Related Experiment Video
Updated: Jun 4, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Enhancing lesion detection in liver and kidney CT scans via lesion mask selection from two models: A main model and a
Abdullah F Al-Battal1, Van Ha Tang2, Quang Duc Tran2
1Electrical and Computer Engineering Department, UC San Diego, La Jolla, CA, USA; Electrical Engineering Department, King Fahd University of Petroleum and Minerals, Saudi Arabia.
Abstract:
Automated segmentation and detection of tumors in CT scans of the liver and kidney have a significant potential in assisting clinicians with cancer diagnosis and treatment planning. However, current approaches, including state-of-the-art deep learning ones, still face many challenges. Many tumors are not detected by these approaches when tested on public datasets for tumor detection and segmentation such as the Kidney Tumor Segmentation Challenge (KiTS) and the Liver tumor segmentation challenge (LiTS). False negative rates by lesion as high as 50% are commonly observed, and this rate is even higher for smaller lesions as they exhibit a high degree of variability (heterogeneity) among themselves. Additionally, in numerous instances, these lesions share similarities (homogeneity) in intensity, size, and shape with other anatomical structures as well as blurriness and blending with surrounding tissue. To improve the detection and segmentation accuracy of lesions in CT scans of the liver and kidney, we propose a selective ensemble approach that uses the predictions of two models to select the best possible mask for lesions. Both models are based on the UNet architecture and use the ConvNext convolutional block in both the encoder and decoder. The first model is trained on lesion segmentation regardless of size, while the second is designed and fine-tuned to segment and detect small lesions. Once the segmentation mask is predicted from both models we extract intensity-based features from within the lesion, contrast them with features from surrounding tissue, and select the mask that maximizes features' separation between the two. We test our approach on three different datasets for lesion segmentation in the kidney and liver. Our proposed approach achieves an improved detection and segmentation performance and is able to increase the number of lesions detected in all three datasets when compared to current state-of-the-art models.
Insights
This study introduces a novel ensemble method to improve automated tumor detection and segmentation in CT scans, significantly reducing false negatives for liver and kidney cancers. The approach enhances diagnostic accuracy by selecting the best segmentation mask based on feature separation.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Automated tumor segmentation in CT scans aids cancer diagnosis and treatment planning.
- Current deep learning models struggle with detecting small, heterogeneous, or low-contrast lesions, leading to high false negative rates.
Purpose of the Study:
- To enhance the accuracy of automated tumor detection and segmentation in liver and kidney CT scans.
- To address the limitations of existing models in identifying challenging lesions.
Main Methods:
- A selective ensemble approach using two UNet-based models with ConvNext blocks was developed.
- One model segments all lesions; the second is fine-tuned for small lesions.
- A feature-based selection mechanism identifies the optimal segmentation mask by maximizing intensity separation between lesions and surrounding tissue.
Main Results:
- The proposed method demonstrated improved detection and segmentation performance across three datasets.
- It successfully increased the number of detected lesions compared to state-of-the-art models.
- The approach showed particular efficacy in identifying challenging small or heterogeneous tumors.
Conclusions:
- The selective ensemble approach offers a significant advancement in automated tumor analysis for liver and kidney CT scans.
- This method has the potential to improve clinical decision-making by providing more accurate and comprehensive tumor detection.
- Further validation on diverse datasets could solidify its role in clinical practice.
More Related Videos
09:49Dual-phase Cone-beam Computed Tomography to See, Reach, and Treat Hepatocellular Carcinoma during Drug-eluting Beads Transarterial Chemo-embolization
Published on: December 2, 2013
10:25Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019