Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification

Ling Ma; Maysam Shahedi; Ted Shi; Martin Halicek; James V. Little; Amy Y. Chen; Larry L. Myers; Baran D. Sumer; Baowei Fei

doi:10.1117/12.2581046

Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification

Ling Ma, Maysam Shahedi, Ted Shi, Martin Halicek, James V. Little, Amy Y. Chen, Larry L. Myers, Baran D. Sumer, Baowei Fei

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

6 Scopus citations

Abstract

Surgery is a major treatment method for squamous cell carcinoma (SCC). During surgery, insufficient tumor margin may lead to local recurrence of cancer. Hyperspectral imaging (HSI) is a promising optical imaging technique for in vivo cancer detection and tumor margin assessment. In this study, a fully convolutional network (FCN) was implemented for tumor detection and margin assessment in hyperspectral images of SCC. The FCN was trained and tested with hyperspectral images of 25 ex vivo SCC surgical specimens from 20 different patients. The network was evaluated per patient and achieved pixel-level tissue classification with an average AUC of 0.88, 0.83 accuracy, 0.84 sensitivity and 0.70 specificity. The 95% Hausdorff distance of assessed tumor margin in 17 patients was less than 2 mm, and the classification time of each tissue specimen took less than 10 seconds. The proposed method potentially facilitates intraoperative tumor margin assessment and improves surgical outcomes.

Original language	English (US)
Title of host publication	Medical Imaging 2021
Subtitle of host publication	Image-Guided Procedures, Robotic Interventions, and Modeling
Editors	Cristian A. Linte, Jeffrey H. Siewerdsen
Publisher	SPIE
ISBN (Electronic)	9781510640252
DOIs	https://doi.org/10.1117/12.2581046
State	Published - 2021
Event	Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling - Virtual, Online Duration: Feb 15 2021 → Feb 19 2021

Publication series

Name	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume	11598
ISSN (Print)	1605-7422

Conference

Conference	Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling
City	Virtual, Online
Period	2/15/21 → 2/19/21

Keywords

Classification
Hyperspectral imaging
Squamous cell carcinoma
Tumor margin assessment
U-Net

ASJC Scopus subject areas

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

Access to Document

10.1117/12.2581046

Cite this

Ma, L., Shahedi, M., Shi, T., Halicek, M., Little, J. V., Chen, A. Y., Myers, L. L., Sumer, B. D., & Fei, B. (2021). Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification. In C. A. Linte, & J. H. Siewerdsen (Eds.), Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling Article 1159811 (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 11598). SPIE. https://doi.org/10.1117/12.2581046

Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification. / Ma, Ling; Shahedi, Maysam; Shi, Ted et al.
Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling. ed. / Cristian A. Linte; Jeffrey H. Siewerdsen. SPIE, 2021. 1159811 (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 11598).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Ma, L, Shahedi, M, Shi, T, Halicek, M, Little, JV, Chen, AY, Myers, LL , Sumer, BD & Fei, B 2021, Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification. in CA Linte & JH Siewerdsen (eds), Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling., 1159811, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 11598, SPIE, Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling, Virtual, Online, 2/15/21. https://doi.org/10.1117/12.2581046

Ma L, Shahedi M, Shi T, Halicek M, Little JV, Chen AY et al. Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification. In Linte CA, Siewerdsen JH, editors, Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling. SPIE. 2021. 1159811. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). doi: 10.1117/12.2581046

Ma, Ling ; Shahedi, Maysam ; Shi, Ted et al. / Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification. Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling. editor / Cristian A. Linte ; Jeffrey H. Siewerdsen. SPIE, 2021. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).

@inproceedings{c41401d6371d4c4e8f76a720d7a60cb3,

title = "Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification",

abstract = "Surgery is a major treatment method for squamous cell carcinoma (SCC). During surgery, insufficient tumor margin may lead to local recurrence of cancer. Hyperspectral imaging (HSI) is a promising optical imaging technique for in vivo cancer detection and tumor margin assessment. In this study, a fully convolutional network (FCN) was implemented for tumor detection and margin assessment in hyperspectral images of SCC. The FCN was trained and tested with hyperspectral images of 25 ex vivo SCC surgical specimens from 20 different patients. The network was evaluated per patient and achieved pixel-level tissue classification with an average AUC of 0.88, 0.83 accuracy, 0.84 sensitivity and 0.70 specificity. The 95% Hausdorff distance of assessed tumor margin in 17 patients was less than 2 mm, and the classification time of each tissue specimen took less than 10 seconds. The proposed method potentially facilitates intraoperative tumor margin assessment and improves surgical outcomes.",

