Steve Ebdon-Jackson (HPA, UK)
Background
Nuclear Medicine imaging relies on the tracer principle first established in 1913 by Georg de Hevesy. It is used to demonstrate physiological processes and involves the administration of a small amount of a radioactive material to the patient which then distributes within the body and accumulates in particular areas or organs. The distribution depends upon the particular material administered, which is chosen depending on the organ of interest. The radiation (usually gamma rays) emitted by the radionuclide are detected and an image of the distribution of the radioactivity within the body is constructed.
A key challenge for nuclear medicine is to detect an adequate number of gamma rays in order to acquire an image that contains enough information for accurate diagnosis while keeping the radiation dose as low as reasonably practicable. This might be achieved by reducing the administered activity but imaging for a longer time. For procedures where the distribution is fixed, this is possible within the limits of the patients’ ability to lie completely still. Typically imaging times which exceed 30 minutes are prone to patient motion. This is not possible for some procedures where the distribution within the organ may be changing while we are imaging
The primary tool for nuclear medicine imaging is the gamma camera. The gamma rays emitted by the radionuclide are detected by a crystal and an image of the distribution of the radioactivity is built up. Because the gamma rays are emitted from the patient in all directions a collimator is used to acquire an accurate image of their distribution. Gamma rays which are not coming orthogonally from the patient are absorbed by the collimator and eliminated from the final image.
A collimator with small holes will provide better resolution but has lower sensitivity as it absorbs more of the emitted gamma rays. A collimator with larger holes is more sensitive (as it allows more gamma rays to be detected) but has poorer resolution. In nuclear medicine there is always a trade off between resolution and sensitivity. In practice the collimator choice depends upon the organ being imaged and the type of imaging.
Single Photon Emission Computed Tomography Imaging
As with other imaging modalities, it is possible to produce 2 dimensional planar images or, to use planar images acquired from a range of angles to reconstruct a full 3 dimensional distribution. In conventional nuclear medicine this is known as Single Photon Emission Computed Tomography (SPECT) imaging.
All SPECT reconstruction techniques have limitations. There are issues with:
- Attenuation (gamma rays lost due to absorption in the patient)
- Scatter (gamma rays are scattered within the patient before detection)
- Resolution (becomes poorer with increasing distance from patient to camera)
- Noise (becomes higher with reduced counts)
- Computation time (significant for accurate methodology)
In practice, three reconstruction methods have been used:
- Filtered back projection has been the standard approach for many years. It is fast but amplifies noise and attenuation and scatter corrections produce errors.
- 2-D iterative reconstruction techniques are now available with greater computer power and attenuation and scatter correction have become possible. These techniques are slower than filtered back projection, reconstructing each slice separately, but they deal with noise effectively.
- 3-D iterative reconstruction is now available where all slices are reconstructed together. This is even slower than the 2-D approach but has the advantage that an additional correction can be made for the variation of resolution with depth within the patient - resolution recovery.
Resolution Recovery – Implications for ALARA
All manufacturers now offer resolution recovery software packages. These are gamma camera, collimator and procedure specific. Generic products are also available. Each would need to be validated against conventional techniques. If expected performance is verified, resolution recovery software should be able to change the current balance between image quality, administered activity and scan time.
In most cases these products were developed and marketed with the intention that image quality would be maintained or improved, administered activities would remain unchanged and scan times would be reduced thus improving the efficiency and cost effectiveness of the nuclear medicine service. These products may however offer the potential to maintain image quality and scan times while reducing the administered activity to the patient. This has a positive impact on patient dose but coincidently may also help nuclear medicine services use available 99mTc more effectively, ensuring that costs are reduced and procedures undertaken as required.
A number of concerns and unknowns exist about the routine use of resolution recovery software. Published patient studies have concentrated on the “unchanged activity/ decreased imaging time” approach and moving towards the “decreased activity/routine imaging time” paradigm will require national and local validation.
Pilot Study – Use of Resolution Recovery in Myocardial Perfusion Imaging
To go some way towards addressing these issues, the Administration of Radioactive Substances Advisory Committee (ARSAC), a statutory advisory committee, set up a sub-group in collaboration with the Institute of Physics and Engineering in Medicine (IPEM) Nuclear Medicine Special Interest Group (NMSIG) and Software Validation Working Party. The aims of this group were:
- To establish what products are available, how they work, what is required in order to use them and what costs are associated with each product
- To evaluate at least one of the available products, as a pilot study, to establish whether RR can maintain or improve image quality, and hence image interpretation, and compensate for a reduction in administered activity, when compared to conventional imaging protocols
- To develop a wider study protocol for use in validation of a full range of products.
The pilot study considered myocardial perfusion imaging (MPI), the second most common nuclear medicine procedure in the UK. It is a high dose procedure which offers the potential for significant dose reduction. The diagnostic reference level DRL) for MPI with 99mTc is 1600 MBq (for patients who have both stress and rest components of the study). The principal objectives of the pilot study were:
- To determine whether the interpretation of images obtained with half the normal administered activity and processed with resolution recovery software can be the same as the interpretation from that obtained with normal activity and processed in the standard way
- To determine whether objective quantitative parameters calculated from gated images obtained with half the normal administered activity and processed with resolution recovery software are the same as those obtained with normal activity and processed in the standard way.
The pilot study was carried out using GE Evolution for Cardiac resolution recovery software in the Central Manchester Nuclear Medicine Centre.
Results
The study involved rest and stress data from 44 patients. Each patient was administered the routine activity (1600MBq in total) and gated images acquired but data from the studies were collected, stored and processed to enable full count data to be compared to half count data with resolution recovery applied ie a study using half the administered activity was simulated.
Double reporting resulted in only 2 of the 44 cases having a clinically significant report and of these only one resulted in different patient management. Quantitative left ventricular function analysis showed no significant difference in the LVEF values calculated from the full-count and half-count data at both stress and rest. Further details of the pilot study are included in a report by the ARSAC on the impact of 99mMo shortages on nuclear medicine services, published in November 2010 (www.arsac.org.uk).
Summary and Conclusions
Resolution recovery software was developed to reduce imaging time in busy nuclear medicine departments. Taking an ALARA perspective, this software may be used instead to reduce administered activity and hence patient dose while keeping scan times the same.
Initial results are promising and in MPI show that this approach produces images of accepted image quality from half the administered activity.. Further work will be required to validate this at a local level, for a range of procedures, equipment and software combinations.