Due to overwhelming demand, the free web interface for metal artifact reduction has been disabled.

You can get a free copy of the MDT software (for internal non-commercial use). The software can be integrated with PACS: studies are sent to your server from PACS or from the scanner, and the processed images are automatically sent back to PACS as a new series under the same accession number. First, review the software manual. Next, complete and sign the licensing agreement and return it by email to revisionrads@gmail.com. Commercial licensing terms are available upon request.

How MDT works

We start by forward projecting the reconstructed image, to get an estimate of the raw data from the scanner. Next, we use all of the projection data to reconstruct metal pixels, but only use the high quality non-metal data to reconstruct bone and soft tissue. This means that, instead of trying to look through the metal to see soft tissue, we look around the metal. We forward project the reconstructed image (with the metal pixels deleted), and use this to replace the inaccurate metal data. We perform this iteratively to improve the results.

In the diagram below, metal pixels and projection data are shown in red. The original projection data are plotted with a thick line, and revised projection data are plotted with a thin line.

Features

1. MDT works with DICOM files from any CT scanner.
2. MDT works with a wide range of small and large metal implants, including: orthopedic implants, dental fillings, surgical clips, moving pacer wires, and embolization coils. It works with multiple implants. It reduces artifacts caused by beam hardening, motion, scatter, and photon starvation.
3. Reducing metal artifacts with MDT can change the diagnosis.

In a blinded comparison against 3 other methods, two radiologists judged MDT to have the best image quality in 11 out of 11 scans (with a total of 178 slices containing metal) [1]. This is the first metal artifact reduction technique shown to have statistically significantly better image quality than another technique for clinical scans. In 2 of 11 scans, the improved image quality actually changed the diagnosis. This includes a case of rectal cancer that was missed with the current technology. These are the first reported cases of a metal artifact reduction algorithm changing the diagnosis. This study used raw projection data from the CT scanner.

A follow up study of 80 patients showed that MDT improved image quality 73% of the time for small metal implants, and 75% of the time for large metal implants [3]. MDT had better image quality than 3 other metal artifact reduction techniques. This study used projection data that was simulated from images reconstructed by the scanner.

More recent data shows that in 14% of scans, MDT changed the diagnosis, improved visualization of key findings, or improved diagnostic confidence (presented at RSNA 2012). An analysis by an independent research group showed that MDT outperformed commercial metal artifact reduction techniques from GE and Siemens [4].

Tips for getting the best results

1. Best results are obtained with small stationary implants: hip replacements, aneurysm coils, surgical clips, and dental fillings.
2. Suboptimal results (loss of resolution) are obtained with large implants: pedicle screws and complex orthopedic hardware. In general, if the feature of interest can only be seen by looking through metal, then MDT tends to blur it out.
3. In some cases, MDT decreases resolution or introduces new artifacts. Thus, you should always review MDT images in conjunction with the original images produced by the scanner.
4. You must upload DICOM files. JPEG, GIF, TIFF, or PNG files will not work, even if you convert them to DICOM (these file formats only allow 256 grayscale levels, and they do not specify Hounsfield units).
5. To reduce artifacts on coronal or sagittal images, you should perform MDT on axial slices, then reformat into the desired plane.
6. The metal causing the streaks must be visible on the image. A small field of view with the metal off the edge of the image won't work.
7. MDT uses a fixed Hounsfield unit cutoff of 3000 to detect metal. If the Hounsfield units are less than 3000 due to partial volume effects on small implants, then try thinner slices.

Selected references

1. Boas FE and Fleischmann D. (2011) "Evaluation of two iterative techniques for reducing metal artifacts in computed tomography." Radiology. 259(3): 894-902.

2. Boas FE and Fleischmann D. (2012) "Computed tomography artifacts: Causes and reduction techniques." Imaging in Medicine. 4(2): 229-40.

3. Golden C, Mazin SR, Boas FE, Tye G, Ghanouni P, Gold G, Sofilos M, Pelc NJ. (2011) "A comparison of four algorithms for metal artifact reduction in CT imaging." Proc. SPIE. 7961: 79612Y.

4. Wagenaar D, van der Graaf ER, van der Schaaf A, Greuter MJW. (2015) "Quantitative comparison of commercial and non-commercial metal artifact reduction techniques in computed tomography." PLOS One. 10(6): e0127932.

Talks

Updated slides (2013)

Abstract and slides from RSNA (11/26/2012).

Video and slides from Stanford radiology grand rounds (4/3/2012).

Slides from the International Society for Computed Tomography meeting (8/11/2011).

Video from RSNA (11/29/2010)