Related Experiment Video
Updated: Nov 14, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Diagnostic Performance of Deep Learning-Based Lesion Detection Algorithm in CT for Detecting Hepatic Metastasis from
Kiwook Kim1, Sungwon Kim2, Kyunghwa Han1
1Department of Radiology, Research Institute of Radiological Science and Center for Clinical Image Data Science, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea.
Deep learning algorithms show comparable sensitivity to radiologists for detecting liver metastasis in colorectal cancer patients. However, the algorithm generated more false positives, suggesting its use as an assistive tool.
Area of Science:
- Radiology
- Artificial Intelligence
- Oncology
Background:
- Liver metastasis is a common complication in colorectal cancer (CRC).
- Accurate detection of liver metastasis is crucial for treatment planning and patient outcomes.
- Deep learning-based lesion detection algorithms (DLLD) show promise in medical image analysis.
Purpose of the Study:
- To compare the diagnostic performance of a DLLD with that of human radiologists in detecting liver metastasis.
- To evaluate the sensitivity and false positive rates of DLLD versus radiologists in patients with CRC.
Main Methods:
- A retrospective study utilized CT images from 502 CRC patients for training the DLLD.
- A validation cohort of 85 CRC patients (with and without liver metastasis) was used for performance comparison.
- The DLLD's performance was compared against three abdominal radiologists and three radiology residents using per-lesion binary classification.
Main Results:
- DLLD achieved a sensitivity of 81.82%, comparable to abdominal radiologists (80.81%) and radiology residents (79.46%).
- However, DLLD exhibited a significantly higher false positive rate per patient (1.330) compared to abdominal radiologists (0.357) and radiology residents (0.667).
Conclusions:
- DLLD demonstrates comparable sensitivity to radiologists for liver metastasis detection in CRC patients.
- The higher false positive rate indicates DLLD is better suited as an assistant tool rather than a standalone diagnostic solution.
More Related Videos
07:13Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
04:22Application of Laparoscopic Hepatectomy Combined with Intraoperative Microwave Ablation in Colorectal Cancer Liver Metastasis
Published on: March 3, 2023