scholarly journals Impact of Feature Selection on Clustering Images ofVertebral Compression Fractures

2020 ◽  
Author(s):  
Raquel Candido ◽  
Rafael Lama ◽  
Natália Chiari ◽  
Marcello Nogueira-Barbosa ◽  
Paulo Azevedo Marques ◽  
...  
Keyword(s):  
Feature Selection ◽  
Metastatic Cancer ◽  
Vertebral Compression Fractures ◽  
Features Selection ◽  
Correct Classification ◽  
Compression Fractures ◽  
Vertebral Bodies ◽  
The Impact ◽  
Aided Diagnosis

Non-traumatic Vertebral Compression Fractures (VCFs) are generally caused by osteoporosis (benign VCFs) or metastatic cancer (malignant VCFs) and the success of the medical treatment strongly depends on a fast and correct classification of VCFs. Recently, methods for computer-aided diagnosis (CAD) based on machine learning have been proposed for classifying VCFs. In this work, we investigate the problem of clustering images of VCFs and the impact of feature selection by genetic algorithms, comparing the clustering i)with all features and ii)with feature selection through the purity results. The analysis of the clusters helps to understand the results of classifiers and difficulties of differentiating images of different classes by an expert. The results indicate that features selection improved the separability of clusters and purity. Feature selection also helps to understand which attributes are most important for analysing the images of vertebral bodies.

scholarly journals A New Computer-Aided Diagnosis System with Modified Genetic Feature Selection for BI-RADS Classification of Breast Masses in Mammograms

2020 ◽  
Vol 2020 ◽  
pp. 1-17 ◽  
Author(s):  
Said Boumaraf ◽  
Xiabi Liu ◽  
Chokri Ferkous ◽  
Xiaohong Ma
Keyword(s):  
Feature Selection ◽  
Predictive Value ◽  
Breast Imaging ◽  
Feature Selection Method ◽  
Computer Aided Diagnosis ◽  
Screening Mammography ◽  
Breast Masses ◽  
Computer Aided ◽  
Aided Diagnosis

Mammography remains the most prevalent imaging tool for early breast cancer screening. The language used to describe abnormalities in mammographic reports is based on the Breast Imaging Reporting and Data System (BI-RADS). Assigning a correct BI-RADS category to each examined mammogram is a strenuous and challenging task for even experts. This paper proposes a new and effective computer-aided diagnosis (CAD) system to classify mammographic masses into four assessment categories in BI-RADS. The mass regions are first enhanced by means of histogram equalization and then semiautomatically segmented based on the region growing technique. A total of 130 handcrafted BI-RADS features are then extracted from the shape, margin, and density of each mass, together with the mass size and the patient’s age, as mentioned in BI-RADS mammography. Then, a modified feature selection method based on the genetic algorithm (GA) is proposed to select the most clinically significant BI-RADS features. Finally, a back-propagation neural network (BPN) is employed for classification, and its accuracy is used as the fitness in GA. A set of 500 mammogram images from the digital database for screening mammography (DDSM) is used for evaluation. Our system achieves classification accuracy, positive predictive value, negative predictive value, and Matthews correlation coefficient of 84.5%, 84.4%, 94.8%, and 79.3%, respectively. To our best knowledge, this is the best current result for BI-RADS classification of breast masses in mammography, which makes the proposed system promising to support radiologists for deciding proper patient management based on the automatically assigned BI-RADS categories.

Feature Selection Based on Gaussian Mixture Model Clustering for the Classification of Pulmonary Nodules Based on Computed Tomography

2020 ◽  
Vol 10 (5) ◽  
pp. 1033-1039
Author(s):  
Huihong Duan ◽  
Xu Wang ◽  
Xingyi He ◽  
Yonggang He ◽  
Litao Song ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Feature Selection ◽  
Gaussian Mixture Model ◽  
Mixture Model ◽  
Pulmonary Nodules ◽  
Gaussian Mixture ◽  
Local Optimum ◽  
Computer Aided ◽  
Aided Diagnosis

