is slightly limited in some complex cases because of unusual tumour shape and remote tumour location. In comparison with four other advanced models (including RendUNet [ 57 ], which specializes in nasopharyngeal carcinoma segmentation), our model achieved the best segmentation performance. In addition, our models are more effective with the lower parameters and FLOPs; this enables our model to operate in some small and medium-sized hospitals with limited resources.…”

Section: Discussionmentioning

confidence: 96%

See 1 more Smart Citation

Lightweight Compound Scaling Network for Nasopharyngeal Carcinoma Segmentation from MR Images

Liu

Han

Liu

2022

Sensors

View full text Add to dashboard Cite

Nasopharyngeal carcinoma (NPC) is a category of tumours with a high incidence in head-and-neck. To treat nasopharyngeal cancer, doctors invariably need to perform focal segmentation. However, manual segmentation is time consuming and laborious for doctors and the existing automatic segmentation methods require large computing resources, which makes some small and medium-sized hospitals unaffordable. To enable small and medium-sized hospitals with limited computational resources to run the model smoothly and improve the accuracy of structure, we propose a new LW-UNet network. The network utilises lightweight modules to form the Compound Scaling Encoder and combines the benefits of UNet to make the model both lightweight and accurate. Our model achieves a high accuracy with a Dice coefficient value of 0.813 with 3.55 M parameters and 7.51 G of FLOPs within 0.1 s (testing time in GPU), which is the best result compared with four other state-of-the-art models.

show abstract

“…Considering that NPC is a malignant tumour with a tendency to invade the surrounding tissues, in a complex MRI background, it is difficult to distinguish the signs of invasion on the edge from the closely connected normal tissues. To address the background dominant problem in improving the segmentation accuracy of NPC, Li [ 143 ] proposed a coarse-to-fine deep neural network, which started by predicting a coarse mask based on a well-designed segmentation module, followed by a boundary rendering module, which exploited semantic information from different layers of feature maps to refine the boundary of the coarse mask. The dataset contained 2000 MRI slices from 596 patients, and the DSC of the model was 0.703.…”

Section: Studies Based On DLmentioning

confidence: 99%

A Comprehensive Review on Radiomics and Deep Learning for Nasopharyngeal Carcinoma Imaging

et al. 2021

View full text Add to dashboard Cite

Nasopharyngeal carcinoma (NPC) is one of the most common malignant tumours of the head and neck, and improving the efficiency of its diagnosis and treatment strategies is an important goal. With the development of the combination of artificial intelligence (AI) technology and medical imaging in recent years, an increasing number of studies have been conducted on image analysis of NPC using AI tools, especially radiomics and artificial neural network methods. In this review, we present a comprehensive overview of NPC imaging research based on radiomics and deep learning. These studies depict a promising prospect for the diagnosis and treatment of NPC. The deficiencies of the current studies and the potential of radiomics and deep learning for NPC imaging are discussed. We conclude that future research should establish a large-scale labelled dataset of NPC images and that studies focused on screening for NPC using AI are necessary.

show abstract

Coarse-to-fine Nasopharyngeal Carcinoma Segmentation in MRI via Multi-stage Rendering

Cited by 9 publications

References 22 publications

MFNet: Meta‐learning based on frequency‐space mix for MRI segmentation in nasopharyngeal carcinoma

MFNet: Meta‐learning based on frequency‐space mix for MRI segmentation in nasopharyngeal carcinoma

Lightweight Compound Scaling Network for Nasopharyngeal Carcinoma Segmentation from MR Images

A Comprehensive Review on Radiomics and Deep Learning for Nasopharyngeal Carcinoma Imaging

Contact Info

Product

Resources

About