ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2008.13013
19
28

Lymph Node Gross Tumor Volume Detection in Oncology Imaging via Relationship Learning Using Graph Neural Network

29 August 2020
Chun-Hung Chao
Zhuotun Zhu
Dazhou Guo
K. Yan
T. Ho
Jinzheng Cai
Adam P. Harrison
X. Ye
Jing Xiao
Alan Yuille
Min Sun
Le Lu
D. Jin
ArXivPDFHTML
Abstract

Determining the spread of GTVLN_{LN}LN​ is essential in defining the respective resection or irradiating regions for the downstream workflows of surgical resection and radiotherapy for many cancers. Different from the more common enlarged lymph node (LN), GTVLN_{LN}LN​ also includes smaller ones if associated with high positron emission tomography signals and/or any metastasis signs in CT. This is a daunting task. In this work, we propose a unified LN appearance and inter-LN relationship learning framework to detect the true GTVLN_{LN}LN​. This is motivated by the prior clinical knowledge that LNs form a connected lymphatic system, and the spread of cancer cells among LNs often follows certain pathways. Specifically, we first utilize a 3D convolutional neural network with ROI-pooling to extract the GTVLN_{LN}LN​'s instance-wise appearance features. Next, we introduce a graph neural network to further model the inter-LN relationships where the global LN-tumor spatial priors are included in the learning process. This leads to an end-to-end trainable network to detect by classifying GTVLN_{LN}LN​. We operate our model on a set of GTVLN_{LN}LN​ candidates generated by a preliminary 1st-stage method, which has a sensitivity of >85%>85\%>85% at the cost of high false positive (FP) (>15>15>15 FPs per patient). We validate our approach on a radiotherapy dataset with 142 paired PET/RTCT scans containing the chest and upper abdominal body parts. The proposed method significantly improves over the state-of-the-art (SOTA) LN classification method by 5.5%5.5\%5.5% and 13.1%13.1\%13.1% in F1 score and the averaged sensitivity value at 2,3,4,62, 3, 4, 62,3,4,6 FPs per patient, respectively.

View on arXiv
Comments on this paper