1. NORMAL HPV INFECTION HR-HPV INFECTION HR-HPV PERSISTENCE HG-CIN+ HR-HPV TEST PAP TEST HPV GENOTYPING HPV mRNA, p16-ink4a ScreeningTriageHistologic confirmation IDEAL TEST VACCINATION

2. Technologies for the detection of HPV DNA Ideally, a HPV test should allow the detection of multiple, clinically relevant HR-HPV types, identify individual types, provide quantitative informations about viral load of each individual type, be easy to perform, reproducible, highly specific and sensitive, rapid, automatable and of low cost. The ideal HPV test does not exist yet A limited number of tests is available that perform differently and have a specific clinical applications Translation of HPV testing into clinical practice requires setting rigorous standars of test performance and characteristics Basically, HPV tests can be divided into: Tests for HPV DNA detection and typing Hybrid Capture 2 (HC2) PCR-based assays HPV-based biomarkers

3. Established HPV test technologies and their performances* Type of testDesignationProbes/primers Reaction product Analytical sensitivity, fg Detectable HPV types HybridizationHybrid Capture 2 Mixture of genome lenght RNA probes DNA/RNA hybrids 25-7513 PCRMy09/11 Dot Blot Degenerate primers 450 bp0.1-10039 PCR PGMY09/11 reverse LA Mixture of consensus primers 450 pp0.127 PCR GP5+/GP6+ EIA ELISA Consensus primers 150 bp0.5-1020 PCR GP5+/GP6+ reverse LA Consensus primers 150 bp0.5-1037 PCRSPF-10 LIPA Mixture of consensus primers 65 bp0.1-1043 * Modified from Iftner & Villa, JNCI Monographs, 2003

4. The different performances of HPV tests impact on the estimated prevalences of infections and the clinical significance of HPV status Analytical sensitivity should be clearly differentiated from clinical sensitivity when assessing the clinical relevance of a HPV test Snijders PJF et al. J Pathol, 2003

5. PROs and CONs of available HPV technologies Choice of a test format involves a tradeoff between performance (accuracy, reproducibility and coverage), cost, laboratory requirements, ease of integration in current protocols of screening and diagnosis HC2 is available in large scale format, does not require specific laboratory facilities, does not depend on PCR amplification, provides a semiquantitative estimate of viral load, has been largely validated for screening HC2 does not allow identification of individual types and have some degree of cross-reaction between types PCR-based assays allow individual HPV type identification and are more analytically sensitive than HC2. PCR-based assay vary in sensitivity, specificity and coverage depending mainly on the type of primers and the lenght of amplicons; in general, they show good agreement but their performance in detecting distinctive types especially in multiple infections can be different.

6. Full spectrum genotyping Detection of individual types in multiple infection Distinction of incident from prevalent HPV infections Post-vaccination surveillance Monitoring of HPV type burden in different populations or settings Prediction of type-specific oncogenic risk and/or persistence Triage of HPV-positive women with negative or low-grade cytology or follow-up of women with previous SIL Persistence of any HPV type for 1 yr or more increases the RR of CIN3+ 42.9 times (Koshiol J et al, Am J Epidemiol 2008) Cytologicaly negative women testing positive for HPV16/18 are at higher risk for prevalent and incident CIN3+ than HPV16/18-positive women with L-SIL (Castle PE et al. JNCI 2005) Baseline positivity for HPV16/18 is associate with 4-6x risk of CIN3+ at 10 years compared to other HPV types (Khan MJ et al, JNCI 2005) Women with abnormal cytology and multiple HPV infections are at greater risk for CIN2+ (Dal Bello B et al. J Med Virol in press)

