Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Large Scale MRI Collection and Segmentation of Cirrhotic Liver

Debesh Jha, Onkar Kishor Susladkar, Vandan Gorade, Elif Keles, Matthew Antalek, Deniz Seyithanoglu, Timurhan Cebeci, Halil Ertugrul Aktas, Gulbiz Dagoglu Kartal, Sabahattin Kaymakoglu, Sukru Mehmet Erturk, Yuri Velichko, Daniela Ladner, Amir A. Borhani, Alpay Medetalibeyoglu, Gorkem Durak, Ulas Bagci

TL;DR

This work addresses the lack of large-scale, annotated MRI data for automated cirrhotic liver segmentation by introducing CirrMRI600+, a publicly released dataset comprising 628 abdominal MRI scans (310 T1-weighted and 318 T2-weighted) from 339 patients with liver cirrhosis, plus 55 non-cirrhotic controls. The authors establish a semi-automated annotation pipeline using MRSegmentor followed by expert refinement, yielding 39,954 ground-truth segmentation slices and diverse modality coverage, with data standardized to NIfTI and hosted on OSF for reproducibility. They perform comprehensive benchmarking of 11 state-of-the-art 3D segmentation networks across both T1W and T2W modalities to create robust baselines and reveal modality-specific challenges, with transformer-based approaches like nnSynergyNet3D achieving top performance. The dataset and benchmarks enable rapid development of automated cirrhosis assessment tools and set the stage for future multi-organ MR studies in abdominal imaging.

Abstract

Liver cirrhosis represents the end stage of chronic liver disease, characterized by extensive fibrosis and nodular regeneration that significantly increases mortality risk. While magnetic resonance imaging (MRI) offers a non-invasive assessment, accurately segmenting cirrhotic livers presents substantial challenges due to morphological alterations and heterogeneous signal characteristics. Deep learning approaches show promise for automating these tasks, but progress has been limited by the absence of large-scale, annotated datasets. Here, we present CirrMRI600+, the first comprehensive dataset comprising 628 high-resolution abdominal MRI scans (310 T1-weighted and 318 T2-weighted sequences, totaling nearly 40,000 annotated slices) with expert-validated segmentation labels for cirrhotic livers. The dataset includes demographic information, clinical parameters, and histopathological validation where available. Additionally, we provide benchmark results from 11 state-of-the-art deep learning experiments to establish performance standards. CirrMRI600+ enables the development and validation of advanced computational methods for cirrhotic liver analysis, potentially accelerating progress toward automated Cirrhosis visual staging and personalized treatment planning.

Large Scale MRI Collection and Segmentation of Cirrhotic Liver

TL;DR

This work addresses the lack of large-scale, annotated MRI data for automated cirrhotic liver segmentation by introducing CirrMRI600+, a publicly released dataset comprising 628 abdominal MRI scans (310 T1-weighted and 318 T2-weighted) from 339 patients with liver cirrhosis, plus 55 non-cirrhotic controls. The authors establish a semi-automated annotation pipeline using MRSegmentor followed by expert refinement, yielding 39,954 ground-truth segmentation slices and diverse modality coverage, with data standardized to NIfTI and hosted on OSF for reproducibility. They perform comprehensive benchmarking of 11 state-of-the-art 3D segmentation networks across both T1W and T2W modalities to create robust baselines and reveal modality-specific challenges, with transformer-based approaches like nnSynergyNet3D achieving top performance. The dataset and benchmarks enable rapid development of automated cirrhosis assessment tools and set the stage for future multi-organ MR studies in abdominal imaging.

Abstract

Liver cirrhosis represents the end stage of chronic liver disease, characterized by extensive fibrosis and nodular regeneration that significantly increases mortality risk. While magnetic resonance imaging (MRI) offers a non-invasive assessment, accurately segmenting cirrhotic livers presents substantial challenges due to morphological alterations and heterogeneous signal characteristics. Deep learning approaches show promise for automating these tasks, but progress has been limited by the absence of large-scale, annotated datasets. Here, we present CirrMRI600+, the first comprehensive dataset comprising 628 high-resolution abdominal MRI scans (310 T1-weighted and 318 T2-weighted sequences, totaling nearly 40,000 annotated slices) with expert-validated segmentation labels for cirrhotic livers. The dataset includes demographic information, clinical parameters, and histopathological validation where available. Additionally, we provide benchmark results from 11 state-of-the-art deep learning experiments to establish performance standards. CirrMRI600+ enables the development and validation of advanced computational methods for cirrhotic liver analysis, potentially accelerating progress toward automated Cirrhosis visual staging and personalized treatment planning.

Paper Structure

This paper contains 6 sections, 4 figures, 4 tables.

Table of Contents

  1. MRI acquisition:
  2. Patient Data:
  3. Properties of T1W and T2W:
  4. Data Standardization:
  5. Segmentation Annotation:
  6. Data split and evaluation metrics:

Figures (4)

  • Figure 1: Different stages of fibrosis in liver MRI are shown in T1W and T2W samples, respectively. Varying levels of fibrotic tissues are observed across different scans, indicating the high variability and extreme challenge in texture and shape changes. To highlight the difference in fibrotic tissue diversity, MRI images in the dotted boxes (both red and green) are shown, they are having the smallest vulnerability to fibrotic tissue while still indicating cirrhosis. The first two rows are T1W, and the second two rows are T2W.
  • Figure 2: Qualitative results of different models on segmenting mild and severe cirrhosis from abdominal T1W MRI scans. The white bounding circles show major errors made by the models.
  • Figure 3: Qualitative results of different models on segmenting mild and severe cirrhosis from abdominal T2W MRI scans. The white bounding circles show major errors made by the models.
  • Figure 4: The figure shows the qualitative results examples of different models on segmenting mild and severe cirrhosis from abdominal T2W MRI scans. From the figure, it can be observed that MedSegDiff is the best choice.