1 Vaccines and Related Biological Products Advisory Committee Meeting Preventive HPV Vaccine Endpoints November 28-29, 2001 Bethesda, MD Karen L. Goldenthal,

1

Vaccines and Related Biological Products Advisory Committee Meeting

Preventive HPV Vaccine EndpointsNovember 28-29, 2001Bethesda, MD

Karen L. Goldenthal, M.D.Division of Vaccines and Related Products ApplicationsUS Food and Drug Administration

2

Acknowledgements

Douglas R. Pratt, M.D., M.P.H. Antonia Geber, M.D. Donna K. Chandler, Ph.D. Jerry P. Weir, Ph.D. Gopa Raychaudhuri, Ph.D. Jerry D. Willett, M.D. Rebecca L. Sheets, Ph.D. Robert W. Anderson, Ph.D. Rakesh Pandey, Ph.D. Karen Midthun, M.D. Theresa M. Finn, Ph.D.

3

Cervical Cancer Global Perspective

• Worldwide, 3rd most common CA in women* – After breast and colon/rectum cancer– Developing countries: 2nd most common CA– Developed countries: 6th most common CA

• Worldwide, ~400,000 to 500,000 new cases/yr• Disturbing trend, Finland: Between 1991-95, 60%

in cervical CA incidence for women <55 yrs of age#

• Worldwide, 190,000 deaths per yr**– ~78% in developing countries

*Parkin DM, et al. Int J Cancer 1999;80:827–41. (Worldwide CA incidence, 1990)**Pisani P, et al. Int J Cancer 1999;83:18–29. (Worldwide CA mortality, 1990)#Dillner J, 2000; Anttila et al., 1999.

4

Cervical Cancer U.S.A. Perspective

• In 1930s, cervical CA most common cause of CA deaths in U.S. women

• Incidence/mortality rates for cervical CA declined dramatically following Pap screening & intervention

• For 2001, ~12,900 new cases of cervical CA and 4,400 deaths due to cervical CA projected for U.S.*

• Woman's lifetime risk of cervical CA dx currently ~0.85%; risk of dying from disease is ~0.30%**

*Janicek MF & Averette HE, 2001; Greenlee RT et al., 2001 **Ries LAG et al., 2000 (Surveillance, Epidemiology, & End Results [SEER] data,

NCI)

5

Precursor Lesions of Cervical Carcinoma

From Figure 6.13, DeMay RM. The Art and Science of Cytopathology. CD-ROM. ASCP. 1999.

Wright TC, Kurman RJ, Ferenczy A: Precancerous Lesions of the Cervix. In Kurman RJ, ed: Blaustein's Pathology of the Female Genital Tract. 4th ed. New York: Springer-Verlag NY Inc, 1994.

LSIL = Low Grade Squamous Intraepithelial Lesion CIN = Cervical Intraepithelial NeoplasiaHSIL = High Grade Squamous Intraepithelial Lesion

LSIL HSIL

CIN 1 CIN 2 CIN 3

Normal Mild Moderate Severe Carcinoma Dysplasia Dysplasia Dysplasia in situ

6

Above, right. CIN 3, cervix. Note adjacent koilocytes (bottom right). http://hsc.virginia.edu/med-ed/path/gyn/cervix4.html

Above, left. Squamous carcinoma in situ (CIS), cervix. To the right is normal squamous epithelium with its basal cell layer & orderly maturation upwards. The left side shows CIS characterized by lack of maturation, altered cell polarity, nuclear pleomorphism & increased nuclear / cytoplasmic ratio. As is characteristic of CIS, neoplastic cells are confined to the epithelial layer & have not invaded through the epithelial basement membrane, under which, the submucosal stroma contains chronic inflammatory cells. Cornell Medical School http://www.pathinfo.com/alphlit.htm

7

Cervical Lesions: “Pyramid” of Findings (U.S.)*

*From ALTS: The ASCUS/LSIL Triage Study Newsletter, Vol 1, Issue 1**More recent projection: 12, 900 cases/yr for 2001Janicek MF & Averette HE, 2001; Greenlee RT et al., 2001 ASCUS = Atypical Squamous Cells of Undetermined Significance. A recommendation from The Bethesda 2001 Workshop: Replace term ASCUS with ASC (Atypical Squamous Cells).

