1. Screening stories: avoiding the hard sell Prof. dr. Elke Van Hoof Belgian Cancer Center

4. “The greatest need we have today in the human cancer problem, except for a universal cure, is a method of detecting the presence of cancer before there are any clinical signs of symptoms.” - Sidney Farber, letter to Etta Rosensohn, November 1962 - (The Emperor of All Maladies, Siddhartha Mukherjee) Sidney Farber (1903-1973) Paediatric pathologist and “father” of modern chemotherapy. The Dana-Farber Cancer Institute in Boston is partly named after him.

5. Overview 1.Historical background 2.Europe: Council recommendation on cancer screening (2003) 3.Current screening programmes 4.When to screen-which cancer sites to screen? 5.Controversy in breast cancer 6.Summary

6. 1. Historical background George Papanicolaou (1883 – 1962), Greek cytologist Studied the menstrual cycles of guinea pigs Microscopical examination of cells of the cervix, removed by a cotton swab Changes in morphology of the cells due to hormonal changes Similar observations in women (he took a daily vaginal smear of his wife) Importance of his work was not recognised (“ a useless invention”) Focused on pathological smears observed abnormal cells in smear of women with cervical cancer still received lots of criticism: better methods of diagnosis already available In 1950 (more than 20 years later!) it dawned on him that PAP smears might be useful to detect not cancer, but the precursor stages

7. PAP smears as a means for early detection of cervical cancer? 1952: Papanicolaou persuades the National Cancer Institute to set up a large clinical trial in secondary prevention PAP smears were taken of 150 000 women + follow-up detection of 555 invasive cervical cancer cases ! detection of 557 cases of pre-invasive and pre-cancerous lesions easily removable mean age of these women: about 20 years younger than the mean age of women with cervical cancer confirmed the prolonged development of cancer = opportunity to intervene early in the process Historical background

8. Other attempts to visualise cancer Albert Salomon (1883 – 1976), German surgeon Executed lots of mastectomies Tried to visualise the breast cancer tumours by use of X-rays on the amputated breasts = foundation of mammography, but his work was interrupted by second world war and disinterest of surgeons in screening Robert Egan, pioneer in radiology in Houston Experimented with films, angles, positions and other settings Succeeded in visualising tumours of a few millimetres Historical background

9. Fundamental principles of best practice in early detection of cancer Shared commitment by Member States to implement cancer screening programmes 2. Europe: Council Recommendation on cancer screening (2003)

10. = early diagnosis of non-symptomatic cancer aiming at the reduction of morbidity and mortality Population-based screening: offered systematically to all individuals in the defined target group within a framework of agreed policy, protocols, quality management, monitoring and evaluation Opportunistic screening: offered to an individual without symptoms of the disease when they present to a health care practitioner for reasons unrelated to that disease. Cancer screening

11. European recommendations 3. Current screening programmes Breast cancer screening: 2-yearly Mammography screening for women aged 50 to 69 in accordance with European guidelines on quality assurance in mammography. minimum screening participation rate of 70% recommended Current issues: Allowed rate of overdiagnosis (5%? 10%? 50%?) lower age limit? (40? 45?) upper age limit? dense breast tissue: mmx -> ultrasound? http://eu-cancer.iarc.fr/ (2007)

13. European recommendations Current screening programmes Cervical cancer screening: Pap smear screening for cervical cancer precursors starting not before the age of 20 and not later than the age of 30 with three to five-year intervals. minimum screening participation rate of 85% recommended Current issues: HPV testing instead of pap smear (higher sensitivity) HPV vaccination (primary prevention) http://eu-cancer.iarc.fr/ (2007)

16. European recommendations Current screening programmes Colorectal cancer screening: Faecal occult blood (FOB) screening for colorectal cancer in men and women aged 50 to 74. 95% of the target group should be invited; A minimum of 45% of invitees should be examined, but it is recommended to aim for a rate of at least 65% Current issues: 3 tests: FOBT (guaiac and immunological), sigmoidoscopy, colonoscopy Low participation http://eu-cancer.iarc.fr/ (2007)

17.  IMPORTANT DISEASE?  TEST AVAILABLE?  IMPACT ON DISEASE OUTCOME?  COST-EFFECTIVE?  CONSEQUENCES? 4. When to screen – which cancer sites to screen?

