Comparative Analysis of Logistic Regression, Support Vector Machine and Artificial Neural Network for the Differential Diagnosis of Benign and Malignant Solid Breast Tumors by the Use of Three-Dimensional Power Doppler Imaging

Shou-Tung Chen,Yi-Hsuan Hsiao,Yu-Len Huang,Shou-Jen Kuo,Hsin-Shun Tseng,Hwa-Koon Wu,Dar-Ren Chen 대한영상의학회 2009 Korean Journal of Radiology Vol.10 No.5

Objective: Logistic regression analysis (LRA), Support Vector Machine (SVM) and a neural network (NN) are commonly used statistical models in computeraided diagnostic (CAD) systems for breast ultrasonography (US). The aim of this study was to clarify the diagnostic ability of the use of these statistical models for future applications of CAD systems, such as three-dimensional (3D) power Doppler imaging, vascularity evaluation and the differentiation of a solid mass. Materials and Methods: A database that contained 3D power Doppler imaging pairs of non-harmonic and tissue harmonic images for 97 benign and 86 malignant solid tumors was utilized. The virtual organ computer-aided analysis-imaging program was used to analyze the stored volumes of the 183 solid breast tumors. LRA, an SVM and NN were employed in comparative analyses for the characterization of benign and malignant solid breast masses from the database. Results: The values of area under receiver operating characteristic (ROC) curve, referred to as Az values for the use of non-harmonic 3D power Doppler US with LRA, SVM and NN were 0.9341, 0.9185 and 0.9086, respectively. The Az values for the use of harmonic 3D power Doppler US with LRA, SVM and NN were 0.9286, 0.8979 and 0.9009, respectively. The Az values of six ROC curves for the use of LRA, SVM and NN for non-harmonic or harmonic 3D power Doppler imaging were similar. Conclusion: The diagnostic performances of these three models (LRA, SVM and NN) are not different as demonstrated by ROC curve analysis. Depending on user emphasis for the use of ROC curve findings, the use of LRA appears to provide better sensitivity as compared to the other statistical models. Objective: Logistic regression analysis (LRA), Support Vector Machine (SVM) and a neural network (NN) are commonly used statistical models in computeraided diagnostic (CAD) systems for breast ultrasonography (US). The aim of this study was to clarify the diagnostic ability of the use of these statistical models for future applications of CAD systems, such as three-dimensional (3D) power Doppler imaging, vascularity evaluation and the differentiation of a solid mass. Materials and Methods: A database that contained 3D power Doppler imaging pairs of non-harmonic and tissue harmonic images for 97 benign and 86 malignant solid tumors was utilized. The virtual organ computer-aided analysis-imaging program was used to analyze the stored volumes of the 183 solid breast tumors. LRA, an SVM and NN were employed in comparative analyses for the characterization of benign and malignant solid breast masses from the database. Results: The values of area under receiver operating characteristic (ROC) curve, referred to as Az values for the use of non-harmonic 3D power Doppler US with LRA, SVM and NN were 0.9341, 0.9185 and 0.9086, respectively. The Az values for the use of harmonic 3D power Doppler US with LRA, SVM and NN were 0.9286, 0.8979 and 0.9009, respectively. The Az values of six ROC curves for the use of LRA, SVM and NN for non-harmonic or harmonic 3D power Doppler imaging were similar. Conclusion: The diagnostic performances of these three models (LRA, SVM and NN) are not different as demonstrated by ROC curve analysis. Depending on user emphasis for the use of ROC curve findings, the use of LRA appears to provide better sensitivity as compared to the other statistical models.

A Study of User Experience Resulting From AI Artwork Production - Focus On Character Creation

Shou Yi Yu,Jun Ping Xu,장수민,반영환 대한인간공학회 2024 大韓人間工學會誌 Vol.43 No.1

Objective: The aim of this study is to improve the performance of AI artists from a technical perspective by using AI platforms and training methods that encompass aspects such as image quality, text alignment, text generation, inference time, and training efficiency. Background: With the rise and rapid development of Artificial Intelligence (AI), the art market has embraced its use. The rise of AI in the art market has simultaneously increased the productivity of designer characterization. AI painting is developing rapidly, It would also be interesting to study attribution in this area. Method: By analyzing traditional and AI-based character creation methods using questionnaires and conducting in-depth interviews with 14 participants. Results: Anthropometry was the most dominant area all the time. Health and safety area has been steadily increasing in publication amount. Conclusion: The research aims to improve the performance of AI artists in terms of image quality, text alignment, text generation, inference time and training efficiency through AI platforms and training methods. Users tend to use AI to improve artist efficiency and believe that images generated through AI have commercial uses. However, AI cannot replace the joy and exploration of artists in creation, but serves as a valuable tool to improve creativity and painting efficiency. AI tools will completely change our work and lives when the technology matures. Application: Research contributes to the application of AI in character creation and promotes sustainable development in the art field. This study contributes to the process of AI in character creation, which can help to promote the sustainable development of AI in the field of art, and provides further expansion of the IP of painted characters in the possible space.

