CDISC Europe Interchange 2019 - FDA Presentation

CDISC Interchange 2019

Introduction and Overview

Lilliam Rosario, Ph.D.Director, Office of Computational Science

CDISC Interchange 2019

3

Introduction

4

Mission

5

IND Application

6

Phases of Clinical Testing

7

The Review Process

8

New Drug Application

9

The Review Timeline

10

Meet the Team

The Impact of SEND on FDA Review of Nonclinical Study Data

CDISC Europe Interchange 2019

Matthew Whittaker, Ph.D.Nonclinical Reviewer

FDA Center for Drug Evaluation and Research, Office Of New DrugsSilver Spring, MD

May 9, 2019

Disclaimer

This presentation reflects the views of the presenter and should not be construed to

represent FDA’s views or policies

Outline• Overview of nonclinical review at FDA• Janus Nonclinical• Impact of SEND on nonclinical review• Ongoing and future development initiatives

FDA: OND & OCS RelationshipOffice of Medical Products and

Tobacco

Center for Biologics Evaluation and Research

(CBER)

Center for Tobacco Products

Center for Drug Evaluation &

Research (CDER)

Center for Devices & Radiological Health

(CDRH)

Office of New Drugs (OND)

Review Divisions (by Indication)

Clinical reviewers (medical officers)Nonclinical reviewers (Pharm/Tox)

Office of Translational Sciences

Office of Computational Science (OCS)

Tools & training for review of study data

Define functionality requirements for review tools

Typical FDA Review Team for IND/NDADiscipline Role Office

Regulatory project manager

Project management-communicate with sponsor

Office of New Drugs

Clinical Primary reviewer Office of New Drugs

Team Leader

Nonclinical Primary reviewer Office of New Drugs

Team Leader

Chemistry, Manufacturing & Controls (CMC)

Primary Reviewer Office of Pharmaceutical Quality

Team Leader

Clinical Pharmacology Primary Reviewer Office of Translational Sciences –Office of Clinical Pharmacology

Team Leader

Biostatistics Primary Reviewer Office of Translational Sciences –Office of Biostatistics

Team Leader

Nonclinical Reviewers in OND

• 247 Reviewers• 17 review divisions • All are Ph.D.s

– Pharmacology (areas of research vary widely)– Toxicology

• Come from post-docs, academic research, or industry positions

Nonclinical toxicology studies

• Objectives oDefine toxic effects of a drug that could

potentially be seen in humans

oDefine the doses/exposures at which these effects might be expected to occur

Nonclinical studies in support of clinical development

PRE-CLINICAL (before drug is tested in humans)

CLINICAL(testing of drug in humans)

IND submitted

In vitro Pharmacology Genetic Toxicology Safety Pharmacology

o CNSo Cardiovascularo Respiratory

“IND-enabling” toxicology studieso Rat o Non-rodent (dog or monkey) o Dose animals up to the duration of

the proposed opening clinical study (i.e. 14-days or 28-days)

30 d safety review

NDA submitted

10-month review period

Approval

Chronic toxicology o 6 month rat o 9 month dog/monkey

Reproductive toxicology studies o Fertilityo Embryofetal

development

Carcinogenicity Pre & postnatal

development

IND Phase (several years)

Major review principles

(1) Look for findings that show a dose-response relationship

(2) Of those findings, which ones are considered adverse?

• These 2 components are main factors in limit-dose determination

Challenges with current approach to review of nonclinical studies

• Nonclinical study reports (NCSRs) submitted in pdf formato Data analysis requires re-typing of values from summary

tables into Excel o Variable formats used for organization of NCSRs - 1000+

page documents very difficult to navigateo Nonclinical study data is not readily searchable across

studies

• SEND datasets can address each of these issues

Electronic study data regulations• FD&C Act Section 745A(a): Sponsors must use the data standards

defined in the FDA Data Standards Catalog starting 24-months after final guidance is issued for a specific submission type

Standard Application typeDates that standard is

acceptedDates that standard is

required*

SEND 3.0 NDA/BLA/ANDA 6/13/11 – 3/15/19 12/17/16

Commercial IND 6/13/11 – 3/15/20 12/17/17

SEND 3.1 NDA/BLA/ANDA 8/21/17 - 3/15/19

Commercial IND 8/21/17- 3/15/20

*Requirement dates refer to the nonclinical study initiation date (not the date that the study is submitted to the Agency)