18. When to screen – which cancer sites to screen?  IMPORTANT DISEASE? Top 10 cancers in European men and women WSR

19. When to screen – which cancer sites to screen? Important health problem for the general population Natural history well known Accurate diagnostic assessment Effective treatment options Earlier treatment improves disease outcome/prognosis  IMPORTANT DISEASE?

20. When to screen – which cancer sites to screen? Acceptable to the population Test characteristics Cancer process: initation – promotion – abnormal growth – invasion – metastases symptoms diagnosis and treatment  long interim period: window for screening  SUITABLE TEST? Schiffman. N Engl J Med. 2005

21. When to screen – which cancer sites to screen? Preclinical detectable period depends on: cancer (site, tumour volume doubling time, morphology, agressivity) age (slower growth – longer asymptomatic phase: length time bias) test (characteristics, improvement)  Mean lead time amount of time by which the diagnosis has been advanced by screening longer lead time → higher risk on overdiagnosis  SUITABLE TEST?

22. Lead time bias a monozygotic twin “Hope” and “Prudence” 2000: both sisters develop an identical type of cancer (without knowing it) Hope: Participates in screening 2005: tumour detected by screening => surgery + CT 2010: relapse + decease Prudence: Refuses to participate in screening 2009: tumour detected (1st symptoms) => surgery + CT 2010: relapse + decease At first sight, it seems that Hope lived longer, but In fact, both sisters had the same lifetime, but Hope had to live longer with the disease (The Emperor of All Maladies, Siddhartha Mukherjee)

23. Length time bias more slowly growing tumors, with less capacity to prove fatal longer presymptomatic screen-detectable period more likely to be screen-detected.  artificial survival advantage to screen-detected cases Duffy et al. Am J Epidemiol. 2008

24. Interval cancers Definition: Cancer diagnosed after a negative screening test and before the following screening round Concerns tumours not detected during screening test OR tumour developped after the negative screening test Higher proportion of invasive tumours in group of interval cancers and non-screened patients

25. When to screen – which cancer sites to screen? Sensitivity: Ability of the test to identify positive results Proportion of actual positives which are correctly identified as such (i.e. the percentage of people with cancer who are correctly identified as having cancer) TRUE POSITIVE rate Never 100% Specificity Ability of the test to identify negative results Proportion of negatives which are correctly identified (i.e. the percentage of healthy people who are correctly identified as not having cancer) TRUE NEGATIVE rate  TEST CHARACTERISTICS

26. When to screen – which cancer sites to screen?  TEST CHARACTERISTICS 2 x 2 table

27. When to screen – which cancer sites to screen? Positive predictive value (PPV): The probability to have cancer following a positive test result Proportion of positive test results which are TRUE POSITIVE Negative predictive value (NPV): The probability to be healthy following a negative test result Proportion of negative test results which are TRUE NEGATIVE BUT: PPV and NPV vary with prevalence  TEST CHARACTERISTICS

28. When to screen – which cancer sites to screen?  TEST CHARACTERISTICS e.g. fecal occult blood (FOB) screen test in 2030 people to look for colorectal cancer

29. When to screen – which cancer sites to screen? Likelihood ratio (+): The ratio of the probability of a positive test result when having cancer over the probability of a positive test result when not having cancer The higher LR (+), the higher the positive predictive power of the test Likelihood ratio (-): The ratio of the probability of a negative test result when having cancer over the probability of a negative test result when not having cancer The lower LR (-), the higher the negative predictive power of the test Odds ratio: The overall power of a test to discriminate between an ill and healthy condition Ratio of LR(+)/LR(-) The higher the odds ratio the higher the predictive power of the test  TEST CHARACTERISTICS

