Introduction
Pembrolizumab, an anti–PD-1 monoclonal antibody, has established a broad and evidence-based role across multiple high-impact oncology indications. Pivotal trials—including KEYNOTE-024, KEYNOTE-042, KEYNOTE-189, KEYNOTE-407, KEYNOTE-671, KEYNOTE-048, KEYNOTE-177, KEYNOTE-716, and KEYNOTE-689—demonstrate consistent overall survival (OS) and event-free survival benefits with a manageable immune-related toxicity profile. Recent updates through early 2026 reinforce pembrolizumab's position as a first-line standard of care in PD-L1–expressing NSCLC, adjuvant and perioperative melanoma and HNSCC, urothelial carcinoma (in combination with enfortumab vedotin), and pan-tumor MSI-H/dMMR disease. Biomarker-driven patient selection—particularly PD-L1 expression level, MSI-H/dMMR status, and ECOG performance status—remains essential to optimizing the benefit–risk balance 2122.
1. Clinical Efficacy Across Major Indications
Non-Small Cell Lung Cancer (NSCLC)
Pembrolizumab's efficacy in first-line NSCLC is well-established across PD-L1 expression strata. In patients with PD-L1 tumor proportion score (TPS) ≥50%, monotherapy (KEYNOTE-024) demonstrated a median OS of 26.3 versus 13.4 months (HR 0.62), with 5-year OS rates of 31.9% versus 16.3% 21. KEYNOTE-042 extended benefit to TPS ≥1% (OS HR 0.81), with the greatest gain at TPS ≥50% (HR 0.69) 21. A real-world 5-year observational study confirmed a median OS of 19.1 months (95% CI 13.3–24.9) and a 5-year OS rate of 24.8% in PD-L1–high patients, with squamous histology (HR 2.23) and ≥3 metastatic sites (HR 4.65) as independent adverse prognostic factors 12.
Chemo-immunotherapy combinations extend benefit across all PD-L1 levels. KEYNOTE-189 (nonsquamous) achieved 5-year OS of 22.0 versus 10.6 months (HR 0.60; ORR 48.3% vs 19.9%), and KEYNOTE-407 (squamous) demonstrated OS HR 0.71 with ORR 62.2% versus 38.8% 21. A meta-analysis published in JNCCN (March 2026) noted that Bayesian network meta-analysis favored pembrolizumab-chemotherapy over monotherapy (HR 0.79, 95% CI 0.68–0.93) in PD-L1 ≥50% patients, cautioning that the evidence relies substantially on retrospective cohorts 22.
In the perioperative setting, KEYNOTE-671 demonstrated substantially improved median event-free survival (EFS; 47.2 vs 18.3 months, HR 0.59), OS HR 0.72, pathologic complete response (pCR) of 18.1% versus 4.0%, and major pathologic response (mPR) of 30.2% versus 11.0% 21. Adjuvant pembrolizumab (PEARLS/KEYNOTE-091) improved disease-free survival (HR 0.76; median 53.6 vs 42.0 months) across PD-L1 strata 21.
A real-world Norwegian cohort (n=410) confirmed that combination therapy reduced early mortality risk (adjusted HR for death within 6 months: 0.41, p=0.004), with particularly pronounced benefit in females, stage IVB disease, PD-L1 ≥75%, and ECOG 0–1 patients 1.
Melanoma
In metastatic melanoma, pembrolizumab surpasses ipilimumab with 10-year OS of 34.0% versus 23.6% (OS HR 0.71; PFS HR 0.64; median OS 32.7 vs 15.9 months) 21. In the adjuvant setting, pembrolizumab achieves RFS HR 0.57 in stage III disease 21, and KEYNOTE-716 demonstrated RFS HR 0.62 (48-month RFS 71% vs 58%) and DMFS HR 0.59 in high-risk stage II disease 21. Neoadjuvant-adjuvant pembrolizumab (SWOG S1801) yielded a 2-year EFS of 72% versus 49% (HR 0.58), suggesting a practice-shifting role in resectable stage III–IV disease 21.
Head and Neck Squamous Cell Carcinoma (HNSCC)
KEYNOTE-048 established pembrolizumab ± chemotherapy as first-line standard of care in recurrent/metastatic HNSCC, with 5-year OS advantages maintained across CPS strata and longer duration of response (DoR) versus the EXTREME regimen 21. KEYNOTE-040 showed OS HR 0.79 (ITT), with greater benefit at higher PD-L1 expression 21.
Crucially, KEYNOTE-689 (published 2025) demonstrated that neoadjuvant plus adjuvant pembrolizumab added to standard-of-care (surgery ± adjuvant CRT) significantly improved 36-month EFS: HR 0.66 (CPS ≥10), HR 0.70 (CPS ≥1), and HR 0.73 (total population; all p≤0.008) 25.
