G Currie, J Wheat, R Pearce
G Currie, J Wheat, R Pearce. Planar Image Fusion. The Internet Journal of Nuclear Medicine. 2006 Volume 4 Number 1.
Image fusion of SPECT and PET data has been shown to improve diagnostic accuracy in a variety of clinical circumstances. While the use of SPECT/CT and PET/CT image fusion has been well documented in the literature, the potential role of planar image fusion is not well documented. This brief report presents several circumstances in which planar image fusion might be useful in clinical practice.
Physiological imaging using single photon emission computed tomography (SPECT) is a valuable tool in the diagnosis of many diseases but may be limited, on occasion, by poor anatomic localization (1). This limitation might be overcome using image fusion, a process of data superimposition using multiple image types; typically computed tomography (CT) or magnetic resonance imaging (MRI) for anatomic data superimposed with SPECT or positron emission tomography (PET) for functional data (1). It is not uncommon for the combined image fusion data to provide additional clinical information that was not evident from either anatomic or functional datasets; interpreted in isolation or in tandem without fusion.
Image fusion of SPECT and PET data has been shown to improve diagnostic accuracy in many clinical circumstances (2). There are two main methods of performing SPECT/CT image fusion. The first employs hybrid scanner technology where a CT scanner is integrated in the gamma camera gantry to allow sequential imaging of anatomy and physiology. An alternative approach would be to perform image fusion on data acquired on individual dedicated SPECT and CT systems.
While the use of SPECT/CT and PET/CT image fusion has been well documented in the literature, the potential role of planar image fusion (e.g. planar scintigraphy and radiograph) is not well documented. Merge version 2.0 is a fully functional freeware program readily downloadable over the internet. Merge 2.0 is a simple graphic utility designed to merge graphic images, including batch merging to a reference image. While these features were initially conceived for manipulation of graphics and watermarking images more easily, they readily lend themselves to merging planar scintigraphy with radiographs and applying an anatomic reference image to a dynamic scintigraphic data set. This brief report presents several circumstances in which planar image fusion might be useful.
Scintigraphy to Xray
Perhaps the most useful application of planar image fusion is the merging of bone scans with xrays. Combining the high spatial resolution of xrays with the physiologically sensitive bone scan assist on overcoming the limitations of each. Figure 1 presents a 14 year old male with a history of persistent tenderness of the left wrist following a fall (three weeks). The xray was reported as normal. The bone scan demonstrated increased tracer accumulation extending across the distal growth-plate of the left radius as well as the area surrounding the distal portion of the left scaphoid.
Scintigraphy to Scintigraphy
It is not uncommon for MIBG and DTPA to be summed for better localization of adrenal glands. Summation of DTPA and MIBG images requires rigorous adherence to protocol. On the final day of imaging and after all other images have been collected, a posterior lumbar spine image is typically acquired for MIBG. Subsequently and without the patient moving, DTPA is administered and an identical image acquired. As a result, the DTPA image is only summed with a single image.
Image fusion would permit DTPA to be administered more conveniently. Indeed, the patient may be administered DMSA and return hours later for imaging to avoid the confounding sometimes associated with DTPA clearance. The DTPA or DMSA image need not be acquired on the same gamma camera or even using the same acquisition parameters (e.g. matrix) (Figure 2). Moreover, DTPA or DMSA images may be acquired in multiple projections and merged with corresponding MIBG projections acquired, not just in the final imaging session, but across all imaging session. This might include the merging the DTPA or DMSA planar image with the whole body MIBG sweep (Figure 3).
Similar applications might include, without being limited to: merging blood flow, blood pool and delayed bone images, merging whole body gallium, whole body thallium and whole body bone images, merging previous scans with current scans, merging lung ventilation and perfusion studies, merging white blood cell studies with bone or gallium studies.
Scintigraphy to Photograph
On occasion it is difficult to conceptualise the imaged biodistribution to the original object. While the patient might be standing before us, connecting an area of increased tracer accumulation directly to a point of pain or discomfort may be imperfect. Merging planar scintigraphy with a photograph of the patient, however, may provide more of a promotional tool for referring doctors than a tool for improving diagnostic utility. Certainly if colour images are visually appealing to referring doctors, then having a photograph of their patient merged with the scintigraphy study (localized static or whole body) might increase referrals. More importantly, any number of research studies performed in Nuclear Medicine departments might be assisted by fusion of planar images with photographs. This might be particularly useful if the research involves specialists from outside the Nuclear Medicine domain (Figure 4).
A key limitation of merging scintigraphy with photographs is the ‘pin hole' effect. Like the pinhole collimator, the photograph may distort (parallax error) towards the edge of images. While not generally perceptible to the eye in photographs, high resolution digital photographs taken close to an object or of large objects may be obviously misaligned post fusion. This is despite every effort to ensure the photograph is taken in the same plane as the scintigraphic image.
Planar image fusion may provide a useful tool to improve diagnostic utility and / or interpretive certainty. Planar image fusion is easy to perform with cost neutrality. A number of fusion solutions may provide superior results or greater flexibility than traditional ‘summation' of data sets.
Geoff Currie School of Biomedical Sciences Locked Bag 588 Charles Sturt University Wagga Wagga 2678 Australia Telephone: 61 2 69332822 Facsimile: 61 2 69332587 Email: email@example.com