Abstract:Total Kidney Volume (TKV) is essential for analyzing the progressive loss of renal function in Autosomal Dominant Polycystic Kidney Disease (ADPKD). Conventionally, to measure TKV from medical images, a radiologist needs to localize and segment the kidneys by defining and delineating the kidney’s boundary slice by slice. However, kidney localization is a time-consuming and challenging task considering the unstructured medical images from big data such as Contrast-enhanced Computed Tomography (CCT). This study aimed to design an automatic localization model of ADPKD using Artificial Intelligence. A robust detection model using CCT images, image preprocessing, and Single Shot Detector (SSD) Inception V2 Deep Learning (DL) model is designed here. The model is trained and evaluated with 110 CCT images that comprise 10,078 slices. The experimental results showed that our derived detection model outperformed other DL detectors in terms of Average Precision (AP) and mean Average Precision (mAP). We achieved mAP = 94% for image-wise testing and mAP = 82% for subject-wise testing, when threshold on Intersection over Union (IoU) = 0.5. This study proves that our derived automatic detection model can assist radiologist in locating and classifying the ADPKD kidneys precisely and rapidly in order to improve the segmentation task and TKV calculation.

What problem does this paper attempt to address?

This paper attempts to address the problem of automatically locating Autosomal Dominant Polycystic Kidney Disease (ADPKD) in contrast-enhanced computed tomography (CCT) images. Specifically, the paper aims to design a method that utilizes artificial intelligence technology to automatically detect and classify ADPKD kidneys, in order to improve the accuracy and efficiency of Total Kidney Volume (TKV) measurement. ### Background and Problem 1. **Diagnosis Needs of ADPKD**: - Autosomal Dominant Polycystic Kidney Disease (ADPKD) is one of the most common hereditary kidney diseases, with a prevalence of approximately 1:1000 to 1:2500. - ADPKD is characterized by the continuous proliferation and growth of cysts in both kidneys, leading to compression and damage of healthy kidney tissue, eventually progressing to end-stage renal disease. - To predict the gradual loss of kidney function, it is necessary to examine two important biomarkers: Glomerular Filtration Rate (GFR) and Total Kidney Volume (TKV). 2. **Limitations of Traditional Methods**: - Traditional TKV measurement methods require radiologists to manually locate and segment the kidneys, which is a time-consuming and challenging task, especially when dealing with large amounts of unstructured medical imaging data. - In CCT images, due to the contrast differences between the kidneys and surrounding structures, manual localization and segmentation of ADPKD kidneys are particularly difficult. ### Paper Objective - Design a deep learning-based automatic localization model for detecting and classifying ADPKD kidneys in CCT images. - By improving the accuracy and speed of localization and classification, assist radiologists in more efficiently performing kidney segmentation and TKV calculation. ### Methods and Results - **Dataset**: Used 110 CCT images, containing 10,078 slices. - **Model**: Employed the Single Shot Detector (SSD) Inception V2 deep learning model. - **Experimental Results**: - In image-level testing, the model achieved a mean Average Precision (mAP) of 94%. - In subject-level testing, when the Intersection over Union (IoU) threshold was 0.5, the mAP reached 82%. ### Conclusion - This study demonstrates that the proposed automatic detection model can accurately and quickly locate and classify ADPKD kidneys, thereby improving the efficiency of kidney segmentation tasks and TKV calculation.

Deep Learning Assisted Localization of Polycystic Kidney on Contrast-Enhanced CT Images