Towards automatic 3D bone marrow segmentation

Chuong T. Nguyen, Joseph P. Havlicek, Jennifer Holter Chakrabarty, Quyen Duong, Sara K. Vesely

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

Current noninvasive evaluation of bone marrow proliferation in leukemia treatment is limited to manually examining marrow tissue in multiple regions of interest (ROIs). The statistics extracted from these ROIs often fail to provide an accurate global characterization of the patient's marrow. We propose an automatic framework for segmenting spinal marrow compartments to characterize the bone marrow from full-body joint PET/CT scans acquired subsequent to bone marrow transplantation. We first apply a graph-cut algorithm to the CT volume to obtain a 3D full-body bone map. We then isolate the spinal column in a single sagittal plane where connected components labeling and iterative thresholding are used to segment the vertebral bodies. This fully automated approach achieves an average accuracy of 91.7% and a worst case accuracy of 80.4% in testing on 51 scans of 17 patients. Finally, we outline a method for rejecting the cortical bone in transverse planes that can be combined with the sagittally segmented vertebral bodies to obtain a 3D map of the vertebral body medullary cavities for the entire spine.

Original languageEnglish (US)
Title of host publication2016 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2016 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages9-12
Number of pages4
ISBN (Electronic)9781467399197
DOIs
StatePublished - Apr 25 2016
EventIEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2016 - Santa Fe, United States
Duration: Mar 6 2016Mar 8 2016

Publication series

NameProceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation
Volume2016-April

Conference

ConferenceIEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2016
CountryUnited States
CitySanta Fe
Period3/6/163/8/16

Keywords

  • 3D bone segmentation
  • bone marrow extraction
  • PET/CT
  • SUV

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Computer Science Applications

Fingerprint Dive into the research topics of 'Towards automatic 3D bone marrow segmentation'. Together they form a unique fingerprint.

Cite this