7. 65% of women with abnormal PAP have multiple HPV infections HPV typeAll women (N. 912) Women with multiple infections (N. 585) N. OR 95% CI CIN1CIN2+ 6162 (17.8)123 (21)5 0.80 (0.28-2.26)0 undefined 11106 (11.6)87 (14.9)2 0.81 (0.16-4.06)1 0.94 (0.11-7.95) 16385 (42.2)305 (52.1)35 8.89 (4.49-17.58)10 5.14 (2.12-12.48) 18115 (12.6)109 (18.6)2 undefined3 undefined 31245 (26.9)197 (33.7)23 17.13 (5.78-50.79)14 26.77 (8.37-85.64) 4548 (5.3)37 (6.3)5 12.65 (1.46-109.52)3 17.85 (1.82-175.41) 51163 (17.9)135 (23.1)8 4.10 (1.31-12.81)4 4.80 (1.24-18.52) 52171 (18.7)142 (24.3)11 7.27 (2.27-23.34)5 7.69 (1.99-29.70) 5361 (6.7)39 (6.7)3 3.72 (0.61-22.54)1 2.85 (0.25-32.02) 5649 (5.4)39 (6.7) 5834 (3.7)29 (5) 4212 (1.3)6 (1) 3335 (3.8)33 (5.6) 3518 (2)15 (2.6) 3926 (2.9)18 (3.1) 4033 (3.7)31 (5.3) 596 (0.7)5 (0.9) 434 (0.4)4 (0.7) 4413 (1.4)12 (2.1) 6621 (2.3)19 (3.2) 6814 (1.5)11 (1.9) 703 (0.3)2 (0.3) 741 (0.1)1 (0.2)

8. The number of HPV types is correlated with CIN severity Outcome Number of infecting HPV types 0123>3Total Cytology* Negative132 (32%)61 (19%)61 (15%)22 (14%)3 (10%)279 L-SIL169 (41%)196 (60%)238 (59.5%)80 (52%)18 (56%)701 H-SIL8 (1,8%)19 (6%)42 (10%)31 (19%)9 (28%)109 ASC/AGC102 (25%)51 (15%)59 (15%)20 (13%)2 (6%)234 Total411327400153331323 Histology** Negative231 (94%)56 (26.5%)40 (13.5%)7 (5%)0 (0%)334 CIN19 (4%)109 (51%)160 (53%)39 (30%)3 (10%)320 CIN2-31 (0,5%)41 (19.5%)95 (31.5%)75 (57%)28 (90%)240 Carcinoma4 (1.5%)5 (2.5%)6 (2%)10 (8%)0 (0%)25 Total24521130113131919 * Spearman Rho = 0.094, p = 0.001 ** Sperman Rho = 0.67, p <.001

9. The number of HPV types is correlated with CIN severity NegativeReference CIN10 HPVReference 1 HPV50,7824,0 – 107,49 2 HPV107,6650,13 – 231,22 ≥ 3 HPV177,0261,48 – 510,0<.001 Age0,950,94 – 0,97 Previous SIL/CIN0,430,19 – 0,94 CIN≥2 -0 HPVReference carcinoma1 HPV38.41,.5 – 101,4 2 HPV121,446,3 – 318,4 ≥ 3 HPV817251 – 2655<.001 Age0,970,95 – 0,99 Previous SIL/CIN0,460,20 – 1,07 Odds ratios and 95% confidence intervals (CI) of histological outcomes according to number of infecting HPV types, age and previous SIL/CIN

10. Current vaccines target less than 50% of CIN2+ associated HPVs VariablesAll w/o HPV-16w/o HPV-6,11, 16, 18 NegativeReferenceReferenceReference CIN1Single infectionReferenceReferenceReference Multiple infection11.5* (7.5-17.5)16.15* (9.6-27.0) 14.40*(7.4-27.9) Age0.95* (0.94-0.97)0.96* (0.94-0.97)0.96* (0.93-0.98) Previous SIL/CIN0.34* (0.15-0.77)0.62 (0.22-1.79)1.01 (0.35-2.90) CIN≥2/Single infectionReferenceReferenceReference carcinomaMultiple infection27.6* (17.2-44.2)19.28*(10.9-34.1) 21.53*(10.5-43.9) Age0.97** (0.95-0.99)0.99 (0.97-1.01)1.00 (0.98-1.03) Previous SIL/CIN0.36** (0.15-0.82)1.79 (0.67-4.76)1.36 (0.43-4.27) * p<0.005, ** p  0.05 The effect of the number of infecting HPV types on the risk of CIN is independent of the main oncogenic types targeted by current vaccines Assuming that the observed distribution of viral types is representative of the general population, current vaccines would prevent an expected fraction of any CIN and of CIN≥2, of 30% and 43%, respectively

