ADaM Reviewer’s Guide

(Pre-filled Example – Oncology Phase II)

Study ID: ONC-201
Indication: Metastatic Non-Small Cell Lung Cancer (NSCLC)
Submission Type: NDA
ADaM Standard: ADaM v2.1
Date: 15-Mar-2026

---

1. Purpose of This ADaM Reviewer’s Guide

This ADaM Reviewer’s Guide provides an overview of the analysis datasets, their structure, and key derivations, and describes traceability from SDTM to ADaM.
It is intended to help reviewers understand how the ADaM datasets were constructed and how they support statistical analyses, without restating the SAP.

---

2. Overview of ADaM Datasets

ADaM Dataset	Description	Granularity
ADSL	Subject-level analysis dataset	One record per subject
ADAE	Adverse event analysis dataset	One record per AE per subject
ADLB	Laboratory analysis dataset	One record per test per visit per subject
ADTTE	Time-to-event analysis dataset	One record per subject per parameter

---

3. Subject-Level Analysis Dataset (ADSL)

3.1 Dataset Description

ADSL contains one record per subject and includes treatment assignments, population flags, and baseline demographic variables.
All other ADaM datasets are linked to ADSL via USUBJID.

3.2 Key Variables

Variable	Description
USUBJID	Unique subject identifier
TRT01A	Actual treatment received
ITTFL	Intent-to-treat population flag
SAFFL	Safety population flag
PPFL	Per-protocol population flag
AGE, SEX	Demographics
RFSTDTC	Reference start date

---

4. Adverse Events Analysis Dataset (ADAE)

4.1 Dataset Description

ADAE is derived from the SDTM AE domain and includes one record per adverse event per subject.
It supports all safety analyses.

4.2 Key Variables

Variable	Description
AETERM	Adverse event term
AESEV	Severity
AESER	Serious AE flag
ASTDT	AE start date (numeric)
AENDT	AE end date (numeric)
TRTEMFL	Treatment-emergent AE flag

4.3 Key Derivations

• ASTDT and AENDT are derived from AESTDTC and AEENDTC
• TRTEMFL = 'Y' if AE start date is on or after first dose and on or before last dose + follow-up window

---

5. Laboratory Analysis Dataset (ADLB)

5.1 Dataset Description

ADLB is derived from the SDTM LB domain and includes one record per laboratory test per visit per subject.
Both numeric and character results are retained.

5.2 Key Variables

Variable	Description
PARAMCD	Laboratory test code
PARAM	Laboratory test name
AVAL	Numeric analysis value
AVALC	Character analysis value
BASE	Baseline value
CHG	Change from baseline
AVISIT	Analysis visit

5.3 Baseline Definition

Baseline is defined as the last non-missing laboratory value prior to first dose.

---

6. Time-to-Event Analysis Dataset (ADTTE)

6.1 Dataset Description

ADTTE supports time-to-event analyses such as Progression-Free Survival (PFS).

6.2 Key Variables

Variable	Description
PARAMCD	Parameter code (e.g., PFS)
AVAL	Analysis time (days)
CNSR	Censoring indicator
EVNTDESC	Event description

---

7. Analysis Populations

Population flags are defined in ADSL and propagated to all other ADaM datasets.

Population	Flag	Definition
ITT	ITTFL='Y'	All randomized subjects
Safety	SAFFL='Y'	Subjects receiving ≥1 dose
PP	PPFL='Y'	ITT subjects without major protocol deviations

---

8. Traceability from SDTM to ADaM

8.1 SDTM → ADaM Mapping Examples

ADaM Dataset	ADaM Variable	SDTM Domain	SDTM Variable(s)	Derivation
ADSL	TRT01A	DM	ARM	Assigned
ADAE	AETERM	AE	AEDECOD	Copied
ADAE	ASTDT	AE	AESTDTC	Converted
ADLB	AVAL	LB	LBSTRESN	Copied
ADTTE	AVAL	AE, RS	Event dates	Derived

---

9. Value-Level Metadata Highlights

Value-level metadata are used where variable attributes differ by parameter.

Dataset	Variable	Condition	Description
ADLB	AVAL	PARAMCD=ALB	Numeric, 1 decimal
ADLB	AVALC	PARAMCD=HBsAg	Character POS/NEG

---

10. Known Data Considerations

Issue ID	Description	Impact
ADAM-01	Missing baseline labs for some subjects	Subjects excluded from baseline analyses
ADAM-02	Partial AE end dates	Conservative imputation applied

---

11. Relationship to Other Documents

• Define.xml (ADaM): Technical metadata for variables
• Statistical Reviewer’s Guide: Statistical methods and implementation
• SAP: Statistical methodology
• CSR: Analysis results and interpretation

---

One-sentence takeaway

The ADaM Reviewer’s Guide explains how the analysis datasets are structured, derived, and linked to SDTM, enabling reviewers to understand and trace the analysis data without relying on statistical code.