3rd International Conference on Cancer Screening

17 Nov 2006

By Prof S Satku

Venue: HPB Auditorium

Professor Tony Chen

President of IACCS Steering Committee;

Mr Lam Pin Woon,

CEO, Health Promotion Board;

A/Prof Yeoh Khay Guan and Dr Shyamala Thilagaratnam,

Chairpersons of the Organising Committee,

Distinguished Guests,

Colleagues,

Ladies and Gentlemen;

It gives me great pleasure to join you this morning at the opening of the 3rd International Asian Conference on Cancer Screening. I would like to begin by extending a warm welcome to our overseas delegates and faculty.

I would also like to thank the members of the overseas faculty for taking time off their busy schedule, to share their expertise with us.

Pre-requisites for Screening

The idea of cancer screening is intuitively attractive - detect cancers early, treat early and increase the chance of curing patients.

To achieve this, all screening tests must fulfil some pre-requisites.  It should be sensitive and specific for early cancer, and effective interventions must be available so that the act of screening can translate into better outcomes for patients with early cancer.

The rate of false positives and false negatives should be low. This is a major challenge in cancer screening. Because of the relatively low prevalence of many cancers, and because of imperfections in current screening tests, false positives will continue to be a common problem.

Worse some cancers will go undetected.

Yet another issue is the "over"-diagnosis of slow growing cancers of doubtful clinical significance. For example, although PSA screening can be used to detect prostate cancers, a significant number of these cancers are so slow-growing that in older patients, it is arguably better to leave them undetected  - since they may not pose any problems at all during the individual's lifetime.

When evaluating screening tools, analysis of survival data for screen-detected cancers must always take into account known pitfalls such as lead-time and length-time biases.

This was exemplified by a recent survival-study of patients with lung cancers detected through CT-scan screening. The study did not take into account lead-time bias, and it was pointed out by a professor from Johns Hopkins that: 

"Everyone knows we can pick up things better with screening. But is picking up early the same thing as curing? If I pick up a tumor today and you live five years or I pick it up four years later and you live one year, is it not the same thing"

Similarly, in any cancer, aggressive or rapidly-growing tumors have a short potential screening period.  Unless screening tests are repeated frequently, aggressive tumors would more likely present first as symptoms, rather than through cancer screening.

In contrast, slow-growing tumors have a longer potential screening period and will more likely be detected when patients are still asymptomatic. As a result, slow-growing tumors are more likely to be over-represented in the group of patients with the screen-detected cancer. Known as "length-time bias", this causes an apparent improvement in survival in our evaluation of screening tools.

Well-designed studies, such as large-scale randomised controlled screening trials, are needed to avoid lead-time bias and length-time bias.

Considering the intuitive attractiveness of cancer screening, as well as the commercial pressures, it is all too easy to rush to judgement without carefully considering the science. Screening is carried out in the normal population. We must remember the fundamental ethical principle of medicine and public health - "First, Do No Harm".

Consider the screening for neuroblastoma, the commonest extra-cranial solid tumour during early childhood. This was carried out nationally in Japan from 1984, but was stopped recently, 20 years later. This was because evidence from studies in Canada  and Germany  had showed that although screening picked up early stage cancers, there was no benefit in late-stage cases; or in terms of death rates from neuroblastoma when compared with unscreened controls. 

This is an example of "over"-diagnosis: the early stage cancers detected were "extra" ones that posed no threat to life and would have spontaneously regressed. In one study5, out of 149 cancers detected in the screening group, 99 were "extra" ones that would have posed no threat to life. Yet 3 out of these 99 died from treatment related causes!

Much earlier, although the Japanese had done their own local studies - showing promising results for neuroblastoma screening, these studies suffered from methodological limitations such as the lack of good control groups and the failure to account for lead-time, length-time and "over"-diagnosis biases.  The lesson learned is that we must insist on robust evidence that screening works, lest we do more harm than good for our patients.

Based on our assessment of the scientific evidence, Singapore has put in place national cancer screening programmes for breast and cervical cancer, which rank as the top and 5th most common cancers respectively in women in Singapore.  BreastScreen Singapore was implemented in January 2002 to encourage women aged 50-69 years to go for regular mammography. CervicalScreen Singapore encourages women aged 25-69 to go for Pap smears. 

Quality Assurance Programmes

Quality assurance is yet another important aspect in screening. Maintaining quality standards for screening, assessment and diagnosis is crucial to the success of any screening programme. The outcomes achieved for screening in a clinical trial setting may not always translate well into an actual programme setting unless quality is assured.

Some of you may recall that in 1988, the New Zealand Minister of Health had to appoint a Committee of Inquiry in response to the growing public concern that Pap smear tests carried out as part of the New Zealand Cervical Screening programme were being misread - with the result that cervical abnormalities were under-reported .

Clearly, if the rate of false negatives or "missed cancers" go up as a result of a poorly managed screening programme, the public consequences can be tragic. In addition, recall rates due to false positives or "false alarms" may also increase, resulting in needless anxiety and unnecessary diagnostic assessment procedures for those affected.

Therefore, in Singapore, a centralized quality assurance programme has been established for both the national breast and cervical cancer screening programmes. This ensures that stringent quality measures are in place. For example, we have instituted independent double-reading of mammograms with a 3rd reader to arbitrate discordant reads. This increases the detection rate while decreasing the recall rate. The rate of unsatisfactory Pap smears is monitored and regular audits are also carried out biennially on all centres under the programmes.

Informed-choice Strategy 

This said, even with the best quality assurance programme, there will be false positives and false negatives. As I have mentioned earlier, these can cause unnecessary distress, or in some instances, actual harm in the form of additional invasive testing or false reassurance and denial of the signs and symptoms of cancer.

While individual patients and their values may differ: some are willing to accept substantial risk even for modest benefits, whereas others are averse to all risk, it is therefore important to involve the patient as a partner in informed-decision making.

With an increasingly educated population, we should aspire to provide enough information about the limitations and risks of screening, as well as the potential benefits, in an easily understood form. This can help our patients make an informed choice.

Lastly, adopting an informed-choice strategy does not excuse the need for clinicians and policy-makers to exercise discerning professional judgment in deciding whether or not to offer particular types of screening in the first place.

Conclusion

It is clear that all screening tests involve a balance between the potential benefits and harms, and because these benefits and risks may be viewed differently by different patients - we must pursue cancer screening carefully - giving due consideration to the science, the operational issues, as well as respecting the autonomy of our patients.

On this note, the theme of this year's conference is particularly relevant. I would like to congratulate the organizing committee for putting together this excellent conference. Your discussions on quality, trends and policy issues must delve deeply into the twin thrusts of quality and informed-choice. The knowledge and insight gained through this conference will be useful in making our screening programmes better and more successful.

Thank you and wish you all a pleasant day ahead.