screening & prevention Folkert van Kemenade Dept. Pathology, - - PowerPoint PPT Presentation

Cervical carcinoma: pathogenesis, screening & prevention

Folkert van Kemenade

• Dept. Pathology, Erasmus MC

Rotterdam

Disclosures

Disclosure interests speaker Potential Conflicts of interests None Voor bijeenkomst mogelijk relevant relaties met bedrijven Bedrijfsnamen

• Sponsoring of
• nderzoeksgeld
• Honorarium of andere

financiele vergoeding

• Aandeelhouder
• Andere relatie, namelijk

This presentation has 3 parts

• 1. Cervical cancer: epidemiology - pathogenesis
• 2. Population based screening
• 3. New screening & primary prevention

Cx ca incidence in the world varies

http://gco.iarc.fr/today/home

As incidence increases, mortality does..

Cervical carcinoma in NL (HIC):

• Almost a rare disease (6-7/ 100.000)
• LIC & LMIC much higher incidence. Eg Malawi has around

80-90 per 100.000

• Highest incidences have been noted: in HIV infected

patients in Tanzani (around 2000) this was a staggering 200 per 100.000

•  Now we zoom in on incidence dynamics in NL

Trends in NL and Finland over 60 years

• Trend:

decreasing incidence since 50-ties / after WO II.

In UK, mortality lowered in older women but rose in younger women. After the 70-ies. .

• Incidence in women

(arrow 1).

• Mortality rose form 0,7 per

100.000 (UK) in ‘63-’67  to 2,2 per 100.000 in ‘83- ’87: the rise occurred in young women. In older women, mortality decreased in the same period (‘83-’87).

• Peto ea Lancet 2003.

1

Cervical carcinoma (“ cxca “)

• In summary: even in high income countries, there have

been times with higher mortality but never as high as currently in lower income countries

• Incidence varies widely in the world.
•  now for pathogenesis in detail.

Cervix Uteri (‘ baarmoederhals’) contains a transition of 2 epithelia with a squamocolumnar junction

Squamo Columnar junction is a sharp transition under microscope.

~

SCJ is dynamic: it can move up, making columnar cells squamous (estrogen influence)

SCJ is visible with a colposcope. Then it is called the transition zone (it’s not a line..)

Squamous transition of columnar cells =

• metaplasia. This is vulnerable for transformation

Transformation via precursor phases. May take 10-20 yrs to develop cancer (histology)

• Normal

LSIL HSIL Cxca

Recap: precursor phases are preceded by squamous metaplasia and may transform.

• Epithelium changes from columnar to squamous =

metaplasia (squamous metaplasia)

• As a results, the SCJ ‘moves’ down or up (depending the

availability of estrogens)

• Metaplasia may then undergo cellular morphological

changes (also recognizeable by a colposcope)

Precursor phases in colscopic view

CIN3 /HSIL CIN2 /HSIL

CIN1LSIL Schiffman et al. Review. JNCI. 2011;103:368–383

Morphologic terminology of laesions of the metaplastic transformation zone

New terminology

• Normal maturation can be

perturbed slightly: low grade squamous intra- epithelial lesions (LSIL)

• moderately to fully

perturbed (HSIL) Older terminology

• There perturbed

maturations are called ‘cervical intra-epithelial neoplasia’s’ or CIN’s.

• CIN1 or Low SIL
• CIN2
• CIN3 (+CIS).

Model: metaplasia transforming into an autonomous, transformed lesion

So far, phenomenology.

Precursor laesions can be also easily detected with cytology..

• Easy test. Provides an opportunity to detect precursors,

remove them and prevent transformation (Papanicolaou’s smear*’ or PAPtest)

• This method is appealing but not without problems.
• *the cervix lies only a few centimeters from the

external world (accessible by speculum examination)  exfoliative cytology

Taking a Pap smear requires skill. It can not be self sampled.

