With logistic regression analysis and C-statistics, we found that the differentiating performance of our logistic model using the clinical and thin-section CT features as well as the texture features of PSNs was excellent, and by adding texture analysis of whole PSNs to the classical clinical and thin-section CT features analysis, we were able to significantly increase the differentiating performance between transient and persistent PSNs. We believe that unnecessary procedures such as additional diagnostic tests or invasive procedures may be obviated using this additional analysis method although follow-up CT might not be able to be skipped for definite confirmation of a lesion’s persistency. Our study had several limitations. First, our study was of retrospective design, and was performed on 
a relatively small number of patients. Second, we retrospectively searched for individuals with pulmonary PSNs identified at low dose thinsection CT using the electronic medical records and Senegenin radiology information system of our hospital. Thus, there is a possibility that nodules might have been Cryptochlorogenic-acid unreported and therefore missed with this search method. Third, the time interval between an initial CT showing PSNs and follow-up CT used for determination of lesions’ transiency was not uniform and varied widely. Fourth, the texture features in this study were derived from lesions, which can be significantly influenced by observers’ subjective trend or bias. However, manual segmentation may be the gold standard for lesion segmentation particularly in the case of GGNs as their margins are usually indistinct or unclear from the normal lung parenchyma and thus difficult to segment automatically. Nonetheless we believe that a reliable and robust automatic boundary extraction method should be further developed to solve the variability issue. In conclusion, texture analysis of PSNs has the potential to improve the differentiation of transient from persistent PSNs when used in addition to clinical and CT features analysis. Most cancer patients show disease recurrence that rapidly progresses to the advanced stages with vascular invasion and their 5-year relative survival rate is only 7%. Therefore new therapies and new detection methods for this aggressive disease are extremely needed. Angiogenesis is required for invasive tumor growth and metastasis, so it plays an important role in the control of cancer progression. The rapid growth of the tumor needs a large amount of nutrients and oxygen, which prompted the growth of blood vessels. Moreover, HCC tumors depend on a rich blood supply, therefore, inhibition of angiogenesis has constituted a crucial point in liver cancer therapy. Preclinical studies have shown that endostatin could shrink existing tumor blood vessels effectively. Endostar is a novel recombinant human endostain expressed and purified in Escherichia coli with a modified N-terminal, it has been shown to inhibit endothelial cell proliferation, migration, and vessel formation. Based on systemic preclinical studies and clinical trials, the State Food and Drug Administration of China approved the regimen for the treatment of non-small-cell lung cancer in 2005. In this study, we explored the antitumor effects of Endostar on liver cancer in vivo. The traditional approach for anti-neoplastic research such as histopathological analysis is accurate but time-consuming and cannot provide 3-dimensional structural information. Furthermore, some anti-angiogenic agents usually inhibit tumor progression rather than tumor volume shrinkage. Therefore, the evaluation of therapeutic response only through tumor volume measurement is no longer comprehensive, and it is an urgent need to develop a more sensitive and effective detection method.