keywords = "Classification, Hyperspectral imaging, Squamous cell carcinoma, Tumor margin assessment, U-Net",

author = "Ling Ma and Maysam Shahedi and Ted Shi and Martin Halicek and Little, {James V.} and Chen, {Amy Y.} and Myers, {Larry L.} and Sumer, {Baran D.} and Baowei Fei",

note = "Funding Information: This research was supported in part by the U.S. National Institutes of Health (NIH) grants (R01CA156775, R01CA204254, R01HL140325, and R21CA231911) and by the Cancer Prevention and Research Institute of Texas (CPRIT) grant RP190588. Publisher Copyright: {\textcopyright} 2021 SPIE.; Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling ; Conference date: 15-02-2021 Through 19-02-2021",

year = "2021",

doi = "10.1117/12.2581046",

language = "English (US)",

series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

publisher = "SPIE",

editor = "Linte, {Cristian A.} and Siewerdsen, {Jeffrey H.}",

booktitle = "Medical Imaging 2021",

}

TY - GEN

T1 - Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification

AU - Ma, Ling

AU - Shahedi, Maysam

AU - Shi, Ted

AU - Halicek, Martin

AU - Little, James V.

AU - Chen, Amy Y.

AU - Myers, Larry L.

AU - Sumer, Baran D.

AU - Fei, Baowei

N1 - Funding Information: This research was supported in part by the U.S. National Institutes of Health (NIH) grants (R01CA156775, R01CA204254, R01HL140325, and R21CA231911) and by the Cancer Prevention and Research Institute of Texas (CPRIT) grant RP190588. Publisher Copyright: © 2021 SPIE.

PY - 2021

Y1 - 2021

N2 - Surgery is a major treatment method for squamous cell carcinoma (SCC). During surgery, insufficient tumor margin may lead to local recurrence of cancer. Hyperspectral imaging (HSI) is a promising optical imaging technique for in vivo cancer detection and tumor margin assessment. In this study, a fully convolutional network (FCN) was implemented for tumor detection and margin assessment in hyperspectral images of SCC. The FCN was trained and tested with hyperspectral images of 25 ex vivo SCC surgical specimens from 20 different patients. The network was evaluated per patient and achieved pixel-level tissue classification with an average AUC of 0.88, 0.83 accuracy, 0.84 sensitivity and 0.70 specificity. The 95% Hausdorff distance of assessed tumor margin in 17 patients was less than 2 mm, and the classification time of each tissue specimen took less than 10 seconds. The proposed method potentially facilitates intraoperative tumor margin assessment and improves surgical outcomes.

AB - Surgery is a major treatment method for squamous cell carcinoma (SCC). During surgery, insufficient tumor margin may lead to local recurrence of cancer. Hyperspectral imaging (HSI) is a promising optical imaging technique for in vivo cancer detection and tumor margin assessment. In this study, a fully convolutional network (FCN) was implemented for tumor detection and margin assessment in hyperspectral images of SCC. The FCN was trained and tested with hyperspectral images of 25 ex vivo SCC surgical specimens from 20 different patients. The network was evaluated per patient and achieved pixel-level tissue classification with an average AUC of 0.88, 0.83 accuracy, 0.84 sensitivity and 0.70 specificity. The 95% Hausdorff distance of assessed tumor margin in 17 patients was less than 2 mm, and the classification time of each tissue specimen took less than 10 seconds. The proposed method potentially facilitates intraoperative tumor margin assessment and improves surgical outcomes.

KW - Classification

KW - Hyperspectral imaging

KW - Squamous cell carcinoma

KW - Tumor margin assessment

KW - U-Net

UR - http://www.scopus.com/inward/record.url?scp=85105542671&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85105542671&partnerID=8YFLogxK

U2 - 10.1117/12.2581046

DO - 10.1117/12.2581046

M3 - Conference contribution

C2 - 35755403

AN - SCOPUS:85105542671

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2021

A2 - Linte, Cristian A.

A2 - Siewerdsen, Jeffrey H.

PB - SPIE

T2 - Medical Imaging 2021: Image-Guided Procedures, Robotic Interventions, and Modeling

Y2 - 15 February 2021 through 19 February 2021

ER -

Pixel-level tumor margin assessment of surgical specimen with hyperspectral imaging and deep learning classification

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this