Deep Attentive Panoptic Model for Prostate Cancer Detection Using Biparametric MRI Scans
Multi-parametric MRI (mp-MRI) has recently been established in major guidelines as a first-line diagnostic test for men suspected of having prostate cancer (PCa) primarily to detect and classify clinically significant lesions. However, widespread utilizat
- PDF / 1,274,371 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 59 Downloads / 143 Views
Digital Technology and Innovation, Siemens Healthineers, Princeton, NJ, USA [email protected] 2 Universit¨ atsspital Basel, Basel, Switzerland 3 Diagnostic Imaging, Siemens Healthcare, Erlangen, Germany
Abstract. Multi-parametric MRI (mp-MRI) has recently been established in major guidelines as a first-line diagnostic test for men suspected of having prostate cancer (PCa) primarily to detect and classify clinically significant lesions. However, widespread utilization is still challenged by 1) the difficulty of interpretation specifically for radiologists less experienced in reading mp-MRI scans, and 2) decreased productivity associated with increased time spent per case for reading these complex scans. Deep learning based lesion detection and segmentation methods have been proposed for radiologists to perform their tasks more accurately and efficiently. In this work, we present a novel panoptic lesion detection and segmentation method with both semantic and instance branches as well as an attention module to optimally incorporate both local and global image features. In a free-response receiver operating characteristics (FROC) analysis for lesion sensitivity on an independent dataset with 243 patients, our method has achieved 89% sensitivity and 85% with 0.94 and 0.62 false positives per patient, respectively. Using the proposed method, we have achieved an unprecedented area under ROC curve (AUC) of 0.897 in identifying clinically significant cases. Keywords: Prostate cancer detection Biparametric MRI
1
· Panoptic segmentation ·
Introduction
Prostate Cancer (PCa) ranks as the second most frequent cancer and fifth leading cause of death in men worldwide [1]. Early PCa detection plays a crucial role in cancer treatment. Multi-parametric MRI (mp-MRI) and biparametric MRI (bpMRI) based PCa diagnosis have been shown to be effective and lead to a superior ProstateAI Clinical Collaborators–A list of members and affiliations appears at the end of the paper. c Springer Nature Switzerland AG 2020 A. L. Martel et al. (Eds.): MICCAI 2020, LNCS 12264, pp. 594–604, 2020. https://doi.org/10.1007/978-3-030-59719-1_58
Deep Attentive Panoptic Model for Prostate Cancer Detection
595
biopsy yield rate as compared to standard systematic biopsy based on ultrasound guidance [7,14,15,23]. Computer-aided diagnosis (CAD) of PCa using mp-MRI or bp-MRI scans have become an active research area and many methods have recently been proposed [19,24]. However, there are still some major challenges for PCa detection. For example, there is a relatively high feature similarity between benign prostatic hyperplasia (BPH) and high grade PCa, specifically in the central/transition zone. Furthermore, for some clinically significant cases, there are only subtle differences between lesion features and those from normal prostate parenchyma. These together with varying shapes and sizes of lesions, and alignment errors across various imaging contrasts, make it very hard for a lesion detection system to capture all possible lesions (high sensitivit
Data Loading...
