Roberto Lo Gullo MD, Presenter: Nothing to Disclose

Mannudeep K. S. Kalra MD, Abstract Co-Author: Nothing to Disclose

Alexi Otrakji MD, Abstract Co-Author: Nothing to Disclose

Lecia Sequist, Abstract Co-Author: Nothing to Disclose

Jo-Anne O. Shepard MD, Abstract Co-Author: Consultant, Agfa-Gevaert Group

Subba Rao Digumarthy MD, Abstract Co-Author: Nothing to Disclose

Zofia Piotrowska, Abstract Co-Author: Nothing to Disclose

To assess the difference in CT texture, among histological subtypes of primary non small cell lung cancer (NSCLC) and to identify the most helpful metrics of CT histogram analysis.

This IRB approved study included 94 consecutive patients with biopsy proven primary NSCLC (45M; 49F; mean age 65 years; range26-86). The subtypes included 69 adenocarcinoma ( 27 EGFR mutants and 42 nonmutants) and 25 squamous cell carcinoma. CT examinations of patients were deidentified and exported to an offline server with a commercial texture analysis software (TexRAD limited, UK). The image analysis was performed on a single image with manually drawn large region of interest on dominant lung lesions after excluding areas of cavitation and contrast and beam hardening artifacts. The analyzed features included mean HU values, percent positive pixels (PPP), mean value of positive pixels (MPP), standard deviation (SD), normalized SD, skewness, kurtosis, and entropy. Data were analyzed with non parametric ANOVA Kruskal- Wallis test and T-test with Welch correction.

Significant differences in CT texture features were found between squamous and adenocarcinoma for entropy (p-value=0.002) and normalized SD (p=0.0073). There is significant difference in kurtosis (p=0.01) between EGFR mutation positive adencarcinoma and non EGFR mutant adenocarcinoma. There was no significant difference for skewnes and mean values for positive pixels.

Entropy, normalized SD and kurtosis are the most useful texture parameters to differentiate histological subtypes of NSCLC . Squamous and adenocarcinoma (EGFR mutant and nonmutant subtypes) have disntict CT texture.

Image analysis using CT histogram allows tissue characterization and has potential clinical applications to choose therapy and supplement other diagnostic tests.

Lo Gullo, R, Kalra, M, Otrakji, A, Sequist, L, Shepard, J, Digumarthy, S, Piotrowska, Z, Non Small Cell Lung Cancer: Efficacy of CT Texture Analysis in Assessment of Histological Subtype.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14008214.html