The high-resolution requirements of medical image modalities such as mammography delay the implementation of digital imaging systems. Increases in both processing and transmission time and storage capacity cost are associated with these images. These
The high-resolution requirements of medical image modalities such as mammography delay the implementation of digital imaging systems. Increases in both processing and transmission time and storage capacity cost are associated with these images. These demands have also hindered the use of computers as a second opinion in the detection of breast cancer.
While compression may provide a solution to this problem, clinical evaluations have yet to validate whether lossy compression can be applied to digital mammograms before interpretation.
A new study (IEEE Trans Med Imaging 2003;22[10]:1288-1296) describes a wavelet compression scheme that can reduce the amount of data, and thus storage costs and transmission times, without significantly degrading image quality.
"It is clear that technological advances in storage and transmission are not sufficient to solve the problem," said lead author Monica Penedo-Ojea, Ph.D., a professor of radiology at University of Santiago de Compostela in Spain.
Penedo-Ojea presented an adaptation for high-resolution digital mammography of two region of interest (ROI) wavelet-based compression methods.
"Our region-based methods code only the breast region where the most important information is contained," she said.
Identifying those regions can be done by radiologists or by computer-aided diagnosis.
In the study, those ROI techniques have been evaluated against state-of-the-art compression methods, such as SPIHT and JPEG 2000, by means of PSNR (peak signal-to-noise ratio), a widely used parameter for measuring reconstruction quality of compressed images.
Penedo-Ojea found that region-based techniques performed substantially better than conventional SPIHT (set partitioning in hierarchical trees) and JPEG 2000, showing much higher quality in the breast region for the same compressed file size.
"For digital mammography, region-based compression methods represent an improvement in compression efficiency from full-image methods, also providing the possibility of encoding multiple regions of interest independently," she said.
Regions inside the mammogram containing primary radiological signs of breast cancer can be determined by an expert radiologist or a CAD system as a preliminary step in a compression scheme for digital mammography. These areas can then be compressed at different ratios from the rest of the breast. The image background can be discarded.
"ROI coding techniques are particularly suitable for digital mammography and, generally, for medical imaging," Penedo-Ojea said.
Earlier studies applying lossy compression methods to digital mammography have determined that loss of information does not affect diagnostic interpretation when compression rates are limited to certain ranges. Compression in these studies had been applied to the least rectangular area containing the breast region.
Study: AI Bolsters Sensitivity for Pneumothorax on CXR and Significantly Reduces Reporting Time
October 30th 2024For clinically actionable pneumothorax, an artificial intelligence algorithm demonstrated a 93 percent AUC and a 96 percent specificity rate in a study involving chest X-rays from over 27,000 adults.
AI Segmentation, Intraprostatic Tumor Volume and Metastases: What a New mpMRI Study Reveals
October 29th 2024For patients who received radiotherapy for localized prostate cancer, total intraprostatic tumor volume derived from AI segmentation had a 10 percent higher AUC for predicting seven-year metastases in comparison to a risk model from the National Comprehensive Cancer Network (NCCN).