Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque: An automated classification method

Jennifer Phipps, Yinghua Sun, Nisa Hatami, Michael C. Fishbein, Amit Rajaram, Ramez Saroufeem, Laura Marcu

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

2 Citations (Scopus)

Abstract

The objective of this study was to develop an automated algorithm which uses fluorescence lifetime imaging microscopy (FLIM) images of human aortic atherosclerotic plaque to provide quantitative and spatial information regarding compositional features related to plaque vulnerability such as collagen degradation, lipid accumulation, and macrophage infiltration. Images were acquired through a flexible fiber imaging bundle with intravascular potential at two wavelength bands optimal to recognizing markers of vulnerability: F 377: 377/55 nm and F460: 460/50 nm (center wavelength/bandwidth). A classification method implementing principal components analysis and linear discriminant analysis to correlate FLIM data sets with histopathology was validated on a training set and then used to classify a validation set of FLIM images. The output of this algorithm was a false-color image with each pixel color coded to represent the chemical composition of the sample. Surface areas occupied by elastin, collagen, and lipid components were then calculated and used to define the vulnerability of each imaged location. Four groups were defined: early lesion, stable, mildly vulnerable and extremely vulnerable. Each imaged location was categorized in one of the groups based on histopathology and classification results; sensitivities (SE) and specificities (SP) were calculated (SE %/SP %): early lesion: 95/96, stable: 71/97, mildly vulnerable: 75/94, and extremely vulnerable: 100/93. The capability of this algorithm to use FLIM images to quickly determine the chemical composition of atherosclerotic plaque, particularly related to vulnerability, further enhances the potential of this system for implementation as an intravascular diagnostic modality.

Original languageEnglish (US)
Title of host publicationProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume7548
DOIs
StatePublished - 2010
EventPhotonic Therapeutics and Diagnostics VI - San Francisco, CA, United States
Duration: Jan 23 2010Jan 25 2010

Other

OtherPhotonic Therapeutics and Diagnostics VI
CountryUnited States
CitySan Francisco, CA
Period1/23/101/25/10

Fingerprint

Optical Imaging
vulnerability
Microscopy
Microscopic examination
Fluorescence
microscopy
Imaging techniques
life (durability)
fluorescence
Atherosclerotic Plaques
collagens
Collagen
Color
lesions
Lipids
lipids
chemical composition
Sensitivity and Specificity
Elastin
elastin

Keywords

  • atherosclerosis
  • classification
  • endoscopy
  • FLIM

ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics
  • Electronic, Optical and Magnetic Materials
  • Biomaterials
  • Radiology Nuclear Medicine and imaging

Cite this

Phipps, J., Sun, Y., Hatami, N., Fishbein, M. C., Rajaram, A., Saroufeem, R., & Marcu, L. (2010). Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque: An automated classification method. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE (Vol. 7548). [754839] https://doi.org/10.1117/12.842724

Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque : An automated classification method. / Phipps, Jennifer; Sun, Yinghua; Hatami, Nisa; Fishbein, Michael C.; Rajaram, Amit; Saroufeem, Ramez; Marcu, Laura.

Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 7548 2010. 754839.

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

Phipps, J, Sun, Y, Hatami, N, Fishbein, MC, Rajaram, A, Saroufeem, R & Marcu, L 2010, Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque: An automated classification method. in Progress in Biomedical Optics and Imaging - Proceedings of SPIE. vol. 7548, 754839, Photonic Therapeutics and Diagnostics VI, San Francisco, CA, United States, 1/23/10. https://doi.org/10.1117/12.842724
Phipps J, Sun Y, Hatami N, Fishbein MC, Rajaram A, Saroufeem R et al. Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque: An automated classification method. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 7548. 2010. 754839 https://doi.org/10.1117/12.842724
Phipps, Jennifer ; Sun, Yinghua ; Hatami, Nisa ; Fishbein, Michael C. ; Rajaram, Amit ; Saroufeem, Ramez ; Marcu, Laura. / Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque : An automated classification method. Progress in Biomedical Optics and Imaging - Proceedings of SPIE. Vol. 7548 2010.
@inproceedings{6c0bd05ca1bb4684833920f9ee547818,
title = "Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque: An automated classification method",
abstract = "The objective of this study was to develop an automated algorithm which uses fluorescence lifetime imaging microscopy (FLIM) images of human aortic atherosclerotic plaque to provide quantitative and spatial information regarding compositional features related to plaque vulnerability such as collagen degradation, lipid accumulation, and macrophage infiltration. Images were acquired through a flexible fiber imaging bundle with intravascular potential at two wavelength bands optimal to recognizing markers of vulnerability: F 377: 377/55 nm and F460: 460/50 nm (center wavelength/bandwidth). A classification method implementing principal components analysis and linear discriminant analysis to correlate FLIM data sets with histopathology was validated on a training set and then used to classify a validation set of FLIM images. The output of this algorithm was a false-color image with each pixel color coded to represent the chemical composition of the sample. Surface areas occupied by elastin, collagen, and lipid components were then calculated and used to define the vulnerability of each imaged location. Four groups were defined: early lesion, stable, mildly vulnerable and extremely vulnerable. Each imaged location was categorized in one of the groups based on histopathology and classification results; sensitivities (SE) and specificities (SP) were calculated (SE {\%}/SP {\%}): early lesion: 95/96, stable: 71/97, mildly vulnerable: 75/94, and extremely vulnerable: 100/93. The capability of this algorithm to use FLIM images to quickly determine the chemical composition of atherosclerotic plaque, particularly related to vulnerability, further enhances the potential of this system for implementation as an intravascular diagnostic modality.",
keywords = "atherosclerosis, classification, endoscopy, FLIM",
author = "Jennifer Phipps and Yinghua Sun and Nisa Hatami and Fishbein, {Michael C.} and Amit Rajaram and Ramez Saroufeem and Laura Marcu",
year = "2010",
doi = "10.1117/12.842724",
language = "English (US)",
isbn = "9780819479440",
volume = "7548",
booktitle = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