SEND studies submitted to FDA

# of SEND Studies (# of applications)

Application type Total* 2016 2017 2018

Commercial IND 369 (174) 3 (2) 47 (30) 319 (145)

NDA 54 (32) 8 (3) 8 (6) 38 (23)

BLA 3 (3) 0 1 (1) 2 (2)

*As of 12/28/18

• All SEND submissions to date have been in SEND 3.0

Janus Nonclinical

Janus Nonclinical

• FDA-specific, web-based application that allows users to analyze and visualize SEND datasets

• Pharm/tox reviewers from OND are guiding development of Janus NC by OCS

•Why?

Janus Nonclinical• Current

o Table & figure format – consistent with general FDA reviewer practices

• Review documents – consistent formatting across reviewers & review divisions

• Ongoingo Interface with IND Smart Template

• Futureo Cross-study analysis (nonclinical)o Nonclinical – clinical data correlation

Janus NC Dashboard• User assigns group alias, control group, set type in

the Study Sets Tableo 5 possible Set Types: interim, terminal, recovery,

toxicokinetic, satellite

Janus NC Summary Table - BW

Janus Nonclinical: Now• Janus NC summary tables reflect summary tables in NCSR

o Mean calculations based on Set Types defined in Dashboardo Highly customizable

• Dynamically filter & sort summary table datao Greatest impact: LB & MI

• Switch from Mean ± SD view to % change from control in 2 clickso Eliminates need for re-typing & analysis in Excel

• Individual animal data available from summary table (click on mean)

Impact of SEND data on nonclinical review

Current impact: Time Savings

Data table Traditional methods – pdf Janus NC

Clinical Chemistry 3+ hrs 30 – 45 mins

Histopathology 6 – 8 hrs 1 – 2 hrs

• Summary table preparation time

Improved review efficiency: benefits

• More time to critically consider observed findings in tox studieso Risk assessment o Sponsor’s perspective & potential explanations for findings

• Potentially earlier engagement with sponsors on specific review issues (rather than at the last minute before the 30 d review clock is about to expire)

Improved review efficiency: benefits• P/T Reviewers: 50+ INDs/NDAs/BLAs in their portfolio at

any given timeo Submissions are constantly coming in to each INDo P/T reviewers: Must keep up with new clinical protocols,

clinical protocol amendments, nonclinical inquiries, new study reports…

• Reality: Prioritize PDUFA deadline-related assignments

• More efficient review of tox studies more time to keep up with ongoing portfolioo More timely responses to inquiries with non-PDUFA

mandated timelines?

Improved review efficiency: benefits

• More time to engage in non-review activities– Subcommittees within FDA

• Examine topics of scientific interest

– Engagement with industry/academic community

Future impact

• More informed regulatory decision-makingo Janus: Warehouse of nonclinical datasets

• Robust historical control databases• Within a single product

o Follow specific findings across studies of different durations

oCompare specific findings across species

• Across productsoUnderstand effects of drug classes (i.e. products

targeting the same molecular pathway) on different nonclinical species

OCS Support for Pharm/tox Reviewers

Integrated approach• Tools

o For analysis and visualization of SEND data• Janus Nonclinical• Other commercially available software applications for specific uses

• Trainingo SEND standard (how nonclinical study data is modeled in SEND)o Practical use of Janus Nonclinicalo Formats: Videos, instructor-led seminars, text documents

• Resourceso KickStart Service

• Conducted at the request of pharm/tox reviewer• In-person seminar

Data quality assessment of SEND datasets Orient reviewer to how their SEND submissions are organized Instruction on use of Janus Nonclinical to identify key findings

o OCS Service Desk

Ongoing & future development initiatives

• Training – for all nonclinical reviewers in ONDo SEND standard (how nonclinical study data is

modeled in SEND)o Practical use of Janus Nonclinical

• Janus Nonclinicalo Optimize summary table displays for all domainso Graphing/ data visualization enhancementso Cross-study analyses

Summary

• Nonclinical reviewers in OND work closely with staff in OCS to guide development of Janus Nonclinical (FDA-specific software application for visualization/analysis of SEND data)

• Janus Nonclinical – increased review efficiencyo Much faster identification of treatment-related findings than

traditional methods with pdf NCSRs

• Increased review efficiency - expected to benefit both FDA and Sponsors

A Clinical Reviewer’s Approach to the Evaluation and Interpretation of Structured Data

Alan M Shapiro, MD, PhD, FAAPHelena Sviglin, MPH

CDISC Interchange 2019

40

Disclaimer

“This presentation reflects the views of the author and should not be construed to represent FDA’s views or policies.”

