Computer‐aided diagnosis of prostate cancer using a deep convolutional neural network from multiparametric MRI

Song, Yang; Zhang, Yudong; Xu, Yan; Liu, Hui; Zhou, Minxiong; Hu, Bingwen; Yang, Guang

doi:10.1002/jmri.26047

Cited by 156 publications

(130 citation statements)

References 18 publications

Supporting

Mentioning

127

Contrasting

Unclassified

Order By: Relevance

“…Recently, several studies have evaluated machine‐learning‐based techniques for automated PCa detection using mp‐MRI. Song et al developed a computer‐aided tool for PCa diagnosis from mp‐MRI using a convolutional neural network (CNN) . They aligned and resampled low‐resolution DWI/ADC images to high‐resolution T 2 W images using rigid and nonrigid transformation to create a registered image with three channels.…”

Section: Discussionmentioning

confidence: 99%

“…Song et al developed a computer-aided tool for PCa diagnosis from mp-MRI using a convolutional neural network (CNN). 30 They aligned and resampled low-resolution DWI/ADC images to high-resolution T 2 W images using rigid and nonrigid transformation to create a registered image with three channels. The images were then artificially augmented via random rotation, shift, flip, and stretching strategies to train a deep CNN.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Zabihollahy

Ukwatta

Krishna

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Background Accurate detection and localization of prostate cancer (PCa) in men undergoing prostate MRI is a fundamental step for future targeted prostate biopsies and treatment planning. Fully automated localization of peripheral zone (PZ) PCa using the apparent diffusion coefficient (ADC) map might be clinically useful. Purpose/Hypothesis To describe automated localization of PCa in the PZ on ADC map MR images using an ensemble U‐Net‐based model. Study Type Retrospective, case–control. Population In all, 226 patients (154 and 72 patients with and without clinically significant PZ PCa, respectively), training, and testing was performed using dataset images of 146 and 80 patients, respectively. Field Strength 3T, ADC maps. Sequence ADC map. Assessment The ground truth was established by manual delineation of the prostate and prostate PZ tumors on ADC maps by dedicated radiologists using MRI‐radical prostatectomy maps as a reference standard. Statistical Tests: Performance of the ensemble model was evaluated using Dice similarity coefficient (DSC), sensitivity, and specificity metrics on a per‐slice basis. Receiver operating characteristic (ROC) curve and area under the curve (AUC) were employed as well. The paired t‐test was used to test the differences between the performances of constituent networks of the ensemble model. Results Our developed algorithm yielded DSC, sensitivity, and specificity of 86.72% ± 9.93%, 85.76% ± 23.33%, and 76.44% ± 23.70%, respectively (mean ± standard deviation) on 80 test cases consisting of 41 and 39 instances from patients with and without clinically significant tumors including 660 extracted 2D slices. AUC was reported as 0.779. Data Conclusion An ensemble U‐Net‐based approach can accurately detect and segment PCa in the PZ from ADC map MR prostate images. Level of Evidence: 4 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2020;51:1223–1234.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Zabihollahy

Ukwatta

Krishna

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

show abstract

“…While a vast body of research has been proposed to analyzing natural images, only a handful has dealt with the problem of prostate cancer classification with deep learning methods. [26][27][28][29][30][31] Reda et al 26 trained a stacked auto-encoder network with non-negativite constraint algorithm with a logistic regression classifier to distinguish the prostate tumor as either benign or malignant with ADC images. Kiraly et al 28 proposed multichannel image-to-image convolutional encoderdecoders to localize lesions and then output different tumor classes.…”

Section: A Existing Prostate Cancer Diagnosis Methodsmentioning

confidence: 99%

“…Le et al 30 proposed a new similarity loss function in the multimodal convolutional neural networks for prostate cancer diagnosis to enforce the features extracted from ADC and T2W images consistent. However, these methods [26][27][28][29][30][31] utilized existing deep learning models to conduct prostate cancer classification task. The specific problems (such as limited data) associated with the medical images were tacitly ignored, which may lead to failing or overfitting in the training procedure.…”

Section: A Existing Prostate Cancer Diagnosis Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Prostate cancer classification with multiparametric MRI transfer learning model

et al. 2019

View full text Add to dashboard Cite

Purpose Prostate cancer classification has a significant impact on the prognosis and treatment planning of patients. Currently, this classification is based on the Gleason score analysis of biopsied tissues, which is neither accurate nor risk free. This study aims to learn discriminative features in prostate images and assist physicians in classifying prostate cancer automatically. Methods We develop a novel multiparametric magnetic resonance transfer learning (MPTL) method to automatically stage prostate cancer. We first establish a deep convolutional neural network with three branch architectures, which transfer pretrained model to compute features from multiparametric MRI images (mp‐MRI): T2w transaxial, T2w sagittal, and apparent diffusion coefficient (ADC). The learned features are concatenated to represent information of mp‐MRI sequences. A new image similarity constraint is then proposed to enable the distribution of the features within the same category in a narrow angle region. With the joint constraints of softmax loss and image similarity loss in the fine‐tuning process, the MPTL can provide descriptive features with intraclass compactness and interclass separability. Results Two cohorts: 132 cases from our institutional review board‐approved patient database and 112 cases from the PROSTATEx‐2 Challenge are utilized to evaluate the robustness and effectiveness of the proposed MPTL model. Our model achieved high accuracy of prostate cancer classification (accuracy of 86.92%). Moreover, the comparison results demonstrate that our method outperforms both hand‐crafted feature‐based methods and existing deep learning models in prostate cancer classification with higher accuracy. Conclusion The experiment results showed that the proposed method can learn discriminative features in prostate images and classify the cancer accurately. Our MPTL model could be further applied in the clinical practice to provide valuable information for cancer treatment and precision medicine.

show abstract

PI‐RADS: Past, present, and future

Gupta

Mehta

Türkbey

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Prostate cancer (PCa) is extremely prevalent and is the most common noncutaneous malignancy and second‐most common cause of cancer death in men. In the last decade, there has been dramatic growth in the use of multiparametric magnetic resonance imaging (mpMRI) for diagnosis and characterization of PCa. With the recent and marked surge in popularity in prostate imaging and, specifically, mpMRI, there has been an increased focus on structured reporting as a means by which to provide more actionable information to the referring clinician as well as to improve diagnostic performance with this technique. This work focuses on the evolution of the major structured reporting system in prostate mpMRI, Prostate Imaging Reporting And Data System (PI‐RADS), from its initial proposal and establishment in 2012 as PI‐RADS v. 1 to its most current iteration, PI‐RADS v. 2.1. This will highlight the key elements that have changed between the versions as well as provide context and rationale for these changes. In addition, this work explores what future iterations of PI‐RADS could look like based on current limitations of the system as well as explore areas for future growth of prostate mpMRI, including use of the system in active surveillance populations and in the posttreatment setting.Level of Evidence: 5Technical Efficacy: Stage 5J. Magn. Reson. Imaging 2020;52:33–53.

show abstract

Computer‐aided diagnosis of prostate cancer using a deep convolutional neural network from multiparametric MRI

Abstract: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018;48:1570-1577.

Cited by 156 publications

References 18 publications

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Fully automated localization of prostate peripheral zone tumors on apparent diffusion coefficient map MR images using an ensemble learning method

Prostate cancer classification with multiparametric MRI transfer learning model

PI‐RADS: Past, present, and future

Contact Info

Product

Resources

About