Our results, that should be considered in an ordinal perspective rather than in a cardinal one, show that for island locations with population sizes justifying the existence of one NM department with PET, there is a specific curve (Fig. 1) predicting the least costly solution, which is transport of all radiopharmaceuticals from the mainland for low, SCb for middle and HC for high number of annual PET examinations. We show that Corsica, Sardinia, Crete and Cyprus would minimize their costs using local radiopharmaceutical production. Although our results show small advantages for HC over SCb, it might prove even more advantageous in the medium term due to its capability of producing radiopharmaceuticals also for therapeutic NM purposes (27) and better control over inflation-driven increase of local transportation costs against externally dependent delivery costs (of SPECT radiopharmaceuticals).
The point after which local production becomes less costly depends equally on annual examination numbers and difference between delivery and local transportation costs. This means that results are very sensitive to these elements, should one be either very high or very low, as is the case of annual examinations in France and Cyprus respectively (30-fold difference). The latter depends, among others, on the NM culture of each country (28). However, the existence of SCb or HC will either way augment local examination numbers via the supply-driven demand procedure (29,30).
Additionally, it may seem obvious (31) that for extremely isolated island regions, such as the Canary Islands, New-Caledonia, French Polynesia, Azores, etc., delivery of short-lived radiopharmaceuticals is impossible, making the existence of SCb or HC of paramount importance, at least in theory. However, most studies suggest no more than 2 to 2.5 PET machines per million people (32), so one per 400,000 or 500,000 people. Existence of some kind of cyclotron is only relevant if accompanied by at least one PET machine, meaning that it might not make sense to have any type of cyclotron if the local population is less than the above-mentioned limit. But annual examinations are expected to be on the rise, mainly due to the rise of cancer cases (33), dementia (34) and heart (35) imaging, and, as already mention, the existence of a cyclotron in island locations will push annual examination numbers up. Thus, local cyclotron existence in isolated island regions with medium population sizes may be worthy.
SCb or HC existence in an island location will also offer indirect advantages to local society, the most important being the brain gain, via the highly specialized personnel that will be working there, not only to the cyclotron installation, but also to the NM departments. Additionally, SCb or HC can offer higher health services independency due to partial or full local radiopharmaceutical production respectively, no longer dependent on weather or strikes. It is to be noted that in a recent study concerning Corsica almost 50% of local cancer patients did not have any PET examinations during the whole course of their disease, simply because there was no PET in Corsica (23), although this travel to the mainland would be fully reimbursed by the French social security services. SCb or HC existence locally argues towards the installation of PET machines and vice versa, improving the overall diversity and availability of offered medical services to the local population.
On the other hand, having a cyclotron in remote places will take away the (not only economic) advantages of the transporting companies that until recently fed the NM departments of island regions with radiopharmaceuticals and, at the same time, will highlight the possible insufficiency in isolated location transportation infrastructures: it may be that a helicopter transporting radiopharmaceuticals between a given mainland area and island location would need less time to reach a NM department than a truck from the local cyclotron being driven on the usually rather winding roads. The existence of a cyclotron may, thus, need to be accompanied by either local infrastructure upgrading or, as is the current trend, use of medical equipment transportation drones, solution, however, that is, in its turn, weather dependent.
Our research entails limitations. First, the exact number of daily doses for either PET or SPECT examinations needed in island regions is unknown, since not only some NM examinations are omitted due to scarcity of radiopharmaceuticals and/or PET machines, but also because NM market can present, at least at first, a supply-driven demand (37,38), meaning that more NM examinations than normally needed would take place, should radiopharmaceuticals be produced locally. This results in unknown numbers of needed production runs and personnel shifts. Additionally, we have set a limit of six 2-hour runs per working day, which is reached twice for Sardinia (second cyclotron needed on year 8, third one on year 15). This means that the investment costs for Sardinia are tripled, so maybe a Medium-size Cyclotron (see Appendix A) might be more suitable from the very beginning. Either way, more SPECT and PET machines and, maybe, NM departments are needed, at least in Corsica and Sardinia, resulting in extra personnel, something extremely scarce in island regions (39), not to mention the possibility of (cyclotron) production stall due to even one personnel member being absent (40). On top of this, our research does not consider therapeutic NM, where other radiopharmaceuticals are needed, most of which can be produced by HC, but not SCb (41). Furthermore, we did not take into consideration the situation where patients would need to go to the mainland in order to have their PET examinations, as is the case for Corsica and Crete, since this would require a cost-benefit analysis due to the different outcome regarding the NM examination accessibility. Moreover, our analysis is restricted to the positive direct effects that a SCb or HC could bring to the local health economy, ignoring the indirect ones, such as the short-term economic growth brought by the building labors’ presence and the medium-term creation of extra jobs (local transportation service being a good example). Finally, our analysis does not include the SPECT and PET scanner acquisition and running costs, but these will be the same per country for any of the three scenarios under investigation, except for the respective PPP, and, thus, would not alter our results.