ECRI Institute CT/PET Technology

---

Technology

Introduction

PET is unequalled in its ability to stage and monitor a number of cancers. Over the last 5 years Positron Emission Tomography (PET) has rapidly moved from the research domain into accepted clinical practice. According to the market research company IMV, the number of PET studies in the USA increased by 58% in 2003 {IMV, 2004). However, PET is not new, having been under development since the mid 1970s. Two events prompted the sudden and dramatic rise in interest: the development of a combined PET/CT scanner developed by Siemens Medical Solutions (Erlangen, Germany) and CTI Inc. (Nashville, USA) and the wider provision of reimbursement in the USA. The development of PET/CT has added to the usefulness of PET.

The rapid growth in the USA can be attributed to the highly competitive nature of the healthcare business. The same data from IMV shows that there are approximately 3.6 PET scanners per million of population in the USA. Most scanners sold in the USA today are hybrid scanners. The culture of healthcare provision in Europe is very different, with central governments controlling expenditure rather than competing independent hospitals. This led to significant discrepancies in the availability of PET imaging throughout Western Europe. A recently published survey (Dietlein and Schicha 2003) found between 0.1 and 2.5 PET scanners per million in the UK and Austria respectively. Politics and payment structures have more to do with these differences rather than any fundamental differences in clinical need. In the past some have questioned the need for such expensive technology. However, the clinical value of PET is now widely recognized and, increasingly, the value of hybrid scanners. So, it is vitally important that hospital decision makers understand the technical issues of PET imaging as its use is likely to grow significantly in Europe.

So what makes PET so desirable and what further developments can be expected?

PET Technology

Tracers

The biologic specificity of tracers is key to all nuclear medicine studies. Most tracers used in nuclear medicine rely on high atomic number isotopes. These isotopes emit gamma rays that are detected by a gamma camera. Synthesizing these isotopes into molecules that preferentially target specific anatomical areas is the real technology in nuclear medicine. Unfortunately, the high atomic number means that the atoms are large and difficult to synthesize into metabolically active analogues. Also, the gamma cameras require heavy collimators to distinguish spatial information. Conversely, positron emitting isotopes tend to be low atomic number and, very importantly, cause two gamma photons to be emitted in opposite directions almost exactly 180 degrees apart following the annihilation of the positron. Therefore, if two events are recorded coincidentally, then by connecting those two events by a straight line it is reasonable to assume that the anatomical region of interest is somewhere along that line. Heavy collimators are not required and the isotopes can be more readily synthesized into metabolic analogues.

The most commonly used tracer is fluorine-18 synthesized into a glucose analogue known as 2-fluoro-2-deoxyglucose, or simply FDG. Being a glucose analogue it collects in areas of high metabolism, such as tumours, the heart, and brain. The ability to find and gauge the activity in metastases is FDG’s most useful characteristic. For a positron emitter fluorine-18 has a relatively long half life (~2 hours), so it has time to be transported from a cyclotron to an imaging centre, synthesized into FDG and administered to a patient.

Other positron isotopes, such as oxygen-15, have much shorter half lives so cannot be easily used in a clinical setting. Therefore, fluorine-18 is likely to remain the mainstay in PET imaging for sometime, though research continues.

Crystals

Although it is possible to modify a standard gamma camera for positron coincidence imaging, better image quality can be achieved with a detector specifically designed for positron imaging. Perhaps the single most important component of any PET scanner is the crystal.

The crystal is a scintillator that converts gamma photons into visible light. Ideally the crystal should have a high detection efficiency (absorption), high light output, narrow energy resolution, short decay time and be easy to manufacture. In dedicated PET scanners today three crystal types are used: bismuth germinate (BGO), gadolinium orthosilicate (GSO) and lutetium orthosilicate (LSO). Another material that has excited the interest of PET designers recently is lanthanum bromide (LaBr3). The pertinent characteristics are listed in Table 1. This data shows that no one crystal has the perfect overall characteristics. For example, the crystal with the highest absorption efficiency (BGO) has the longest decay time. Overall, LSO is generally considered to have the most attractive properties of the crystals presently available. However, acceptable image quality can be achieved with any of the crystals listed. In PET imaging, image quality is highly dependent on exam time, so the required exam time is different for each crystal.