Urothelial Carcinoma
KEYNOTE-905 (Phase 3) evaluated perioperative enfortumab vedotin plus pembrolizumab versus surgery alone in cisplatin-ineligible muscle-invasive bladder cancer (median follow-up 25.6 months): 2-year EFS 74.7% versus 39.4% (HR 0.40; p<0.001), 2-year OS 79.7% versus 63.1% (HR 0.50; p<0.001), and pCR 57.1% versus 8.6% 23. The NCCN Bladder Cancer Guidelines (Version 1.2026) incorporate these findings 24.
Pan-Tumor MSI-H/dMMR Solid Tumors
KEYNOTE-177 established pembrolizumab as first-line therapy for metastatic MSI-H/dMMR colorectal cancer (PFS HR 0.60; OS HR 0.73; median OS 40.8 vs 36.7 months; TRAE rate 21.6% vs 67.1%) 26. KEYNOTE-158 confirmed a 33.8% ORR across noncolorectal MSI-H/dMMR solid tumors, median OS 19.8 months, and median DoR of 63.2 months, with grade ≥3 TRAEs in 13% of patients 4.
2. Safety and Tolerability
The key safety differentiator for pembrolizumab is its immune-related adverse event (irAE) profile. In monotherapy settings, grade ≥3 TRAEs occur in approximately 10–18% of patients 21, compared with 72–75% in chemo-immunotherapy arms of KEYNOTE-189/407 21. A real-world comparative study (n=406) found pembrolizumab's grade ≥3 severe irAE rate of 8.2%—notably lower than nivolumab (29.6%), avelumab (22.2%), and atezolizumab (13.8%), but higher than durvalumab (5.8%) 18.
Common irAEs include hypothyroidism (~10–15%), hyperthyroidism (~4–10%), pneumonitis (~2–8%), and dermatologic events; rare but serious events include colitis, hepatitis, and endocrinopathies. In elderly patients (≥75 years), real-world data show all-grade irAEs in 40.6% and grade ≥3 irAEs in 11.7%, with a 19.6% discontinuation rate 2. Combination therapy significantly increases toxicity: grade ≥3 irAEs of 19.0% (combination) versus 10.1% (monotherapy), and discontinuation rates of 33.9% versus 14.5% 2.
A critical emerging safety concern is long-term or persistent irAEs following treatment cessation. These are classified as delayed/late-onset, chronic active, or chronic inactive subtypes and require vigilant post-treatment monitoring beyond the active treatment period 1317.
Practical management requires systematic grading using CTCAE criteria, early corticosteroid initiation for grade ≥2 irAEs, and multidisciplinary involvement. Pembrolizumab should be held for grade 2 and permanently discontinued for most grade 4 irAEs. Patients with pre-existing autoimmune disease or transplant history require heightened monitoring.
3. Biomarker-Driven Patient Selection
| Biomarker | Indication | Key Finding |
|---|---|---|
| PD-L1 TPS ≥50% | NSCLC monotherapy | OS HR 0.62; 5-yr OS 31.9% 21 |
| PD-L1 TPS ≥1% | NSCLC (KEYNOTE-042) | OS HR 0.81; greater benefit at higher TPS 21 |
| PD-L1 TPS <1% | NSCLC | meta-analysis HR 0.66 (wide CI) 2 |
| PD-L1 CPS ≥10 | HNSCC (KEYNOTE-689) | EFS HR 0.66 25 |
| MSI-H/dMMR | CRC and pan-tumor | KEYNOTE-158: ORR 33.8%; median DoR 63.2 months 426 |
| ECOG 0–1 vs ≥2 | All NSCLC | RWE HR 0.42 for OS; ECOG ≥2 shows poor outcomes even with high PD-L1 29 |
| PD-L1 ≥75% | NSCLC combination | Early death HR 0.31 favoring combination 1 |
ECOG performance status is a critical determinant: in a Japanese real-world cohort, patients with ECOG 2–3 receiving pembrolizumab monotherapy had median PFS of 0.7 months and OS of 1.6 months, versus 10.5 and 21.7 months for ECOG 0–1 9. The GFPC 08-2015 ENERGY trial (nivolumab + ipilimumab vs chemotherapy) confirmed that ECOG 2 patients fared worse with dual immunotherapy (OS HR 1.32), while ECOG 0–1 elderly patients benefited (HR 0.64), highlighting performance status as the key triage criterion 8.
A meta-analysis of 35 Phase 3 RCTs in elderly NSCLC patients found significant OS benefit for age 65–74 years (HR 0.75), but no statistically significant benefit for patients ≥75 years (HR 0.89, 95% CI 0.77–1.04) in RCT data, though real-world outcomes were comparable across age groups 2.
4. Comparative and Guideline Context
A network meta-analysis of 37 RCTs (31,779 patients) across 10 PD-1/PD-L1 inhibitors found that pembrolizumab monotherapy significantly improved OS versus chemotherapy (HR 0.72), ranking favorably alongside tislelizumab (HR 0.66) and nivolumab (HR 0.76) 3. For chemo-IO combinations, nivolumab + chemotherapy demonstrated superior OS (HR 0.57) compared with other combinations. Pembrolizumab showed a relatively higher risk of severe irAEs (OR 8.40) compared with other agents in the irAE-specific analysis 3.
