TYTsung-Lung Yang

Tsung-Lung Yang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Tsung-Lung Yang has a total of 38 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Multi-Scale Attention Convolutional Network for Masson Stained Bile Duct Segmentation from Liver Pathology Images

    by Chun-Han Su, Pau-Choo Chung, Sheng-Fung Lin, Hung-Wen Tsai, Tsung-Lung Yang, Yu-Chieh Su

    In clinical practice, the Ishak Score system would be adopted to perform the evaluation of the grading and staging of hepatitis according to whether portal areas have fibrous expansion, bridging with other portal areas, or bridging with central veins. Based on these staging criteria, it is necessary to identify portal areas and central veins when performing the Ishak Score staging. The bile ducts have variant types and are very difficult to be detected under a single magnification, hence pathologists must observe bile ducts at different magnifications to obtain sufficient information. This pathologic examinations in routine clinical practice, however, would result in the labor intensive and expensive examination process. Therefore, the automatic quantitative analysis for pathologic examinations has had an increased demand and attracted significant attention recently. A multi-scale inputs of attention convolutional network is proposed in this study to simulate pathologists’ examination procedure for observing bile ducts under different magnifications in liver biopsy. The proposed multi-scale attention network integrates cell-level information and adjacent structural feature information for bile duct segmentation. In addition, the attention mechanism of proposed model enables the network to focus the segmentation task on the input of high magnification, reducing the influence from low magnification input, but still helps to provide wider field of surrounding information. In comparison with existing models, including FCN, U-Net, SegNet, DeepLabv3 and DeepLabv3-plus, the experimental results demonstrated that the proposed model improved the segmentation performance on Masson bile duct segmentation task with 72.5% IOU and 84.1% F1-score.


  • Pathologic liver tumor detection using feature aligned multi-scale convolutional network

    by Tsung-Lung Yang, Hung-Wen Tsai, Wei-Che Huang, Jung-Chia Lin, Jia-Bin Liao, Nan-Haw Chow, Pau-Choo Chung


  • Open Access

    Feasibility of Breast MRI as the Primary Imaging Modality in a Large Asian Cohort

    by Fang-Ying Li, Alan Hollingsworth, Wai-Tak Lai, Tsung-Lung Yang, Liang-Juan Chen, Wei-Teng Wang, Jing-Lung Wang, Abraham N. Morse

    Contrast-enhanced MRI has repeatedly demonstrated significantly enhanced sensitivity compared to mammography and ultrasound in breast cancer detection. The purpose of this study was to evaluate the feasibility and outcomes of using breast MRI as the initial imaging study for screening and diagnosis. In this retrospective review of a cohort of 10,374 breast MRI scans in 7967 patients in Taitung County, Taiwan, a total of 5619 participants met inclusion criteria and were included in our analysis. We reviewed all biopsies that were performed subsequent to MRI studies in women (screening vs. diagnostic). The primary outcomes were false-positive (FP) biopsy rates and positive predictive value (PPV) of MRI - parameters that have historically been associated with performance that restricts more widespread use of MRI. False-positive rate based on benign biopsies (FPR-3) and the positive predictive value (PPV-3) were calculated. Without complementary imaging or follow-up to identify false negatives, the study of performance characteristics was limited to false positives and PPV. There were 351 benign biopsies generated by MRI out of the cohort of 5555 participants (5619 minus the malignant biopsies), generating a false-positive rate of 6.3%. Sixty-four patients out of 415 biopsies were malignant, generating a PPV-3 of 15.4%. In this Asian cohort, utilizing breast MRI as the initial study for screening and/or diagnosis appears to be limited more by practical considerations such as cost and patient flow efficiency than by feasibility based on performance characteristics. With well-established superior sensitivity, coupled with improved interpretive skills and techniques that allow for low false-positive rates, MRI should be further studied for its role as the primary imaging modality in breast screening and diagnosis.


  • Open Access

    Kinetic patterns of benign and malignant breast lesions on contrast enhanced digital mammogram

    by Jer-Shyung Huang, Huay-Ben Pan, Tsung-Lung Yang, Bao-Hui Hung, Chia-Ling Chiang, Meng-Yuan Tsai, Chen-Pin Chou

    To evaluate the kinetic patterns of benign and malignant breast lesions using contrast-enhanced digital mammogram (CEDM). Women with suspicious breast lesions on mammography or ultrasound were enrolled. Single-view mediolateral oblique (MLO) CEDM of an affected breast was acquired at 2, 3, 4, 7, and 10 min after injection of contrast agent. Three readers visually and semi-quantitatively analyzed the enhancement of suspicious lesions. The kinetic pattern of each lesion was classified as persistent, plateau, or washout over two time intervals, 2–4 min and 2–10 min, by comparing the signal intensity at the first time interval with that at the second. There were 73 malignant and 75 benign lesions in 148 patients (mean age: 52 years). Benign and malignant breast lesions showed the highest signal intensity at 3 min and 2 min, respectively. Average areas under receiver operating characteristic (ROC) curve for diagnostic accuracy based on lesion enhancement at different time points were 0.73 at 2 min, 0.72 at 3 min, 0.69 at 4 min, 0.67 at 7 min, and 0.64 at 10 min. Diagnostic performance was significantly better at 2, 3, and 4 min than at 7 and 10 min (all p < 0.05). A washout kinetic pattern was significantly associated with malignant lesions at 2–4 min and 2–10 min frames according to two of the three readers’ interpretations (all p ≤ 0.001). Applications of optimal time intervals and kinetic patterns show promise in differentiation of benign and malignant breast lesions on CEDM.