The excitement in LaBr3 is due to its short decay time, narrow energy resolution and high light output. These characteristics compensate for the lower absorption efficiency. A recent report based on computer simulations found that LaBr3 theoretically produces similar overall image quality to LSO (Surti et al. 2004). Of particular interest is the possibility that LaBr3 could be used for time of flight imaging. Time of flight makes use of the small difference in time between the detection of the two photons emitted from the same positron annihilation event. The time difference is simply due to the different distances traveled (light travels about 30 cm in 1 ns). This is made possible due to the short decay time of LaBr3. Using time of flight allows the image quality to be improved without increasing the overall exam time, which is normally the case in PET studies. Image quality in PET can be a problem with larger patients. So, LaBr3 could be the crystal of choice in the future. However, the manufacturer, Saint Gobain (Paris, France), has indicated some problems with its manufacture. Despite this, it is expected that LaBr3 will have lower manufacturing costs on account of its lower melting point.

Table 1

PET crystal properties (Surti et al. 2004)

Detection electronics

The light emitted from the crystal is usually detected with an array of photomultiplier tubes (PMT). How these PMTs are arranged with respect to the crystals depends on the characteristics of the crystal. Manufacturers take very different approaches to the design. This makes it very difficult to compare specifications between manufacturers. The most important consideration is the spatial resolution achieved. The spatial resolution specifications are similar between each manufacturer.

One question that has been asked is whether it is possible to change the type of crystal in a PET scanner. The simple answer is no since the crystal and detection electronics must all be designed together. Once the crystal and electronics have been removed, then there is not much left to the PET scanner.

PET + CT Hybridization

When the first commercial PET/CT was shown in 2000 it elicited two common reactions: 1) Wow! Just what is needed or 2) Why? It is a very expensive imaging system with a relatively small number of patients that could benefit. Since then PET/CT scanners have come to dominate the PET marketplace. This is despite a price tag of about $3 million. (€ 2.3 million). The earliest PET/CT systems incorporated a high specification PET scanner with a low specification CT scanner. Today, it is possible to have the latest 64 slice CT scanner with a PET. So, what are the attractions of hybridization?

Image fusion was the primary driver for hybridization. Normal PET images are sparse in terms of anatomical landmarks. Although the physics of PET imaging limits the resolution to about 5 to 6 mm, it is still necessary to be as precise as possible. The location can have particular significance on a patient’s prognosis and treatment choices. However, since PET is primarily used to stage and monitor treatment progress and not to diagnose, almost without exception every patient having a PET study would have already had a CT study earlier. Software exists to fuse images taken on different scanners. The fusion algorithms available today are fast and can allow for small amounts of patient movement. However, other problems exist such as patient positioning. For example, it is common for patients to have their arms above their heads during a CT exam. A PET study is much longer and patients cannot comfortably hold their arms in such a position. Also, CT exams are usually carried out with the patient holding their breath, which is simply not possible in a PET study. Hybrid scanners should ameliorate this problem.

Very few clinical users have had the opportunity to quantify the clinical impact hybridization. In most studies the fused results from separate scanners are adequate. For example, Lemke et al. (2004) using separate CT and PET scanners found that fusion was successful 96% of the time. Reinartz et al. (2004), again using separate scanners, estimated that 6.7% of patients may have benefited from a combined scanner. However, it is hard to surpass the almost 100% success record that would be expected with a hybrid scanner. Another factor to bear in mind is that future PET tracers may be so specific that they will provide almost no anatomical information. If this happens, then hybrid systems will be essential.

