Optimal Lossy Compression Scheme of 3D Volumetric Ultrasound Images For Remote Breast Cancer Screening

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Lossy compression is underutilized in medical imaging due to a lack of implementation guidelines.
• Remote breast cancer screening requires efficient medical image compression.

Purpose of the Study:

• To assess the feasibility and diagnostic reliability of lossy compression for 3D whole-breast ultrasound (WBUS) images.
• To establish guidelines for remote breast cancer screening in resource-limited settings.

Main Methods:

• Retrospective collection of WBUS images from 400 patients.
• Compression using JPEG 2000 (JP2K), JPEG 3D (JP3D), and High-Efficiency Video Coding (HEVC) at ratios from 5:1 to 50:1.
• Objective image quality assessment (PSNR, SSIM) and radiologist review for visual lossless quality and diagnostically acceptable compression ratio (DAS).

Main Results:

• HEVC demonstrated superior objective image quality and diagnostic accuracy compared to JP2K and JP3D.
• At 5:1 compression, images were visually indistinguishable from originals, with unaffected diagnostic performance (AUC > 0.85).
• Compression up to 20:1 maintained diagnostic accuracy, with high inter-reader agreement.

Conclusions:

• Lossy HEVC compression is a practical solution for remote breast cancer screening.
• Compression ratios of 5:1 were visually lossless, and up to 20:1 maintained diagnostic performance in WBUS images.
• Findings support the use of HEVC in tele-ultrasound workflows, warranting further validation.