[CT colonography: evaluation of two 3D algorithms in a screening population]

Area of Science:

• Radiology
• Medical Imaging
• Gastroenterology

Background:

• Computed Tomography Colonography (CTC) is an imaging technique for colorectal cancer screening.
• Effective visualization of CTC datasets is crucial for accurate polyp detection.
• Different 3D visualization methods, such as endoluminal view and dissection display, are available for CTC analysis.

Purpose of the Study:

• To compare the effectiveness of a dissection display versus an endoluminal display for CT colonography.
• To evaluate differences in polyp detection rates and evaluation time between the two visualization methods.
• To assess these visualization methods in a screening population.

Main Methods:

• A multicenter study involving 4 blinded readers evaluating CTC datasets from 42 patients with 55 confirmed polyps.
• Datasets were analyzed using two 3D visualization methods: endoluminal view and dissection display.
• Per-lesion sensitivity, per-patient sensitivity, per-patient specificity, and evaluation time were calculated.

Main Results:

• Overall per-lesion sensitivity ranged from 55% to 67% for the dissection display and 53% to 71% for the endoluminal view.
• Per-patient sensitivity was high for both methods, ranging from 80% to 95%.
• Per-patient specificity varied, with the endoluminal view showing slightly higher specificity in some readers (up to 77% vs. 68%).
• Experienced readers were significantly faster when using the dissection display.

Conclusions:

• A dissection display does not offer superior polyp detection rates compared to the endoluminal view when CTC datasets are prepared with stool and fluid tagging.
• The dissection display allows for significantly faster evaluation times for most experienced readers.
• The choice of visualization method may depend on balancing detection accuracy with efficiency in clinical practice.