Precision Medicine and Imaging

The First-week Proliferative Response ofPeripheral Blood PD-1þCD8þ T Cells Predicts theResponse to Anti-PD-1 Therapy in Solid TumorsKyung Hwan Kim1, Jinhyun Cho2, Bo Mi Ku3, Jiae Koh3,4, Jong-Mu Sun2, Se-Hoon Lee2,Jin Seok Ahn2, Jaekyung Cheon5, Young Joo Min5, Su-Hyung Park1, Keunchil Park2,Myung-Ju Ahn2,4, and Eui-Cheol Shin1

Abstract

Purpose: To investigate blood-based dynamic biomarkersthat predict responses to anti–programmed cell death protein1 (PD-1) therapy in solid tumors.

Experimental Design: Preplanned biomarker analysis wasperformed as part of a phase II clinical trial (NCT02607631) inpatients withmetastatic or refractory thymic epithelial tumors(TETs; n ¼ 31) who received pembrolizumab. The biomarkerwas further tested in an independent cohort of prospectivelyrecruited patients with metastatic non–small cell lung cancer(NSCLC) who received pembrolizumab or nivolumab(NSCLC cohort 1; n ¼ 33) and validated in an independentcohort of patients with NSCLC (NSCLC cohort 2; n ¼ 46).Peripheral blood samples were obtained immediately beforetreatment (D0) and 7 days after the first dose (D7) andanalyzed using multi-color flow cytometry.

Results: A higher fold-change in the percentage of Ki-67þ

cells among PD-1þCD8þ T cells 7 days after the first dose

(Ki-67D7/D0) significantly predicted durable clinical benefit(DCB; P < 0.001) and prolonged progression-free survival(PFS; P ¼ 0.027) in patients with TETs. Ki-67D7/D0 � 2.8 wasalso associated with better DCB, PFS, and overall survival(OS) in NSCLC cohort 1 (all P < 0.05). Ki-67D7/D0 wassubsequently validated in NSCLC cohort 2, and Ki-67D7/D0

� 2.8 significantly predicted better DCB (P ¼ 0.001), PFS(P ¼ 0.002), and OS (P ¼ 0.037). Ki-67D7/D0 had a lowcorrelation with tumor PD-L1 expression and combiningboth factors did not improve the predictive power ofKi-67D7/D0.

Conclusions: The proliferative response of peripheralblood PD-1þCD8þ T cells, measured as the fold-change inthe percentage of Ki-67þ cells 7 days after treatment (Ki-67D7/D0), may be a useful surrogate biomarker for predict-ing the response and prognosis to anti-PD-1 therapy in solidtumors.

IntroductionAntibody-based therapies that block programmed cell death

protein 1 (PD-1), such as pembrolizumab and nivolumab, exerttheir antitumor activity by reinvigorating exhausted CD8þ cyto-

toxic T cells (1, 2). A number of clinical trials have demonstratedthe therapeutic efficacy of anti-PD-1 therapies against varioushuman cancers, including malignant melanoma and non–smallcell lung cancer (NSCLC; refs. 3–11). Because the clinical benefitof anti-PD-1 therapy is limited to a small proportion of patients, agreat deal of effort has been made to identify biomarkers thatpredict the treatment response to this therapy (12, 13).

In clinical settings, programmed death-ligand 1 (PD-L1)expression in tumor tissues is widely used to select patient sub-groups that have a high chance of responding to treatment (7, 8,14). However, some trials have demonstrated that PD-L1 expres-sion has an unsatisfactory power of prediction (5, 15). Forexample, in a phase III trial of patientswith squamous cellNSCLC,the treatment response and survival rateswere independent of PD-L1 expression (5). Other biomarkers, including tumor mutationburden (16), mismatch repair deficiency (17), immune microen-vironment (18, 19), RNA signatures in tumor tissues (20), com-binations ofmultiple clinicopathologic factors (21), and baselinemonocyte frequency (22), have also been proposed as biomarkersfor predicting the responsiveness to anti-PD-1 therapy, but nonehave shown sufficient predictive value for routine clinical use.