Interpretation of exofoliative cytology also requires a skill …

For many years, a cytopathic effect was observed, reminiscent of a virus

• C. Crum. Gynaecopatholgy

Perinuclear cytoplasm cleared- enhanced rim, shrunken, nucleus shrunken = koilocytosis

In the 80-ies: Harold zur Hausen proved human papillomavirus as a agent

• With DNA probes, viral sequences could be detected in

both precursor laesion and cases

• It was not Herpes but Papillomavirus
• It was different from skin warts and genital warts…

Human papillomaviruses (HPV)

• Currently > 300 different HPV types have been

identified (50 nm)

• Most strictly epitheliotropic (squamous/columnar) and

some ~40 mucosal HPV types are known:y8u,..

• low risk types (HPV 6, 11) associated with genital warts
• high-risk types (n=15) associated with cervical cancer

With knowledge of HPV…the pathogenesis makes more sense

1. Is metaplasia more sensitive to hrHPV infection? Can an innocent, low grade precursor phase arise in the ‘ junctional area’. 2. However, most HPV infection resolves spontaneously: the host isn’t even aware of it and only a minority develop into high grade precursor or eventually cancer 3. Two hrHPV types (16, 18) cause most cancers (75%) of

• all. Another 15 types cause the rest

A closer look at the virus: the HPV genome organization

• Circular, double stranded

DNA genome (8 kb/9-10 genes);

• 6 early ORFs (E1-E7),
• 2 late ORFs (L1, L2)
• And viral oncogenes:
• E6: inactivates p53
• E7: inactivates pRb

HPV life cycle

Recap: HPV-induced carcinogenesis is a 10-20 year multistep process

Transient infection latent infection?

intraepithelial neoplasia invasive cancer grade 1 2 3

Woodman et al., Nature Reviews Cancer 2007; Steenbergen et al. Nature Reviews Cancer 2014

HPV-mediated cervical carcinogenesis

Accumulation of (epi)genetic changes in host cell DNA

Steenbergen et al., Nature Reviews Cancer 2014, 14: 395-405

Going a little bit deeper…

• Lisa Mirabello et al had a publication in Cell, volume 170,

Issue 6, Pages 1164-1174 e6 (September 2017)

• HPV16 E7 Genetic Conservation Is Critical to

Carcinogenesis

They sequenced 5570 viral genomes from 4 cohorts

Compiled four cohorts

• In all four cohorts, A1 or

A2 were in majority (85%)

• Extremely variable: only

24% of genomes were shared between women

• Different sites with HPV:

72% were the same controls vs cxca cases

• Cumulative variant

analysis: controls had a significant higher number

• f variants compared to

precancer/cancer cases

• E7 ORF had statistically
• signif. fewer non

synonymous nonsense variants.

Cell 2017 170, 1164-1174.e6DOI: (10.1016/j.cell.2017.08.001)

Summarizing

• E7 for some reasons, remains preserved (still tentative)
• Cxca develops after a long period with hypermethylation
• f genomic and viral genes (not shown)
• HPV integrates into the genomes (after episomal phase)

and this is usually associated with E2 disruption (not shown)

This presentation has 3 parts

• 1. Cervical cancer: epidemiology - pathogenesis
• 2. Population based screening
• 3. New screening & primary prevention

Screening a population is NOT easy

• 1. There are many more reversible precursor laesions or

HPV infected women than irreversible laesions. You must screen for clinically undetecteable lesions

• 2. You must find laesions prior to development of cancer (if

you screen for cancer: find low stage ones!)

• 3. Overdiagnosis bias, length bias & lead time bias

Yet, everybody in the western world started screening with cytology

• Cumulative wordwide: > 80million Paptests
• Commoditized with liquid based cytology
• Computer aided recognition
• But still..

• Lead time bias: tumor are

detected prior to symptoms.

• An apparent longer survival

between diagnosis and death

A screening progamme is fraught with pitfalls

• Length time bias: a more likely

detection of tumors with a long preclinical phase

23-feb-07

early diagnosis of cervical cancer

• Overdiagnosis bias: you detect tumors people would never have

died from. An apparent better survival of patiente with tumor detected by a screening programme

Overdiagnosis bias: not all precursors lead to cancer

This limitations hold for most cancer screening programs, but certainly for cytology

• The WHO has made

criteria (in 1968) and modified since.

