Related Experiment Videos
Estimation or quantification of tumour volume? CT study on irregular phantoms
M Tiitola1, L Kivisaari, P Tervahartiala
1Department of Radiology, Helsinki University Central Hospital, Finland.
Acta Radiologica (Stockholm, Sweden : 1987)
|February 13, 2001
Summary
Computerised volume quantification of tumour phantoms in spiral CT scans is more accurate than subjective size estimation by radiologists. Using multiple observers or computer analysis improves accuracy for tumour volume assessment.
Area of Science:
- Radiology
- Medical Imaging Analysis
Background:
- Accurate tumor volume quantification is crucial for effective cancer treatment and monitoring.
- Subjective size estimation by radiologists can be prone to variability.
Purpose of the Study:
- To compare the accuracy of subjective size estimation versus computerised volume quantification for tumor-like phantoms in spiral CT.
- To evaluate observer agreement in both methods.
Main Methods:
- Eight tubular phantoms with irregular silicone inclusions (8.7-31.6 ml) were imaged using spiral CT.
- Radiologists and residents performed subjective size estimation and computerised volume quantification.
- Accuracy and observer agreement (kappa values) were calculated for both methods.
Main Results:
- Subjective size estimation had 51% accuracy (mean weighted kappa, Kqw=0.73), improved slightly with multiple observers.
- Computerised volume quantification achieved 70% accuracy (mean Kqw=0.85).
- Intra-observer agreement was better than inter-observer agreement for subjective estimation.
Conclusions:
- Computerised volume quantification is more reliable than subjective estimation for irregular tumor volumes in CT.
- Repeated subjective assessments by the same observer or using mean/median estimates from multiple observers can improve accuracy.
- Computer-based methods are recommended, especially for less experienced readers, with minimal accuracy loss even when measuring every fourth slice.