Cancer ~15,000 cases/yr**

HSIL ~300,000 cases/yr

LSIL ~1,250,000 cases/yr

ASCUS ~2,000,000 cases/yr

8

Natural History of CIN

Table 7 (Modified) from Östör AG: Natural history of cervical intraepithelial neoplasia: A critical review Int. J Gynecol. Pathol. 12:186-92, 1993. (Literature Review)

Regress Persist Progression toCIN 3 Invasion

CIN 1 57% 32% 11% 1%

CIN 2 43% 35% 22% 5%

CIN 3 32% <56% >12%

9

Natural History of CIN

Limitations of this literature review, e.g.:*• Biopsy may alter natural history• 1950s-1990; Non-uniform diagnostic criteria• Cytology does not always predict histology • “Jigsaw puzzle effect” e.g., CIN 1 & 3 in

same cervix• Lack of long-term follow-up• Estimate CIN 3 to ICC “>12%” likely too low,

may be 20-30% within 5-10 yrs**

* Östör AG: Int J Gynecol Pathol 12:186-92, 1993. (Lit Review)**Chang AR: CIS of the cervix & its malignant potential. A lesson from New Zealand. Cytopathology 1:321-8, 1990.

10

Natural History of Cervical Dysplasiaa Cytology Progressionb Progression to Regression

Invasive Cancer to Normalc

6 mos 24 mos 6 mos 24 mosASCUS 2% 7.2% 0.06% 0.25% 68.2%LSIL 6.6% 20.8% 0.04% 0.15% 47.4%HSIL 6.8% 23.4% 0.15% 1.44% 35.0%

aData source: Melnikow J et al., Natural history of cervical squamous intraepithelial lesions: a meta-analysis. Obstet Gynecol 1998;92:727-735. Also, from Table 3 in: Bristow RE et al. Workup of the Abnormal Pap Test. Clinical Cornerstone 3:12-24, 2000.bFollow-up smears or biopsies showed a higher grade lesion than theentry cytology. For HSIL, progression to CIN 3 (from grade 2) or CIS. cNeg. result of F/U cytology or biopsy. No relationship found between proportion of subjects regressing to normal & length of follow-up.

11

HPV Vaccine Composition

• 93% HPV positive by PCR*• 5 most common HPV types:

HPV 16 in 50% HPV 31 in 5.3 % HPV 33 in 2.8%HPV 18 in 14% HPV 45 in 8.4%

• Subsequently, 97% HPV positive with improved retesting (e.g., more sensitive primers); 99.7 % w/author-defined adequate specimens**

* Bosch FX et al., J Natl Cancer Inst 1995;87:796-802.

** Walboomers, J.M.M. J. Pathol 1999;189:12¯19.

Meijer C et al.: Commentary in Histopathology 1998;33:83-6.

Worldwide survey >1000 invasive cervical cancers*

12

Squamous Cell Carcinoma* - HPV 16 Most Common Type

HPV Bosch IwasawaType N=881 N=352

HPV 16 51% 78%

HPV 18 12% 16%

Bosch FX et al. JNCI 87:796-802, 1995Iwasawa A et al. Cancer 1996;77:2275–9.*Cervix

13

Adenocarcinoma (Cervix) - HPV 18 - Most Common Type

Bosch FX et al. JNCI 1995; 87:796-802. (Global cases)Iwasawa A et al. Cancer 1996;77:2275–9. (Finnish cases)Andersson S et al. Eur J Cancer 2001;37:246-50. (Swedish cases)Pirog EC et al. Am J Pathol 2000;157:1055–62. (US & Polish cases) *Excludes 6 HPV neg cases of nonmucinous adenocarcinomas (clear cell, serous & mesonephric).

HPV Bosch Iwasawa Andersson Pirog Type N=25 N=108 N=131 N= 67*

HPV 16 28% 17% 24% 36%

HPV 18 56% 56% 37% 36%

14

Incidence of Squamous Cell Carcinoma (SCC) vs. Adenocarcinoma (ADC) of the Cervix

• Cytology screening - far less impact on incidence of ADC compared to SCC*

• In U.S., review of SEER database has shown – A clear decrease in incidence of SCC, and

invasive cervical cancer (ICC) overall; – However, an increase in the rate of ADC

• Overall, similar findings in Scandinavia*Smith HO, et al. The rising incidence of adenocarcinoma relative to squamous cell carcinoma of the uterine cervix in the U.S.--a 24-year population-based study. Gynecol Oncol 2000;78:97-105.

