Task-based optimization of flip angle for fibrosis detection in T1-weighted MRI of liver

Area of Science:

• Medical Imaging
• Hepatology
• Biomedical Engineering

Background:

• Chronic liver disease is a global health concern, with hepatic fibrosis (HF) as a key indicator.
• Current HF diagnosis relies on liver biopsy, which has limitations like sampling errors and procedural risks.
• Pathological HF diagnosis involves identifying textural changes indicative of collagen deposition within liver lobules.

Purpose of the Study:

• To develop and validate a non-invasive method for detecting hepatic fibrosis using MRI.
• To optimize MRI pulse sequences for enhanced detection of HF-related textural changes.
• To train model observers for accurate HF detection based on MRI phantoms.

Main Methods:

• Utilized ex vivo formalin-fixed human liver samples as phantoms to mimic in vivo Gd-MRI textural contrast.
• Developed a local texture analysis algorithm applied to phantom MRI images.
• Trained model observers using phantom image analysis to detect HF, assessing performance with AUROC.
• Optimized MRI pulse sequences by varying flip angles to maximize AUROC for HF detection.

Main Results:

• A local texture analysis method was successfully developed and applied to MRI phantoms.
• Model observers trained on phantom data demonstrated the potential for detecting HF.
• Optimized MRI pulse sequences, determined by maximizing AUROC, enhance HF detection capabilities.
• The study established a correlation between MRI textural features and HF presence.

Conclusions:

• MRI-based texture analysis shows promise as a non-invasive tool for hepatic fibrosis detection.
• Optimized MRI pulse sequences can improve diagnostic accuracy for HF.
• This approach may offer a safer and more reliable alternative to liver biopsy for HF assessment.