Screening for cervical cancer

1
Screening for cervical cancer

• Human Papillomavirus DNA versus Papanicolaou
  Screening Tests for Cervical Cancer
• Marie-Helene Mayrand, M.D., Eliane Duarte-Franco,
  M.D., Isabel Rodrigues, M.D., Stephen D. Walter,
  Ph.D., James Hanley, Ph.D., Alex Ferenczy, M.D.
  Sam Ratnam, Ph.D., Francois Coutlee, M.D., and
  Eduardo L. Franco, Dr.P.H., for Canadian Cervical
  Cancer Screening Trial Study Group NEJM. October
  18, 2007.
• Ivania Rizo
• February 19, 2008

2
Cervical Cancer

• Second most common cause of cancer-related
  morbidity and mortality among women in developing
  countries
• - 371,200 new cases annually with a 50 percent
  mortality rate
• Past 50 years 75 decrease in incidence and
  mortality in developed countries due to Pap tests
  and the detection of precursor lesions and
  earlier-stage cancers
• However in the US, invasive cervical cancer
  remains the cause of death for almost 4000 women
  each year, approximately 1.3 of cancer deaths in
  women.
• Squamous cell carcinomas account for
• approximately 80 percent of cervical
• cancers and adenocarcinomas
• 15 percent.

3
Background

• Risk factors include early onset of sexual
  activity, multiple sexual partners, smoking, and
  immunosuppression
• Mean age of diagnosis of invasive cervical cancer
  is 47 in the US, and the incidence increases with
  age
• Human papillomavirus (HPV) is central to the
  development of cervical neoplasia and can be
  detected in almost all cervical cancers

4
Current Screening guidelines

• Annual Pap testing approximately 3 years after
• onset of sexual activity or at age 21 ( ACS
  and USPSTF)
• Cease screening after hysterectomy for benign
  conditions or after age 65 (USPSTF) or 70 ( ACS)
  in women with adequate recent screening, who are
  not otherwise at high risk of cervical cancer
• Max screening interval should be 3 years, with
  more frequent screening at onset and in high-risk
  situations (HIV, prior dysplasia, chronic
  immunosuppression)

5
Current Screening Guidelines

• HPV testing as an adjunct to cervical cytology
  for screening is an acceptable strategy when used
  in women aged 30 years or older. Combined
  screening should only be done once every 3 years
  since combined tests have a high NPV (99-100)
  for high grade lesions.
• Preferred management of ASCUS in most situations
  is reflex HPV testing ( high risk types) on the
  original LBC specimen collected for Pap test
• Women with positive ASC-H and LSIL tests should
  undergo colposcopy
• Minimal follow up for high grade squamous
  intraepithelial lesion, atypical glandular cells,
  and AIS interpretations includes colposcopy and
  endocervical curettage

6
Objective

• To determine whether testing for DNA of oncogenic
  HPV is superior to the Pap test for
  cervical-cancer screening

7
Participants

• Women ages 30 69 years. Per Mayrand et al., the
  study was limited to women older than 30, since
  in sexually active women younger than 30 years,
  transient HPV infections are common, which gives
  HPV DNA testing an unacceptably low specificity
  in identifying cervical cancer precursors.
• Women who sought screening tests for cervical
  cancer in any of 30 clinics in Montreal and St.
  Johns, Canada

8
Exclusion Criteria

• Currently being followed for a cervical lesion
• Lack of cervix
• Pregnant
• Hx of cervical cancer

• Undergone Pap testing in the previous year
• Unable to provide consent

9
Study design

• Randomized, controlled trial
• Participants were randomized 11 to 1 of 2 arms
  the Pap screening arm or the HPV screening arm.
  Participants were blinded to arm allocation.
• There was insufficient evidence regarding the
  efficacy of HPV DNA testing as a stand-alone
  screening test to withhold Pap cytology from
  women in the trial. Therefore, each arm included
  both tests and the authors randomized the order
  in which the test samples were collected.
• A random sub sample of 10 of the women in St.
  Johns and 20 of the women in Montreal with
  negative index tests were invited to undergo
  colposcopy. Per Mayrant et al. this allows for
  correction of verification bias.
• Cytotechnologists and cytopathologists were not
  aware of the patients status as participants in
  the study or HPV test results. Colposcopists and
  pathologists were not aware of the screening-
  test results.
• Supported by a grant from the Canadian Institute
  of Health Research and partially by unrestricted
  grant from Merck Frosst Canada

10
Outline of Study Design Mayrand MH et al.
Randomized controlled trial of human
papillomavirus testing versus Pap cytology in the
primary screening for cervical cancer precursors
design, methods, and preliminary accrual results
of the Canadian Cervical Cancer Screening Trial.
International Journal of Cancer 2006 119615-23
11
Screening Tests

• Conventional ( slide smear) Pap test- results
  reported according to the Bethesda System
  terminology
• Results of ASCUS, AGC (atypical glandular cells),
  or worse were considered positive
• Hybrid Capture 2 test ( HC2 probe B, Digene) was
  used for HPV testing and tested for the 13
  high-risk HPV types most commonly associated with
  high-grade dysplasia and cancer
• Positive HPV test if specimen had at least 1 pg
  of HPV DNA per milliliter

12
Terminology Used in Reporting Cervical
cytology
Bundrick et al. Screening for cervical cancer and
initial treatment of patients with abnormal
results from papanicolaou testing. Mayo clinic
proceedings.2005 80(8) 1063-1068.
13
Screening Tests

