Purpose
Conventional chest X-ray is still the most common radiological diagnostic imaging.
CAD-application might be benefitial to improve reliability of nodule detection.
Nevertheless nodule detection at specific locations is still difficult,
thus a bone reducing software might be benefitial at these specific lesions.
Study aimed to analyze the diagnostic potential of bone reducing and nodule detecting CAD-software.
Methods and Materials
100 prone chest X-rays with histologically proven and CT-verified intrapulmonary nodule underwent a CAD-base bone reduction (Riverain,USA) and nodule detection.
Size groups were built: S1: <10mm (n=23); S2: <20mm (n=44); S3: <30mm (n=16); S4: <40mm (n=7); S5: >40mm (n=10).
Mean density and SD of the nodule and the surrounding with/without bone reduction were measured using a strict reliable protocol.
Contrast was calculated as the ratio nodule density/surrounding density.
Nodule detection was measured,
FP-values calculated and character of FP-markers analyzed.
Results
Mean density of the detection lesion varies size-dependent from 1649/1688 (1),
1857/1895 (2),
1863/1917(3),
1928/1996 (4) to 1859/1906 (5).
Contrast from lesion vs.
surrounding alters from 0,98; 0,98; 0,97; 0,96; 0,97 without CAD-application to 1,23; 1,34; 1,54; 1,60; 1,60 with CAD suggesting that there is a size dependent contrast enhancement induced by bone reduction.
Malignant lesions revealed an overall density of 1856 whereas benign lesions showed a higher density (1985).
Size-dependence was: 1633/1905 (1); 1883/2022 (2); 1847/2221 (3); 2037/1755 (4); 1906/1910 (5).
Location of the...
Conclusion
Bone reduction is a valuable tool to induce a better contrast nodule/surrounding and thus supports significantly the detection of intrapulmonary nodules.
This effect is running best at lesions 2-3cm in size but is also valuable at subtle lesions.
Malignant lesions of up to 2 cm are less dense vs.
benign lesions of same size,
software thus could allow a better discrimination of benign vs.
malignant lesions.
Best effect of applied software was measured at lesions with localization close to the pleura.
Detection rate of lung...
References
•Boo,
Diederick W.
de; Uffmann,
Martin; Weber,
Michael; Bipat,
Shandra; Boorsma,
Eelco F.; Scheerder,
Maeke J.
et al.
(2011): Computer-aided detection of small pulmonary nodules in chest radiographs: an observer study.
In Acad Radiol 18 (12),
pp.
1507–1514.•Chen,
Joseph J.; White,
Charles S.
(2008): Use of CAD to evaluate lung cancer on chest radiography.
In J Thorac Imaging 23 (2),
pp.
93–96.•Fraioli,
F.; Serra,
G.; Passariello,
R.
(2010): CAD (computed-aided detection) and CADx (computer aided diagnosis) systems in identifying and characterising lung nodules on chest...