15

Risk of CIN 2 or 3: Relation to HPV Type 16 and 18

• 241 women presenting for STD evaluation & with negative cervical cytology*

• Followed q4 months (ave. F/U 25 mos): – Cytology, colposcopy, HPV DNA, STDs

• CIN 2 or 3 (biopsy) in 28 women (11.6%)– If + HPV type 16 or 18, adj. RR 11 (4.6-26) for CIN

2 or 3 compared to those w/o HPV– Attributable proportion to types 16 & 18 = 52% – Only 36% (10/28) had LSIL Pap before HSIL Pap

• Subset HPV DNA neg. on 1st 2 visits – In subsequent 2 yrs, 8.6% (9/105)** had CIN 2/3 dx

*Koutsky L, et al.: A cohort study of the risk of CIN 2 or 3 in relation to papillomavirus infection. NEJM 327:1272-8, 1992. (STD clinic, Seattle, WA)**8 of 9 cases occurred among 35 women who developed HPV infection

16

Risk of CIN 2 or 3: Relation to HPV Type 16

• 1075 women with normal cytology and HPV negative tests at baseline*

• Followed q6 months: Pap smear & HPV DNA (median FU = 26 months)

• If abn. cytology, referred to colposcopy/bx• CIN 2 or 3 (biopsy) in 28 women (2.6%)

– Median time to dx = 36 months from study entry– 20 of these CIN 2/3 cases diagnosed during work-

up of the 1st episode of abn. cytology– If + HPV type 16, adj. RR 8.5 (3.7 to 19) for CIN 2

or 3 compared to those w/o HPV

*Woodman CB, et al. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet 2001;357:1831–36.

17

Stages of Review and Regulation

Clinical Investigational Plan

Phase 1SafetyImmuno-genicity

Phase 2Immuno-genicitySafetyDose Ranging

Phase 3EfficacySafetyImmuno-genicity

BLAData to support approval;Inspection

Phase 4InspectionSafetyEfficacyLot Release

BLA SupplementPost-approvalChanges:New IndicationsDosingManufactureEquipment/Facilities

IND

IND = Investigational New Drug Application; BLA = Biologics License Application

18

“Theoretical” Endpoints - HPV Preventive Vaccines

Types 16, 18 and other “high risk” types• Virology (various definitions of incident or

persistent infection)• LSIL cytology (w/o biopsy confirmation) or

worse with virology• CIN 1 histology, AIS*, or worse with virology

(Hereafter = CIN 1 or worse with virology)• CIN 2/3 histology, AIS, or worse with virology

(Hereafter = CIN 2/3 or worse w/virology)• Invasive cervical cancer with or w/o virology

*AIS = Adenocarcinoma in situ (of the cervix)

19

Possible Endpoints – HPV Vaccines

Virology (only) - Advantages• Feasibility• Smaller Trial, i.e., a few thousand• Populations exist in many countries

w/sufficient incident HPV “infection” rate, e.g., for Type 16 – U.K. - 10.5% cum for 3 yrs (Woodman CB et al., 2001)– U.S. - 7% cum for 2 yrs (Ho GYF et al., 1998)– U.S. - 4.7 / 100 person yrs (Thomas KK et al., 2000)– Brazil - 1.68 / 100 person yrs (Franco EL et al., 1999)

20

Possible Endpoints – HPV Vaccines

Virology (only) - Advantages• HPV strongly associated with HSIL cytology,

CIS and cervical cancer*• Determine vaccine versus non-vaccine

serotypes• Might be able to determine immune correlate

*Koutsky LA, et al. N Engl J Med 1992;327:1272-8. (CIN 2/3 and HPV infection)Woodman CBJ, et al. Lancet 2001;357:1831–36. (CIN 2/3 and HPV infection)Ylitalo N, et al. Lancet 2000;355:2194-8. (HPV Viral load and CIS)Josefsson AM, et al. Lancet 355:2189, 2000. (HPV Viral load and CIS)Ylitalo N, et al. Cancer Res 2000;60:6027-32. (Long term HPV infection and CIS)Bosch FX, et al. J Natl Cancer Inst 1995; 87:796-802. (HPV and cervical cancer)

21

Possible Endpoints - HPV Vaccines

Virology (only) - Disadvantages• HPV infection is not a clinical disease• Most HPV infection resolves• Durability of protection uncertain

– ? Change in lifetime risk for HSIL histology/cancer• Not as proximal to cancer as other endpoints

– e.g., may overestimate vaccine efficacy with regard to high grade lesions/cancer

• May want definitive high grade clinical endpoint data before extensive deployment of a new HPV vaccine

22


Virology (only) - Disadvantages• Uncertainty in existence of, or detection of

latent infections in the cervix– Detection in superficial vs. basal layer cells?

