Mednotes

L11 Cancer Screening for cancer

Properties2
tagscancer, lecture
aliasesL11 Cancer Screening for cancer

20 min read

Overview

This lecture covers the principles of cancer screening (ethics, screening criteria, test performance measures, benefits/harms, and evaluation biases), then applies these to the current NZ cancer screening programmes for cervical, breast and colorectal cancer, and to the unresolved cases of prostate and lung cancer screening.

Ethics and definitions of screening

  • Cochrane & Holland (1971): there is an ethical difference between everyday clinical practice and screening. In everyday practice a doctor responds to a patient’s request and is not responsible for gaps in medical knowledge. In screening, because the practitioner initiates the process, they should have conclusive evidence that screening can alter the natural history of disease in a significant proportion of those screened.
  • Screening is performed on healthy, asymptomatic people, distinct from diagnostic testing (performed on people with symptoms/suspected disease).
  • Screening programmes are organised activities with quality control that try to screen an entire population, versus opportunistic screening, which reaches people ad hoc (often at the GP).

NZ criteria for assessing a screening programme

A programme should be assessed against:

  • The condition is suitable for screening.
  • There is a suitable test.
  • There is an effective and accessible treatment/intervention for the condition once detected early.
  • High-quality evidence, ideally from RCTs, that the programme reduces mortality or morbidity.
  • Potential benefit outweighs potential physical and psychological harm (from the test, diagnostic procedures and treatment).
  • The health care system can support the whole screening pathway (diagnosis, follow-up, evaluation).
  • Social and ethical issues are considered.
  • Cost-benefit is considered.

Sensitivity, specificity and predictive values

A screening test divides people into four groups based on whether further testing confirms disease: true positive, false positive, true negative, false negative.

Using a 2x2 table (screening test result x true disease status present/absent), with cells a (true positive), b (false positive), c (false negative), d (true negative):

  • Sensitivity = ability to correctly identify those with disease (“detection rate”) = = true positives / (true positives + false negatives).
  • Specificity = ability to correctly identify those without disease = = true negatives / (true negatives + false positives).
  • Positive predictive value (PPV) = probability that a person with a positive test is a true positive = .
  • Negative predictive value (NPV) = probability that a person with a negative test does not have the disease = .
  • Predictive values are especially useful to clinicians, who must interpret test results to patients.

Benefits and harms of screening

Benefits (Chamberlain, 1984):

  • Improved prognosis for some people whose disease is detected by screening.
  • Less radical treatment, curing some people with early disease.
  • Reassurance for those with negative results.

Disadvantages (Chamberlain, 1984):

  • Longer morbidity for those whose prognosis is unaltered.
  • Over-treatment of questionable abnormalities.
  • False reassurance for false-negative results.
  • Anxiety and sometimes morbidity for those testing false positive.
  • Unnecessary intervention for false-positive results.
  • Hazards of the test itself (e.g. venepuncture, radiation risks).
  • Resource costs; diversion of scarce resources to screening.
  • Possible stigma for those testing positive.
  • Insurance implications.

Biases in evaluating screening

  • Lead time bias: screening advances the date of diagnosis, extending the interval between diagnosis and death even if the actual time of death is unchanged. People whose disease was screen-detected appear to survive longer than those diagnosed by symptoms, without any true survival benefit. For screening to be genuinely effective, early treatment must extend life by more than this lead time.
  • Length bias: fast-growing tumours progress rapidly through the preclinical phase and are less likely to be caught by infrequent screening. Screening at infrequent intervals therefore disproportionately detects slow-growing tumours with a good prognosis, biasing apparent screening benefit upward.
  • Selection bias: people who take up screening may differ in underlying risk of disease/mortality from non-participants, so their prognosis would have differed even without screening.
  • Over-diagnosis bias: screening may detect abnormalities of questionable malignancy that would never have been diagnosed (or caused harm) without screening.
  • The only study design not affected by these biases is a randomised controlled trial (RCT) with mortality as the outcome measure: participants are randomised to an intervention group or control group, and mortality is compared between groups.