Hybridization has one other significant advantage over a standalone PET. Studies with a hybrid system can be significantly faster, so more patients can be scanned. Why is this? To achieve acceptable image quality it is necessary to correct for the attenuation caused by the patient’s anatomy, so called attenuation correction. This is normally achieved with a long half life gamma ray source that is rotated around the patient to generate low quality CT like images. This process can typically take 2 to 3 minutes on top of the 5 minutes for the PET scan at each bed position. Most PET scanners can scan  ~15 cm at each position, so a whole body scan requires the scan to be repeated about 6 times to cover the full anatomy. Adding these together means the exam time is about 30 minutes for the PET and 15 minutes for the attenuation correction. In a hybrid scanner the CT scanner can be used to obtain the attenuation correction data. Therefore the total exam time can be significantly reduced and more patients scanned. It is important to note that exam time depends on other factors such as the patient’s size and the performance of the PET scanner.

Other restrictions, such as the logistics of tracer supply, mean that if patient numbers are expected to exceed 4 to 5 a day then a hybrid scanner is required. A hybrid scanner should be able to comfortably scan 12 to 14 patients per day. Multiple deliveries of the FDG and adequate patient handling facilities would also be required.

Cost Considerations

A review of the ECRI’s database shows that the prices paid by hospitals for hybrid PET/CT systems is in the region of $2 to $2.5 million (€ 1.53 to € 1.9 million) compared to $1 to $1.3 million (€766,000 to €996,000) for standalone PET scanners. The type of CT scanner attached to the PET has a significant bearing on the price. Despite this, most interest of ECRI’s customers has been for 16 slice CT scanners rather than a more modest number of slices. The number of slices will not significantly affect the exam time or clinical utility of the PET/CT in oncology applications. However, one theory is that PET/CT may gain more use in cardiac imaging. In fact PET has been used for cardiac imaging for some time; however, it is significantly more expensive compared to other modalities. So, if cardiac imaging does become more important, a 16 slice (or greater) CT would be preferable. Therefore, most buyers opt to pay more to guard against obsolescence and the additional cost is quite small when considering the full cost of installing a PET/CT.

In addition to the cost for the scanner, considerable renovations will probably be necessary when installing a PET/CT. For a start the radiation protection requirements for PET are considerably more than conventional nuclear medicine. The energy of the emitted radiation is higher, so thicker shielding is required and the short half life means higher activities have to be administered. After the tracer has been injected into a patient it is very important the patient is kept in a quiet place while the tracer is taken up. The patient is very hot during this time so the area must be well shielded. Before the exam the patient must void, approximately 15% of the administered activity will be contained in the patient’s urine, so a separate toilet facility may be necessary. The radiopharmacy must also be equipped to protect personnel when preparing the radiopharmaceutical. The thickness of shielding, therefore cost, will largely depend on the expected workload. So, realistic estimations are important to control costs. Having said that, it is better to design for a higher workload than to install additional shielding at a later date.

Some institutions may want to install their own cyclotron. Cyclotrons are used to produce the isotopes. While FDG may be available by delivery, other PET isotopes have too short a half life. If research into new tracers is foreseen, then a cyclotron would be needed. Generally, however, it is more cost effective to buy FDG as needed. Traditionally cyclotrons have been very large devices needing significant amounts of expert care. However, cyclotrons are now available specifically designed for PET imaging. These devices are much smaller and do not require the same levels of expertise to use. Despite this buying and installing a cyclotron will cost ~ €3 million.

One other significant issue facing PET is the training of personnel. Radiographers, nuclear pharmacists, nurses, physicists and doctors may all require additional training. The rapid growth of PET means that qualified individuals may be hard to find.

Other Hybrids

The interest in PET/CT has re-ignited interest in other hybrid options. In 2004 a number of manufacturers launched SPECT cameras with diagnostic quality CT. Earlier versions have included a CT, but not of diagnostic quality. Although SPECT images tend to include more anatomical information than PET there are still good reasons for adding CT. For example, Horger et al. 2004 found that the specificity of skeletal SPECT exams increased from 19% to 81% when a CT was added.

