RSNA 2017 CT Preview

Study refines lung hamartoma diagnosis on CT

By Abraham Kim, staff writer

November 2, 2017 --

Monday, November 27 | 11:20 a.m.-11:30 a.m. | SSC03-06 | Room S504CD
A team of researchers from Israel is proposing an increase in the current threshold for identifying fat in suspected cases of lung hamartoma to improve accuracy.

The present method for detecting lung hamartomas involves spotting popcorn calcifications or fat within a threshold of -40 to -120 Hounsfield units (HU). The decades-old technique often leads to incorrectly diagnosed lung hamartomas, according to the researchers from Chaim Sheba Medical Center in Ramat Gan.

"This inability to diagnose most hamartomas with imaging resulted in too many unnecessary surgical interventions," Dr. Michal Eifer told

A lack of updates to the criteria pushed the researchers to apply different Hounsfield unit thresholds to hamartomas and malignant nodules on CT. They found that the most accurate HU range for diagnosing lung hamartomas was -40 HU to -20 HU, which increased sensitivity from 9.3% to 18.6% while maintaining a positive predictive value and specificity of 100%.

"We hope broadening the threshold this way will lead to better diagnosis and decrease the amount of patients who get unnecessary surgery for hamartomas," Eifer said.