Naohiro Sakashita, Presenter: Nothing to Disclose

Yumi Onishi, Abstract Co-Author: Nothing to Disclose

Chie Miyake, Abstract Co-Author: Nothing to Disclose

Sachi Katayama, Abstract Co-Author: Nothing to Disclose

Kenji Ugusa, Abstract Co-Author: Nothing to Disclose

Shodayu Takashima, Abstract Co-Author: Nothing to Disclose

We studied the most useful imaging factor in discriminating two categories of lung neoplasm of replacement growth pattern.

We evaluated thin-section CT findings (presence or absence of lobulation, spiculation, air bronchogram of lesion and nodule pattern (solid and semi-solid or ground glass opacity), standard uptake values (SUVs) on FDG-PET, and nodule volume and the ratio of solid part to nodule volume that were semi-automatically measured using originally developed software on CT images of 64 solitary pulmonary nodules of less than 2cm in 60 consecutive patients (24 men and 36 women with a mean age of 65 years). We performed stepwise logistic modeling using all the CT findings, 3D volumetric data, and PET SUVs as independent variables to estimate the statistically significant factors to discriminate adenocarcinoma with replacement growth pattern (lesions with poor prognosis, n=42) from the other neoplasm, including atypical adenomatous hyperplasia (n=3) and localized bronchioloalveolar cell carcinoma (lesions with good prognosis, n=19). Then, optimal diagnostic accuracy was calculated with the most significant factor.

The prevalence of pleural indentation (p=0.01) and solid nodule pattern (p<0.001), PET SUVs (p<0.001), nodule volume (p=0.02), and the ratio of solid part to nodule volume (p<0.001) were significantly greater in adenocarcinoma with replacement growth pattern than in the other lesions. Logistic regression modeling revealed that the ratio was the only significant factor (p=0.04) for discriminating two categories. Using the ratio for predicting adenocarcinoma with replacement growth pattern, 0.23 or more showed the highest accuracy of 84% with 91% sensitivity and 73% specificity.

The ratio that was semi-automatically measured using originally developed software was useful to discriminate two categories of lung neoplasm of replacement growth pattern.

Prognosis of patients with small lung neoplasm of replacement growth pattern may be predicted with 3D volumetric analysis on thin-section CT.

http://media.rsna.org/media/abstract/2010/9005101/wjmp192_9005101_RSNA_presentation_ver6.pptx

Sakashita, N, Onishi, Y, Miyake, C, Katayama, S, Ugusa, K, Takashima, S, Discrimination of Lung Neoplasm of Replacement Growth Pattern: Comparison of Computer-aided 3D Volumetric Analysis with Thin-section CT Findings and FDG-PET.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9005101.html