30. Lower disease-specific mortality Less morbidity Lower cancer incidence E.g.: cervical and colorectal cancer – Detection + removal of pre-cancerous lesions => progression towards cancer is stopped Higher cancer incidence – but shift towards lower stages = smaller tumours, not metastasised E.g.: breast, prostate and lung cancer Remark: at the start-up of a screening programme, prevalent tumours will be detected Programme should be evaluated when it’s active during several years. Otherwise mortality rates will be biased by “old” = prevalent cases. When to screen – which cancer sites to screen?  IMPACT OF EARLY DETECTION ON DISEASE OUTCOME?

31. Favourable versus unfavourable effects Advantages Decrease of cancer mortality Healthy life-years gained (or Quality Adjusted LifeYears if in good quality (QUALY)) Prevention of metastasis (more early stages, less advanced stages detected) Disadvantages Earlier and additional diagnoses More years lived with disease and follow-up after treatment People worry about the risk that they might have a cancer Unpleasant test False positives False negatives => false reassurance Financial costs, time loss When to screen – which cancer sites to screen?  COST-EFFECTIVENESS OF SCREENING PROGRAMMES

32. A large benefit for a few and relatively small unfavourable effects for many The main benefit, which is prevention of deaths, and the main harm, with is the over-detection, is not know to the individual participant On the other hand, individual participants are confronted with less serious harms, false positive and false negative test results. Screening programmes will always cause harm as well – but all efforts should be made to minimise them as much as possible! Physical harm: e.g. invasive interventions Psychological harm: e.g. anxiety, additional years of living with a disease,… Social harm: e.g. family relations, employment, insurance, financial implications,… When to screen – which cancer sites to screen?  COST-EFFECTIVENESS OF SCREENING PROGRAMMES

33. If a screening programme is well organised, with high quality and if participation is high  screening might be beneficial Population  Lower cancer-specific mortality  Life-years saved  Less advanced disease stages Individual  May be not dying from disease  May be life-years gained  Less severe diagnostics and treatment needed  May have a higher quality of life When to screen – which cancer sites to screen?  COST-EFFECTIVENESS OF SCREENING PROGRAMMES

34. Screening programmes People must be enabled to make a deliberated choice by informing them on all aspects: harms and benefits => Informed choice Whether people decide to participate or not in screening, providing sufficient and correct information will increase the awareness of the public (+) Objective is to detect in an early stage to increase prognosis and decrease late effects of treatment (less agressive treatment) When to screen – which cancer sites to screen?  CONSEQUENCES

35. When becomes screening acceptable? Correct test: proven effectivenes – preferably in several well set-up randomised clinical trials Positive balance between fabourable and unfavourable effects Correct frequency: periodical screening, but not too often (costs ↗) Correct risk group: broad age range, but not too young and not too old (=> identification of target population) Optimal quality of organisation and performance of screening Continual evaluation is essential When to screen – which cancer sites to screen?  CONSEQUENCES

36. 5. Controversy Breast cancer screening:  Tabár et al. Radiology. 2011 Swedish Two-county Trial: longest-follow- up (29 years) of any breast screening trial a highly significant decrease in breast cancer-specific mortality at 29 years of follow-up: 1 death prevented for every 414 or 519 women screened for a 7-year period 42 years of life saved per 1000 women ASP = active study population PSP = passive study population RR = relative risk  Autier et al. BMJ. 2011 Comparison of breast cancer mortality within 3 pairs of neighbouring European countries with different levels of screening Despite time differences in implementation of mammography screening between the country pairs → similar reductions in mortality  suggest that screening did not play a direct part in the reductions in breast cancer mortality

37. Proven effectiveness and acceptable unfavourable side-effects => population-based screening more efficient than ad hoc screening of individual patients Screening always implicates negative effects => balanced information on both advantages and disadvantages is indispensable Population-based screening aims to improve public health. This can collide with interests of individual participants Organising a screening programme is complex. Effects only visible after a long period. 6. Summary