Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers

Vivek Yadav; Amine Geahchan; Valentin Fauveau; Kazuya Yasokawa; Bachir Taouli; Hayit Greenspan

doi:10.1117/12.3047021

Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers

Vivek Yadav
, Amine Geahchan
, Valentin Fauveau
, Kazuya Yasokawa
, Bachir Taouli
, Hayit Greenspan

Icahn School of Medicine at Mount Sinai

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

Abstract

Hepatocellular carcinoma (HCC) surveillance primarily relies on ultrasound (U/S), which often exhibits decreased sensitivity in high-risk populations, such as individuals with cirrhosis or obesity. Abbreviated magnetic resonance imaging (AMRI) offers a potential alternative by employing targeted MRI sequences to enhance HCC detection. AMRI encompasses three primary strategies: non-contrast, dynamic contrast-enhanced, and hepatobiliary phase imaging, showing potential for overcoming U/S limitations in these populations. This study investigates the application of deep learning (DL) techniques to automate HCC tumor detection and segmentation within dynamic contrast-enhanced (Dyn-AMRI) protocols. Specifically, we leverage the capabilities of Vision Transformers (ViTs) to analyze complex image data and extract relevant features. Additionally, a novel heuristic is introduced to enhance the segmentation performance of the MedNeXt architecture. Our aim is to develop a robust DL pipeline for accurate HCC detection and segmentation on Dyn-AMRI, ultimately improving diagnostic outcomes.

Original language	English
Title of host publication	Medical Imaging 2025
Subtitle of host publication	Computer-Aided Diagnosis
Editors	Susan M. Astley, Axel Wismuller
Publisher	SPIE
ISBN (Electronic)	9781510685925
DOIs	https://doi.org/10.1117/12.3047021
State	Published - 2025
Externally published	Yes
Event	Medical Imaging 2025: Computer-Aided Diagnosis - San Diego, United States Duration: 17 Feb 2025 → 20 Feb 2025

Publication series

Name	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume	13407
ISSN (Print)	1605-7422

Conference

Conference	Medical Imaging 2025: Computer-Aided Diagnosis
Country/Territory	United States
City	San Diego
Period	17/02/25 → 20/02/25

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Deep Learning
Dynamic AMRI
HCC Detection
HCC Segmentation
Vision Transformers

Access to Document

10.1117/12.3047021

Cite this

Yadav, V., Geahchan, A., Fauveau, V., Yasokawa, K., Taouli, B., & Greenspan, H. (2025). Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers. In S. M. Astley, & A. Wismuller (Eds.), Medical Imaging 2025: Computer-Aided Diagnosis Article 134071T (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 13407). SPIE. https://doi.org/10.1117/12.3047021

@inproceedings{b906f0c7f0d24e4f81759e479b56b0db,

title = "Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers",

abstract = "Hepatocellular carcinoma (HCC) surveillance primarily relies on ultrasound (U/S), which often exhibits decreased sensitivity in high-risk populations, such as individuals with cirrhosis or obesity. Abbreviated magnetic resonance imaging (AMRI) offers a potential alternative by employing targeted MRI sequences to enhance HCC detection. AMRI encompasses three primary strategies: non-contrast, dynamic contrast-enhanced, and hepatobiliary phase imaging, showing potential for overcoming U/S limitations in these populations. This study investigates the application of deep learning (DL) techniques to automate HCC tumor detection and segmentation within dynamic contrast-enhanced (Dyn-AMRI) protocols. Specifically, we leverage the capabilities of Vision Transformers (ViTs) to analyze complex image data and extract relevant features. Additionally, a novel heuristic is introduced to enhance the segmentation performance of the MedNeXt architecture. Our aim is to develop a robust DL pipeline for accurate HCC detection and segmentation on Dyn-AMRI, ultimately improving diagnostic outcomes.",

keywords = "Deep Learning, Dynamic AMRI, HCC Detection, HCC Segmentation, Vision Transformers",

author = "Vivek Yadav and Amine Geahchan and Valentin Fauveau and Kazuya Yasokawa and Bachir Taouli and Hayit Greenspan",

note = "Publisher Copyright: {\textcopyright} 2025 SPIE.; Medical Imaging 2025: Computer-Aided Diagnosis ; Conference date: 17-02-2025 Through 20-02-2025",

