초록접수 현황
| 14F-061 | 구연 발표 |
Which is better way for Measuring Proportions of Ground Glass Opacity (GGO) in Thin-section Chest Computed Tomography – diameter, Width or Volume?
이창영¹, 홍새롬², 유우식¹, 정희석¹, 이진구¹, 김대준¹, 정경영¹
연세대학교 의과대학 세브란스병원 흉부외과학교실¹, 연세대학교 의과대학 세브란스병원 영상의학교실²
Background : Proportions of ground glass opacity (GGO) in thin section chest computed tomography (CT) has been an emerging prognostic factor for lung adenocarcinoma. We measured proportions of GGO one (1D, diameter), two (2D, width) and three dimensionally (3D, volume) and investigated which method is better way to predict oncologic outcomes in surgically resected stage I adenocarcinoma.
Methods : One hundred fifty patients who underwent surgically complete resection for pathologic stage I adenocarcinoma between January 2004 and December 2010 were included. Data of thin section (1mm thickness) chest CT for each patient were acquired and all images were evaluated on an off-line workstation. Diameter, width and volume of the total nodule, GGO portion and solid portion were calculated after 3D segmentation of the nodule. We reviewed clinico-pathological records and evaluated their relation to least invasiveness disease (LID), which was defined as no lymphovascular invasion, no pleural invasion, no perineural invasion and no recurrence.
Results : There were 52 cases in LID group and 98 cases in invasive disease (ID) group. Clinico-radiological parameters are listed and compared between two groups in Table 1. Kaplan-Meier survival curve showed GGO 2D dominance (less than 50% of solid portion in width) made cancer specific survival differences with more statistical significance (p=0.002) compared to GGO 1D dominance (less than 50% of solid portion in diameter, p=0.026) and GGO 3D dominance (less than 50% of solid portion in volume, p=0.012). The specificity and sensitivity of the diagnosis for LID was as follows; 100% and 44.1% in GGO 1D dominance, 94.1% and 49.5% in GGO 2D dominance, 90.5% and 52.9% in GGO 3D dominance.
Conclusion : In this study, proportion of GGO measured two-dimensionally is statistically most significant prognostic factor in terms of cancer specific survival and least invasiveness. However, proportion of GGO measured one-dimensionally had highest specificity for least invasiveness. Herein, GGO 1D proportion might be an appropriate parameter to select optimal candidates for limited resection.
Methods : One hundred fifty patients who underwent surgically complete resection for pathologic stage I adenocarcinoma between January 2004 and December 2010 were included. Data of thin section (1mm thickness) chest CT for each patient were acquired and all images were evaluated on an off-line workstation. Diameter, width and volume of the total nodule, GGO portion and solid portion were calculated after 3D segmentation of the nodule. We reviewed clinico-pathological records and evaluated their relation to least invasiveness disease (LID), which was defined as no lymphovascular invasion, no pleural invasion, no perineural invasion and no recurrence.
Results : There were 52 cases in LID group and 98 cases in invasive disease (ID) group. Clinico-radiological parameters are listed and compared between two groups in Table 1. Kaplan-Meier survival curve showed GGO 2D dominance (less than 50% of solid portion in width) made cancer specific survival differences with more statistical significance (p=0.002) compared to GGO 1D dominance (less than 50% of solid portion in diameter, p=0.026) and GGO 3D dominance (less than 50% of solid portion in volume, p=0.012). The specificity and sensitivity of the diagnosis for LID was as follows; 100% and 44.1% in GGO 1D dominance, 94.1% and 49.5% in GGO 2D dominance, 90.5% and 52.9% in GGO 3D dominance.
Conclusion : In this study, proportion of GGO measured two-dimensionally is statistically most significant prognostic factor in terms of cancer specific survival and least invasiveness. However, proportion of GGO measured one-dimensionally had highest specificity for least invasiveness. Herein, GGO 1D proportion might be an appropriate parameter to select optimal candidates for limited resection.
책임저자: 정경영
연세대학교 의과대학 세브란스병원 흉부외과학교실
연락처 : 이창영, Tel: 02-2228-2140 , E-mail : cyleecs@yuhs.ac
