Improved visualization of artificial pulmonary nodules with a new subvolume rendering technique

Area of Science:

• Medical Imaging
• Radiology
• Computer-Aided Diagnosis

Background:

• Detecting small pulmonary nodules in multidetector computed tomography (MDCT) datasets is challenging.
• Improved visualization techniques are needed to enhance reader sensitivity and efficiency.
• A novel subvolume rendering technique, 'Softslice,' was developed for improved nodule differentiation.

Purpose of the Study:

• To compare the efficacy of the Softslice subvolume rendering technique against Maximum Intensity Projection (MIP) rendering.
• To evaluate visualization of artificial pulmonary nodules within MDCT datasets using these techniques.

Main Methods:

• Virtual pulmonary nodules (2-7 mm diameter) with signal intensity similar to vessels were inserted into MDCT datasets.
• Renderings included MIP (10/20 mm thickness) and Softslice (20/30 mm thickness) with radiologists using free scrolling.
• Readings were performed with 1-, 4-, or 8-mm increments between subvolume positions.

Main Results:

• No significant difference in nodule detection was observed with 1- or 4-mm increments between renderings.
• At an 8-mm increment, 30-mm Softslice rendering significantly outperformed MIP (10/20 mm) and 20-mm Softslice (P < 0.021).
• A marked reduction in nodule detection occurred for 10-mm MIP when increments increased from 4 mm to 8 mm (P = 0.018).

Conclusions:

• The Softslice rendering technique shows potential for improved pulmonary nodule detection in MDCT.
• Softslice may offer more efficient reading compared to standard MIP subvolume techniques.
• Nodule detection rates, particularly with 10-mm MIP, are significantly influenced by subvolume increment distance.