Abstract
Uterine cervical cancer is a leading cause of women’s mortality worldwide. Cervical tissue ablation is an effective surgical excision of high grade lesions that are determined to be precancerous. Our prior work on the Automated Visual Examination (AVE) method demonstrated a highly effective technique to analyze digital images of the cervix for identifying precancer. Next step would be to determine if she is treatable using ablation. However, not all women are eligible for the therapy due to cervical characteristics. We present a machine learning algorithm that uses a deep learning object detection architecture to determine if a cervix is eligible for ablative treatment based on visual characteristics presented in the image. The algorithm builds on the well-known RetinaNet architecture to derive a simpler and novel architecture in which the last convolutional layer is constructed by upsampling and concatenating specific RetinaNet pretrained layers, followed by an output module consisting of a Global Average Pooling (GAP) layer and a fully connected layer. To explain the recommendation of the deep learning algorithm and determine if it is consistent with lesion presentation on the cervical anatomy, we visualize classification results using two techniques: our (i) Class-selective Relevance Map (CRM), which has been reported earlier, and (ii) Class Activation Map (CAM). The class prediction heatmaps are evaluated by a gynecologic oncologist with more than 20 years of experience. Based on our observation and the expert’s opinion, the customized architecture not only outperforms the baseline RetinaNet network in treatability classification, but also provides insights about the features and regions considered significant by the network toward explaining reasons for treatment recommendation. Furthermore, by investigating the heatmaps on Gaussian-blurred images that serve as surrogates for out-of-focus cervical pictures we demonstrate the effect of image quality degradation on cervical treatability classification and underscoring the need for using images with good visual quality.
Highlights
Uterine cervix cancer is the fourth most common cancer in women with nearly570,000 new cases reported by the World Health Organization (WHO) in 2018 [1]
Treatment eligibility is determined through expert visual assessment of the cervix following a positive screening and histological confirmation of high grade Cervical Intraepithelial Neoplasia (CIN)
Inspired by the above studies, in this work we investigate the effectiveness of object detection networks, such as RetinaNet, in classifying a digitized cervix image, into: (1) eligible for thermal ablation, denoted as “treatable”, or (2)
Summary
570,000 new cases reported by the World Health Organization (WHO) in 2018 [1]. The disease is singularly caused by persistent infection with certain oncogenic types of the. Single Lens Reflex (SLR) cameras), environmental factors (e.g., lighting, provider training), and imaging procedures (e.g., handheld smartphones, adapter mounted or colposcopeattached SLR cameras, and time of exposure after application of weak acetic acid) These image datasets differ with respect to image quality factors such as partial/full absence of the anatomical region of interest (cervix), or other factors such as illumination or focus, which could impact classification prediction performance. CRM takes both positive and negative contributions (i.e., contribution toward an increase in the output score for the desired class and a decrease in the scores for the other classes) for each element in the feature maps into consideration Note that these visualization methods have been developed mainly for the interpretation of deep learning networks, which assign a class label to an entire input image. To evaluate the performance of the two visualization techniques on our cervix images, we generate the heat maps by applying CRM and CAM methods on our best-performing classification model. The rest of the paper is organized as follows: Section 2 describes the details of datasets used in this study, Section 3 describes the network architecture and two visualization methods; Sections 4 and 5 present the experiments, the results and the conclusion
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