Dynamic biomarkers that can be evaluated during early treat-ment were recently suggested to have superior predictive powercompared to static biomarkers that aremeasured at baseline (23).Dynamic biomarkers have been studied in serially acquired tumortissue biopsies or peripheral blood samples from patients treatedwith anti-PD-1 (24–27). Although tumor tissue better reflects

1Graduate School of Medical Science and Engineering, Korea Advanced Instituteof Science and Technology, Daejeon, Republic of Korea. 2Department of Med-icine, Division of Hematology-Oncology, Samsung Medical Center, Sungkyunk-wan University School of Medicine, Seoul, Republic of Korea. 3Samsung Bio-medical Research Institute, Samsung Medical Center, Sungkyunkwan UniversitySchool of Medicine, Seoul, Republic of Korea. 4Department of Health Sciencesand Technology, SAIHST, Sungkyunkwan University, Seoul, Republic of Korea.5Department of Internal Medicine, Division of Hematology and Oncology, UlsanUniversity Hospital, University of Ulsan College of Medicine, Ulsan, Republic ofKorea.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corrected online November 6, 2020.

Corresponding Authors: Eui-Cheol Shin, Korea Advanced Institute of Scienceand Technology, 291 Daehak-ro, Yuseong-gu, KAIST, Biomedical Research Bldg.,Daejeon 34141, Republic of Korea. Phone: 824-2350-4236; Fax: 824-2350-4240;E-mail: ecshin@kaist.ac.kr; and Myung-Ju Ahn, Samsung Medical Center, Sung-kyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul,06351, Republic of Korea. Phone: 82-2-3410-3438; Fax: 82-2-3410-1754; E-mail:silkahn@skku.edu

doi: 10.1158/1078-0432.CCR-18-1449

�2019 American Association for Cancer Research.

ClinicalCancerResearch

Clin Cancer Res; 25(7) April 1, 20202144

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

tumor biology, biopsies are invasive and repetitive biopsies forsequential examination are challenging. As blood samples areeasier and less invasive to obtain, dynamic biomarkers in periph-eral blood that significantly predict the response to anti-PD-1therapy would be helpful.

Dynamic changes in CD8þ T cells after anti-PD-1 therapy areassociated with treatment outcomes, which is expected on thebasis of themode of action of anti-PD-1 blocking antibodies (18).Early proliferation of CD8þ T cells has been assessed in peripheralblood as a biomarker predicting treatment response, but itshowed limited predictive ability (24, 25). One study found thatthe percentage of Ki-67þ cells among PD-1þCD8þ T cells 3 weeksafter the first dose of pembrolizumab correlated with the tumorresponse in patients with melanoma, but only in combinationwith the initial tumor burden (24). Another study reported thatCD8þ T cells in the peripheral blood of patients with NSCLCproliferated early after anti-PD-1 therapy, but that study did notdemonstrate predictive ability (25).

In this study,wemeasured thepercentage ofKi-67þ cells amongperipheral blood PD-1þCD8þ T cells in samples collected beforetreatment and 7 days after the first dose, when the proliferativeresponse of PD-1þCD8þ T cells peaked. We tested whether thefold-change in the percentage of Ki-67þ cells after the first week oftreatment could predict treatment outcomes in 3 independentcohorts of patients with thymic epithelial tumors (TETs) andNSCLC. In addition, we examined factors that correlate with theproliferative response of PD-1þCD8þ T cells.

Materials and MethodsPatients and sample collection

Patients with histologically confirmed thymoma or thymiccarcinoma who were not amenable to curative treatment andhad progressed after at least 1 line of platinum-based chemother-apy were enrolled in a phase II trial (NCT02607631; n¼ 33). Theaim of the trial was to evaluate the safety and efficacy of pem-