• Jungner and Wilson

criteria *(www.who.int)

• Before you start: model!
• In the US, 50 million of

PAP smears are done yearly (NL: 750.000 until 2016).

• In 1989 Leopold Koss:

The US experience is NOT unique but showed how screening can harm

George Sawaya (1998) did the counting

• Koss (JAMA ’89): cytology is

screening a triumph because it saved many lifes but a tragedy:

•  Initial clinical examination /

the taking of the smear samples/ Laboratory errors in screening and interpretation can all go wrong and do go wron

• guidance of the physician.

Before you start: you need to model..

early diagnosis of cervical cancer

• In the Netherlands, they simply started to screen and then

did the modelling..

• After the screening programme was arrested in 1993 and

restarted in 1996 after the modelling

Before you start, model many scenario’s and find optimal cost-effective programme

• In NL cytology

based screening had started prior to 1993 (bottom up).

• In 1993 they

modelled, chose a programme and centralized invitations

• Quality control

<1996 > 1996

Screening in NL was remodelled in 1996 but then rolled out accordingly

George Sawaya (1998) < 1996 > 1996* Startage –last age 35 – 55 30 – 60 Screening interval 3 5 Deaths prevented 3900 4563 Programme costs* 1025 x106 990 x106 # smears 7 7

In summary: cytology based programmes

• Seem to reduce cxca incidence, based on much
• bservational, (epidemiologic) research
• It’s capacity to reduce mortality is much harder to prove
• A formal trial powered for incidence/mortality would be

prohibitedly large (about a million women to invite; screen vs not-screen and watch cancer incidence & mortality… ).

• Mortality is a ‘messy’ endpoint because treatment can

also improve

Two ‘ incidental’ proofs from European countries on cytology screening

• Compare e.g. West-

Germany with East- Germany

• West Germany started

screening in 1970. East Germany started to catch up from1990.

• NB they didn’t measure
• vertreatment.

The second one from the UK: mortality started to rise in younger women but it stopped.

• Incidence in women (arrow

1).

• Mortality rose form 0,7 per

100.000 (UK) in ‘63-’67  to 2,2 per 100.000 in ‘83-’87: the rise occurred in young women.

• Cervical screening prevented

an epidemic is a famous statement by Sir Richard Peto (Lancet 2003).

1

How about NL after the restructuring?

• In contrast, the decrease

in incidence and mortality seemed to slow down after 2000, despite screening.

• And in NL, 25% of all

women is NOT screened

Nationwide profile. Green = normal

• result. Orange : repeat. Red: refer.

25000 100000 200000 300000 400000 475000 25000 100000 200000 300000 400000 475000

Big problem 1 : participation!

caption

= 3000 women

Big problem 2: missed cases Big problem 3: false positivity

= CIN2+

True negatives false positive

false positiv e True positive

False negative Not tested, possibly CIN2+_

= 21000 women per yr

This presentation has 3 parts

• 1. Cervical cancer: epidemiology - pathogenesis
• 2. Population based screening
• 3. New screening & primary prevention

What to do? Switch to hrHPV based screening, offer self sampling & vaccinate

• 1. A vaccin or primary

prevention became available (a Virus Like Particles against type 16 & 18 (bivalent) or 16, 18, 6 and 11 (6 and 11 cause genital warts) could do thejob

• 2. Switch to hrHPV based

screening in the meantime

• 3. Participation is a problem in

screening : offer self sampling

First : hrHPV based screening. In 2000-2007

• Large randomized trials demonstrated hrHPV to have

better sensitivity than cytology. Pooled analysis showed reduction of incidence of cxca in the test group.

• hrHPV test is more objective and reproducible than the

more subjective cytology reading

• However, it also gives more false positive results  more

referrals for colposcopy  more harm.

• How to find a balance?

HPV testing is more upstream than cytology

Schiffman JNCI 2013.

23-feb-07

early diagnosis of cervical cancer

The NL sceening programme had 2 groups that modelled

• Since 2017 NL has an

hrHPV test based screening programme

• A brush is still needed to

sample to cervix

• DK, S, N and AU also

have hrHPV based screening

Programme in NL since ’17: hrHPV based screening

New developments:

• Introduction of self sample method (no more need to

brush). This may possibly recruit some non-responders

• Reduction in the number of labs that do the test (5 rather

than 44)

• From 40-60, women are only screened 2x instead of 4x

with cytology: longer protection.