Background: In the pulmonary nodules computer aided diagnosis systems (CAD), feature selection plays an important role in reducing the false positive rate and improving the system accuracy. To solve the problem of feature selection techniques by which the diversity of features was damaged in the process of distinguishing malignant pulmonary nodules from benign pulmonary nodules, this study developed a novel feature selection algorithm for improving the accuracy of traditional computer-aided differential diagnosis for benign and malignant classification of pulmonary nodules. Method: Firstly, we divided the extracted features of nodules into several groups by using Gaussian mixture model (GMM). Secondly, we applied Relief and sequential forward selection (SFS) algorithm to find local optimum features dataset for each group. Afterwards, we used the optimumpath forest (OPF) classifier with the found features dataset to obtain the classification results. Finally, the local optimum features dataset with the highest area under curve AUC in all groups were added into the final selected set. Results: According to collected pulmonary nodules on computed tomography (CT) scans, tested with two set of samples, we achieved an average accuracy of 89.5%, sensitivity of 87.1% and specificity of 90.9% on the first set of samples, and 90.1%, 88.7% and 92.1% on the second set of samples. The areas under the receiver operating characteristic (ROC) curves based on these two sample sets were 95.2%, and 96.3% respectively. Conclusions: This study shows that the proposed method was promising for improving the pulmonary nodules computer aided diagnosis systems performance of benign and malignant pulmonary nodules.

Preoperative chemoradiotherapy for pancreatic cancer encountered vertebral compression fractures.

2013 ◽  
Vol 31 (4_suppl) ◽  
pp. 227-227
Author(s):  
Keisuke Otani ◽  
Teruki Teshima ◽  
Kinji Nishiyama ◽  
Yuri Ito ◽  
Yoshifumi Kawaguchi ◽  
...  
Keyword(s):  
Pancreatic Cancer ◽  
De Novo ◽  
Preoperative Chemoradiotherapy ◽  
Incidence Rates ◽  
Mesenteric Arteries ◽  
Vertebral Compression Fractures ◽  
Resectable Pancreatic Cancer ◽  
Compression Fractures ◽  
Dose Volume ◽  
Vertebral Bodies

227 Background: Preoperative chemoradiotherapy (CRT) with gemcitabine (GEM) for pancreatic cancer at our institute achieved excellent cure rates, while treated patients encountered vertebral compression fractures (VCFs) frequently. Methods: From January 2006 to December 2011, 220 patients (male/female: 134/86, median age: 66 (range: 33-84)) with resectable pancreatic cancer have been treated with preoperative CRT with GEM. This method consisted of concurrent radiotherapy (50-60 Gy in 25 fractions over 5 weeks) and chemotherapy (GEM 1,000 mg/m2intravenous, weekly for 12-15 weeks, 3 times during 4 weeks). Three-dimensional conformal radiotherapy was used and its fields covered both primary pancreatic tumor and retropancreatic tissues with 50 Gy level. Boost irradiation to the roots of celiac and superior mesenteric arteries with 2.4 Gy/fr using field-within-field technique has been started since 2009. The risk factors for VCF and its causes were retrospectively analyzed for 1,308 Th10-L3 vertebral bodies. Results: Follow-up time from the initiation of CRT ranged from 3.4-73.9 months (median: 17.9). Median overall survival time of these patients was 40.6 months. Radical surgery was completed in 164 patients (75%). Twenty-five patients (11%) diagnosed as VCF (CTCAE v4.0 grade1/2: 12/13) and 22 of them were at 3.6-23 months from CRT. Cumulative incidence rates at 2 years were 18.9% in total, 8.5% in male and 38.4% in female (p = 0.0002). Corresponding rates were 5.1% in age < 60 and 24.9% in age ≥ 60 (p = 0.0107). On their courses 37 de novo VCFs (Th10/11/12/L1/2/3: 2/3/11/13/7/1) were observed at first and 9 patients repeated VCFs near the initial VCFs. Dose-volume factors of each vertebral bodies also significantly correlated with VCF; vertebrae whose mean dose were < 38 Gy/ ≥ 38 Gy developed VCFs in 2.7%/10.0% at 2 years (p < 0.0001), vertebrae whose V30 were < 80%/≥80% developed VCFs in 2.4%/10.6% (p < 0.0001). Conclusions: Besides women and higher age, dose-volume factors of radiotherapy were risk of developing VCF in patients with pancreatic cancer after preoperative CRT with GEM. These findings should be considered to avoid VCFs as late adverse event after CRT for pancreatic cancer, especially in IMRT era.