brolizumab (200 mg every 3 weeks). Patients were enrolled fromMarch 2016 through June 2016 at Samsung Medical Center(Seoul, Republic of Korea; ref. 28). Peripheral bloodwas collectedfrom 31 patients before treatment (day 0) and 7 days after theyreceived the first dose of pembrolizumab. Peripheral bloodmononuclear cells (PBMCs) were isolated from whole bloodusing standard Ficoll-Paque (GE Healthcare) density gradientcentrifugation. We also prospectively recruited 2 independentcohorts of patients with metastatic NSCLC who received pem-brolizumab (200mg every 3weeks) or nivolumab (2mg/kg every2 weeks). NSCLC cohort 1 included patients enrolled betweenApril 2016 and August 2017 at SamsungMedical Center (n¼ 33).NSCLC cohort 2 included patients enrolled when the KoreanNational Health Insurance Program started to cover the use ofpembrolizumab and nivolumab in NSCLC patients, from Sep-tember 2017 to April 2018 at Ulsan University Hospital (Ulsan,Republic of Korea; n ¼ 11) and at Samsung Medical Center (n ¼35). Peripheral blood was also collected from these patients ondays 0 and 7 and the PBMCs were isolated. The biomarkerdiscovered in the TET cohort was confirmed in NSCLC cohort1 and subsequently validated in NSCLC cohort 2. On days 0, 7,and 21, we collected the peripheral blood from 9 patients withNSCLC to monitor the kinetics of immunologic responses. Thisstudywas approved by the Institutional ReviewBoard of SamsungMedical Center andUlsanUniversityHospital andwas conductedin accordance with the Declaration of Helsinki. All patientsprovided informed consent before inclusion in the study.

EvaluationPhysical examination and laboratory tests were performed

every 2 or 3 weeks. The tumor response was assessed by CT orMRI every 8 or 9 weeks according to the Response EvaluationCriteria In Solid Tumors (RECIST), version 1.1. Patients with TETsand NSCLC were followed with the same schedule. Durableclinical benefit (DCB) was defined as a complete response (CR),partial response (PR), or stable disease (SD) lasting longer than 6months. Progression-free survival (PFS) was defined as the timefrom the start of anti-PD-1 therapy to either disease progression(according to RECIST v1.1) or death from any cause. Overallsurvival (OS) was defined as the time from the start of anti-PD-1therapy to death from any cause.

Multi-color flow cytometryMulti-color flow cytometry of PBMCs was performed at Korea

Advanced Institute of Science andTechnology (Daejeon, Republicof Korea),with the investigators blinded to the clinical outcomeofthe patients. The following fluorochrome-conjugated mAbs wereused in multi-color flow cytometry: anti-CD8 (SK1 and RPA-T8),anti-CD3 (HIT3a), anti-CD4 (SK3), and anti-CD28 (CD28�2)from BD Biosciences; anti-PD-1 (EH�12�2H7) and anti-Ki-67(Ki-67) from BioLegend; anti-CD14 (61D3) and anti-CD19(HIB19) from eBioscience; and anti-human IgG4 Fc (HP6025)from Southern Biotech. The therapeutic binding of pembrolizu-mab or nivolumab (human IgG4) to PD-1þ cells interfered withPD-1 staining of posttreatment specimens. Therefore, we per-formed antihuman IgG4 Fc staining in addition to anti-PD-1staining to define PD-1þ cells. Dead cells were identified using theLIVE/DEAD Fixable Red Dead Cell Stain Kit (Invitrogen). Intra-cellular staining of Ki-67 was performed using a FoxP3 transcrip-tion factor staining buffer set (eBioscience) and specific antibo-dies. Tumor antigen–specific CD8þ T cells were detected by a PE-

Translational Relevance

Current biomarkers for predicting the response to anti–programmed cell death protein 1 (PD-1) therapy require tissuebiopsy and have not shown sufficient predictive value. Bystudying blood-based dynamic biomarkers, we suggest thatthe first-week proliferative response of peripheral blood PD-1þCD8þ T cells (Ki-67D7/D0) could predict the response toanti-PD-1 therapy in solid tumors. We discovered that, in anti-PD-1–treated patients with metastatic or refractory thymicepithelial tumors (TETs), Ki-67D7/D0 � 2.8 predicts superiordurable clinical benefit (DCB) and survival rates. This resultwas validated in anti-PD-1–treated patients with non–smallcell lung cancer (NSCLC). In particular, 89% of patients withKi-67D7/D0 < 2.8 progressed within 6 months, indicating thatthis biomarker could effectively screen for nonresponders toanti-PD-1 therapy. The high predictive value of Ki-67D7/D0 inidentifying nonresponders may help clinicians to providealternative treatments to patients who are unlikely to benefitfrom anti-PD-1 therapy. Further prospective studies are need-ed to validate Ki-67D7/D0 as a useful biomarker in patientswithsolid tumors treated with anti-PD-1 therapy.

