Improvement in Visibility of Simulated Lung Nodules on Computed Radiography (CR) Chest Images by Use of Temporal Subtraction Technique

Purpose: To compare the performance of various types of unsharp mask filter applied in storage-phosphor-based computed radiography (SR), and to improve the detection of faint nodules of the lung cancer type. Material and Methods: A total of 120 SR radiographs were obtained by means of an anthropomorphous chest phantom and a combination of 3 types of small simulated nodule (5-mm sphere-shaped, and 5-mm and 10-mm hemisphere-shaped) placed on the phantom's surface. Eight combinations of nodule site were selected from 16 predetermined chosen sites, and 5 types of parameter were used for unsharp mask filtering. Eight observers evaluated the images, and the detectability of the lung nodules was evaluated from the images by a ROC analysis. Results: The visibility of the 10-mm hemispherical nodules was nearly equivalent at each site when 5 types of unsharp mask filter were used. The detection of the 5-mm nodules with mid-frequency suppressing and very-low-frequency enhancing filters was better than with a conventional (department standard) mid-frequency enhancing filter. Conclusion: Mid-frequency suppressing versions of the filter helped to demonstrate faint nodular opacity, which is often shown by early bronchogenic carcinoma. This filter could replace conventional filters in the detection of lung nodules.

Download Full-text

Clinical usefulness of temporal subtraction technique for detection of interval changes on digital chest radiographs

CARS 2002 Computer Assisted Radiology and Surgery ◽

10.1007/978-3-642-56168-9_115 ◽

2002 ◽

pp. 689-694

Author(s):

S. Katsuragawa ◽

T. Uozumi ◽

S. Kakeda ◽

H. Watanabe ◽

H. Nakata ◽

...

Keyword(s):

Chest Radiographs ◽

Clinical Usefulness ◽

Subtraction Technique ◽

Temporal Subtraction ◽

Digital Chest

Download Full-text

Reduction of FPs for Lung Nodules in MDCT by Use of Temporal Subtraction with Voxel-Matching Technique

Advances in Neuro-Information Processing - Lecture Notes in Computer Science ◽

10.1007/978-3-642-02490-0_62 ◽

2009 ◽

pp. 504-512 ◽

Cited By ~ 3

Author(s):

Yoshinori Itai ◽

Hyoungseop Kim ◽

Seiji Ishikawa ◽

Shigehiko Katsuragawa ◽

Kunio Doi

Keyword(s):

Lung Nodules ◽

Temporal Subtraction ◽

Matching Technique

Download Full-text

Detection of Lung Nodules in Thoracic MDCT Images Based on Temporal Changes from Previous and Current Images

Journal of Advanced Computational Intelligence and Intelligent Informatics ◽

10.20965/jaciii.2011.p0707 ◽

2011 ◽

Vol 15 (6) ◽

pp. 707-713 ◽

Cited By ~ 4

Author(s):

Shinya Maeda ◽

◽

Yasuyuki Tomiyama ◽

Hyoungseop Kim ◽

Noriaki Miyake ◽

...

Keyword(s):

Neural Network ◽

Temporal Change ◽

Lung Nodules ◽

Detection Accuracy ◽

Multidetector Row Computed Tomography ◽

Neural Network Classifier ◽

Shape Features ◽

Temporal Subtraction ◽

Nodule Detection ◽

Positive Results

Temporal subtraction enhances temporal change by subtracting images captured at different times. Medical images captured currently (current images) and in previous examination (previous images) are subtracted to enhance new lesions and temporal change in existing lesion shadows. Temporal subtraction using chest MultiDetector-Row Computed Tomography (MDCT) images and currently being developed is to be applied to nodule detection in pulmonary regions. Nodule detection using conventional temporal subtraction, however, yields many false-positive results for those 20 mm or less in diameter, requiring improvement. We discuss improvements in nodule detection accuracy using temporal subtraction, first extracting rough nodules from temporal subtraction images as candidate shadows. Features are then acquired from current, previous, and temporal subtraction images. We use intensity features in previous images and shape features in the current images and in features used in conventional methods. Using acquired features, we build a neural network classifier, then extract final pulmonary candidates in unknown shadows.

Download Full-text