• In 2023 the first vaccinated girls enter the programme in NL

New Cervical Cancer screening programme since 2017

Stop cytologie screening november 2016 Start hrHPV screening january 2017

15 march 15 february 23 january 1 february 1 march

Summing up so far

• hrHPV is necessary but not sufficient to cause cervical

cancer ( primary prevention)

• The long preclinical precursor phase allows for

screening prevention programme (= secondary screening)

• Maintenance of WHO criteria are essential to prudent

roll out a programme (Anttila. Eur J cancer 2015: towards better implementation in Europe).

• Participation remains the main problem

So, how are we doing after 1 year? First results.

• DATA PROVIDED BY HELEEN

VAN AGT.

• ERASMUS MC UNIVERSITY

MEDICAL CENTER, DEP. OF PUBLIC HEALTH,

• THIS STUDY WAS FUNDED BY

NATIONAL INSTITUTE FOR PUBLIC HEALTH AND THE ENVIRONMENT

• 15 march

15 februa 23 january 1 february 1 march

Participation rate as yet in new vs old screening programme (SSK: self sample kit).

8% of attenders chose self-sampling device

More detection of precursos lesions (from direct referral) despite less coverage

More CIN2+ & more irrelevant findings (CIN1 or lower) in new programme

Conclusions

• Results were as expected but not the decrease in

participation neither the increase in positivity

• Higher number of referrals and higher number of

cervical lesions detected than in the previous cytology- based screening programme, - as predicted

• Higher number with normal histology or CIN1: we have

to reduce the number of over referral

And for a vaccin with Viral Like Particles (VPN)

• More on that later..

L1 pentamer

22-1-08 SBBHK- NBNL– info

HPV types in baarmoederhalskanker

Cancer cases attributed to the most frequent HPV genotypes (%)

HPV genotype 2.3 2.2 1.4 1.3 1.2 1.0 0.7 0.6 0.5 0.3 1.2 4.4 2.6 6.7 2.9 10 20 30 40 50 60 70 80 90 100 X Other 82 73 68 39 51 56 59 35 58 52 33 31 45 18 16 Vaccine types 53.5 17.2 53.5% 70.7% 77.4% 80.3%

Munoz N et al. Int J Cancer 2004;111:278–85.

Vaccin efficacy trials.

Participants randomised VACCINE PLACEBO (sham vaccin) Number of CIN3+ Number of CIN3+

FUTURE 2 Q4 vaccin (n=5305) Placebo (n=5260)

• bserved efficacy

Geen HPV bij start CIN 2 or worse

1 58 98 %

Vaccin trial: Future 2 (16-23 yrs of age)

3 jaars follow up; NB CIN3 kan zich ontwikkelen in 3 jr

Vaccin will have a slow effect on reducing cx ca

• Takes time before full

impact is materialized

• In countries with a

programme, the effect is mitigated

• Optimal scenario with

100% stable ab levels and optimale vaccin uptake

• Goldie S et al. J Natl Can-

cer Inst 2004; 96: 604–15.

10 20 30 40 50 60 70 80 90 100 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 no vaccine 98% Efficacy vaccination

Vaccine simulation:prevention of persistent HPV 16/18

Age Incidence of carcinoma (per 100,000)

Vaccin has quicker impact on ‘precursorprofile’ in cytology screening

Cytological abnormalities Vaccinatie 16,18 Cytological abnormalities

h i s t

• l
• g

i e Pap 2/3a1 / CIN 1 Pap 2/3a1/ CIN1

>=Pap 3a2/ cin2-3

>=Pap 3a2

In Australia they started vaccination in 2008

• They already notice the disappearance of warts (hrHPV 6

and 11)

• CIN 2 and CIN3 are decreasing
• CxCa is becoming a rare disease in Australia
• The combination vaccination and 1 or 2 hrHPV screens at

eg 35 / 45 years allows for eliminatino of cxca

• The WHO had adopted this: world wide elimination before

2080

Thanks for your attention