NSCLC tumor shrinkage prediction using quantitative image features. (April 2016)

• Record Type: Journal Article
• Authors: Hunter, Luke A.; Chen, Yi Pei; Zhang, Lifei; Matney, Jason E.; Choi, Haesun; Kry, Stephen F.; Martel, Mary K.; Stingo, Francesco; Liao, Zhongxing; Gomez, Daniel; Yang, Jinzhong; Court, Laurence E.
• Is Part Of: Computerized medical imaging and graphics. Volume 49(2016)
• Page Start: 29
• Page End: 36
• Publication Date: 2016-04
• Subjects: Quantitative image feature -- texture -- tumor shrinkage -- prediction -- NSCLC; Diagnostic imaging -- Periodicals; Imaging systems in medicine -- Periodicals; Diagnosis, Radioscopic -- Data processing -- Periodicals; Diagnostic Imaging -- Periodicals; Imagerie pour le diagnostic -- Périodiques; Diagnostic imaging; Periodicals; Electronic journals; Electronic journals; 616.0754
• DOI: 10.1016/j.compmedimag.2015.11.004
• Languages: English
• ISSNs: 0895-6111

Abstract

Highlights: Lung tumors shrink during radiotherapy, with much variation between patients. Pre-treatment CT images can be used to predict tumor shrinkage after treatment. Potential uses include identifying patients who will benefit from adaptive radiation therapy. Abstract: The objective of this study was to develop a quantitative image feature model to predict non-small cell lung cancer (NSCLC) volume shrinkage from pre-treatment CT images. 64 stage II-IIIB NSCLC patients with similar treatments were all imaged using the same CT scanner and protocol. For each patient, the planning gross tumor volume (GTV) was deformed onto the week 6 treatment image, and tumor shrinkage was quantified as the deformed GTV volume divided by the planning GTV volume. Geometric, intensity histogram, absolute gradient image, co-occurrence matrix, and run-length matrix image features were extracted from each planning GTV. Prediction models were generated using principal component regression with simulated annealing subset selection. Performance was quantified using the mean squared error (MSE) between the predicted and observed tumor shrinkages. Permutation tests were used to validate the results. The optimal prediction model gave a strong correlation between the observed and predicted tumor shrinkages with r = 0.81 and MSE = 8.60 × 10 −3 . Compared to predictions based on the mean population shrinkage this resulted in a 2.92 fold reduction in MSE. In conclusion, this study indicated that quantitative image features extracted from existing pre-treatment CT images can successfully predict tumor shrinkage and provide additional information for clinical decisions regarding patient risk stratification, treatment, and prognosis.