year = "2025",

doi = "10.1117/12.3047021",

language = "אנגלית",

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

publisher = "SPIE",

editor = "Astley, \{Susan M.\} and Axel Wismuller",

booktitle = "Medical Imaging 2025",

address = "ארצות הברית",

}

Yadav, V, Geahchan, A, Fauveau, V, Yasokawa, K, Taouli, B & Greenspan, H 2025, Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers. in SM Astley & A Wismuller (eds), Medical Imaging 2025: Computer-Aided Diagnosis., 134071T, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 13407, SPIE, Medical Imaging 2025: Computer-Aided Diagnosis, San Diego, United States, 17/02/25. https://doi.org/10.1117/12.3047021

Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers. / Yadav, Vivek; Geahchan, Amine; Fauveau, Valentin et al.
Medical Imaging 2025: Computer-Aided Diagnosis. ed. / Susan M. Astley; Axel Wismuller. SPIE, 2025. 134071T (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 13407).

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

TY - GEN

T1 - Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers

AU - Yadav, Vivek

AU - Geahchan, Amine

AU - Fauveau, Valentin

AU - Yasokawa, Kazuya

AU - Taouli, Bachir

AU - Greenspan, Hayit

PY - 2025

Y1 - 2025

N2 - Hepatocellular carcinoma (HCC) surveillance primarily relies on ultrasound (U/S), which often exhibits decreased sensitivity in high-risk populations, such as individuals with cirrhosis or obesity. Abbreviated magnetic resonance imaging (AMRI) offers a potential alternative by employing targeted MRI sequences to enhance HCC detection. AMRI encompasses three primary strategies: non-contrast, dynamic contrast-enhanced, and hepatobiliary phase imaging, showing potential for overcoming U/S limitations in these populations. This study investigates the application of deep learning (DL) techniques to automate HCC tumor detection and segmentation within dynamic contrast-enhanced (Dyn-AMRI) protocols. Specifically, we leverage the capabilities of Vision Transformers (ViTs) to analyze complex image data and extract relevant features. Additionally, a novel heuristic is introduced to enhance the segmentation performance of the MedNeXt architecture. Our aim is to develop a robust DL pipeline for accurate HCC detection and segmentation on Dyn-AMRI, ultimately improving diagnostic outcomes.

AB - Hepatocellular carcinoma (HCC) surveillance primarily relies on ultrasound (U/S), which often exhibits decreased sensitivity in high-risk populations, such as individuals with cirrhosis or obesity. Abbreviated magnetic resonance imaging (AMRI) offers a potential alternative by employing targeted MRI sequences to enhance HCC detection. AMRI encompasses three primary strategies: non-contrast, dynamic contrast-enhanced, and hepatobiliary phase imaging, showing potential for overcoming U/S limitations in these populations. This study investigates the application of deep learning (DL) techniques to automate HCC tumor detection and segmentation within dynamic contrast-enhanced (Dyn-AMRI) protocols. Specifically, we leverage the capabilities of Vision Transformers (ViTs) to analyze complex image data and extract relevant features. Additionally, a novel heuristic is introduced to enhance the segmentation performance of the MedNeXt architecture. Our aim is to develop a robust DL pipeline for accurate HCC detection and segmentation on Dyn-AMRI, ultimately improving diagnostic outcomes.

KW - Deep Learning

KW - Dynamic AMRI

KW - HCC Detection

KW - HCC Segmentation

KW - Vision Transformers

UR - https://www.scopus.com/pages/publications/105004410479

U2 - 10.1117/12.3047021

DO - 10.1117/12.3047021

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:105004410479

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

BT - Medical Imaging 2025

A2 - Astley, Susan M.

A2 - Wismuller, Axel

PB - SPIE

T2 - Medical Imaging 2025: Computer-Aided Diagnosis

Y2 - 17 February 2025 through 20 February 2025

ER -

Yadav V, Geahchan A, Fauveau V, Yasokawa K, Taouli B, Greenspan H. Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers. In Astley SM, Wismuller A, editors, Medical Imaging 2025: Computer-Aided Diagnosis. SPIE. 2025. 134071T. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). doi: 10.1117/12.3047021

Automated Hepatocellular Carcinoma Detection and Segmentation in Abbreviated MRI using Vision Transformers

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this