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.

Survival Benefit of Tamoxifen in Estrogen Receptor-Negative and Progesterone Receptor-Positive Low Grade Breast Cancer Patients

Li-Heng Yang,Hsin-Shun Tseng,Che Lin,Li-Sheng Chen,Shou-Tung Chen,Shou-Jen Kuo,Dar-Ren Chen 한국유방암학회 2012 Journal of breast cancer Vol.15 No.3

Purpose: This study aimed to analyze the efficacy and prognostic significance of adjuvant tamoxifen in breast cancer patients with various hormone receptor statuses. Methods: Typically, 1,260 female breast cancer patients were recruited in this study. The correlation between estrogen receptor (ER)/progesterone receptor (PR) phenotypes and clinical characteristics was investigated, and the survival rate was assessed after 5-year follow-up. Results: The 5-year overall survival (85%) was better in women under the age of 50 years. Patients with ER+/PR+ tumors had a better 5-year survival rate (94%); those with ER-/PR- tumors experienced the worst outcome (74% survival rate); whereas singlepositive cases were in between. In 97 out of 128 patients with ER-/PR+ tumors, tamoxifen was given as adjuvant hormonal therapy, and it increased the survival benefit in the lower grade group in terms of overall survival and disease-free survival (p=0.01 and p=0.03, respectively). Conclusion: For high-grade tumors with ER-/PR+, adjuvant tamoxifen therapy may have no survival benefit, whereas for the patients with low-grade ER-/PR+ tumors, adjuvant tamoxifen therapy is highly suggestive.

A New Reference Pixel Prediction for Reversible Data Hiding with Reduced Location Map

( Jeanne Chen ),( Tung-shou Chen ),( Wien Hong ),( Gwoboa Horng ),( Han-yan Wu ),( Chih-wei Shiu ) 한국인터넷정보학회 2014 KSII Transactions on Internet and Information Syst Vol.8 No.3

In this paper, a new reversible data hiding method based on a dual binary tree of embedding levels is proposed. Four neighborhood pixels in the upper, below, left and right of each pixel are used as reference pixels to estimate local complexity for deciding embeddable and non-embeddable pixels. The proposed method does not need to record pixels that might cause underflow, overflow or unsuitable for embedment. This can reduce the size of location map and release more space for payload. Experimental results show that the proposed method is more effective in increasing payload and improving image quality than some recently proposed methods.

Reversible Data Hiding in Block Truncation Coding Compressed Images Using Quantization Level Swapping and Shifting

( Wien Hong ),( Shuozhen Zheng ),( Tung-shou Chen ),( Chien-che Huang ) 한국인터넷정보학회 2016 KSII Transactions on Internet and Information Syst Vol.10 No.6

The existing reversible data hiding methods for block truncation coding (BTC) compressed images often utilize difference expansion or histogram shifting technique for data embedment. Although these methods effectively embed data into the compressed codes, the embedding operations may swap the numerical order of the higher and lower quantization levels. Since the numerical order of these two quantization levels can be exploited to carry additional data without destroying the quality of decoded image, the existing methods cannot take the advantages of this property to embed data more efficiently. In this paper, we embed data by shifting the higher and lower quantization levels in opposite direction. Because the embedment does not change numerical order of quantization levels, we exploit this property to carry additional data without further reducing the image quality. The proposed method performs no-distortion embedding if the payload is small, and performs reversible data embedding for large payload. The experimental results show that the proposed method offers better embedding performance over prior works in terms of payload and image quality.

An Optimized Iterative Semantic Compression Algorithm And Parallel Processing for Large Scale Data

( Ran Jin ),( Gang Chen ),( Anthony K H Tung ),( Lidan Shou ),( Beng Chin Ooi ) 한국인터넷정보학회 2018 KSII Transactions on Internet and Information Syst Vol.12 No.6

With the continuous growth of data size and the use of compression technology, data reduction has great research value and practical significance. Aiming at the shortcomings of the existing semantic compression algorithm, this paper is based on the analysis of ItCompress algorithm, and designs a method of bidirectional order selection based on interval partitioning, which named An Optimized Iterative Semantic Compression Algorithm (Optimized ItCompress Algorithm). In order to further improve the speed of the algorithm, we propose a parallel optimization iterative semantic compression algorithm using GPU (POICAG) and an optimized iterative semantic compression algorithm using Spark (DOICAS). A lot of valid experiments are carried out on four kinds of datasets, which fully verified the efficiency of the proposed algorithm.