Cervical cancer screening (NZ)

  • About 164 registrations per year (2022, excluding in-situ) and 60 deaths per year (2018); much higher incidence and mortality in Maori.
  • Incidence was stable until 1991, then fell by about 50% after the National Cervical Screening Programme (NCSP) began in 1990. Mortality declined slowly (improved treatment plus the 1990s fall in incidence), and its decline has tapered off in the most recent 10-year period; mortality was 1.7 per 100,000 in 2018.
  • Cervical cancer is highly preventable with early detection.
  • Policy: aims to reduce incidence and mortality by detecting relevant HPV types and by early cytological detection/treatment of precancerous squamous changes. Delivered via primary care, with a Cervical Screening Register and reminders.
  • HPV primary screening is recommended for people with a cervix aged 25-69.
  • Regular screening can reduce risk of developing cervical cancer by 90%+. Only persistent infection with a high-risk HPV type carries cervical cancer risk; high-risk HPV types cause >95% of cervical cancers. Genital warts are caused by low-risk HPV types not associated with cervical cancer. There is no treatment for persistent HPV infection itself, but there is treatment for premalignant/malignant cervical lesions.
  • From 2023 the primary test changed to HPV testing, with a self-testing option. This also began free screening for people aged 30+ who have never screened or not screened in 5+ years. Two options must be offered: a vaginal HPV swab (self-taken or by a health professional), or a cervical smear sample (liquid-based cytology, taken by a health professional).
  • Self-testing: a negative HPV test means very low risk of abnormal cells over the next 5 years, so routine screening is only needed every 5 years if HPV is not detected (every 3 years if immune-deficient). Women still consult a healthcare provider even when self-testing; mail-out self-test kits are being considered for the future.
  • Self-swabbing occurs in privacy at clinics, doctors’ surgeries, community clinics, or Marae and Pasifika community events; some centres offer a take-home option. As of 2025, >80% of participants choose the self-test; estimated overall coverage is 73%.
  • If a vaginal swab detects HPV, follow-up testing for cell changes is needed. People aged 25-69 who have ever had intimate skin-to-skin or sexual contact should be screened every 5 years (3 years if immune-deficient).
  • The HPV test does not fully replace the smear test, which remains needed to check whether HPV has caused cell abnormalities. People with HPV types 16 or 18 detected are referred to a specialist colposcopist.
  • Current efforts target the lower screening rate among Pacific, Asian and Maori women, aiming to improve access for unscreened/under-screened women. The HPV immunisation programme (introduced 2008, funded for ages 9-26) will have a future preventive effect; vaccinated women should still be screened, since the vaccine does not cover all high-risk HPV types and no vaccine is 100% effective.
  • Cost: the NCSP is not fully funded; free screening is available for people who are Maori or Pacific (any age); aged 30+ and under-screened or never screened; requiring follow-up testing (any age); or holding a Community Services Card (any age). All follow-up tests/treatment are free for programme participants.
  • Monitoring: anonymised NCSP Register data are regularly analysed against indicators/targets covering participation, clinical outcomes, provider performance and overall programme performance, with corrective action taken as needed.
  • Risks and harms: psychosocial impact of an abnormal result and referral for colposcopy/treatment; discomfort/unpleasantness of the test; possible unnecessary treatment of lesions that would have regressed (though it is not possible to identify who will progress, so all women with high-grade changes are treated); colposcopy treatment carries a small risk of preterm labour or miscarriage in the second trimester; some cancers are missed by screening.

Breast cancer screening (NZ)

  • Breast cancer is the most commonly registered cancer in women (approx. 30% of all women’s cancer registrations, 3,660 registrations per year, 2022); Maori women have higher registration and death rates.
  • Policy: aims to detect invasive breast cancer while small and without spread, reducing mortality and morbidity. Ages 45-74, 2-yearly, two-view mammography (mediolateral and oblique views), free of charge, delivered via 8 lead providers (35 fixed and mobile units), with double reading. Results go to women and GPs within 2 weeks.
  • Aim/benefits: to reduce morbidity and mortality via earlier, less extensive treatment. Target coverage is 70% of eligible women screened 2-yearly; 2024 actual coverage was 59% Maori, 63% Pacific, 67% other women. In women aged 45-69, mammograms find 8-9 out of 10 cancers. Regular mammograms can reduce risk of dying from breast cancer by more than a third.
  • Harms: anxiety, inconvenience, discomfort, radiation exposure, false positive and false negative results, interval cancers, overdiagnosis. The programme is monitored to minimise harms.

Colorectal cancer screening (NZ)

  • 3,545 registrations and >1,200 deaths per year (2022). Male incidence is significantly higher; non-Maori age-standardised incidence exceeds Maori. Overall age-standardised incidence has been decreasing in recent decades, but the ageing population means annual case numbers are increasing. Early-onset CRC (age <50) incidence has been increasing, especially in Maori.

