Premium
Pulmonary hamartomas: CT pixel analysis for fat attenuation using radiologic–pathologic correlation
Author(s) -
Gleeson Tadhg,
Thiessen Rennae,
Hannigan Ailish,
Murphy Darra,
English John C.,
Mayo John R.
Publication year - 2013
Publication title -
journal of medical imaging and radiation oncology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.31
H-Index - 43
eISSN - 1754-9485
pISSN - 1754-9477
DOI - 10.1111/1754-9485.12083
Subject(s) - medicine , hamartoma , lesion , attenuation , radiology , pixel , histology , nuclear medicine , pathology , physics , computer science , optics , computer vision
To assess the accuracy of CT pixel analysis for fat attenuation in pulmonary hamartomas. Methods Retrospective review identified 32 patients in three separate groups; pathologically proven hamartoma ( n = 11), hamartoma diagnosed on imaging ( n = 9) and a control group ( n = 14) of pathology‐proven non‐hamartomatous smoothly marginated solitary pulmonary nodules. All lesions were assessed using: visual assessment for fat, pixel analysis of the inner 2/3rds and mean attenuation of the entire lesion, using an internal reference for fat. Fat percentages on CT and at histology were compared. Results Visual assessment for macroscopic fat was the most reliable method for diagnosing pulmonary hamartoma. Combining percentage of fat‐attenuation pixels in the inner 2/3rds of the lesion improved specificity to 100%. Mean attenuation or pixel analysis in isolation were not helpful in lesional characterization. Conclusion Combining percentage fat‐attenuating pixels in the inner 2/3rds with visual assessment for macroscopic fat improves specificity for diagnosing pulmonary hamartomas.