Ryoji Iwamoto, Presenter: Nothing to Disclose

Kiminori Fujimoto MD, PhD, Abstract Co-Author: Nothing to Disclose

Junko Sadohara MD, Abstract Co-Author: Nothing to Disclose

Tatsuyuki Tonan MD, Abstract Co-Author: Nothing to Disclose

Shuji Nagata MD, Abstract Co-Author: Nothing to Disclose

Toshi Abe MD, Abstract Co-Author: Nothing to Disclose

Kenji Izuhara, Abstract Co-Author: Nothing to Disclose

Periostin, a matrix protein, was originally identified in osteoblasts. Periostin is believed to promote various types of tumor growth, in addition to migration and epithelial-mesenchymal transition of the malignant cells. This study aimed to compare the thin-section CT features with histological assessment of small lung invasive adenocarcinoma based on the immunohistochemistry stain using periostin to determine the CT features useful for the prediction of patient outcomes.

Seventy-seven consecutive patients with small lung invasive adenocarcinoma (< 3 cm in diameter) who underwent surgical resection from 2000 to 2009 were enrolled. The thin-section CT features were evaluated for 7 categories (size, contour, percentage of ground-glass attenuation scale [%GGA score], presence of air-bronchogram and/or bubble-like appearance, number of involved vessels, shape of pleural indentation, and number of pleural indentation). Two independent expert chest radiologists assessed these findings. Periostin expression was evaluated on the basis of strength and the range of staining. A univariate and multivariable analyses were performed using the Cox proportional hazards model.

There was substantial to perfect agreement between the observers for classification of each thin-section CT finding (kappa value = 0.73- 1.0). Histological WHO typing, pathologic lymphatic invasion, pathologic nodal metastasis, periostin expression, and %GGA score were statistically significant for recurrence (P = 0.001, 0.032, 0.003, < 0.0001, and 0.011, respectively) on univariate analysis. Multivariable analysis revealed that both the periostin high expression and the lower %GGA score were relative risk factors for recurrence (P = 0.026 and 0.012, respectively).

The results of this study suggested that periostin high-expression might be a significant predictor of postoperative recurrence. Moreover, preoperative thin-section CT findings (%GGA score) could also predict the likelihood of postoperative recurrence.

Evaluation of %GGA score before surgical resection and assessment of pathological periostin expression may be predictors for future tumor recurrence in patients with small invasive adenocarcinoma.

Iwamoto, R, Fujimoto, K, Sadohara, J, Tonan, T, Nagata, S, Abe, T, Izuhara, K, Small Invasive Lung Adenocarcinoma: Periostin Expression and Ground-glass Percentage Scale Predict Tumor Recurrence.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14009039.html