PD-1 Inhibitors Plus Lenvatinib Versus Lenvatinib Alone for Unresectable Hepatocellular Carcinoma: A Systematic Review and Meta-Analysis

**English Title**: PD-1 Inhibitors Plus Lenvatinib Versus Lenvatinib Alone for Unresectable HCC: A Systematic Review and Meta-Analysis

**Generated**: 2026-06-22
**Researcher**: John Doe
**Affiliation**: XX Medical School
**Research field**: Hepatocellular carcinoma immunotherapy

1. Background and rationale

Unresectable hepatocellular carcinoma (HCC) has a poor prognosis; lenvatinib monotherapy is one standard first-line option. In recent years, multiple RCTs have explored PD-1 inhibitors combined with lenvatinib, but their conclusions are inconsistent; evidence synthesis is needed to guide clinical decision-making.

2. PICO/PECO decomposition

P — Population

Adults (>=18 years) with histologically or radiologically confirmed unresectable HCC, BCLC stage B-C, Child-Pugh A, no prior systemic therapy; exclude active HBV/HCV, HIV, other malignancy history.

I — Intervention

PD-1 inhibitor (pembrolizumab 200 mg q3w / nivolumab 240 mg q2w / tislelizumab 200 mg q3w) combined with lenvatinib (>=60 kg: 12 mg qd; <60 kg: 8 mg qd), treated until PD/toxicity/24 months.

C — Comparator

Lenvatinib monotherapy (same dose), or sorafenib 400 mg bid.

O — Outcome

Type	Outcome name	Measurement tool/definition	Timepoint	Effect size
Primary	OS	Date of death	>=12 months follow-up	HR
Primary	PFS	RECIST 1.1	>=12 months follow-up	HR

**Research-question statement**:

In adults with unresectable HCC (BCLC B-C, Child-Pugh A, treatment-naive), does PD-1 inhibitor combined with lenvatinib compared with lenvatinib monotherapy improve OS and PFS?

3. Meta-analysis type and rationale

**Type**: Traditional pairwise meta-analysis

**Rationale**: Intervention and comparator are both fixed two-arm comparisons with a simple evidence network; no NMA needed; all data are aggregate; no IPD needed.

**Methodological details**:

Effect size: Time-to-event outcomes use HR + 95% CI; binary outcomes use RR + 95% CI
Synthesis model: Random-effects model (DerSimonian-Laird), due to expected high heterogeneity
Heterogeneity assessment: I2, tau2, Q test; I2 > 50% triggers subgroup analysis
Publication bias: When >=10 studies are included, Egger test + funnel plot + Trim-and-fill

4. Four-dimension topic assessment

Dimension	Score (0-5)	Reason
Clinical value	5	Directly addresses first-line treatment controversy; may change NCCN/CSCO guideline recommendations. Matches anchor 5: directly addresses a guideline-level controversy; expected to change clinical practice
Methodological feasibility	4	Standard methods directly applicable; note potential heterogeneity across PD-1 inhibitors. Matches anchor 4: light methodological adaptation needed, but mature toolchain
Data availability	4	Pre-search estimates 6-8 RCTs; outcomes mostly consistently reported. Matches anchor 4: 6-10 studies; outcomes mostly consistent
Novelty	3	A same-topic meta-analysis was published in 2024; must clearly state the new-study increment. Matches anchor 3: prior meta-analysis <3 years old, but with a clear evidence increment (pending Stage 4 dedup confirmation)

**Total**: 16 / 20

**Recommendation**: Recommend proceeding — register PROSPERO after shoring up risk dimensions

**Cross-check result**:

✅ None of the 6 cross-check rules triggered (R1-R6)

5. Dedup search report

PROSPERO search

Query: `PD-1 AND lenvatinib AND HCC`
Search date: 2026-06-22
Hits: 2
Key hits:

1. CRD42024567890 — Immunotherapy + Lenvatinib in HCC — status: Ongoing — registered: 2024-09-01

Cochrane Library search

Query: `PD-1 AND lenvatinib`
Search date: 2026-06-22
Hits: 0

PubMed published meta-analysis search

Query: `(PD-1[Title/Abstract]) AND (lenvatinib[Title/Abstract]) AND (HCC[Title/Abstract]) AND (meta-analysis[Publication Type])`
Search date: 2026-06-22
Hits: 3
Key hits (by relevance):

