Interstitial lung disease (ILD) refers to a group of various abnormal inflammations of lung tissues and early diagnosis of these disease patterns is crucial for the treatment. Yet it is difficult to make an accurate diagnosis due to the similarity among the clinical manifestations of these diseases. In order to assist the radiologists, computer-aided diagnosis systems have been developed. Besides, the potential of deep convolutional neural networks (CNNs) is also expected to exert on the medical image analysis in recent years. In this paper, we design a new deep convolutional neural network (CNN) architecture to achieve the classification task of ILD patterns. Furthermore, we also propose a novel two-stage transfer learning (TSTL) method to deal with the problem of the lack of training data, which leverages the knowledge learned from sufficient textural source data and auxiliary unlabeled lung CT data to the target domain. We adopt the unsupervised manner to learn the unlabeled data, by which the objective function composed of the prediction confidence and mutual information are optimized. The experimental results show that our proposed CNN architecture achieves desirable performance and outperforms most of the state-of-the-art ones. The comparative analysis demonstrates the promising feasibility and advantages of the proposed two-stage transfer learning strategy as well as the potential of the knowledge learning from lung CT data. Graphical Abstract The framework of the proposed two-stage transfer learning method.

Objective To investigate the characteristics of breast neoplasms on contrast-enhanced ultrasonography(CEUS) and its clinical value.Methods Two hundred and twenty-five patients with breast masses unable to be diagnosed by conventional ultrasonography were examined with CEUS.The characteristics of these masses on CEUS were analyzed and compared with the results of pathology examination.Results The typical features of breast cancers on CEUS were enlarged maximum diameter of the lesions on CEUS compared to pre-contrast ( P ＜0.05),irregular shapes,heterogeneous distribution of contrast enhancement with perfusion defect or local retention of contrast signals,tortuous,massive or penetrating vessels rapidly entering and exporting from the lesions.The sensitivity and specificity of perfusion defect for breast cancer on CEUS were 89.0％ and 91.8％,respectively; the sensitivity and specificity of local retention of contrast signals for breast cancer on CEUS were 93.4％ and 92.5％,respectively.Conclusions It is valuable for CEUS in the diagnosis and differential diagnosis of breast neoplasms clinically.

To explore the clinical value and characteristics of contrast-enhanced ultrasound(CEUS) in diagnosis of gallbladder carcinoma. 384 patients with benign and malignant gallbladder disease were examined by CEUS. The characteristics of CEUS were analyzed and compared with pathological examination. (1) The CEUS patterns of gallbladder carcinomas showed quick and heterogeneous hyper-enhancement at the early arterial phases. The CEUS shape of the gallbladder carcinomas were irregular. The wall of gallbladder was irregular thicken and interrupted by the mass. The basement of lesions were wide and connected with the gallbladder wall. Almost all the gallbladder carcinomas showed washout from hyper-enhancement to hypo-enhancement quickly after contrast agent administration. (2) It was significant different between benign and malignant gallbladder diseases of CEUS characteristics(p ＜ 0.05). (3) Compared with pathological examination, the sensitivity, specificity and accuracy of CEUS in gallbladder carcinomas diagnosis was 96.6%(28/29), 99.4%(353/355), 99.2%(381/384), respectively. CEUS has an important clinical value in diagnosis of gallbladder carcinoma.

Background/Aims: We evaluated the long-term efficacy of the combination of transcatheter arterial chemoembolization (TACE) using cisplatin-lipiodol suspension, transultrasonic portal vein chemoembolization (SPVE), radiofrequency ablation (RF), percutaneous ethanol injection (PEI) for treatment of advanced small hepatocellular carcinoma (HCC).Methodology: A total of three hundred and eighteen patients with HCC were enrolled in this study. According to the blood supply characteristics to the tumor, individual combined therapy models were adopted: one hundred and fifty-nine patients with HCC less than 5cm were treated with a combination of RF and PEI (RF/PEI group) and one hundred and one patients with HCC greater than 5cm were treated with a combination of TACE, RF and PEI (TACE/RF/PEI group). One hundred and eleven HCC nodules confirmed to be hypervascular by color Doppler flow imaging were treated with a combination of TACE, RF, SPVE and PEI (TACE/RF/SPVE/PEI group).Results: The combination treatment of RF and PEI (RF/PEI group), the TACE/RF/PEI group, TACE/RF/SPVE/PEI group, the 1-year survival rates and the 3-year survival rates were 97.3% and 82.4%; 73.5% and 44.9%; 74.1% and 37.9%, respectively; The vanishing rate of blood flow around and within the tumor, the tumor size decrease rate, AFP transformed, to negative rate, were significantly raised compared to those in the TACE treatment only group.Conclusions: The individual combined therapy models combination of TACE, PEI, SPVE, RF appears to prolong survival, compared with one treatment alone (TACE). This combination therapy method is an effective way for treating HCC, and color Doppler can provide important information to verify the therapeutic effects.