Su Yeon Ahn, Presenter: Nothing to Disclose

Chang Min Park MD, PhD, Abstract Co-Author: Nothing to Disclose

Sang-Joon Park MD, Abstract Co-Author: Nothing to Disclose

Hak Jae Kim, Abstract Co-Author: Nothing to Disclose

Chang Hoon Song, Abstract Co-Author: Nothing to Disclose

Jin Mo Goo MD, PhD, Abstract Co-Author: Research Grant, Guerbet SA

To investigate whether CT texture parameters of primary tumors are associated with progression-free survival (PFS) and overall survival (OS) in non-small cell lung cancer (NSCLC) patients undergoing definitive concomitant chemoradiotherapy (CCRT).

Our institutional review board approved this retrospective study with waiver of patients’ informed consents. From January 2006 to December 2009, 72 patients with non-operable NSCLCs (stage IIIA, 31; stage IIIB, 40; stage IV, 1) underwent definitive CCRT at our institution. For all patients, CT texture parameters of primary tumors including entropy, homogeneity, skewness, kurtosis, mean Hounsfield unit (HU), standard deviation (SD) and volume were extracted from contrast-enhanced chest CT taken prior to CCRT using an in-house software program. Thereafter, each parameter was dichotomized based on optimal cutoff values calculated from receiver-operating characteristics curve analysis. PFS and OS were compared between the dichotomized subgroups via Kaplan-Meier analyses with log rank test. Multivariate Cox regression analyses were performed to determine significant prognostic factors for survival.

Median OS and PFS were 23 months (range, 2-86 months) and 10 months (range, 8-12 months), respectively. There were no significant differences in OS and PFS according to tumor stage. As for PFS, homogeneity (p=0.019) and kurtosis (p=0.044) were significantly associated with PFS on univariate analysis. Multivariate Cox regression analysis revealed that homogeneity (Adjusted hazard ratio (HR)=2.33, p=0.014) and kurtosis (Adjusted HR=2.76, p=0.026) were both significant independent predictors. As for OS, homogeneity (p=0.018), skewness (p=0.001) and kurtosis (p=0.036) were significantly associated with OS. Multivariate Cox regression analysis revealed that homogeneity (Adjusted HR=5.49, p =0.002), skewness (Adjusted HR=2.52, p=0.004), mean HU (Adjusted HR=5.19, p=0.001), and SD (Adjusted HR=2.24, p=0.03) were independent significant predictors.

CT texture features were associated with PFS and OS in NSCLC patients undergoing definitive CCRT and may potentially be utilized as prognostic biomarkers. 

Texture features of pretreatment contrast enhanced CT images in NSCLC patients may potentially be utilized as prognostic biomarkers of PFS and OS.

Ahn, S, Park, C, Park, S, Kim, H, Song, C, Goo, J, Prognostic Value of CT Texture Features in Non-small Cell Lung Cancers Treated with Definitive Concomitant Chemoradiotherapy.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14010523.html