• Colposcopists followed standard protocol
• - endocervical curettage
• - ectocervical biopsies of all
  abnormal-appearing cervical regions
• - at least one biopsy of normal-appearing
  ectocervical epithelium aiming at an aceto-white
  transformation zone
• Biopsy first and then treatment afterwards if
  warranted

14
Study Endpoint

• Primary endpoint Histological CIN2, CIN3, and
  cervical cancer
• 2 Case definitions
• Liberal definition all cases of grade 2 or 3
  CIN, adenocarcinoma in situ, or cervical cancers
  that were histologically confirmed on the basis
  of any of the histologic specimens
• Conservative definition met liberal criteria and
  in addition were confirmed in the LEEP specimen
  or confirmed on biopsy when ablation was done

15
Statistical analysis

• Differences in categorical data were assessed by
  Fishers test and the chi-square test
• Differences in continuous data was assessed by
  Kruskal-Wallis test
• Mayrand et. al. obtained verification in a random
  sample of participants with negative screening
• Verification bias caused by verification of the
  lesion only in participants with a positive
  result, can lead to overestimates of sensitivity

16
Enrollment and Outcomes
Mayrand M et al. N Engl J Med 20073571579-1588
17
Characteristics of Participants According to
Study Group and Center
Mayrand M et al. N Engl J Med 20073571579-1588
18
Results

• According to the conservative definition
  sensitivity of the Pap test (55.4) was
  significantly lower than the sensitivity of the
  HPV test ( 94.6, P0.01)
• Specificity of the Pap test after the correction
  for verification bias was 96.8 versus 94.1 in
  the HPV test ( Plt0.001) using the conservative
  definition
• Difference in corrected sensitivity between the
  liberal and the conservative definitions
• There were 4 cases of high-grade intraepithelial
  neoplasia among participants in whom both
  screening tests were negative according to the
  liberal definition. According to conservative
  definition there were no cases. These 4 cases
  influenced how the authors extrapolated the
  occurrence of high-grade cervical intraepithelial
  neoplasia lesions among all women with negative
  tests. The authors state that this explains the
  difference in corrected sensitivity between the
  liberal and conservative definitions.
• The results of the tests were not influenced by
  the order in which they were collected
  therefore, the authors pooled the two groups to
  access different screening algorithms.

19
Group-Specific Comparison of Pap and HPV Testing
to Identify High-Grade Cervical Intraepithelial
Neoplasia and Cancer
Mayrand M et al. N Engl J Med 20073571579-1588
20
Comparison of Pap Testing and HPV DNA Testing
Using Combined Study Groups According to
Different Positivity Thresholds and Test
Combinations
Mayrand M et al. N Engl J Med 20073571579-1588
21
Discussion

• According to the CONSERVATIVE definition the HPV
  test was more sensitive (39.2 difference) and
  only 2.7 less specific than Pap testing
• Corrections for verification bias provided
  absolute rather than relative estimates
• Per Mayrand et al. the protocol mandated
  endocervical curettage and biopsies in all
  participants who underwent colposcopy, therefore
  reducing the possibility of missing CIN. But the
  protocol reveals more indolent lesions that would
  probably be missed in routine colposcopy.
• Per authors, the liberal definition detected
  lesions not routinely found in community
  screening and the conservative definition reduced
  overdiagnosis.
• The authors believed that some of the specimens
  classified as CIN negative per LEEP but positive
  per cytology or colposcopy were false positives,
  due to misclassification of squamous metaplasia
  as cervical intraepithelial neoplasia.

22
Weaknesses

• Outcome measure (conservative versus liberal
  outcome definitions) perhaps chosen after the
  fact
• Alternative outcome measure such as
  histologically verified, high-grade cervical
  intraepithelial neoplasia in either the
  colposcopically targeted biopsy specimen or the
  final excision.
• Negative HPV tests that were consequently
  positive on biopsy and negative on excision were
  interpreted as false positive results versus
  false negative results
• Selection bias and inability to generalize
  results to the screening population
• - Exclusion of young women and pregnant women
  perhaps causes a bias against Pap tests and for
  HPV testing
• - Authors did state they excluded women under
  30 because HPV DNA testing would result in low
  specificity

23
Weaknesses

• Per Berkhof and Meijer, in order to reliably
  state the sensitivity of HPV tests, the sample
  size of women with two negative tests and
  histological verification needs to be larger.
• They state that a very small number of positive
  CIN 2 or worse makes a big impact on the
  sensitivity of the HPV test.

24
So what can we conclude?

• We are not ready to use HPV tests for primary
  screening.
• Need rapid, simple, accurate , and affordable HPV
  tests
• At this time, we may choose to use HPV DNA test
  as adjunct to cytology for screening women over
  30 years of age and no more frequently than every
  3 years.

25
References

• Bundrick J, Cook D, Gostout B. Screening for
  Cervical Cancer and Initial Treatment of Patients
  With Abnormal Results From Papanicolaou Testing.
  Mayo Clinic Proceedings. 2005801063-1068.
• Lytwyn et al. Correspondence to Human
  Papillomavirus DNA versus Papancolaou Screening
  Tests for Cervical Cancer. 2008358 (2) 641-
  644.
• Mayrand MH, Duarte-Franco E, Coutlee F, et al.
  Randomized controlled trial of human
  papillomavirus testing versus Pap cytology in the
  primary screening for cervical cancer precursors
  design, methods and preliminary accrual results
  of the Canadian Cervical Cancer Screening Trial
  (CCCaST). Int J Cancer 2006119615-623.