APPROACH TO CONTROL METHOD OF PLANNING LEVEL FOR EXCAVATING ROBOT

Sun Shou Qian,Xu Hai Ming,Wang Yi Jun,He Bao Liang 대한전자공학회 1992 대한전자공학회 Conference Vol.1992 No.2

The Addressing of Distributed FBG Based on FMCW

Li Shou-duo,Xiong Yan-ling,Yang Wen-long,Li Qiao-yi 보안공학연구지원센터 2013 International Journal of Future Generation Communi Vol.6 No.4

Expression and Significance of TSGF, CEA and AFP in Patients Before and after Radical Surgery for Colon Cancer

Hu, Yi,Wang, Jing-Liang,Tao, Hai-Tao,Wu, Bai-Shou,Sun, Jin,Cheng, Yao,Dong, Wei-Wei,Li, Rui-Xin Asian Pacific Journal of Cancer Prevention 2013 Asian Pacific journal of cancer prevention Vol.14 No.6

Objective: To explore the expression and significance of tumor specific growth factor (TSGF), carcinoembryonic antigen (CEA) and alpha fetoprotein (AFP) in cancer tissue and serum of patients with colon cancer. Materials and Methods: Radical surgery for colon cancer was performed on 43 patients with laparoscopu under conditions of general anesthesia. The Elisa method was used to detect the levels of serum TSGF, CEA and AFP before and after radical operation, and cancer tissue underwent TSGF, CEA and AFP immunohistochemistry staining after laparoscopic surgery. The decreased conditions of serum TSGF, CEA and AFP in patients with colon cancer at different levels of differentiation and clinical stagings were analyzed, and the relationships of expression rates between histological types, colon cancer morphology, lymph node metastasis and TSGF, CEA as well as AFP in cancer tissue were assessed. Results: Compared with before radical surgery, the levels of serum TSGF, CEA and AFP decreased notably in patients after operations (p<0.01). The decreased degree of TSGF and CEA was the largest in patients with poorly differentiated cancer tissue (p<0.01), while that of AFP was noted in patients with moderately differentiated cancer tissue (p<0.01). The decreased degree of TSGF and AFP was the largest in patients at phase Dukes A (p<0.01), while that of CEA in patients at phase Dukes C (p<0.01). There were no significant differences among the positive expression rates of TSGF, CEA and AFP with different histological types and colon cancer morphologies (p>0.05). The positive expression rates of TSGF and CEA in patients with lymph node metastasis were significantly higher than those without lymph node metastasis (p<0.01). Conclusions: TSGF, CEA and AFP can be used to evaluate the effect of radical operation for colon cancer, and the changed levels of different markers are associated with tumor differentiation, clinical stating and presence or absence of lymph node metastasis.

Analysis on the Characteristics and Prognosis of Pulmonary Neuroendocrine Tumors

Wu, Bai-Shou,Hu, Yi,Sun, Jing,Wang, Jin-Liang,Wang, Peng,Dong, Wei-Wei,Tao, Hai-Tao,Gao, Wen-Juan Asian Pacific Journal of Cancer Prevention 2014 Asian Pacific journal of cancer prevention Vol.15 No.5

Objective: To retrospectively review the clinical characteristics and analyze the prognostic factors of Chinese patients with pulmonary neuroendocrine tumors. Materials and Methods: The clinical data of 176 patients with pulmonary neuroendocrine tumors in Chinese PLA General Hospital from Mar., 2000 to Oct., 2012 were retrospectively analyzed. The parameters were evaluated by univariate and multivariate analysis, including the gender, age, smoking history, family history, TNM staging, localization (central or peripheral), tumor size, nodal status, histological subtype and treatment (operation or non-operation). Results: There were 23 patients with typical carcinoids (TC) (13.1%), 41 with atypical carcinoids (AC) (23.3%), 10 with large cell neuroendocrine carcinoma (LCNEC) (5.7%) and 102 with small cell lung cancer (SCLC) (57.9%). The median follow-up time was 64.5 months for AC, 38 months for LCNEC and 27 months for SCLC. The typical carcinoid censored data was 18 (more than 50% of the patients), so the median follow-up time was not obtained, and actuarial 5-year survivals for TC, AC, LCNEC and SCLC were 75.1%, 51.7%, 26.7% and 38.8%, respectively. COX univariate analysis revealed that the age (P=0.001), histological subtype (P=0.005), nodal status (P=0.000), treatment (P=0.000) and TNM staging (P=0.000) were the prognostic factors of the patients with pulmonary neuroendocrine tumors, whereas its multivariate analysis showed that only the age(P=0.001), TNM staging (P=0.002) and treatment (P=0.000) were independent prognostic factors. Conclusions: Radical surgery remains the treatment of choice, and is the only curative option. The age, TNM staging and treatment are confirmed to be the independent prognostic factors in multivariable models for pulmonary neuroendocrine tumors.

APPROACH TO CONTROL METHOD OF PLANNING LEVEL FOR EXCAVATING ROBOT

Qian, Sun Shou,Ming, Xu Hai,Jun, Wang Yi,Liang, He Bao 대한전자공학회 1992 HICEC:Harbin International Conference on Electroni Vol.1 No.1

The method of path planning and trajectory planning are studied. The control of planning level for whole work cycle process and excavating process based on predetermined path and force monitor is realized on an experimental prototype. The software of three dimensional dynamic simulation is developed to examine the control method of planning level.