• Could a vaccine-induced immune response make HPV DNA more difficult to detect?

• Uncertainty distinguishing new infection from re-infection (although this problem would tend to negatively bias efficacy estimate)

• Various definitions of persistent infection, e.g., infection X 2 at 4, 6 or 12 months, not validated in previous clinical trials

23


Virology (only) - Disadvantages• May not identify unanticipated vaccine-

associated problems with virology endpoint: – Alteration in HPV natural history, difficulty in

detecting cervical dysplasia, enhanced disease (dysplasia, cancer)

– Post-approval, could be difficult to detect outside context of a randomized, well-controlled trial

• Small efficacy trial – Provides less well-controlled safety data – Could be addressed w/supplemental trials

24


LSIL cytology or worse w/Virology - Advantages

• Feasibility• Smaller trial, i.e., a few thousand• LSIL leads to many clinical work-ups

in developed countries

25


LSIL cytology or worse w/Virology - Disadvantages

• Clinical relevance: About half of LSIL resolves w/o therapy

• LSIL cytology, by itself, does not represent a definitive diagnosis – Need histologic diagnosis for therapy– Without colposcopy and biopsy, if needed,

cannot R/O higher grade lesions – CIN 2/3 found at work-up in a % of LSIL cases,

even w/o prior hx of cervical disease

26


LSIL cytology or worse w/ Virology - Disadvantages (cont.)

• Not as proximal to cancer as other endpoints• May want definitive high grade clinical

endpoint data before extensive deployment of a new HPV vaccine

• May be easier to identify unanticipated vaccine-associated problems with high grade disease endpoint (see earlier slide)


27


CIN 1* or worse w/ Virology - Assumptions• CIN 1 (biopsy) or worse (CIN 1 +) = 10 endpoint• Virology used to classify CIN 1 + cases as

vaccine serotype or not– Prespecify HPV testing on cervical samples vs

histology to classify cases • Cases would be identified from colposcopic

work-up of (mostly) ASCUS & LSIL cytology• Plan/algorithm for colposcopy will affect # of

endpoints (more frequent = more endpoints)• Can’t use economics as basis for US licensure * Histology

28


CIN 1 histology or worse w/ Virology - Advantages

• Feasibility• Smaller trial, i.e., a few thousand• Populations exist in many countries with

sufficient incident CIN 1 • Necessitates definite work-up

– Generates histologic dx that guides therapy• Data are available on the natural history of

CIN 1 from initial diagnosis

29


CIN 1 histology or worse w/Virology - Disadvantages

• At least half of CIN 1 resolves w/o therapy

• Not as proximal to cancer as other endpoints• May want definitive high grade clinical

endpoint data before extensive deployment of a new HPV vaccine

30


CIN 1 histology or worse w/Virology - Disadvantages (cont.)



31


CIN 2/3 or worse w/ Virology - Assumptions• Virology used to classify CIN 2/3 + cases as

vaccine serotype or not– Prespecify HPV testing on cervical samples vs

histology to classify cases • Many/most CIN 2/3 + cases would be identified

from colposcopic work-up of ASCUS and LSIL* • Plan/algorithm for colposcopy critical to # of

endpoints (more frequent = more endpoints)• Many/most CIN 2/3+ cases could be those

found at 1st work-up of abnormal cytology* Lonky NM, et al. The clinical significance of the poor correlation of cervical dysplasia & cervical malignancy with referral cytologic results. Am J Obstet Gynecol 1999;181:560-6.

32


CIN 2/3 histology or worse w/Virology - Advantages

• More proximal to cervical cancer • Provides definitive high grade clinical

endpoint data before, before wide-spread public health use

• Natural history data - Prevent lesions that need therapy (US standard of care)

33


CIN 2/3 histology or worse w/Virology - Advantages (cont.)


• Larger efficacy trial – Implications for randomized safety database

34


CIN 2/3 or worse w/Virology - Disadvantages

• Feasibility - Subject #s may not differ from most vaccine efficacy trials; however, type of follow-up likely to be more resource intensive

• Uncertainty trial size/ duration estimate??– Little natural hx data to est. trial size: women

w/ baseline neg. HPV, normal cytology – 12,000 women, randomized 1:1 (vaccine to

placebo), vaccine efficacy > 80%, 3 1/2 yrs*

* Preliminary estimate based on data from: Woodman CBJ, et al. Lancet 2001;357:1831–36.

