In 1995, a massive earthquake hit Kobe, Japan. With inadequate foundations, many buildings simply toppled over. Subsequently, building codes were revised to ensure that buildings were constructed with foundations capable of withstanding similar seismic events. Ten years later, a scandal broke when it was found that an architect had been flouting these codes and authorizing as compliant buildings that he knew to have been constructed to inferior engineering standards. Although no building had collapsed, the public were justifiably outraged at the potential risk to life.
The foundation on which modern oncology is built is the accurate detection of cancer, determination of the extent of disease, and post-treatment assessments of the success of treatment. This process of diagnosis, staging, and therapeutic response assessment intimately involves imaging.
Unfortunately, in very many cancers, we know that conventional structural imaging performs relatively poorly in all these roles. With such a poor footing, is it surprising that the superstructure of oncological therapy so often collapses? If there were a technique that manifestly improved the ability of oncologists to make informed decisions regarding patient management, wouldn’t it be scandalous not to embrace it in order to prevent unnecessary loss of life and inappropriate expenditure of scarce healthcare resources? Yet this is what has happened over the past 10 to 15 years in the case of PET.
Numerous studies validated by pathological findings, serial imaging follow-up, and most importantly in the case of cancer the ability to stratify survival, have compellingly demonstrated that FDG-PET generally outperforms not only individual diagnostic tests, but also a combination of tests in many cancers.
In particular, the great strength of FDG-PET is its capability to detect remote systemic and regional nodal metastases. In diseases where the management options extend from surgery for local tumor to combination therapies, often including radiotherapy, for regional disease and to systemic therapies for disseminated cancer, this ability can have a major impact on patient management.
Economies of scale
One of the major impediments to the wider application of PET has been the perception of its high cost. In many regards, this has been due to self-fulfilling policy! Because of high scan costs, access to funding was restricted in many parts of the world and consequently relatively few scans were referred. This, in turn, stifled sale of PET scanners and the number of scans performed per scanner.
Fast-forward to 2007 and we now have many hundreds of PET/CT scanners being delivered each year by each of the major manufacturers, efficiently manufactured and benefiting from continual quality and design improvement. The large installed base allows a larger spare-parts inventory to be available and more effective use of maintenance engineers, decreasing servicing costs.
On the operational side, regional cyclotrons are producing and distributing FDG in bulk—allowing a marked reduction in the unit dose cost. The advent of PET/CT has allowed the acquisition time for typical whole-body scans from more than one hour to less than 30 minutes. This, in turn, has allowed higher throughput and hence, greater amortization of capital equipment costs, more productive use of staff, and more efficient use of FDG, which can rapidly decay during the scanning of preceding patients.
In 1996-1997, my facility performed slightly more than 700 scans on a single scanner; whereas in 2006-2007, we have performed more than 4,500 scans on two PET/CT scanners. The quality of these scans has improved dramatically and the real cost of scans has decreased.
Nevertheless, it needs to be understood that scan costs are critically dependent on practice models and that if throughput is constrained by bureaucratic or logistical constraints, or the production and use of FDG from a cyclotron is inefficient, then PET will remain substantially more expensive than competing diagnostic modalities used in cancer evaluation.
PET in perspective
However, the unit cost of PET scans is not the critical issue in justifying the wider use of this technology. Cancer care is amongst the most expensive areas of therapeutics. The surgery involved is often complex and may require major tissue reconstruction. No longer is surgery confined to those with local disease, with major hepatic and lung resections being increasingly performed for patients presumed to have limited, and therefore resectable, metastatic disease.
Radiotherapy also is becoming increasingly complex and expensive to deliver due to advancements aimed at better confining dose to disease sites and sparing toxicity in normal tissues. Furthermore, a move away from cheap older-style cytotoxics to targeted therapies has substantially increased chemotherapy costs. Moreover, the days of single modality therapies are rapidly passing, if not already gone. Most patients will now receive adjuvant chemotherapy or radiotherapy in addition to “curative” surgery.
Radiotherapy is often used in combination with chemotherapy and, more recently, with biological-targeting agents. Monoclonal antibodies and small molecule tyrosine kinase inhibitors are being added to more conventional chemotherapeutic agents in the hope of improving progression-free and overall survival, but are massively increasing the cost of treatment.
Despite this explosion in therapeutic options and hugely burgeoning costs, reductions in cancer mortality have been disappointing. In 2004, more than 557,000 people died of cancer in the United States, but this number was reduced by a bare 3,000 or so compared with the year before.
The total cost of cancer care in 2004 was $72.1 billion (U.S.), an increase of 75 percent compared with expenditures in 1995. These costs exclude the cost of screening programs and don’t reflect the societal costs, which have been estimated to be in excess of $190 billion in the same year. These ever-expanding health costs are clearly not sustainable with an aging population. So, why should it even be considered that an expensive test such as FDG-PET could be added to existing costs?
Well, let us consider lung cancer, the most common cause of cancer mortality. This disease costs approximately $9.6 billion in 2004 in the U.S. Medicare payments alone cost $24,700 per patient in the first year of diagnosis, ignoring out-of-pocket expenses, which may add in excess of 10 percent to this.
Assuming that the majority of these costs represent attempted curative treatments in the approximately 75 percent of patients who don’t already have clear evidence of metastatic disease at presentation, the scope for more accurate diagnosis to impact health expenditure can clearly be appreciated if one considers the incremental diagnostic value of FDG-PET in published series.
Evidence-based medicine
In a prospective randomized trial from Holland, performed in patients otherwise being considered for surgical treatment, FDG-PET reduced futile thoracotomies by 25 percent. In a prospective series of patients being considered for curative chemoradiotherapy, my group in Australia demonstrated that FDG-PET altered management in more than 50 percent of cases. Importantly, approximately 25 percent of patients were found to have distant metastases that had not been recognized on conventional staging.
Obviously, money spent on expensive and potentially morbid attempts at cure in such patients is illogical since these patients are doomed to fail and will still require the palliative or systemic therapies that they ought to have received up-front. Furthermore, another 25 percent of radiotherapy patients had their radiation treatment fields altered, primarily to include FDG-avid nodal metastases that were not enlarged on CT and therefore would not otherwise have been included in the treatment volume.
Again, clearly it is a complete waste of a valuable resource such as radiotherapy to miss macroscopic disease sites. Although not all patients are cured with radiotherapy, even when all their disease is included in the radiation volume, it is a forlorn hope that patients with inadequate disease coverage will be cured. Those patients who miss the opportunity for cure will go on to require palliative treatment and contribute to the loss of productivity and the other societal costs associated with cancer mortality. It is said that cancer doesn’t touch an individual, it attacks a whole family. To go on using clearly inferior techniques for staging lung cancer is being penny wise but dollar foolish.
Beyond the implications for an individual patient, there are important considerations for resource allocation rendered by more appropriate selection of patients for curative treatment. In many parts of the world, access to radiation oncology services is severely constrained and waiting times are long. During this delay, there is the potential for disease progression, making prior planning studies irrelevant to the current extent of locoregional disease or converting some patients from curative to only palliative candidates.
By identifying a greater number of patients who are, by virtue of the extent of disease diagnosed only on PET, suitable only for palliative treatment, the course of radiotherapy is reduced from approximately six weeks to one week. This resource opportunity allows those in the queue to get their radiotherapy earlier. By better targeting locoregional disease, early local recurrences ought to also be reduced, minimizing the call on salvage radiotherapy and opening additional time on valuable linear accelerators. Similar arguments can be made about futile surgical procedures.
Perhaps the greatest opportunity for PET to influence healthcare costs is in the arena of chemotherapy. Based on current response assessment criteria, which are related to changes in the dimensions of index lesions as measured on morphological imaging studies, there needs to be a significant increase in tumor burden before a chemotherapeutic regime can be considered to have failed due to progressive disease.
However, particularly in cancer that has a relatively indolent natural history or that presents with relatively large tumor deposits, progression may be slow or reflect a huge increase in tumor volume before ineffective treatment is ceased. During this time, the cost and toxicity of this treatment is accumulating. The opportunities for a favorable response to subsequent therapies are compromised by the increased number of cancer cells that must now be obliterated and the compromised physical, physiological, emotional, and, often, fiscal reserves of the patient. Earlier response assessment has been a key feature of the majority comparative studies using PET and conventional response assessment and most also have demonstrated superior stratification of survival.
Molecular therapeutics
A final note must be made regarding the potential for PET to speed and simultaneously reduce the cost of cancer drug development. This is a huge topic, and well beyond the scope of this discussion. However, it is being increasingly recognized that metabolic imaging using PET can provide unique insights into drug effects. More robust proof-of-mechanism studies, selection of target populations, and assessment of response rates in better stratified groups will allow ineffective drugs to be culled at an earlier stage and effective drugs to be fast-tracked to market.
Hopefully, this will reduce the cost of these novel therapies and allow them to also be more rationally prescribed. The current cost of development of each new cancer drug has been estimated to be up to $1 billion and to take up to 14 years. With a current patent life of only 15 years, it is clearly difficult to provide shareholder returns on such an investment unless there is a very large market and a high cost for the drug. At the same time, there is a move away from so-called “block-buster” drugs toward personalized medicine with targeted therapies that will have niche applications. Unless new strategies for drug development are found, this new era of medical care will fail. I share the view that molecular imaging has a key role to play in drug development.
Each individual brings his or her own emotional, physical, intellectual, experiential, supporting and fiscal resources to bear when diagnosed with cancer. Patients need to be informed not only of the cost of FDG-PET, but the implications of not having the most accurate disease evaluation available for many cancer situations. If we can discover the definitive blood test and a universally effective cancer therapy, we won’t need any imaging at all. Until that day, we can’t afford not to use PET.
Professor Rodney J Hicks, MD, is professor, Department of Medicine, the University of Melbourne, and director of The Centre for Molecular Imaging and Translational Medicine, The Peter MacCallum Cancer Centre in Melbourne, Australia. He also is an editorial advisory board member of Molecular Imaging Insight.