Feasibility of an artificial intelligence system for tumor response evaluation
Nie Xiuli1, Chen Hua2, Gao Peng3
1Department of Radiology, Jinan Central Hospital, Shandong First Medical University, Jinan, China.
BMC Medical Imaging
|October 18, 2024
Summary
Artificial Intelligence (AI) accurately measured tumor long-diameters, improving efficiency and reducing errors in treatment response evaluation. AI measurements showed good agreement with manual methods, unlike 3D measurements.
Area of Science:
- Oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Accurate tumor size measurement is crucial for evaluating treatment response in oncology.
- Manual measurement of tumor dimensions can be subjective and time-consuming.
- AI offers potential for automated and objective image analysis.
Purpose of the Study:
- To assess the feasibility and accuracy of using Artificial Intelligence (AI) for measuring tumor long-diameter.
- To compare AI-based measurements with manual radiological measurements for treatment response evaluation.
Main Methods:
- 48 patients with lung tumors underwent 277 measurements.
- Radiologists manually measured tumor long-diameter in the axial plane.
- AI software measured long-diameter in both axial and 3D planes.
- Statistical analyses included Bland-Altman plots, Spearman correlation, and paired t-tests.
Main Results:
- AI measurements showed good agreement with manual measurements (bias -0.28 mm, P=0.497).
- Three-dimensional (3D) AI measurements did not agree with manual measurements (P<0.001).
- AI measurements demonstrated no statistically significant difference from manual measurements (P=0.497).
Conclusions:
- AI measurement of tumor long-diameter is feasible and reliable for treatment response evaluation.
- AI enhances efficiency and reduces subjective errors in tumor measurement.
- AI-based tumor measurement offers a more convenient and accurate approach.