35


Cervical Cancer w or w/o Virology - AssumptionsOne scenario (developed country)• Randomize subjects; infrastructure in place for routine

F/U & a method of capturing all diagnoses for an area, e.g., cancer registry and death certificates– Scandinavian Countries (Comprehensive CA registries)

• HPV testing: Presence or absence of baseline testing before randomization, & cervical sample/ histology testing will affect efficacy level assessment

• Esp. in U.S., many/most cervical CA cases identified from colposcopic work-up of ASC, AGC and LSIL*

* Lonky NM, et al. The clinical significance of the poor correlation of cervical dysplasia & cervical malignancy with referral cytologic results. Am J Obstet Gynecol 1999;181:560-6.

36


Cervical Cancer with or w/o Virology - Advantages

• The major concern is cervical CA• Provides definitive high grade clinical

endpoint data before, before wide-spread public health use


37


Cervical Cancer with or w/o Virology - Advantages (cont.)

• Give best understanding of impact of vaccine (if Type 18 included) on adenocarcinoma

• Larger efficacy trial– Implications for randomized safety database,

although detailed data not expected

38


Cervical Cancer with or w/o Virology - Disadvantages

• Feasibility• Uncertainty

– Trial size/ duration estimate/ population selection/countries w/o screening program??

39

Efficacy Trial Analysis

• Statistical section of protocol (prospective)– Detailed, especially for 1o endpoint– Interim analysis

• Intent-to-treat vs. per protocol: vaccine serotypes

• Handling of mixed infections, esp. vaccine & non-vaccine types, in 1o endpoint analysis

• Evidence of overall benefit when all endpoints, including ones attributed to non-vaccine high risk HPV types, are included in an analysis

40

Accelerated Approval (AA)

• 21 CFR 601.40s Subpart E: Accelerated Approval of New Biologic Products for Serious or Life-Threatening Illnesses*

• “…provide meaningful therapeutic benefit to patients over existing treatments (e.g., ability to treat patients unresponsive to, or intolerant of, available therapy, or improved patient response over available therapy).”

* Federal Register 57;58942-60, 1992.See 21 CFR 314.500s Subpart H for AA of new drugs

41


• “FDA may grant marketing approval for a biological product on the basis of adequate and well-controlled clinical trials establishing that a biological product has an effect on a surrogate endpoint that is reasonably likely, based on epidemiologic, therapeutic, pathophysiologic, or other evidence, to predict clinical benefit, or on the basis of an effect on a clinical endpoint other than survival or irreversible morbidity.”

42


• Intended to make available promising therapies while definitive confirmatory efficacy studies being completed

• Confirmatory post-marketing studies:– Usually well underway at time of AA– Must be adequate and well-controlled – Must be carried out with due diligence

• Original, and current, intent of AA regulation = serve best interests of public

• Preventive vaccines have not been previously approved using AA regulation

43

FDA Questions 1. Please discuss and identify the most appropriate

endpoints for traditional approval of HPV vaccines intended to prevent cervical cancer: In particular, please discuss the use of the following endpoints in clinical trials intended to demonstrate the efficacy of HPV vaccines for oncogenic types, and the indications (e.g., prevention of HPV infection, etc.) these endpoints would support: a. Incident HPV infection by oncogenic HPV types

(i.e., at least one positive HPV DNA test result).b. Persistent HPV infection by oncogenic HPV types.

- Regarding this endpoint, please also discuss the appropriate number of positive virologic results, and the interval between positive virologic results.

c. LSIL (cytology), associated w/oncogenic HPV types.d. CIN 1, associated with oncogenic HPV types.e. CIN 2/3, associated with oncogenic HPV types.f. Cervical cancer.

44

FDA Questions (continued)2. Please discuss the use of the accelerated approval

regulations for licensure of HPV vaccines for the prevention of cervical cancer: Specifically, please discuss and identify possible surrogate endpoints to support accelerated approval. In particular, consider the following endpoints:a. Incident HPV infection by oncogenic HPV types. b. Persistent HPV infection by oncogenic HPV types. c. LSIL (cytology), associated w/oncogenic HPV types.d. CIN 1, associated with oncogenic HPV types.

In the context of accelerated approval, please discuss and identify possible endpoints for the confirmatory trial.

Documents

1 Vaccines and Related Biological Products Advisory Committee Meeting Preventive HPV Vaccine Endpoints November 28-29, 2001 Bethesda, MD Karen L. Goldenthal,