  • Screening test strategies: faecal occult blood (FOB) testing aims to detect invasive cancer at an early stage; flexible sigmoidoscopy or colonoscopy aims to find and remove preinvasive adenomatous polyps (a different, preventive strategy); other modalities include double-contrast barium enema and combined methods.
  • National programme: free bowel screening every 2 years for eligible people aged 58/60-74. The age range has just been expanded from 60-74 to 58-74 in Northland, Auckland and the South Island, with the rest of NZ following from March 2026. Three health districts ran 2-year pilots (ending 2025) offering screening to Maori and Pacific peoples aged 50-74.
  • Delivery: participants are sent an invitation letter, consent form and test kit (an instructions booklet in multiple languages, a stool collection tube, and a return envelope). The National Coordination Centre sends letters after negative results and notifies GPs electronically of all results. Regional Centres are advised of all results and ensure colonoscopy or other investigation is offered after a positive iFOB result. Hospitals deliver colonoscopies and cancer treatment.
  • Benefits: can detect bowel cancer at its earliest stage, when there is a 90% chance of cure; can prevent some cancers from developing. FOB testing is effective in reducing colorectal cancer incidence and mortality in cohort studies and trials; regular screening reduces the risk of dying from bowel cancer by 16% (Cochrane Database of Systematic Reviews).
  • Harms: a cancer can be missed if it was not bleeding at the time of testing; cancer can develop in the interval between screens; not all detected cancers can be successfully treated; colonoscopy after an abnormal test carries risks such as bleeding or bowel tear; some people find the FOB test unpleasant.
  • Programme indicators (Feb 2025): overall participation 57.8% (target 60%), with disparities by ethnicity: Maori 51.5%, Pacific peoples 39.1%, Asian populations 44.2%, other 62.4%.
  • Increased-risk groups: 2012 NZ guidance covers personal history of colorectal adenomas/CRC, inflammatory bowel disease, and family/whanau history of CRC, with three risk categories (1: slightly above average, 2: moderately increased, 3: potentially high risk). Updated (2023) Te Whatu Ora surveillance recommendations for family history (FHx, FAP, HNPCC etc.) base risk categories, and surveillance recommendations, on the number of affected relatives, whether first- or second-degree, and their age at diagnosis. Source: Guidance on Surveillance for People at Increased Risk of Colorectal Cancer (New Zealand Guidelines Group, 2011).
  • Future directions: follow-up of population-based RCT results; newer methods such as virtual colonoscopy and faecal DNA testing.

Prostate cancer screening

  • Prostate cancer is the most common cancer registration in men (4,330 registrations in 2022) and accounts for about 13% of male cancer deaths.
  • 99% are adenocarcinomas, with uncertain potential for many lesions to progress. About one-third of men in their 40s-50s have histologically evident (mainly well-differentiated, microscopic) prostate cancer, but clinically significant tumours are typically larger and higher grade. This means screening can detect many cancers that would never cause death. Men with a family history are at greatest risk.
  • History: an Advisory Group on prostate cancer screening was convened in 2001 (NZGG); in 2004 the National Health Committee recommended against population screening, while emphasising better advice to men and GPs.
  • Test performance: digital rectal exam (DRE) is a poor screening test (PPV 6-30%); transrectal ultrasound (TRUS, PPV 5-9%) is considered a diagnostic rather than screening test; PSA is a better screening test but of unproven benefit (PPV 28-35%); combined abnormal PSA and DRE gives PPV 38-50%.
  • Policy: population-based screening of asymptomatic men is not recommended, due to low specificity and sensitivity and insufficient evidence that DRE or PSA screening reduces mortality. Ad hoc PSA testing (opportunistic screening, DRE and PSA) is nonetheless widespread. There is potential harm from unnecessary treatment.
  • A survey of NZ GPs (Durham et al., NZMJ 2003) found some form of prostate screening was performed by 98% of GPs, and 50% supported a national population-based programme.
  • Durham (quoted): men must be given accurate, understandable information about screening; it is unacceptable for men to suffer morbidity and regret from learning the harms of screening only through experience, without understanding the small chance of benefit.
  • Possible harms: from biopsy, infection (rare) and bleeding (common); from treatment, incontinence and sexual dysfunction (common). The International Prostate Screening Trial Evaluation Group: decisions on population-based screening cannot be made before RCTs are completed.
  • A meta-analysis of RCTs (Djulbegovic et al., BMJ 2010; incomplete per the slide) of six trials totalling 387,286 participants found screening was associated with an increased risk of being diagnosed with cancer (RR 1.46; 95% CI 1.21-1.77) but only a small, non-significant effect on death from prostate cancer (RR 0.88; 95% CI 0.71-1.09); all trials had major methodological limitations.
  • The question of a formal screening programme remains controversial: at best, screening has a modest effect on prostate cancer mortality, at the cost of substantial over-diagnosis and over-treatment; even experts find balancing this a challenge.
  • Further reading (optional): Armstrong et al. (2017), four expert viewpoints on PSA testing of average-risk men, covering uncertainty, harms, evidence trustworthiness, cost and the potential to limit harm via more conservative treatment; they generally agree men should decide for themselves, with considerable support likely needed to do so.
  • Targeted (rather than widespread) testing may be most appropriate, given tumour biology, testing specificity and complications from intervention (bpac.nz, 2020).
  • The Kupe website (Ministry of Health-sponsored) helps men understand the risks, benefits and implications of prostate testing to support an informed conversation with their doctor.
  • A 2025 Canadian group proposed implementing population-wide, adaptive platform trials embedded in the healthcare system, enabling real-time integration of new technologies, standardised protocols and equitable access (Nguyen et al., Lancet Reg Health Am 2025).

