Iterative reconstruction v filtered backprojection

Iterative reconstruction (IR) is the primary method of reconstruction of SPECT data. However there are some drawbacks to using this process.

IR utilizes the collected data and breaks it down into pixels and then treat each pixel value as an unknown. Using an algorithm values can be assigned to each pixel, permitting the emission and detection data to be accurately modeled.

Noise acquired during the SPECT acquisition, in the absence of attenuation correction, is constant across the reconstructed field. This eliminates streaking artifact in the image and provides a visually superior image to filtered back projection (FB) processing.

The above image of a bone whole body spect, processed with iterative reconstruction shows a large photopenic area where the femurs should be. Using IR the pixel values of a low count area may be treated as background or noise in comparison to the adjoining pixels...

Filtered backprojection, takes the acquired data and fills in the spaces between the projections (azimuths) to create a completed 360-degree image. Although this completes the image it does create blurring, to which filters can be applied for an improved appearance. Unlike IR processing, FB does contain noise and background activity..

Recprocessing the acquired data with filtered back projection shows a complete if not more favorable image. Note the noise or "rays" of activity surrounding the scan, especially around the bladder activity...

References:
1. Image reconstruction - a tutorial, G.L. Zeng, Computerized Medical Imaging and Graphics 2001
2. An introduction to iterative reconstruction, B Hutton
3. Filtering in SPECT image reconstruction, M. Lyra and A. Ploussi, International Journal of Biomedical Imaging,  April 2011

WB Skeletal Imaging - Contamination

Clinical History: 73 year old female with low back pain.

Technique: 3phase and whole body bone scan performed following injection of 23 mCi MDP.

Findings: The dynamic blood flow study and immediate post injection scan of the lower back demonstrate no evidence of a focally increased blood flow of increased vascularity. Bone imagings from the head to toes demonstrate a slightly increased tracer uptake involving the lower lumbar spine at the level of L4 and L5, probably due to degenerative bone and joint disease. Slightly increased tracer uptake involving both shoulder joints, probably non-specific arthritis. No abnormal tracer uptake is noted in the sacrum and coccyx.

Slightly increased tracer uptake involving patient's boots. The patient was on lasix, and urinated through her pantyhose. She went home and changed her hose prior to imaging. She did not change her boots, therefore only the boots show contamination and not her hosiery.

Thanks to Y.H. Park, MD, Meredith Smela, Debbie Reed

Bone Scan - Post Reclast Therapy

Reclast is marketed by Novartis, was was approved by the FDA for treatment of Paget's Disease in 2007. Administered by infusion, the main ingredient, zoledronioc acid, is also marketed by Novartis as Zometa for use in the treatment of skeletal fractures in patients with cancers such as multiple myeloma and prostate. It has also been used to treat osteoporosis. For the treatment of Paget's patients recieve a single dose, annually. Results of studies for Reclast were published in the Sep 1, 2005 issue of the New England Journal of Medicine.

The cause of Paget's Disease is unknown. The disease causes an abnormal breakdown of bone tissue, followed by abnormal bone formation. The new bone is enlarged but weaker. The disease often effects the bones of the arms, collarbones, legs, pelvis, spine and skull.

2011

2013

The above patient studies were obtained at 3 hours following the injection of 25 mCi Tc MDP. Whole body bone scans with post void images of the pelvis were obtained. The study from 2013 demonstrates marked decreased activity in the right hemipelvis, following Reclast therapy, as compared to the study from 2011. 

Whole Body Bone SPECT

Recently we starting looking at whole body bone spect imaging for metastatic evaluation. Metastatic disease in the vertebral bodies can be difficult to detect on planar whole body imaging, and an abnormal finding does not always mean metastatic disease. It is also difficult to detect small abnormalities in the thoracic vertebrae.1 

In order to acquire a whole body bone spect in an acceptable time frame, the time per bed has to be reduced. Studies were performed and determined that a bone spect can be acquired in as little as 5 minutes and still be acceptable as compared to planar imaging.2

Filtered back projection - the most commonly used method of image reconstruction. the back projection assigns values in the projection to all points along the line of acquisition through the image plane from which it is acquired. This can create an star artifact  consisting of radial lines at the edge of the object, to which a filter is then applied to reduce this artifact.3

Iterative reconstruction - is a more computer intensive program, which uses the original projections and models of the acquisition to predict areconstruction. The predicted projections are used to modify the recontstruction. In the OSEM method the projection data are ordered into subsets. The advantage of this is an increase in computer processing speed.3

Iterative reconstruction has proven to be superior to count poor data, which would create excessive noise from the radial lines when filtered back projection is applied. Our data was acquired at 6 secs per view, 64 views for a 12 minute acquisition time, compared to our standard SPECT acquisition time of 40 minutes.

When the initial study was reviewed it was noted that some problems existed. The first bed (patient head) looked good but the rest of the body offered excessive background noise, especially around the bladder. It was discovered that the iterative reconstruction was applied to the first bed, while filtered back projection had been applied to the remainder of the study.

Whole body coronal slices processed with filtered back projection

After re-processing the data with iterative reconstruction applied to all of the beds we noted another problem. The right half of the skeleton has a brighter, crisper look. It was then discovered that the data selector for processing was for 180 degrees. 

Reprocessed coronal slices with iterative reconstruction, but 180 degree processing

We adjusted the processing protocol to used the 360 degree method and re-processed the data again. This time the study came out with excellent results. (I never knew the system could process half a study?!)

Processing page, applying 5 separate acquisitions into one study

References:

1 - Do short-time SPECT images of bone scintigraphy improve the diagnostic value in the evaluation of solitary lesions in the thoracic spine in patients with extra-skeletal malignancies? 

Kobayashi, et al. Annals of Nuclear Medcine vol 19 No 7, p557-566, 2005.

2 - What is the optimal minimum whole body bone SPECT scan duration that can replace planar scintigraphy? 

Mawlawi, et al, The Journal of Nuclear Medicine vol 48 supplement 2, p120, 2007.

3 - Nuclear Medicine and PET/CT, technology and Techniques.

Christian and Watersum-Rich, Mosby Elevier

Skeletal Imaging - Scapula lesion

Skeletal Imaging - positioning for a scapula lesion.

  

Clinical history: Female with a history of breast cancer. Bone scan requested to rule out metastases. Patient complained of left shoulder pain.

Technique: The patient received 25 mCi Tc MDP and whole body and spot shots were obtained at 3 hours post injection.

Findings: The study was compared to a previous bone scan acquired 1-year earlier. Findings demonstrate an area of increased uptake at the inferior angle of the left scapula. The finding was considered to be a stable metastatic lesion that was also seen on a CT and described as mixed sclerotic and lytic lesion.

No other abnormalities were noted..

Posterior whole body view shown.

Additional images were acquired in the left anterior oblique view, to demonstrate the lesion was truly in the scapula, not in the ribs. In this case the lesion still overlies the ribs making true determination difficult.

An additional posterior view was obtained with the patients arms raised over her head. This movement moves the scapula outward providing separation of the scapula from the posterior rib cage.