41

Overview

• Part I – The life of the Clinical Reviewer– Alan Shapiro

• Part II – How OCS impacts review – Helena Sviglin

42

Part I Overview

• Approach to safety evaluation 80 – 100% manual work

• Considerations for safety evaluation• Example of using safety matrix for a product

Safety Evaluation• Summary

43

Integrative Safety Review

Safety review requires comprehensive integrative approach utilizing knowledge from product indication, product class and its development history along with the product’s safety findings

This information needs to be organized in a fashion to help reviewers integrate and make safety determinations

One approach is the ‘Safety Matrix’ which will be discussed in this presentation

44

EVALUATION OF SAFETY RESULTS USING CLINICAL REVIEWER ASSEMBLED INFORMATION

45

Pre-NDA or Pre-BLA Submission Information

45

1 Drugs and biologics can have primary and secondary targets

2Product or product class findings• Literature search• Labels search (e.g. FDALabel)• Post-marketing data from FAERS

3 Pre-clinical and IND phase (Phase 1 –Phase 3) findings• Includes all forms of data• Evaluation usually done with support from pharmacology-toxicology

reviewer• Initial IND submission with pre-clinical reports• Reports and other information from clinical IND phase

o Additional sources of Informationo Study investigational brochureo Development safety update Report (DSUR)

46

Product Indication Information

Underlying disease safety related findings– Drug indication signs and symptoms

• Medical textual resources • Text mining tools are needed to summarize

this information

47

NDA or BLA Submission Safety ComponentsInitial Overview

• Proposed product labeling• Clinical ‘Summary of Safety’• Tabular listing of clinical studies• Study by study Information

– Clinical study report (CSR) for each study and addendums

– Study data reviewer’s guide (SDRG)– Analysis data reviewer’s guide (ADRG)

48

Tools Used for Adverse Event Analyses

• MedDRA Adverse Event Diagnosis (MAED)• MedDRA at a Glance• Internally developed scripts • Commercial tools that integrate and analyze multiple domains

Initial Evaluation of Safety Results

Comparative Clinical Trials Used for Analysis

• Product versus Placebo• Product versus comparison• Dose-safety analyses

49

Considerations for Safety Analysis

Verify findings identified by the

Applicant

Identify product class issues not found so far for

the product

Design custom MedDRA queries to identify safety

signals

Identify new safety signal not

previously identified(not found in IND phase

or other products in class)

50

Example of Safety Evaluation of a Hypothetical Product

• Therapeutic Monoclonal Used Against a Respiratory Viral Disease in Pediatric Patients administered sub-cutaneously

• Respiratory virus infection leads to fever, wheezing, apnea, morbilliform rash, diarrhea and occasionally pneumonia

• Humanized monoclonal antibody targets virus to block infectivity

• Even though monoclonal is humanized, there is a risk of hypersensitivity reactions to this protein product– Hypersensitivity can be manifested by various clinical findings

• Rash and angioedema• Respiratory difficulties or compromise in upper and lower airways• Alteration in liver laboratories

51

Blood and lymphatic system disorders Leukocytosis Neutropenia NeutropeniaCardiac disorders TachycardiaGastrointestinal disorders DiarrhoeaGeneral Disorders and Administration Site Conditions Injection site reaction Injection site reactionHepatobiliary disorders and related Investigations Transaminases increased Transaminase increasedImmune system disorders PneumoniaInfections and infestations Hypersensitivity