Use of Magnetic Resonance Imaging for Evaluating Residual Breast Tissue After Robotic-Assisted Nipple-Sparing Mastectomy in Women With Early Breast Cancer

Wu Wen-Pei,Lai Hung-Wen,Liao Chiung-Ying,Lin Joseph,Huang Hsin-I,Chen Shou-Tung,Chou Chen-Te,Chen Dar-Ren 대한영상의학회 2023 Korean Journal of Radiology Vol.24 No.7

Objective: Prospective studies on postoperative residual breast tissue (RBT) after robotic-assisted nipple-sparing mastectomy (R-NSM) for breast cancer are limited. RBT presents an unknown risk of local recurrence or the development of new cancer after curative or risk-reducing mastectomies. This study investigated the technical feasibility of using magnetic resonance imaging (MRI) to evaluate RBT after R-NSM in women with breast cancer. Materials and Methods: In this prospective pilot study, 105 patients, who underwent R-NSM for breast cancer at Changhua Christian Hospital between March 2017 and May 2022, were subjected to postoperative breast MRI to evaluate the presence and location of RBT. The postoperative MRI scans of 43 patients (age, 47.8 ± 8.5 years), with existing preoperative MRI scans, were evaluated for the presence and location of RBT. In total, 54 R-NSM procedures were performed. In parallel, we reviewed the literature on RBT after nipple-sparing mastectomy, considering its prevalence. Results: RBT was detected in 7 (13.0%) of the 54 mastectomies (6 of the 48 therapeutic mastectomies and 1 of the 6 prophylactic mastectomies). The most common location for RBT was behind the nipple-areolar complex (5 of 7 [71.4%]). Another RBT was found in the upper inner quadrant (2 of 7 [28.6%]). Among the six patients who underwent RBT after therapeutic mastectomies, one patient developed a local recurrence of the skin flap. The other five patients with RBT after therapeutic mastectomies remained disease-free. Conclusion: R-NSM, a surgical innovation, does not seem to increase the prevalence of RBT, and breast MRI showed feasibility as a noninvasive imaging tool for evaluating the presence and location of RBT.

Evaluating genetic variants associated with breast cancer risk in high and moderate-penetrance genes in Asians

Han, Mi-Ryung,Zheng, Wei,Cai, Qiuyin,Gao, Yu-Tang,Zheng, Ying,Bolla, Manjeet K.,Michailidou, Kyriaki,Dennis, Joe,Wang, Qin,Dunning, Alison M.,Brennan, Paul,Chen, Shou-Tung,Choi, Ji-Yeob,Hartman, Mikae Oxford University Press 2017 Carcinogenesis Vol.38 No.5

<P>Over the past 20 years, high-penetrance pathogenic mutations in genes BRCA1, BRCA2, TP53, PTEN, STK11 and CDH1 and moderate-penetrance mutations in genes CHEK2, ATM, BRIP1, PALB2, RAD51C, RAD50 and NBN have been identified for breast cancer. In this study, we investigated whether there are additional variants in these 13 genes associated with breast cancer among women of Asian ancestry. We analyzed up to 654 single nucleotide polymorphisms (SNPs) from 6269 cases and 6624 controls of Asian descent included in the Breast Cancer Association Consortium (BCAC), and up to 236 SNPs from 5794 cases and 5529 controls included in the Shanghai Breast Cancer Genetics Study (SBCGS). We found three missense variants with minor allele frequency (MAF) < 0.05: rs80358978 (Gly2508Ser), rs80359065 (Lys2729Asn) and rs11571653 (Met784Val) in the BRCA2 gene, showing statistically significant associations with breast cancer risk, with P-values of 1.2 x 10(-4), 1.0 x 10(-3) and 5.0 x 10(-3), respectively. In addition, we found four low-frequency variants (rs8176085, rs799923, rs8176173 and rs8176258) in the BRCA1 gene, one common variant in the CHEK2 gene (rs9620817), and one common variant in the PALB2 gene (rs13330119) associated with breast cancer risk at P < 0.01. Our study identified several new risk variants in BRCA1, BRCA2, CHEK2, and PALB2 genes in relation to breast cancer risk in Asian women. These results provide further insights that, in addition to the high/moderate penetrance mutations, other low-penetrance variants in these genes may also contribute to breast cancer risk.</P>