Bevacizumab Concomitant with Chemotherapy is Effective in Treating Chinese Patients with Advanced Non-Squamous Non-Small Cell Lung Cancer

Sun, Jing,Hu, Yi,Wu, Bai-Shou,Wang, Jin-Liang,Tao, Hai-Tao,Zhang, Su-Jie Asian Pacific Journal of Cancer Prevention 2014 Asian Pacific journal of cancer prevention Vol.15 No.14

Objectives: To retrospectively review the safety and clinical efficacy of bevacizumab concomitant with chemotherapy in Chinese patients with advanced non-squamous non-small cell lung cancer (NSNSCLC). Methods: Clinical data for 79 patients with NSNSCLC who received bevacizumab concomitant with chemotherapy in Chinese PLA General Hospital from April 28th 2009 to May 5th 2013 were retrospectively reviewed to analyze the clinical efficacy including disease control rate (DCR), overall response rate (ORR), progression-free survival (PFS), overall survival (OS), the Eastern Cooperative Oncology Group (ECOG) score and the safety. Results: The Eastern Cooperative Oncology Group (ECOG) score was 0-2. By the final cutoff date (June 9, 2013), 54 (68.4%) patients had disease progression and 37 (46.8%) died. The ORR was 32.9% and the DCR was 83.5%. The ORR of the first-, second-, and third- or later-line treatments were 51.4%, 25.0% and 12.5%, while the DCR were 94.3%, 80.0% and 70.8%, respectively. The median OS (mOS) and PFS (mPFS) were 13.5 and 5.83 months, respectively. The mOS of patients with the first-, second-, and third- or later- line treatments were 16.2, 10.9 and 8.30 months, while the mPFS were 7.27, 5.90 and 5.17 months, respectively. Chemotherapy-related adverse events included myelosuppression, vomiting, hepatic dysfunction and renal dysfunction, while the common serious bevacizumab-related adverse events were thromboembolic problems, gastrointestinal perforation and reversible posterior leukoencephalopathy syndrome, which could be well managed. Conclusions: Bevacizumab concomitant with chemotherapy is effective and the related toxicity can be well tolerated in Chinese patients with NSNSCLC.

Rs7574865 polymorphism of STAT4 and risk of anti-tuberculosis drug-induced hepatotoxicity in Chinese Han

Guo Chen,Wei Mao,Shou‑Quan Wu,Yu Wang,Gui‑Yi Ji,Miao‑Miao Zhang,Qian‑Qian Liu,Jian‑Qing He 한국유전학회 2017 Genes & Genomics Vol.39 No.11

Anti-tuberculosis drug-induced hepatotoxicity (ATDH) is the most frequent, and potentially fatal adverse effect in the treatment of tuberculosis (TB). The rs7574865 polymorphism in the signal transducer and activator of transcription 4 gene (STAT4) was reported to be associated with drug-induced liver injury. However, there was no study aimed to this association in Chinese patients. The aim of this study was to investigate the influence of STAT4 polymorphisms on the susceptibility to ATDH in a Chinese Han population. A total of 280 TB patients with the prescription of anti-TB therapy, of Chinese Han origin, were enrolled. They were followed up for 3 months and demographic, clinical, laboratory and therapeutic data at each visit were collected. Two single nucleotide polymorphisms (SNPs) (rs7574865 and rs7582694) of STAT4 were genotyped with the MassARRAY platform. The associations between SNPs and ATDH risk were analyzed by logistic regression adjusting for confounding factors. A total of 24 patients were diagnosed with ATDH and considered as the case group, and 33 patients were lost to follow-up; the remaining 223 subjects without ATDH were considered as the control group. There was strong linkage disequilibrium (LD) between rs7574865 and rs7582694 ( r2 = 0.928 and D′ = 1). No significant association was found between SNPs or haplotypes of STAT4 and ATDH after correction for confounding factors. This prospective study is the first to investigate the association of genetic polymorphisms of STAT4 and ATDH in Chinese individuals. There was no significant association between the rs7574865 of STAT4 and ATDH in a Chinese Han population.

Local conversion of four Einstein-Podolsky-Rosen photon pairs into four-photon polarization-entangled decoherence-free states with non-photon-number-resolving detectors.

Wang, Hong-Fu,Zhang, Shou,Zhu, Ai-Dong,Yi, X X,Yeon, Kyu-Hwang Optical Society of America 2011 Optics express Vol.19 No.25

<P>We propose a linear-optics-based scheme for local conversion of four Einstein-Podolsky-Rosen photon pairs distributed among five parties into four-photon polarization-entangled decoherence-free states using local operations and classical communication. The proposed setup involves simple linear optical elements and non-photon-number-resolving detectors that can only distinguish between the presence and absence of photons, and no information on the exact number of photons can be obtained. This greatly simplifies the experimental realization for linear optical quantum computation and quantum information processing.</P>