Respiratory, thoracic and mediastinal disorders Wheezing, Apnoea Wheezing

Laryngospasm, Apparent life threatening event

Skin and subcutaneous tissue disorders Rash morbilliform urticaria urticaria

MedDRA SOC Disease FindingsProduct and Product Class Safety Issues

Product Safety AE Dataset Findings

Matrix for Evaluation of Safety in MedDRA Preferred Terms

Found in Class and Product safety

Found in Class and not in Product Safety

Found in Disease, Class, and Product Safety

52

Findings from Safety Matrix

• Product class issues are also seen in the product under review yellow– Neutropenia– Injection site reaction– Transaminase increased– Urticaria

pink• Product class issue of hypersensitivity not found in product safety?– NO: findings of laryngospasm and urticaria are consistent

with hypersensitivity reaction• Need safety reviewer with clinical training to associate findings

53

Summary – Part I

Safety review requires comprehensive integrative approach utilizing: knowledge of the indication product chemical and biological properties safety findings in the product class product development history safety findings from submitted datasets

Much of the safety evaluation is manual Clinical safety reviewer needs to be able to comprehensively

identify and characterize safety findings to include in product labeling

Requires clinical expertise and support from others on the review team

Review team needs better data and adequate analytic tools

54

Part II - How does OCS help the Reviewer tackle the safety review?

• Develops review tools that help the Reviewer inquiry and manipulate the study data– MAED– CoreDF

• Provides clinical, review, and data standards expertise– JumpStart– KickStart– OCS Service Desk for additional consult– FDALabel

55

MAED

• OCS developed tool to help reviewers look for potential signals in the tabular adverse event data

• MAED also works on legacy data• SMQ inquiry – broad and narrow• Provides some basic statistics (point estimates,

confidence intervals, etc.) for hypotheses generating

56

MAED

57

CoreDF - Overview

58

CoreDF – Summary Report

59

CoreDF - Coding Report

• Insert slide for CoreDF AE coding report here• Demo the coding report after Alan’s slide as part

of her presentation (CoreDF reports have been prepared for a study similar to the one Alan is talking about. The CoreDF report has been anonymized, but does not plan to release it as part of conference materials)

60

CoreDF – Treatment Arm anamolies

61

CoreDF – Missing Dates

62

CoreDF – Supplemental Domain Contents

63

Summary – Parts I and II

Safety review requires comprehensive integrative approach utilizing: knowledge of the indication product chemical and biological properties safety findings in the product class product development history safety findings from submitted datasets

Much of the safety evaluation is manual Clinical safety reviewer needs to be able to comprehensively

identify and characterize safety findings to include in product labeling

Requires clinical expertise and support from others on the review team

Review team needs better data and adequate analytic tools

CDISC and PhUSE

collaborations

OCS tools and services

64

Additional Resources

• OCS public website• eData@fda.hhs.gov (CDER)• CBER-edata@fda.hhs.gov (CBER)• Study Data Standards Resources website• OCSServiceDesk@fda.hhs.gov

A Statistical Reviewer’s Perspective on Data Review in NDA/BLA Submissions

Weiya Zhang, Ph.D.Statistical Reviewer, CDER/OTS/OB/DBIII

CDISC Europe InterChange, May 9, 2019

Disclaimer

This presentation reflects the views of the speaker and should not be construed to represent FDA’s views or policies

Outline

• Statistical reviewer’s perspective of NDA/BLA review

• Recommendations on NDA/BLA data submissions

Study Data in Clinical Trials

Subjects from study

sitesCRFs SDTM ADaM CSR

TAUGs

• Annotation• Define.xml

• Define.xml• Software

programs

• Software programs

• Follow protocol and SAP

Study Data in Clinical Trials

Subjects from study

sitesCRFs SDTM ADaM CSR

TAUGs

• Annotation• Define.xml

• Define.xml• Software

programs

• Software programs

• Follow protocol and SAP• Reproduce/verify study results• Support product labels?

21st Century Review

https://www.fda.gov/downloads/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CDER/ManualofPoliciesProcedures/UCM218757.pdf

Pre-NDA Communication on Study Data

Topics:

• Study datasets and data standards to be applied

• Software programs

• Study (analysis) data reviewer’s guide (sdrg/adrg)

• ISE/ISS pooling strategies

• …

Suggestions:

• Request pre-NDA meeting

• Understand study data well

• Follow FDA guidance on data submission, communications, etc.

• Be specific on your pre-NDA meeting questions

Pre-NDA Communication on Dataset – Example

Sponsor’s question:

• We will submit study datasets following SDTM 1.3/SDTMIG 3.1.2 for phase 2 trials A1 and A2 and phase 3 pivotal trial A3 and associated define.xml.

