Publication Details
Overview
 
 
Jakub Ceranka, Joris Wuts, Ophélye Chiabai, Frédéric Lecouvet, Jef Vandemeulebroucke
 

Contribution to journal

Abstract 

The confident detection of metastatic bone disease is essential to improve patients{\textquoteright} comfort andincrease life expectancy. Multi-parametric magnetic resonance imaging (MRI) has been successfullyused for monitoring of metastatic bone disease, allowing for comprehensive and holistic evaluationof the total tumour volume and treatment response assessment. The major challenges of radiologicalreading of whole-body MRI come from the amount of data to be reviewed and the scattered distri-bution of metastases, often of complex shapes. This makes bone lesion detection and quantificationdemanding for a radiologist and prone to error. Additionally, whole-body MRI are often corruptedwith multiple spatial and intensity distortions, which further degrade the performance of imagereading and image processing algorithms. In this work we propose a fully automated computer-aideddiagnosis system for the detection and segmentation of metastatic bone disease using whole-bodymulti-parametric MRI. The system consists of an extensive image preprocessing pipeline aiming atenhancing the image quality, followed by a deep learning framework for detection and segmentationof metastatic bone disease. The system outperformed state-of-the-art methodologies, achieving adetection sensitivity of 63% with a mean of 6.44 false positives per image, and an average lesionDice coefficient of 0.53. A provided ablation study performed to investigate the relative importanceof image preprocessing shows that introduction of region of interest mask and spatial registration havea significant impact on detection and segmentation performance in whole-body MRI. The proposedcomputer-aided diagnosis system allows for automatic quantification of disease infiltration and couldprovide a valuable tool during radiological examination of whole-body MRI.

Reference