}

TY - GEN

T1 - Endoscopic fluorescence lifetime imaging microscopy (FLIM) images of aortic plaque

T2 - An automated classification method

AU - Phipps, Jennifer

AU - Sun, Yinghua

AU - Hatami, Nisa

AU - Fishbein, Michael C.

AU - Rajaram, Amit

AU - Saroufeem, Ramez

AU - Marcu, Laura

PY - 2010

Y1 - 2010

N2 - The objective of this study was to develop an automated algorithm which uses fluorescence lifetime imaging microscopy (FLIM) images of human aortic atherosclerotic plaque to provide quantitative and spatial information regarding compositional features related to plaque vulnerability such as collagen degradation, lipid accumulation, and macrophage infiltration. Images were acquired through a flexible fiber imaging bundle with intravascular potential at two wavelength bands optimal to recognizing markers of vulnerability: F 377: 377/55 nm and F460: 460/50 nm (center wavelength/bandwidth). A classification method implementing principal components analysis and linear discriminant analysis to correlate FLIM data sets with histopathology was validated on a training set and then used to classify a validation set of FLIM images. The output of this algorithm was a false-color image with each pixel color coded to represent the chemical composition of the sample. Surface areas occupied by elastin, collagen, and lipid components were then calculated and used to define the vulnerability of each imaged location. Four groups were defined: early lesion, stable, mildly vulnerable and extremely vulnerable. Each imaged location was categorized in one of the groups based on histopathology and classification results; sensitivities (SE) and specificities (SP) were calculated (SE %/SP %): early lesion: 95/96, stable: 71/97, mildly vulnerable: 75/94, and extremely vulnerable: 100/93. The capability of this algorithm to use FLIM images to quickly determine the chemical composition of atherosclerotic plaque, particularly related to vulnerability, further enhances the potential of this system for implementation as an intravascular diagnostic modality.

AB - The objective of this study was to develop an automated algorithm which uses fluorescence lifetime imaging microscopy (FLIM) images of human aortic atherosclerotic plaque to provide quantitative and spatial information regarding compositional features related to plaque vulnerability such as collagen degradation, lipid accumulation, and macrophage infiltration. Images were acquired through a flexible fiber imaging bundle with intravascular potential at two wavelength bands optimal to recognizing markers of vulnerability: F 377: 377/55 nm and F460: 460/50 nm (center wavelength/bandwidth). A classification method implementing principal components analysis and linear discriminant analysis to correlate FLIM data sets with histopathology was validated on a training set and then used to classify a validation set of FLIM images. The output of this algorithm was a false-color image with each pixel color coded to represent the chemical composition of the sample. Surface areas occupied by elastin, collagen, and lipid components were then calculated and used to define the vulnerability of each imaged location. Four groups were defined: early lesion, stable, mildly vulnerable and extremely vulnerable. Each imaged location was categorized in one of the groups based on histopathology and classification results; sensitivities (SE) and specificities (SP) were calculated (SE %/SP %): early lesion: 95/96, stable: 71/97, mildly vulnerable: 75/94, and extremely vulnerable: 100/93. The capability of this algorithm to use FLIM images to quickly determine the chemical composition of atherosclerotic plaque, particularly related to vulnerability, further enhances the potential of this system for implementation as an intravascular diagnostic modality.

KW - atherosclerosis

KW - classification

KW - endoscopy

KW - FLIM

UR - http://www.scopus.com/inward/record.url?scp=79751502139&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79751502139&partnerID=8YFLogxK

U2 - 10.1117/12.842724

DO - 10.1117/12.842724

M3 - Conference contribution

AN - SCOPUS:79751502139

SN - 9780819479440

VL - 7548

BT - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

ER -