Disease Markers

Research Article

The Value of PD-L1 Expression in Predicting the Efficacy of Anti-PD-1 or Anti-PD-L1 Therapy in Patients with Cancer: A Systematic Review and Meta-Analysis

Table 1

Characteristics of the included trials.
StudiesPhasePopulationTreatmentPD-L1 antibody cloneIHC scoring methodCutoff for PD-L1 positivityPatients numberNumber of the patients by PD-L1 expression statusFollow-up (months)Studies
Experimental armControl arm
PD-L1 positivePD-L1 negativePD-L1 positivePD-L1 negative
Motzer (2015) [29]3RCCNivolumab vs. everolimus28-8Type 1≥1%8212769429987NR
Robert (2015) [30]3MelanomaNivolumab vs. dacarbazine28-8Type 1≥5%418741367413416.7
Borghaei (2015) [12]3Nonsquamous NSCLCNivolumab vs. docetaxel28-8Type 1≥1%58210812310112313.2
Brahmer (2015) [28]3Squamous-cell NSCLCNivolumab vs. docetaxel28-8Type 1≥1%2725463525611
Fehrenbacher (2016) [13]2NSCLCAtezolizumab vs. docetaxelSP142Type 2≥1%28793511024113.0
Reck (2016) [33]3NSCLCPembrolizumab vs. chemotherapy22C3Type 1≥50%30515415111.2
Ferris (2016) [16]3Head and neck cancerNivolumab vs. chemotherapy28-8Type 1≥1%361887361385.1
Hodi (2016) [32]2MelanomaNivobumab+ipilimumab vs. ipilimumab28-8Type 1≥5%1425624271124.5
Herbst (2016) [31]2/3NSCLCNivolumab vs. docetaxel22C3Type 1≥1%103369034313.1
Schachter (2017) [37]3MelanomaPembrolizumab vs. ipilimumab22C3Type 1≥1%8344461032254722.9
Rittmeyer (2017) [36]3NSCLCAtezolizumab vs. docetaxelSP142Type 2≥1%85024118022219921
Wolchok (2017) [17]3MelanomaNivolumab+ipilimumab; nivolumab vs. ipilimumab28-8Type 1≥1%94568; 80210; 2087520235.7 vs. 38 vs. 18.6
Larkin (2017) [18]3MelanomaNivolumab vs. chemotherapy28-8Type 1≥5%405134138676624
Bellmunt (2017) [19]3Urothelial carcinomaPembrolizumab vs. chemotherapy22C3Type 2≥10%542741869017614.1
Carbone (2017) [37]3NSCLCNivolumab vs. IC chemotherapy C28-8Type 1≥5%541208632106013.5
Kang (2017) [35]3Gastric cancerNivolumab vs. placebo28-8Type 1≥1%4931141652108.8
Gandhi (2018) [38]3NSCLCChemotherapy+pembrolizumab vs. chemotherapy+placebo22C3Type 1≥1%6162601271286310.5
Shitara (2018) [40]3Gastric or gastroesophageal junction cancerPembrolizumab vs. paclitaxel22C3Type 2≥1%59219699199968.5
Motzer (2018) [39]3RCCNivolumab+ipilimumab vs. sunitinib28-8Type 1≥1%84710028411427825.2
Antonia (2018) [14]3NSCLCDurvalumab after chemodadiotherapy vs. placebo after chemodadiotherapySP263Type 1≥1%71321290915825.2
Barlesi (2018) [41]3NSCLCAvelumab vs. docetacel73-10Type 1≥1%69226412926513218.3
Paz-Ares (2018) [42]3Squamous NSCLCPembrolizumab+chemotherapy vs. placebo+chemotherapy22C3Type 1≥1%55917695177997.8
Schmid (2018) [43]3TNBCAtezolizumab+nabpaclitaxel vs. placebo+nabpaclitaxelSP142Type 2≥1%90218526618426712.9
Socinski (2018) [44]3NSCLCABCP vs. BCPSP142Type 2≥1%69220916619517215.4
NSCLC: nonsmall cell lung cancer; RCC: renal cell carcinoma; TNBC: triple-negative breast cancer; NR: not reached; IHC: immunohistochemistry; Type 1 = membranous staining on tumor cells; Type 2 = membranous or cytoplasmic staining, or both, of tumor cells and tumor-infiltrating immune cells; all the included patients in KEYNOTE-024 were patients with PD-L1 ≥50%; the experimental arm of KEYNOTE-010 comprised two cohorts: pembrolizumab 2 mg/kg () and pembrolizumab 10 mg/kg (), and all the included patients in KEYNOTE-010 were patients with PD-L1 ≥1%; the experimental arm of KEYNOTE-006 comprised two cohorts: intravenous pembrolizumab every 2 weeks () and intravenous pembrolizumab every 3 weeks (); here, we combined these two cohorts as the experimental arm to compare with ipilimumab group (control arm; ); the experimental arm of CheckMate 067 comprised two cohorts: nivolumab-plus-ipilimumab group () and nivolumab group (); these two cohorts were compared with ipilimumab group (), respectively; the IMpower150 had three cohorts: BCP cohort (bevacizumab plus carboplatin plus paclitaxel), ACP cohort (atezolizumab plus carboplatin plus paclitaxel), and ABCP cohort (atezolizumab plus BCP); here, the ABCP cohort was compared with the BCP cohort before the ACP cohort was compared with the BCP cohort.