11. Guidelines for HPV testing requirements in primary screening* * Meijer C et al, Int J Cancer 2009, 124:516-520 Most candidate HPV DNA assay differ in clnical specificity due to differences in the detection rates of transient infections with low viral loads At present, only HC2 and GP5+ /6+ -PCR-EIA have been consistenly validated in large prospective cohorts or RCTs. Clinical sensitivity for CIN2+ not less than 90% of HC2 in women  30 years Clinical specificity for CIN2+ not less than 98% of HC2 in women  30 years Interlaboratory agreement not less than 87% Specific infrastructure for nucleic acid amplification Laboratory accreditation for clnical molecular testing Monitoring of laboratory performances by proficiency testing Standardization requirements extend largely beyond the perfomances of the specific assay

12. Quality control and standardization requirement for HPV testing* * Modified from Gravitt PE et al. Vaccine 2008, 26S:K42-K52 ONGOING QUALITY CONTROL STANDARDIZED PROTOCOLS AND REAGENTS Specimen collection Specimen transport medium DNA extraction method PCR primers and amplification Type-specific probes Interpretation of data

13. The enormous advancement in knowledge on the biology of SIL can be translated into new screening strategies and pathologic criteria HPV-related biomarkers HPV E6-E7 mRNA (NASBA) HPV viral load HPV integration p16-ink4a IHC Methylation profiles TERC-gain Telomerase/topoisomerase IIa Proliferation /cell cycle markers Quiagen’s fast HPV technology E6 strip test

14. Clinical applications of HPV DNA testing Triage of undetermined cytology Test-of-cure after excision of HG-CIN Triage of L-SIL Primary screening Post-vaccination surveillance Detection of persistent infections Inter-individual transmission Epidemiological surveillance

15. HPV testing for the triage of ASCUS A cytological diagnosis of ASCUS underlays CIN2 in 8,7% [6.9-10.5] of cases and CIN3 in 3.9% [2.4-5.5], on average 43% of women are HPV+. Twenty-two studies using HC2 were considered (Vaccine, 2008, 26S, K29-K41). Pooled sensitivity of HPV DNA detection in predicting CIN2 and CIN3 is 93.1% [91.1-95.1] and 95.5% [92.7-98.2], respectively. Pooled specificity for CIN2 and CIN3 is 62,3% [57.6-67.1] and 60.5% [52.9- 68.2], respectively. Repeat Pap smear was taken in 7 studies, HC2 sensitivity was on average 14% higher for the detection of CIN2+ considering a cytological cut-off of ASCUS (ratio 1.14 [1.08-1.20]) with a similar specificity (ratio 0.99 [0.88-1.10]). The ALTS study provides data of 2 year follow-up with serial cytology for an outcome of CIN3 +. Three repeat smears had a sensitivity of 60% for a cytological cut-off of CIN2 with 12% colposcopy and sensitivity of 97% for a cut off of ASCUS with 73% colposcopy. These figures compare with a sensitivity of 92% for HPV test with a 53% colposcopy rate

16. HPV testing for surveillance after the treatment of HG-CIN Following excision of HG-CIN, 10% of women may experience a recurrence over a 4 year period. Risk is recurrence is higher in older women, has increased in recent years and is lifelong (Strander B et al BMJ, 2007). Sixteen studies were considered that were heterogeneous with respect to design, timing of follow-up, type of HPV DNA assay and inclusion criteria. Sensitivity of HPV DNA detection in predicting treatment failure ranges from 67% to 100%, pooled 94.4% [90.9-97.9] (Vaccine, 2006, 24 suppl, S78-89). Specificity varied between 44% and 100%, pooled 75% [68.7-81.4]. Compared with follow-up cytology, HPV testing was more sensitive (ratio 1.16 [1.02-1.33]) with no significant loss of specificity (ratio 0.96 [091-1.0]). Compared with histological assessment of surgical margins, sensitivity was higher (ratio 1.31 [1.11-1.55]) and equally specific (ratio 1.05 [0.96-1.15]). Women with HPV16 infections should be monitored more closely (Murat G et al, Gynecol Oncol 2007).