Shape, texture and statistical features for classification of benign and malignant vertebral compression fractures in magnetic resonance images

2016 ◽  
Vol 73 ◽  
pp. 147-156 ◽  
Author(s):  
Lucas Frighetto-Pereira ◽  
Rangaraj Mandayam Rangayyan ◽  
Guilherme Augusto Metzner ◽  
Paulo Mazzoncini de Azevedo-Marques ◽  
Marcello Henrique Nogueira-Barbosa
Keyword(s):  
Magnetic Resonance ◽  
Magnetic Resonance Images ◽  
Vertebral Compression Fractures ◽  
Statistical Features ◽  
Compression Fractures

Vertebral Augmentation

2017 ◽  
Author(s):  
Siddarth Thakur ◽  
Daniel Rothstein ◽  
Kent H Nouri
Keyword(s):  
Comprehensive Evaluation ◽  
Spinal Metastasis ◽  
Metastatic Cancer ◽  
Compression Fracture ◽  
Public Health Problem ◽  
Economic Cost ◽  
Cement Leakage ◽  
Vertebral Compression Fractures ◽  
Vertebral Augmentation ◽  
Compression Fractures

Vertebral compression fractures are a prevalent and growing public health problem associated with significant morbidity and economic cost. Most commonly, they occur in osteoporotic patients but are also seen in patients with metastatic cancer and secondary to trauma. Appropriate and timely treatment is imperative. When conservative treatments are inadequate, minimally invasive techniques, such as vertebroplasty and kyphoplasty, can provide substantial pain relief, improve function, and enhance quality of life. For appropriate patient selection, a comprehensive evaluation is essential to confirm the presence of concordant pain. Both vertebroplasty and kyphoplasty are performed percutaneously under radiographic guidance, and cement is injected into the collapsed vertebral body to provide strength and stability. Awareness of early and late procedure-related complications is necessary for perioperative planning. Overall, vertebral augmentation is a safe and efficacious procedure for patients suffering from pain related to vertebral compression fractures. This review contains 10 figures, 5 tables, and 74 references. Key words: adjacent level fractures, axial low back pain, cement injection, cement leakage, kyphoplasty, osteoporosis, parapedicular, polymethylmethacrylate, spinal metastasis, transpedicular, vertebral augmentation, vertebral compression fracture, vertebroplasty

CT hounsfield unit and histogram analysis for differentiation of recent versus remote vertebral compression fractures

2021 ◽  
pp. 20210941
Author(s):  
Chang Hyun Ryoo ◽  
Jee Won Chai ◽  
Sung Hwan Hong ◽  
Ja-Young Choi ◽  
Hye Jin Yoo ◽  
...  
Keyword(s):  
Trabecular Bone ◽  
Area Under The Curve ◽  
Hounsfield Unit ◽  
Histogram Analysis ◽  
Vertebral Compression Fractures ◽  
Roc Curve Analysis ◽  
Compression Fractures ◽  
High Area ◽  
Tissue Changes ◽  
Vertebral Bodies

Objectives: The purpose of this study was to analyze the intraosseous tissue changes in recent vertebral compression fractures (VCFs) and to differentiate recent from remote VCFs using CT Hounsfield unit histogram analysis (HUHA). Methods: Sixty-five patients with T11 to L3 VCFs were included. HUHA of 2 vertebral bodies (VBs)— a fractured VB and the closest lower-level unaffected VB—was done. The mean Hounsfield unit (HU) value and HU proportions of 5 ranges (HU ≤ 0, 0 < HU≤50, 50 < HU≤100, 100 < HU≤150, and HU > 150) were obtained. Then, ΔHU value and ΔHU proportion were calculated by subtracting the values from the two vertebrae. Finally, the obtained values were compared between the recent and remote VCF groups and subjected to ROC curve analysis. Results: In recent VCF group, the ΔHU proportion (HU ≤0) corresponding to normal fatty marrow was lower (-0.17 vs 0.01) and the ΔHU proportion (HU >150) representing trabecular bone was higher (0.23 vs 0.04) than in remote VCF group (p < 0.001). In the differentiation of recent from remote VCF, the ΔHU value and ΔHU proportion (HU >150) showed high area under the curve (AUC, 0.939 and 0.912, respectively). Conclusions: CT HUHA demonstrated both trabecular bone and bone marrow changes in recent VCFs, and showed high diagnostic performance in differentiating between recent and remote VCFs. Advances in knowledge: With its vendor neutral applicability, CT HUHA can be used for the differentiation of recent and remote VCFs.

Sign in / Sign up

Export Citation Format

Share Document