• We will submit analysis datasets for the phase 3 pivotal trial A3 and the phase 2 trial A1 following ADaM 2.1/ADaMIG 1.0 as well as supporting data definition files (define.xml).

• We do not plan to submit separate analysis datasets for other clinical trials.

• Analysis datasets created by the multiple imputation procedure will not be submitted, instead the SAS code generating the datasets will be submitted.

• Does the FDA agree with this proposal regarding submitted datasets?

Pre-NDA Communication on Dataset – Example

Review Division’s Responses:

• Your proposal to submit study datasets based on SDTM 1.3/SDTMIG 3.1.2 , is acceptable.

• Your proposal to submit analysis datasets based on ADaM 2.1/ADaMIG1.0, is acceptable.

• In addition to analysis datasets for the Phase 3 trial A3 and the Phase 2 trial A1, you should submit analysis datasets for the Phase 2 trial A2.

• Your proposal to submit the SAS code used to generate the Multiple Imputation datasets is acceptable. In addition, submit the SAS code used to analyze the primary and secondary endpoints using these imputed datasets.

Pre-NDA Communication on Dataset – Example

Review Division’s Additional Comments:

• Each analysis dataset should include the treatment assignments, baseline assessments, and key demographic variables.

• The analysis datasets should include all variables needed for conducting all primary, secondary, and sensitivity analyses included in the study report.

• For endpoints that include imputations, both observed and imputed variables should be included and clearly identified.

• Provide sufficient comments, adequate bookmarks, and hyperlinks in the data define files.

Filing Review

• Datasets

• Data structure (Legacy, SDTM, ADaM)

• Define files sufficiently detailed

• Analysis datasets are sufficiently structured

• Software programs

• Clinical study reports

Filing Review – if Issues Identified

Actions:

• Send information requests (IRs) for clarification

• Can submission related issues be resolved before filing?

– Yes, issue filing letter– No,

o Extend PDUFA clocko Refuse to file

Suggestions:

• Discuss content and format during pre-NDA meeting

• Follow FDA guidance

• Well-prepared data package

• Closely work with your regulatory staff

• Be prompt on IRs

Conduct Review

Check regulatory histories (in the IND stage)

• Protocol/SAP is in alignment with Division’s recommendations?

• Drug specific concerns on efficacy and safety?

• Any issues were identified during trial conduct?

Conduct Review• Check data quality• Check blinding/unblinding procedures• Verify the randomization treatment assignments• Reproduce the primary analysis datasets, particularly the primary endpoint and major secondary

endpoint(s); reproduce study efficacy results – Follow protocol/SAP and do programming

– Data definition files

– sdrg/adrg

– Submitted software programs (analysis result metadata)

• Conduct additional analyses on inconsistent results or clinical concerns• Send IRs if needed

Example – Keep Consistency from SAP to Analysis

Pre-specified in SAP:

• Efficacy population: all randomized subjects who received study medication and had at least one post-treatment efficacy assessment

• Efficacy endpoint: response rate (from Questionnaire X)

• Statistical analysis: logistic regression model with two baseline characteristics

Example – Keep Consistency from SAP to Analysis

Placebo(N=95)

Drug(N=92)

# of responder (%) 46 (48.4) 58 (63.0)

Difference (%) 14.6

p-value 0.043

Sponsor’s Analysis

Example – Keep Consistency from SAP to Analysis

Placebo(N=95)

Drug(N=92)

Placebo(N=95)

Drug(N=93)

# of responder (%) 46 (48.4) 58 (63.0) 46 (48.4) 58 (62.4)

Difference (%) 14.6 14.0

p-value 0.043 0.052

Sponsor’s Analysis Reviewer’s Analysis

Example – Keep Consistency from SAP to Analysis

Placebo(N=95)

Drug(N=92)

Placebo(N=95)

Drug(N=93)

# of responder (%) 46 (48.4) 58 (63.0) 46 (48.4) 58 (62.4)

Difference (%) 14.6 14.0

p-value 0.043 0.052

Sponsor’s Analysis Reviewer’s Analysis

Example – Keep Consistency from SAP to Analysis

Placebo(N=95)

Drug(N=92)

Placebo(N=95)

Drug(N=93)