17. HPV testing for the triage of L-SIL On average 74.4% of women with L-SIL are HPV+, this limits the capacity of the test to stratify women according to disease severity. Following a cytological diagnosis of L-SIL, histologically confirmed CIN2 and CIN3 are present in 17.6% [11.8-23.3] and 7.4% [2.9-12.0]. This proportion can be substantially lower for women aged  35 (Ronco G et al, Eur J Cancer 2007; Cuzick J et al., Lancet 2003). Eleven studies are available for meta-analysis and systematic review. Pooled sensitivity of HC2 triage for an outcome of CIN2 is 97.2% [95.6- 98.8] and 97.1% [94.0-100] for CIN 3 (Vaccine, 2006, 24 suppl, S78-89). Pooled specificity an outcome of CIN2 is 30.6% [22.7-38.6] and 26.1% [15.1-37.1] for CIN 3. The value of HPV testing in this setting depends on management strategies of L-SIL (follow-up with repeat smear or immediate colposcopy), and the accuracy of colposcopy/biopsy (false negative colposcopy rates as high as 30% have been reported in histological diagnoses of CIN2+).

18. HPV testing in primary screening Pooled data for 25 cross-sectional studies and 3 randomized studies (Vaccine, 2008, 26S, K29-K41). 5-15% of women undergoing screening are HPV-DNA positive with a inverse correlation with age Pooled absolute sensitivity of HC2 for CIN2+ is 89.7% [86.4-93.0] with a wide range (50 to 100%) mainly according to geographic areas. Pooled specificity in excluding CIN2+ is 88.2% [86.2-90.1]. Sensitivity and specificity of HC2 for CIN+ in studies for Europe and North America is consistently higher, 98.1% [96.8-99.4] and 91.7% [90.3- 93.1], respectively, with low values of heterogeneity. Relative sensitivity was 33% [20-47%] higher than cytology for a cut-off of ASCUS, whereas relative specificity was 6% [2-8%] lower. Adding cytology to HPV DNA testing adds little or no sensitivity and reduces specificity.

19. HPV testing in primary screening When HPV DNA testing is used as part of the initial screening, either alone or in combination with cytology, lower rates of CIN2  are detected at subsequent rounds. This further increases if HPV-positive women are triaged with cytology In a recent multinational cohort study, the cumulative proportion of CIN3+ at 6 years for HPV-negative women at baseline was one third that for PAP- negative women (0.27% vs. 0.97%). By comparison, the proportion observed at 3 years for PAP-negative women was 0.51% (Dillner J et al, BMJ 2008). In the Portland study, the NPV of CIN3+ for double-negatives over a 5 year period was 99.91% [99-85-99.9] corresponding to 9/10.000 women that will develop CIN3+ compared with 30/10.000 cytology-screened women (Sherman ME et al, JNCI 2003). These arguments support the idea that screening intervals can be safely extended to at least 5 years

20. HPV testing in primary screening: conclusions HPV DNA testing is substantially more sensitive than cytology to detect CIN2+, whereas it is less specific primarily due to the detection of transient infections that do not produce cytologic changes. In this context, basic principles would suggest that the more sensitive test (HPV DNA) is applied first and the more specific (cytology) is used for triage. Several RTCs and mathemathical models support the cost-efficacy and the increased accuracy of HPV DNA testing as a primary screening modality Additional advantages of HPV DNA testing compared to cytology: objective and automable test allowing better quality control and reducing medico-legal litigation cytology can be reserved for triage of a minority of samples allowing the widspread introduction of more accurate test and better quality controls and trained cytologists reduces the need for triage of ASCUS/L-SIL allows longer screning intervals and possibly later onset of screen allows monitoring of vaccinated women

21. Future scenario of primary cervical screening Women aged 25-64 yrs HPV DNA TESTING NORMAL 5-YEAR RECALL CYTOLOGY HPV TEST & CITOLOGY at 6-12 months COLPOSCOPY NORMAL 5-YEAR RECALL COLPOSCOPY HPV TEST & CITOLOGY at 6-12 months HPV-negativeHPV-positive Normal/bordelineL-SIL+ HPV & cytology negative L-SIL+ borderline cytologyHPV-positive and/or Cuzick J et al. Vaccine, 2008, 26S, K29-K41).

22. HPV testing in primary screening: open questions Should be introduced within organized programs rather than used in an opportunistic setting Organized screening should consider separate protocols according to age and vaccination status CIN2 remains an equivocal clinical endpoint The optimal management for HPV-positive cytology-negative women should be defined The role of multiple HPV infections and persistence should be clarified Further biomarkers for truly progressive epithelial lesions are needed for triage of low-grade abnormalities The impact of vaccination on HPV type circulation and performance of HPV DNA testing and cytology requires attention Ultimately, clinical evidences should be translated into new health policies