Prostate Cancer Localization With Multispectral MRI Using Cost-Sensitive Support Vector Machines and Conditional Random Fields

Abstract

Prostate cancer is a leading cause of cancer death for men in the United States. Fortunately, the survival rate for early diagnosed patients is relatively high. Therefore, in vivo imaging plays an important role for the detection and treatment of the disease. Accurate prostate cancer localization with noninvasive imaging can be used to guide biopsy, radiotherapy, and surgery as well as to monitor disease progression. Magnetic resonance imaging (MRI) performed with an endorectal coil provides higher prostate cancer localization accuracy, when compared to transrectal ultrasound (TRUS). However, in general, a single type of MRI is not sufficient for reliable tumor localization. As an alternative, multispectral MRI, i.e., the use of multiple MRI-derived datasets, has emerged as a promising noninvasive imaging technique for the localization of prostate cancer; however almost all studies are with human readers. There is a significant inter and intraobserver variability for human readers, and it is substantially difficult for humans to analyze the large dataset of multispectral MRI. To solve these problems, this study presents an automated localization method using cost-sensitive support vector machines (SVMs) and shows that this method results in improved localization accuracy than classical SVM. Additionally, we develop a new segmentation method by combining conditional random fields (CRF) with a cost-sensitive framework and show that our method further improves cost-sensitive SVM results by incorporating spatial information. We test SVM, cost-sensitive SVM, and the proposed cost-sensitive CRF on multispectral MRI datasets acquired from 21 biopsy-confirmed cancer patients. Our results show that multispectral MRI helps to increase the accuracy of prostate cancer localization when compared to single MR images; and that using advanced methods such as cost-sensitive SVM as well as the proposed cost-sensitive CRF can boost the performance significantly when compared to SVM.

A non-invasive image based system for early diagnosis of prostate cancer.

“…The system had a specificity of 89.58%, sensitivity of 87.50%, accuracy of 89.38%, and a DSC of 62.2%. A similar approach was developed by Artan et al [57] and located cancerous regions using cost-sensitive support vector machine (SVM). Prostate segmentation was performed using a conditional random field and the same three features in [56] were utilized for classification.…”

“…The DSC for prostate localization and segmentation was 0.46 ± 0.26, and the area under the receiver operator characteristic (ROC) curves (A z ) of the classification was 0.79 ± 0.12. Ozer et al [58] also developed a technique that directly segmented prostate cancers using the same three features in [56,57]. Both the SVM and relevance vector machine (RVM) [59] classifiers were used and the system showed a specificity of 0.78 and a sensitivity of 0.…”

Analysis of contrast-enhanced medical images.

“…The system had a specificity of 89.58%, sensitivity of 87.50%, accuracy of 89.38%, and a DSC of 62.2%. A similar approach was developed by Artan et al [228] and located cancerous regions using cost-sensitive support vector ma-chine (SVM). Prostate segmentation was performed using a conditional random field and the same three features in [227] were utilized for classification.…”

“…The DSC for prostate localization and segmentation was 0.46±0.26, and the area under the receiver operator characteristic (ROC) curves (A z ) of the classification was 0.79±0.12. Ozer et al [448] also developed a technique that directly segmented prostate cancers using the same three features in [227,228]. Both the SVM and relevance vector machine (RVM) [449] classifiers were used and the system showed a specificity of 0.78 and a sensitivity of 0.…”

A novel NMF-based DWI CAD framework for prostate cancer.

“…This system Ozer et al [226] also developed a technique that directly segmented prostate cancers using the same three features in [253,254]. Both the SVM and RVM [168] classifiers were used and the system showed a specificity of 0.78 and a sensitivity of 0.74 for RVM and 0.74 and 0.79 for SVM.…”