Related Experiment Videos
Double-contrast barium enema studies: effect of multiple reading on perception error
J B Markus1, S Somers, B P O'Malley
1Department of Radiology, McMaster University Medical Centre, Hamilton, Ont., Canada.
Radiology
|April 1, 1990
Summary
Multiple radiologists reviewing double-contrast barium enema (DCBE) images significantly improved lesion detection rates. Combining readings from two or three radiologists enhanced sensitivity for detecting radiologically visible lesions, reducing errors in colon cancer screening.
Area of Science:
- Radiology
- Gastroenterology
- Medical Imaging
Background:
- Double-contrast barium enema (DCBE) is a common imaging technique for evaluating the colon.
- Accurate detection of colonic lesions is crucial for early diagnosis and treatment.
- Perception errors by single readers can lead to missed diagnoses in DCBE examinations.
Purpose of the Study:
- To evaluate the impact of multiple radiologist readings on the sensitivity and specificity of double-contrast barium enema (DCBE) for detecting colonic lesions.
- To determine if combining interpretations from multiple radiologists can reduce false-negative errors.
Main Methods:
- Sixty double-contrast barium enema (DCBE) examinations were independently reviewed by three radiologists.
- Lesion detection rates were calculated for single, double, and triple reading combinations.
- Sensitivity and specificity were assessed for each reading scenario.
Main Results:
- Single radiologists detected an average of 70.2% of visible lesions.
- Double reading combinations increased detection to 83.3%, and triple reading to 89.0%.
- These increases in sensitivity were statistically significant but accompanied by slight decreases in specificity.
Conclusions:
- False-negative perception error is a significant issue in interpreting DCBE examinations.
- Multiple radiologist readings effectively reduce interpretation errors and improve lesion detection sensitivity.
- This strategy enhances the diagnostic accuracy of DCBE for colonic lesion detection.