Lung cancer screening

  • Lung cancer is the commonest cause of cancer death in Australia and NZ.
  • Targeted screening of people at highest risk aims to detect early-stage disease that is more amenable to curative treatment.
  • Findings from large international trials (US National Lung Screening Trial; Dutch-Belgian Randomised Lung Cancer Screening Trial) and a 2019 systematic review showed benefit, but the benefit depends on the risk profile of the target group. Screening high-risk people with low-dose CT improves outcomes (Manners et al., 2021).
  • NZ’s first lung cancer screening trial (Auckland, announced 19 May 2021), Maori-led and funded with a $2 million grant, will use low-dose CT scans on approximately 500 people at high risk of lung cancer (Maori smokers or ex-smokers, aged 55-74) to assess risk level. Professor Sue Crengle (PI): low survival is largely due to late diagnosis; international trials show early detection with low-dose CT can reduce lung cancer mortality by 20-24% in high-risk groups.

Transcript gaps

Slide 47 (a photo of the national bowel screening test kit) had no caption/title text extractable by pdftotext; the image content is described but any accompanying title could not be confirmed.
Slides 64 and 65 could not be distinguished from slide 63’s content in the native text extraction (text ran continuously from slide 63 into slide 66) and were not rendered, so their content could not be independently verified; they may be build/animation duplicates of slide 63 with no new material.

Self-test

  1. Explain the ethical distinction Cochrane and Holland drew between everyday clinical practice and screening.
  2. Distinguish screening from diagnostic testing, and distinguish an organised screening programme from opportunistic screening.
  3. List the NZ criteria used to assess whether a screening programme is appropriate.
  4. Using the 2x2 table cells a (true positive), b (false positive), c (false negative), d (true negative), give the formulas for sensitivity, specificity, PPV and NPV, and state what each measures.
  5. List the benefits and the disadvantages of screening described by Chamberlain (1984).
  6. Describe lead time bias and explain why it can make screening appear to improve survival even when it does not.
  7. Describe length bias and predict its effect on the tumours detected by infrequent screening.
  8. Distinguish selection bias from over-diagnosis bias in the evaluation of screening.
  9. Why is a randomised controlled trial with mortality as the outcome the only study design unaffected by lead time, length, selection and over-diagnosis biases?
  10. Describe the current NZ cervical screening pathway: eligible ages, testing options, and screening intervals depending on HPV result and immune status.
  11. Explain why a positive HPV self-swab does not replace the smear test, and what happens if HPV type 16 or 18 is detected.
  12. Describe the NZ breast screening programme (target ages, test, frequency, reading protocol) and state its effect on breast cancer mortality.
  13. Distinguish the FOB test strategy from the flexible sigmoidoscopy/colonoscopy strategy in colorectal cancer screening.
  14. Describe how a positive iFOB result is followed up in the National Bowel Screening Programme, from Regional Centre notification through to hospital treatment.
  15. Explain why prostate cancer screening can detect many cancers that would never cause death.
  16. Compare the positive predictive values of DRE, TRUS and PSA for prostate cancer screening, and state NZ policy on population-based prostate screening.
  17. Summarise the findings of the Djulbegovic et al. (2010) meta-analysis of prostate cancer screening RCTs.
  18. Explain why targeted, rather than widespread, testing is considered more appropriate for both prostate and lung cancer screening.

Answers

Graph View

Backlinks