  • Impact of the COVID‐19 pandemic on the volume of mammography examinations in Southern Taiwan

    by Chen‐Pin Chou, Huay‐Ben Pan, Tsung‐Lung Yang, Chia‐Ling Chiang, Jer‐Shyung Huang, Meng‐Yuan Tsai


  • Hash Transformation and Machine Learning-Based Decision-Making Classifier Improved the Accuracy Rate of Automated Parkinson’s Disease Screening

    by Tsung-Lung Yang, Chia-Hung Lin, Wei-Ling Chen, Hsin-Yu Lin, Chen-San Su, Chih-Kuang Liang

    Digitalized hand-drawn pattern is a noninvasive and reproducible assistive manner to obtain hand actions and motions for evaluating functional tremors and upper-limb movement disorders. In this study, spirals and straight lines in polar coordinates are used to extract polar expression features such as the key parameters deviation (cm) and accumulation angle (rad). These parameters are quantitative manner to scale the variations of functional tremors in normal control subjects and patients with Parkinson’s disease (PD) and essential tremor (ET). However, difficulty arises in using nonlinear polar expression features in the two-dimensional feature space to separate normal control subjects from those with PD and ET. To solve the nonlinear separable classification problem, hash transformation is used to map polar expression features to a high-dimensional space using hash weighing function and modulo operation. Then, a machine learning method, such as the generalized regression neural network (GRNN), is implemented to train a decision-making classifier using the particle swarm optimization (PSO) algorithm for possible class assessment. With the enrolled data from 50 subjects, the fivefold cross validation, mean true positive, mean true negative, and mean hit rates of 98.93%, 98.96%, and 98.93%, respectively, are obtained to quantify the performance of the proposed decision-making classifier to identify normal controls and subjects with PD or ET. The experimental results indicate that the proposed screening model can improve the accuracy rate compared with the conventional machine learning classifier.


  • Using Polar Expression Features and Nonlinear Machine Learning Classifier for Automated Parkinson’s Disease Screening

    by Tsung-Lung Yang, Ping-Ju Kan, Chia-Hung Lin, Hsin-Yu Lin, Wei-Ling Chen, Her-Terng Yau

    Analyzing digitalized hand-drawn patterns, such as Archimedes’ spirals, words, and sentences, is one strategy for evaluating functional tremors and upper-limb movement disorders for neurodegenerative diseases. A pattern, such as a spiral or a line, in polar coordinates is a straight line or a curve that can be easily compared to a hand-drawn pattern with the same coordinates. Hence, in this study, using polar expression features, the deviation (cm), the accumulation angle (rad), and the drawing velocity (cm/s) were extracted to scale the variability in different tremor levels associated with Parkinson’s disease (PD) or essential tremor (ET). Then, a nonlinear support vector machine (SVM)-based classifier was used to separate the normal condition from PD or ET. The classifier was trained by the wolf pack search (WPS) optimization method. Tremor level progress was evaluated by integrating an assistive method in a smart mobile device (iPad), along with the nonlinear SVM-based classifier, into a decision-making system for individualized functions. In contrast to the multilayer machine learning method, with the 8-fold cross-validation, the proposed classifier exhibits superior performance in identifying normal controls and PD or ET, with the mean true positive, mean true negative, and mean hit rates being 93.72%, 86.79%, and 90.84%, respectively. The experimental results indicated that the proposed decision-making method is effective for detecting the progression of Parkinson-related diseases. This exami- nation technique is simple, comfortable, and repeatable for helping neurologists with preliminary diagnoses, drug treatments, and at-home monitoring.


  • Open Access

    CT Images of a Malignant-Transformed Ovarian Mature Cystic Teratoma with Rupture: a Case Report

    by Po Chin Wang, Tsung Lung Yang, Huay Ben Pan

    A malignant transformation or a tumor rupture is a rare complication of ovarian mature cystic teratoma (MCT). A tumor rupture in a malignant-transformed MCT has never been reported in the literature. We present the CT images of a 39-year-old woman showing a large, predominantly cystic mass in the lower abdomen, with fat-fluid-level ascites. A contrast-enhanced solid component, with regional discontinuity within the cystic lesion, is also demonstrated. The pathologic diagnosis of the ruptured MCT unveils the malignant transformation (squamous cell carcinoma) and mesenteric carcinomatosis.