Comprehensive study of good model training for prostate segmentation in volumetric MRI

TL;DR

This paper conducts an extensive analysis of training strategies for prostate segmentation in volumetric MRI, emphasizing data resolution effects and using established deep learning architectures.

Contribution

It uniquely focuses on training methodologies and data resolution impacts for prostate MRI segmentation, rather than proposing new architectures.

Findings

Resampling resolution significantly affects segmentation performance.

Optimal training configurations improve model accuracy.

Study provides guidelines for effective prostate MRI segmentation training.

Abstract

Prostate cancer was the third most common cancer in 2020 internationally, coming after breast cancer and lung cancer. Furthermore, in recent years prostate cancer has shown an increasing trend. According to clinical experience, if this problem is detected and treated early, there can be a high chance of survival for the patient. One task that helps diagnose prostate cancer is prostate segmentation from magnetic resonance imaging. Manual segmentation performed by clinical experts has its drawbacks such as: the high time and concentration required from observers; and inter- and intra-observer variability. This is why in recent years automatic approaches to segment a prostate based on convolutional neural networks have emerged. Many of them have novel proposed architectures. In this paper I make an exhaustive study of several deep learning models by adjusting them to the task of prostate prediction. I do not use novel architectures, but focus my work more on how to train the networks. My approach is based on a ResNext101 3D encoder and a Unet3D decoder. I provide a study of the importance of resolutions in resampling data, something that no one else has done before.