# of responder (%) 46 (48.4) 58 (63.0) 46 (48.4) 58 (62.4)

Difference (%) 14.6 14.0

p-value 0.043 0.052

Sponsor’s Analysis Reviewer’s Analysis

Example – Keep Consistency from SAP to Analysis

Placebo(N=95)

Drug(N=92)

Placebo(N=95)

Drug(N=93)

# of responder (%) 46 (48.4) 58 (63.0) 46 (48.4) 58 (62.4)

Difference (%) 14.6 14.0

p-value 0.043 0.052

Sponsor’s Analysis Reviewer’s Analysis

Question: Who is this subject? Included in the analysis or not?

Example – Keep Consistency from SAP to Analysis

Reviewer’s findings• Used PROC COMPARE and identified this subject• Subject’s RESP = missing • Define file:

PARAMCD PARAM DerivationRESP Response = 0 (failure) if a failure in Q1 or Q2 in Questionnaire X;

= missing if there is no failure, but one of the questions is missing;= 1 (success) otherwise

Example – Keep Consistency from SAP to Analysis

PARAMCD PARAM DerivationRESP Response = 0 (failure) if a failure in Q1 or Q2 in Questionnaire X;

= missing if there is no failure, but one of the questions is missing;= 1 (success) otherwise

Define file:

Subject Q1 Q2 RESPxxxx Missing Success

Subject’s data point (only one post-treatment efficacy assessment):

Example – Keep Consistency from SAP to Analysis

PARAMCD PARAM DerivationRESP Response = 0 (failure) if a failure in Q1 or Q2 in Questionnaire X;

= missing if there is no failure, but one of the questions is missing;= 1 (success) otherwise

Define file:

Subject Q1 Q2 RESPxxxx Missing Success Missing

Subject’s data point (only one post-treatment efficacy assessment):

Question: This subject had only one post-treatment record for Q2, shall this subject be included in the efficacy population?

Efficacy population: all randomized subjects who received study medication and had at least one post-treatment efficacy assessment

Example – Keep Consistency from SAP to Analysis

PARAMCD PARAM DerivationRESP Response = 0 (failure) if a failure in Q1 or Q2 in Questionnaire X;

= missing if there is no failure, but one of the questions is missing;= 1 (success) otherwise

Define file:

Subject Q1 Q2 RESPxxxx Missing Success Missing

Question: This subject had only one post-treatment record for Q2, shall this subject be included in the efficacy population? Yes.

Efficacy population: all randomized subjects who received study medication and had at least one post-treatment efficacy assessment

Subject’s data point (only one post-treatment efficacy assessment):

Example – Keep Consistency from SAP to Analysis

Reviewer’s actions:

• Communicate the findings with the review team and the Sponsor

• Conduct additional sensitivity analyses to assess the robustness of the study results

Moving Forward…

• Follow protocol/SAP to define variables

• Follow define files strictly in programming

• Keep consistency from protocol/SAP, define file, programming to analysis results

My NDA/BLA Review Wishlist

• Clear datasets• Sufficient data definition files• Helpful sdrg/adrg (if provided)• Prompt responses to IRs• Close communication

FDA Resource on Data Standards and NDA/BLA Review

Reviewer’s Recommendation on Data Submissions

• Well-structured datasets and associated documentation

• Well-documented software programs

• Internal and external communication

• Follow FDA guidance• Be part of data standard development

Reference

• FDA Study Data Standard Resources: https://www.fda.gov/industry/fda-resources-data-standards/study-data-standards-resources

• FDA Electronic Regulatory Submission and Review: https://www.fda.gov/drugs/forms-submission-requirements/electronic-regulatory-submission-and-review

• FDA: Best Practices for Communication Between IND Sponsors and FDA During Drug Development: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/best-practices-communication-between-ind-sponsors-and-fda-during-drug-development

• FDA: Guidance for Review Staff and Industry Good Review Management Principles and Practices for PDUFA Products• FDA: Guidance for Industry Providing Regulatory Submissions in Electronic Format — Standardized Study Data• FDA: Regulatory Submissions in Electronic Format — Submissions Under Section 745A(a) of the Federal Food, Drug,

and Cosmetic Act• FDA: Guidance for Industry Providing Regulatory Submissions in Electronic Format – Standardized Study Data