These factors have not gone unnoticed and radiopharmaceutical companies are developing new tracers for specific applications. One area in which SPECT/CT is potentially very important is in cardiology. For this reason, a minimum of 16 slices is necessary for the cardiac CT imaging.

SPECT cameras are standard equipment found in most hospitals and the price of a SPECT/CT is about half that of a PET/CT. Add in the other costs and SPECT/CT is far more affordable option than PET/CT. Therefore, it is likely that SPECT/CT will become far more widespread than PET/CT. It should be stressed that at this time the two modalities are complimentary and both will be needed.

Other Considerations

The introduction of hybrid systems introduces some new factors that should be considered during the procurement process. For example, if one half of the system fails, is it still possible to use the remainder? While other CT facilities are probably available within a hospital, it is unlikely that another PET scanner will be available. So, if the CT is temporally out of commission, the PET should be still useable. The implications could be costly. For example, a sudden failure of the CT sub system (eg, failed tube) could deem an expensive supply of FDG worthless.

Another issue is equipment obsolescence. Recent developments in CT technology have been unwarranted; few people were predicting 8 slices systems, let alone 64 just a few years ago. Renewed interest in PET imaging as led to more research and development funding. Therefore, it is impossible to predict what developments will become available next. The issue is, how does a hybrid scanner keep up to date?

PET/CT’s impact with other clinical systems should also be considered. PET/CT provides more information for the planning of radiotherapy. In addition, radiotherapy itself is undergoing a period of significant developments. The full benefit of these advances will only be realised if information can be fully shared between imaging, treatment planning and therapy systems. At this point DICOM does not adequately allow for combined PET and CT images and work rounds are necessary. So, a PET/CT alone will not be sufficient. Updating treatment planning and delivery should be considered in conjunction with the PET/CT.

Conclusion

PET/CT has demonstrated the potential of hybrid imaging. PET/CT is not just a matter of fusing images, it allows more efficient patient throughput. So, the high costs associated with PET mean that the incremental costs of adding a CT can be justified. As treatments are improved better imaging will be required. High quality patient care depends on precise information. Better imaging performance will cost more, and only incremental improvements to the standard performance measures can be expected. PET/CT is just the start of hybridization. So expect more hybrid imaging devices to emerge.

References

Dietlein M, Schicha H. Reimbursement of the PET in oncology in Europe: a questionnaire based survey. Nuklearmedizin. 2003 Jun;42(3):80-5.

Horger M, Eschmann SM, Pfannenberg C, Vonthein R, Besenfelder H, Claussen CD, Bares REvaluation of combined transmission and emission tomography for classification of skeletal lesions.AJR Am J Roentgenol. 2004 Sep;183(3):655-61

IMV. Latest IMV PET Census Shows Fast Growth in PET/CT Installations. www.imvlimited.com, 2004

Lemke AJ, Niehues SM, Hosten N, Amthauer H, Boehmig M, Stroszczynski C, Rohlfing T, Rosewicz S, Felix R. Retrospective Digital Image Fusion of Multidetector CT and 18F-FDG PET: Clinical. Value in Pancreatic Lesions--A Prospective Study with 104 Patients. J Nucl Med. 2004;45:1279-86

Reinartz P, Wieres FJ, Schneider W, Schur A, Buell U. Side-by-side reading of PET and CT scans in oncology: which patients might profit from integrated PET/CT? Eur J Nucl Med Mol Imaging. 2004 Nov;31(11):1456-61

Surti S, Karp JS, Muehllehner G. Image quality assessment of LaBr3-based whole-body 3D PET scanners: a Monte Carlo evaluation. Phys Med Biol. 2004 Oct 7;49(19):4593-610.

Request more information