As of April 2026, major guidelines reflect the following:
- NSCLC: Both pembrolizumab monotherapy (TPS ≥50%) and chemo-IO combinations are NCCN-recommended; JNCCN meta-analysis cautiously suggests combination may offer OS advantage even in high PD-L1 expressers 22
- Melanoma: Pembrolizumab is established in adjuvant stage II–III and metastatic settings
- HNSCC: Neoadjuvant-adjuvant pembrolizumab now supported by KEYNOTE-689 data across PD-L1 strata 25
- Urothelial carcinoma: NCCN v1.2026 incorporates perioperative enfortumab vedotin + pembrolizumab 24
- MSI-H/dMMR CRC: FDA- and EMA-approved first-line therapy; ESMO-MCBS score of 4 26
5. Practical Adoption Considerations
Pembrolizumab is administered at 200 mg Q3W or 400 mg Q6W (flat dosing), with the Q6W schedule reducing infusion visits and potentially improving patient convenience and health-system efficiency. PD-L1 testing (TPS for NSCLC; CPS for HNSCC) and MSI/dMMR testing (IHC or PCR) are essential prerequisites 2122. Assay variability across platforms and turnaround time remain practical barriers, particularly in resource-limited settings.
Health-economics data were not available in the retrieved materials; however, the broad, biomarker-defined patient populations, combined with durable responses—median DoR of 63.2 months in MSI-H/dMMR disease 4—suggest favorable long-term cost-effectiveness in selected populations, though formal reimbursement frameworks vary by country.
6. Benefit–Risk Synthesis and Decision Framework
| Setting | Net Benefit Statement | Key Trade-off |
|---|---|---|
| NSCLC, TPS ≥50%, monotherapy | Substantial OS benefit (HR 0.62); 5-yr OS 31.9%; low grade ≥3 TRAEs (~31.2%) 21 | Possible early mortality risk in high-burden disease; combination may reduce early deaths 1 |
| NSCLC, all PD-L1, chemo-IO | Superior ORR (48–62%) and OS across histologies 21 | Grade ≥3 AEs in 72–75%; 33.9% discontinuation 2 |
| NSCLC, perioperative | EFS HR 0.59; pCR 18.1% (KEYNOTE-671) 21 | Grade ≥3 AEs 45.2%; immune-mediated AEs 25.3% |
| NSCLC, ECOG ≥2 or age ≥75 | Attenuated or absent OS benefit 29 | Disproportionate toxicity; careful case-by-case assessment required |
| Melanoma, metastatic/adjuvant | Long-term OS advantage; RFS HR 0.57–0.62 21 | irAEs generally lower than ipilimumab |
| HNSCC, R/M first-line | Durable OS benefit across CPS strata 2125 | Monotherapy preferred for tolerability; combination for rapid tumor control |
| Urothelial carcinoma | EFS HR 0.40; pCR 57.1% (KEYNOTE-905) 23 | Grade ≥3 AEs 71.3% (enfortumab vedotin combination) |
| MSI-H/dMMR solid tumors | ORR 33.8%; median DoR 63.2 months; TRAE 13% grade ≥3 426 | Best benefit–risk ratio across biomarker-selected populations |
Situations where risks may outweigh benefits: pembrolizumab + SBRT in early inoperable NSCLC (KEYNOTE-867: no EFS/OS benefit; higher fatal AEs) 21; pembrolizumab + ipilimumab in PD-L1–high NSCLC (KEYNOTE-598: no efficacy gain; grade 3–5 AEs 35.1% vs 19.6%) 21; ECOG ≥2 patients with very poor performance 9.
7. Limitations and Clinical Implications
Several important limitations must be acknowledged. Cross-trial comparisons between pembrolizumab and other checkpoint inhibitors (nivolumab, atezolizumab, durvalumab, cemiplimab) remain methodologically constrained by differences in trial populations and designs 3. The survival benefit in patients ≥75 years is not statistically established in RCT data, though real-world evidence suggests comparable outcomes 2. The optimal threshold for choosing monotherapy over chemo-IO in TPS ≥50% patients remains debated—ongoing direct-comparison RCTs are needed 22. Long-term irAE management after treatment cessation is an emerging area lacking standardized protocols 1317. Data for urothelial carcinoma (non-perioperative settings) and TMB as a pan-tumor biomarker were limited in the retrieved materials. Finally, the predictive role of novel biomarkers (circulating tumor DNA, T-cell-inflamed gene expression profile) beyond PD-L1 requires prospective validation 21.
In summary, pembrolizumab offers durable, clinically meaningful survival benefits across NSCLC, melanoma, HNSCC, urothelial carcinoma, and MSI-H/dMMR solid tumors. Optimal clinical adoption requires rigorous biomarker-driven patient selection, performance status assessment, indication-specific benefit–risk weighing, and proactive irAE surveillance—including long-term monitoring after treatment discontinuation.