# Deep-learning-based automatic liver segmentation using computed tomography images in dogs

## TL;DR

This paper introduces a deep learning model for automatically segmenting liver regions in CT scans of dogs, achieving high accuracy and potential for clinical use.

## Contribution

The study presents a validated 3D U-Net model for canine liver segmentation in CT scans, a novel application in veterinary medicine.

## Key findings

- The model achieved a mean Dice similarity coefficient of 0.926 in scans without hepatic masses.
- Segmentation performance was slightly higher at 0.929 when including scans with and without hepatic masses.
- Manual and predicted liver volumes showed strong agreement, supporting clinical applicability.

## Abstract

Deep learning-based automated segmentation has significantly improved the efficiency and accuracy of human medicine applications. However, veterinary applications, particularly canine liver segmentation, remain limited. This study aimed to develop and validate a deep learning model based on a 3D U-Net architecture for automated liver segmentation in canine abdominal computed tomography (CT) scans.

A total of 221 canine abdominal CT scans were analyzed, comprising 159 cases without hepatic masses and 62 cases with hepatic masses. The model was trained and evaluated using two separate datasets: one containing cases without hepatic masses (Experiment 1) and the other combining cases with and without hepatic masses (Experiment 2).

Both experiments demonstrated high segmentation performance, achieving mean Dice similarity coefficients of 0.926 (Experiment 1) and 0.929 (Experiment 2).

The manual and predicted liver volumes showed excellent agreement, highlighting the potential clinical applicability of this approach.