1. 38512345 — PD-1 + Lenvatinib in HCC: A Meta-analysis — 2024 — J Hepatol

Non-English database search

Databases: CNKI, Wanfang, SinoMed
Query: `(PD-1 OR pembrolizumab OR sintilimab) AND (lenvatinib) AND (HCC OR hepatocellular carcinoma) AND ('Meta' OR 'systematic review')`
Search date: 2026-06-22
Hits: 1
Key hits:

1. PD-1 + lenvatinib for advanced HCC: a Meta-analysis — 2024 — Chinese Journal of Hepatobiliary Surgery

Near-duplicate judgment

Near type	Changed element	Counts as duplicate?	Proceed condition
Switch within-class intervention	I: PD-1 -> PD-L1	No	Justify within-class substitution clinically + subgroup by PD-1/PD-L1 type

Innovation judgment

Relationship with existing work: Update type
Evidence-increment type: New-study increment + new-subgroup increment
Increment sufficiency: Sufficient (>=2 new phase III RCTs added after 2024)
Recommendation: Proceed

6. Pre-search strategy and estimated inclusions

Pre-search databases: PubMed, Embase, Cochrane CENTRAL, CNKI
Pre-search query (PubMed): `(PD-1 OR pembrolizumab OR nivolumab OR tislelizumab) AND lenvatinib AND (HCC OR hepatocellular carcinoma)`
Time range: Inception to 2026-06
Estimated total hits: 280
Estimated included studies: 7
Language scope: English + Chinese

7. PRISMA 2020 key-item compliance preview

Item	Content	Preview
Title	Identifiable as systematic review / meta-analysis	PASS
Objectives (PICO)	PICO clear and operational	PASS
Eligibility	Consistent with PICO and explicit	PASS
Information sources	≥3 databases + gray literature	PASS
Search strategy	Complete reproducible search string	PASS
Selection process	Two independent reviewers + arbitration	WARN
Data items	Standardized data-extraction form	PASS
Risk of bias	RoB tool matched to study type	PASS
Effect measures	Prespecified effect-size type and direction	PASS
Synthesis methods	Prespecified synthesis, subgroups, sensitivity	PASS
Heterogeneity	Prespecified I² threshold and analysis path	PASS
Study selection (flow diagram)	Can draw a PRISMA 2020 flow diagram	PASS
Equity (PROGRESS-Plus)	PROGRESS-Plus dimensions considered	WARN

**Overall compliance risk**: MEDIUM (two-reviewer screening and equity dimension need shoring up)

8. Primary outcomes and effect measures

Primary outcomes

Name	Measurement tool	Timepoint	Effect size	Model
OS	Date of death	Minimum 12 months	HR	Random effects
PFS	RECIST 1.1	Minimum 12 months	HR	Random effects

Secondary outcomes

Name	Type	Effect size
ORR	Binary	RR
DCR	Binary	RR
SAE	Binary	RR
Grade 3-4 TRAE	Binary	RR

9. Prespecified subgroup and sensitivity analyses

Prespecified subgroup analyses

**PD-1 inhibitor type** — Different PD-1 inhibitors may differ in efficacy
**BCLC stage** — Stage B vs stage C populations may benefit differently
**Region (Asia vs non-Asia)** — HBV-dominant vs HCV-dominant populations may differ

Sensitivity analyses

Re-pool after excluding high-risk-of-bias studies
Re-pool including only phase III RCTs
Re-pool using a fixed-effect model

10. Potential risks and mitigations

Risk	Level	Mitigation
High heterogeneity across PD-1 inhibitors	Medium	Subgroup by PD-1 type; pool separately if needed
Fewer than 5 included studies	Low	Extend gray-literature search; wait for new RCTs if needed
Two-reviewer screening not guaranteed	Medium	Bring in a methodologist or a second reviewer

11. Recommended next steps

Finalize the search strategy and have a methodologist review it
Complete PROSPERO registration (within 2 weeks)
Draft the full protocol and have the team review it
Execute the three-database search and two-reviewer screening

*This report was auto-generated by the meta-analysis-topic-selector skill. All assessment conclusions are for topic-stage reference only; final methodological decisions should follow the full protocol.*