Dataset Information

Lymph node detection in MR Lymphography: false positive reduction using multi-view convolutional neural networks.

ABSTRACT: Purpose:To investigate whether multi-view convolutional neural networks can improve a fully automated lymph node detection system for pelvic MR Lymphography (MRL) images of patients with prostate cancer. Methods:A fully automated computer-aided detection (CAD) system had been previously developed to detect lymph nodes in MRL studies. The CAD system was extended with three types of 2D multi-view convolutional neural networks (CNN) aiming to reduce false positives (FP). A 2D multi-view CNN is an efficient approximation of a 3D CNN, and three types were evaluated: a 1-view, 3-view, and 9-view 2D CNN. The three deep learning CNN architectures were trained and configured on retrospective data of 240 prostate cancer patients that received MRL images as the standard of care between January 2008 and April 2010. The MRL used ferumoxtran-10 as a contrast agent and comprised at least two imaging sequences: a 3D T1-weighted and a 3D T2*-weighted sequence. A total of 5089 lymph nodes were annotated by two expert readers, reading in consensus. A first experiment compared the performance with and without CNNs and a second experiment compared the individual contribution of the 1-view, 3-view, or 9-view architecture to the performance. The performances were visually compared using free-receiver operating characteristic (FROC) analysis and statistically compared using partial area under the FROC curve analysis. Training and analysis were performed using bootstrapped FROC and 5-fold cross-validation. Results:Adding multi-view CNNs significantly (p < 0.01) reduced false positive detections. The 3-view and 9-view CNN outperformed (p < 0.01) the 1-view CNN, reducing FP from 20.6 to 7.8/image at 80% sensitivity. Conclusion:Multi-view convolutional neural networks significantly reduce false positives in a lymph node detection system for MRL images, and three orthogonal views are sufficient. At the achieved level of performance, CAD for MRL may help speed up finding lymph nodes and assessing them for potential metastatic involvement.

SUBMITTER: Debats OA

PROVIDER: S-EPMC6876485 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Lymph node detection in MR Lymphography: false positive reduction using multi-view convolutional neural networks.

Debats Oscar A OA Litjens Geert J S GJS Huisman Henkjan J HJ

PeerJ 20191122

<h4>Purpose</h4>To investigate whether multi-view convolutional neural networks can improve a fully automated lymph node detection system for pelvic MR Lymphography (MRL) images of patients with prostate cancer.<h4>Methods</h4>A fully automated computer-aided detection (CAD) system had been previously developed to detect lymph nodes in MRL studies. The CAD system was extended with three types of 2D multi-view convolutional neural networks (CNN) aiming to reduce false positives (FP). A 2D multi-v ...[more]

PMID: 31772836

Similar Datasets

Project description:Positron lymphography using 18F-FDG followed by Cerenkov-guided resection of lymph nodes in healthy mice has previously been introduced by our group. Our aim in this study was to further assess the technique's potential beyond merely localizing sentinel lymph nodes. We now aimed to evaluate the potential of positron lymphography to characterize the nodes with respect to their tumor status in order to identify metastatic lymph nodes. We explored whether metastatic nodes could be distinguished from normal nodes via dynamic 18F-FDG lymphography, to then be resected under Cerenkov imaging guidance. Methods: A murine melanoma cell line highly metastatic to lymph nodes (B16F10) was implanted subcutaneously on the dorsal hind paw of C57 mice while the tumor-free contralateral leg served as an intraindividual control. A model of reactive lymph nodes after concanavalin A challenge served as an additional control to provide nonmalignant inflammatory lymphadenopathy. Dynamic PET/CT imaging was performed immediately after injection of 18F-FDG around the tumor or intracutaneously in the contralateral footpad. Furthermore, PET/CT and Cerenkov studies were performed repeatedly over time to follow the course of metastatic spread. In selected mice, popliteal lymph nodes underwent Cerenkov luminescence imaging. Hematoxylin and eosin staining was done to verify the presence of lymphatic melanoma infiltration. Results: Positron lymphography using 18F-FDG was successfully performed on tumor-bearing and non-tumor-bearing mice, as well as on controls bearing sites of inflammation; the results clearly identified the sentinel lymph node basin and delineated the lymphatic drainage. Significantly prolonged retention of activity was evident in metastatic nodes as compared with controls without tumor. On the basis of these results, the contrast in detection and identification of metastatic lymph nodes was distinct and could be used for guided lymph node resection, such as by using Cerenkov luminescence imaging. However, retention after 18F-FDG lymphography was also seen in acute inflammatory lymphadenopathy. Conclusion: In a tumor model, significantly longer retention of the radiotracer during 18F-FDG lymphography was seen in metastatic than nonmetastatic lymph nodes, allowing for differentiation between the two and for selective resection of tumor-bearing nodes using Cerenkov imaging. Inflammation can be better differentiated in a subacute state.

Dataset Information

Lymph node detection in MR Lymphography: false positive reduction using multi-view convolutional neural networks.

Publications

Lymph node detection in MR Lymphography: false positive reduction using multi-view convolutional neural networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets