Cancer Council NSW, Woolloomooloo, New South Wales, (Volunteer).
Although efforts to encourage presentation with early signs of cancer has had some impact in down-staging disease at diagnosis, notably in breast cancer and melanoma, assertive, physician-initiated screening for cancer in asymptomatic patients has demonstrably resulted in earlier diagnosis, as well as reduced mortality in some cancers. Although systematic population-wide screening programs optimise the population benefit of cancer screening, there is a long lag time to adoption of public screening programs, during which physician vigilance fills the void. Furthermore, some cancers will never achieve the prevalence threshold for mass screening, and in some common cancers clinical context and physician judgement are essential in screenee selection. The potential for early diagnosis through screening is broader than the scope of organised mass public cancer screening. Organised professional groups such as the Clinical Oncological Society of Australia need to develop the evidence base for screening in a range of cancers, and develop guidelines to optimise physician-initiated screening and early detection. Doctors at all levels need to apply risk assessment tools in practice, and be alert to opportunities where screening can benefit their patients.
As early stage cancer generally has a better prognosis, efforts to down-stage disease are important. Although screening programs for bowel, breast and cervical cancer delivered on a mass scale are established, the prospects for detecting most cancers at a pre-invasive or early stage rest with the awareness and actions by clinicians and the community. Mass screening programs may eventuate from these efforts, but the prevalence of many cancers will remain below a threshold for mass screening.1
Education of the warning signs and symptoms of cancer has been long established and include important efforts in encouraging detection at an early symptomatic stage of cancer. These include specific education in breast cancer and skin cancer detection, as well as the ‘Seven Warning Signs of Cancer’ developed by the American Cancer Society.2
The Clinical Oncological Society of Australia’s (COSA’s) early years coincided with pioneering efforts to screen for cancer. It was a decade that laid the foundation for more systematic approaches to screening, exemplified by the development, diffusion and evaluation of breast cancer screening in the 1970s and 80s.3
While evidence for the effectiveness of breast cancer screening is solid, the evidence available on cervical cancer screening at the outset of systematic population screening was comparatively weak. Time trends in the relative incidence of carcinoma-in-situ (rising) and invasive cancer (falling) were favourable to the case,4 but it was not until 1978 that research demonstrated at the individual level the protective efficacy of a recent PAP smear.5
In the 1970s, the PAP smear was the poster child for cancer screening, but the program was beset by quality problems,6 and benefits were unequally distributed. Despite this, the incidence of cervical cancer in women by this time had dropped by over 40% in a few states of the US.4 Incidence rates continued to rise in the UK, until the advent of systematic screening, even though 40% of potential cervical cancers in women were being prevented by opportunistic screening in general and obstetric practice.7 Examples of more systematically successful screening rested on commitment and team work between hospital specialists, general practitioners and local authority doctors.8
In contrast to cervical screening, by the 70s there was good evidence to support rectal cancer screening – another cancer with a well-defined precursor lesion. Twenty-five year observational cohort data on 92,650 patient years of periodic screening with rigid sigmoidoscopy and polyp removal demonstrated a reduction of 85% in the expected number of rectal cancers, and all cancers detected were localised.9
The limitation of the instrument, and the arrival of alternative technologies led to the search for more comprehensive solutions to population-wide colorectal cancer screening. However, given that rigid sigmoidoscopy was a tool within the reach of most primary physicians, at least in the US, and 55% of all cancers of the colon lay within the reach of the instrument, it is hard to avoid the conclusion that considerable benefit to those screened may have flowed from the normalisation of screening for rectal cancer among clinicians. Thirty years later we still await the complete roll-out of a bowel cancer screening program in Australia.
On a more positive note, the development in 1971, and subsequent evaluation of the Danish familial adenomatous polyposis registers, demonstrated the remarkable impact on colorectal cancer mortality from compete documentation, registration and follow-up screening of a very high risk population.10 This clinician-driven measure demonstrated that, when high risk populations can be identified in clinical practice, their health expectancy is transformed by effective early detection measures. The first Australian state familial bowel cancer register was established in 1985.11
The success achieved in the dissemination of mammography screening for breast cancer illustrates the importance of well-designed definitive studies at an early-stage.The exemplary foundation trials of mammography in screening have arguably expedited the pathway to mass implementation, though without extensive grass-roots advocacy, the pace of policy development and program funding may have been much slower.
Another lesson from the early years of screening was the failure of trials for early detection of lung cancer. Clinical enthusiasm for screening with chest X-ray and sputum cytology was dissipated by a European trial which showed no reduced mortality benefit from screening, despite an increase in incidence and survival after diagnosis.12,13
Mass screening programs judge their effectiveness by the outcome achieved in populations, such as downstaging, mortality reduction, and interval cancer rates, whereas clinicians focus on individual patient benefit. In mass screening, clinicians are central to screening in cervical cancer, but insufficient emphasis is placed on the impact of their endorsement and promotion of screening for breast and bowel cancers.14
Physicians can also realise the possibility of benefit and reducing the risks associated with screening for other cancers.
Screening for melanoma
Organised screening from community clinics increases participation in whole body skin examination.15 However, Australia remains reliant in melanoma screening on the combined effects of public education and general clinician vigilance, much like the Pap test in the 70s. Only level three evidence exists in support of systematic screening,16 and the failure to mount support for the Queensland melanoma screening trial suggests that the question of the effectiveness of mass screening in melanoma will remain unanswered.17
It is unclear how much Australian success in early detection – either for average risk, or selected higher risk groups – reflects practice level systems for screening, or a high level of community awareness coupled with ready access to general practitioners and skin clinics. The predictable inequity in diagnostic outcome that results from this approach needs to be the focus of greater clinical and educational effort,18 but organised mass screening programs are best placed to erase these differences.
Screening for prostate cancer
After protracted controversy, consensus may settle on the view that a mortality benefit emerges seven years after prostate-specific antigen screening and increases thereafter.19 Screening has driven a very large increase in the incidence of prostate cancer, but many men express dissatisfaction with outcomes, because of long-term side-effects.20–23 The potential to benefit from screening depends not only on age and other comorbidities, but also on individual preferences and acceptance of trade-offs.24 A customised approach to selection for screening is needed to enhance the likelihood of net benefit among those screened. This can only be accomplished if clinicians are at the centre of any organised screening effort and common protocols are employed.
Lung cancer screening
The National Lung Cancer Screening Trial has confirmed the effectiveness of spiral CT screening in lowering both lung cancer specific mortality (by 20.3%) and all-cause mortality (by 6.7%) in smokers and ex-smokers.25 However cost-effectiveness is poor, and worsens if the behavioural impact is to discourage quitting.26 There is also limited capacity in Australia to screen and handle large increases in cases, and false positives.25 Any debate over screening for lung cancer will be further spiced up by the record of tobacco industry support. However, in the US at least, 22% of primary physicians use spiral CT to screen for lung cancer, and one may sympathise with a physician who initiates screening for, say, ex-smokers in their care. While this benefits those patients, it also raises the issue of equity, which can only be solved by an organised program where clinicians are key to risk profiling and assessment.
Less common cancers
Risk profiling and screening for some less common cancers offer potential for benefit from an organised, clinically driven approach, as exemplified by the pioneering work in familial bowel cancer. Early detection in both hepatocellular cancer (HCC) and oesophageal adenocarcinoma (EAC) are discussed, as they have been the focus of research and/or program development in association with Cancer Council.
Since 1982, the incidence of liver cancer in Australia has risen strikingly, particularly in NSW through the weight of migration from HBV endemic regions of East and Southeast Asia and the Pacific. The incidence of HCC among these immigrants is around 10 times that of Australian born.27
The risk of developing HCC in the REVEAL study, 28 of 3613 participants with chronic hepatitis B over an average of 11.3 years, was closely related to the level of HBV-DNA. The risk of HCC rose in a monotonic gradient to 17.7 in those with HBV-DNA copies over a million per ml. with no other markers of liver disease. Effective antiviral treatment for CHB reduces viral load to low levels, and meta-analysis of six treatment studies (of which two were randomised control trials) demonstrated a 70% reduction in long-term liver complications including HCC.29
Robotin et al. have demonstrated that a systematic program to screen people for chronic hepatitis B in general practice, initiating treatment and cancer screening in accordance with guidelines, would be a cost-effective approach to reducing the incidence of and mortality from HCC and associated liver disease.30 A primary care program targeted to at-risk demographic groups is being trialled in South West Sydney,31 and similar efforts are under way in Melbourne.
Effective control of liver cancer should be a priority in multi-cultural Australia. The current epidemic of chronic hepatitis B acquired overseas cannot be addressed by Australia’s domestic immunisation program, and unless effective measures in screening and treatment are introduced, annual deaths from liver cancer may rival those from breast cancer in NSW. For benefit to be realised, clinicians will need to dispense with the out-dated concept of the ‘healthy HBV carrier’, and be mobilised to reach the population at risk.
The incidence of EAC has increased substantially, and continues to rise at a rate of 2.2% per annum (among men).32 EAC is thought to develop through precursor states of intestinal metaplasia in response to oesophageal reflux, with dysplasia developing on the background of Barrett epithelium at a rate of about 1% per year. EAC often presents at an advanced stage with a poor survival rate, so there is a strong prima-facie case to detect and treat precursor lesions or diagnose EAC at a localised stage. However, to date, in only a small proportion of EAC cases has Barrett’s been diagnosed prior.33
Risk factors for EAC are readily identifiable clinically. The Australian Oesophageal Cancer Study identified reflux, elevated body mass index and smoking as independent risk factors.34 The most startling risk elevation resulted from synergy between reflux and obesity, elevating the relative odds of EAC up to 16.5. Ninety-two percent of the population risk of EAC could be attributed to these factors when combined with educational level, acid suppressant medication use and non-steroidal anti-inflammatory drug use.35 This model could conceivably be developed as a clinical tool to assist in identifying patients for further assessment.
Medical and surgical anti-reflux treatments are widely used, but they fail to prevent disease progression,36 and it is unclear how much awareness exists in general practice of the potential future consequences of reflux. Research is also needed to establish how patients should be selected for ongoing surveillance, as the yield from endoscopic surveillance in simple Barrett’s oesophagus is low.37
The economics of screening and surveillance strategy for EAC and its precursors are not yet compelling,38 and reliance on endoscopy limits its wider spread. However, the development of less invasive methods for mucosal sampling,39 better predictive tests and well-targeted case selection may overcome these limitations. The rising burden of illness from this disease provides a strong incentive to accelerate these developments.
Screening generates a remarkable level of controversy, in part because large-scale screening is expensive, and exposes people who don’t directly benefit from screening to unsolicited risk.40 While encouraging symptom awareness for early presentation has a long history, it is not clear that this results in any lower burden of investigation,41 and decades after their launch, community knowledge of the seven warning signs of cancer is low.42
The landscape for screening is expanding, with better understanding of risk factors and precursors for a variety of cancers. Organised medicine needs to be alive to this landscape by supporting and promoting research in risk assessment, early detection and screening. Furthermore, the role of clinicians is crucial in mediating the potential benefits from screening for prostate cancer and melanoma. Organised medicine needs to provide guidelines to motivate and assist doctors to optimise the benefits of screening and early detection for their patients, both within and beyond the three public mass cancer screening programs.
Professor Nico van Zandwijk for information on lung cancer screening.