Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal cancers in the United States with a five-year survival rate of 7.2% for all stages. Although surgical resection is the only curative treatment, currently we are unable to differentiate between resectable patients with occult metastatic disease from those with potentially curable disease. Identification of patients with poor prognosis via early classification would help in initial management including the use of neoadjuvant chemotherapy or radiation, or in the choice of postoperative adjuvant therapy. PDAC ranges in appearance from homogeneously isoattenuating masses to heterogeneously hypovascular tumors on CT images; hence, we hypothesize that heterogeneity reflects underlying differences at the histologic or genetic level and will therefore correlate with patient outcome. We quantify heterogeneity of PDAC with texture analysis to predict 2-year survival. Using fuzzy minimum-redundancy maximum-relevance feature selection and a naive Bayes classifier, the proposed features achieve an area under receiver operating characteristic curve (AUC) of 0.90 and accuracy (Ac) of 82.86% with the leave-one-image-out technique and an AUC of 0.80 and Ac of 75.0% with three-fold cross-validation. We conclude that texture analysis can be used to quantify heterogeneity in CT images to accurately predict 2-year survival in patients with pancreatic cancer. From these data, we infer differences in the biological evolution of pancreatic cancer subtypes measurable in imaging and identify opportunities for optimized patient selection for therapy.

No abstract

Purpose To identify the ability of computer-derived three-dimensional (3D) computed tomographic (CT) segmentation techniques to help differentiate lung adenocarcinoma subtypes. Materials and Methods This study had institutional research board approval and was HIPAA compliant. Pathologically classified resected lung adenocarcinomas (n = 41) with thin-section CT data were identified. Two readers independently placed over-inclusive volumes around nodules from which automated computer measurements were generated: mass (total mass) and volume (total volume) of the nodule and of any solid portion, in addition to the solid percentage of the nodule volume (percentage solid volume) or mass (percentage solid mass). Interobserver agreement and differences in measurements among pathologic entities were evaluated by using t tests. A multinomial logistic regression model was used to differentiate the probability of three diagnoses: invasive non-lepidic-predominant adenocarcinoma (INV), lepidic-predominant adenocarcinoma (LPA), and adenocarcinoma in situ (AIS)/minimally invasive adenocarcinoma (MIA). Results Mean percentage solid volume of INV was 35.4% (95% confidence interval [CI]: 26.2%, 44.5%)-higher than the 14.5% (95% CI: 10.3%, 18.7%) for LPA (P = .002). Mean percentage solid volume of AIS/MIA was 8.2% (95% CI: 2.7%, 13.7%) and had a trend toward being lower than that for LPA (P = .051). Accuracy of the model based on total volume and percentage solid volume was 73.2%; accuracy of the model based on total mass and percentage solid mass was 75.6%. Conclusion Computer-assisted 3D measurement of nodules at CT had good reproducibility and helped differentiate among subtypes of lung adenocarcinoma. (©) RSNA, 2016.