Predictive Peripheral Blood Biomarker for Anti-PD-1 Therapy

www.aacrjournals.org Clin Cancer Res; 25(7) April 1, 2020 2145

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

conjugated MHC-I dextramer (HLA-A�0201: NY-ESO-1157;Immudex). HCMV-specific CD8þ T cells were detected by PE-conjugated MHC-I pentamer (HLA-A�0201: pp65495; Proim-mune). All stained samples were acquired with a LSR II flowcytometer (BD Biosciences), and the data were analyzed byFlowJo software version 10.4.0 (Treestar). The gating strategiesare summarized in Supplementary Fig. S1.

PD-L1 IHC assayPD-L1 expressionwas assessed at baseline using the PD-L1 IHC

22C3 pharmDx Assay Kit (Dako North America) in formalin-fixed, paraffin-embedded tumor tissue obtained before initiationof anti-PD-1 therapy (14). The percentage of tumor cells withmembranous PD-L1 expression [tumor proportion score (TPS)]were evaluated. PD-L1 expression was classified as "high" if atleast 50%of the tumor cells stainedpositive and "low" if 0%–49%of the tumor cells stained positive.

Statistical analysisCategorical variables and continuous variables were compared

as indicated in the figure legends. The optimal cutoff point wasdetermined as the point at which the Youden index was maxi-mized by ROC curve. To assess the predictive power of compositebiomarkers, ROC curves were generated using the predictive scoreof each individual calculated from a logistic regression model.AUC values for paired ROC curves were compared with DeLongtest using pROC package (version 1.13.0). Survival curves weregenerated using the Kaplan–Meier method and comparisonsmade using the log-rank test. To assess whether biomarkers hadan independent significant effect on survival, we performed amultivariate Cox regression analysis that was adjusted for age, sex,

histology, prior chemotherapy history, and tumor burden. Sur-vival analysis was performed using the survival package (version2.42-6). Regarding sample size, the TET cohort had 94% power todetect an effect size of 0.63, NSCLC cohort 1 had 86% power todetect an effect size of 0.53, and NSCLC cohort 2 had 89% powerto detect an effect size of 0.47 all at the 5% significance level. Theeffect sizes were estimated from the rate of DCB in each cohort.Power calculations and effect size estimations were performedusing the pwr package (version 1.2-2). Two-sided P values <0.05were considered significant. All statistical analyses were per-formed with Prism software version 6.0 (GraphPad) and R sta-tistical software (version 3.2.2, The R Foundation for StatisticalComputing).

ResultsPatient characteristics

The baseline characteristics of the patients are summarizedin Table 1. Patients with TETs (n ¼ 31) received a median of 2lines of previous systemic chemotherapy before pembrolizumabtreatment and had no previous treatment with immune checkpointinhibitors. Patients receivedamedianof9 cycles (range,1–35cycles)of pembrolizumab treatment. Abest response of PRwas observed in6 patients (19.4%), SD in 18 patients (58.0%), and PD in 7 patients(22.6%). Eleven patients (35.4%) had DCB. Themedian follow-uptime was 27.5 months (range, 25.3–28.9 months).

Patients in NSCLC cohort 1 (n ¼ 33) received a median of 2lines of previous systemic chemotherapy before pembrolizumabor nivolumab treatment and had no previous treatment withimmune checkpoint inhibitors. Of the 33 patients, 16 receivedpembrolizumab and 17 received nivolumab. Amedian of 4 cycles

Table 1. Baseline characteristics of the TET cohort (n ¼ 31), NSCLC cohort 1 (n ¼ 33), and NSCLC cohort 2 (n ¼ 46)

Discovery ValidationTET cohort NSCLC cohort 1 NSCLC cohort 2

Characteristics (n ¼ 31) (n ¼ 33) (n ¼ 46)

Age, y, n (%)Median (range) 58 (26–78) 65 (35–88) 63 (39–82)<60 18 (58.1) 10 (30.3) 13 (28.3)�60 13 (41.9) 23 (69.7) 33 (71.7)

Sex, n (%)Male 20 (64.5) 25 (75.8) 33 (71.7)Female 11 (35.5) 8 (24.2) 13 (28.3)

ECOG performance status, n (%)0–1 31 (100.0) 20 (60.6) 46 (100.0)2 0 (0.0) 13 (39.4) 0 (0.0)

Histological analysis of TETs, n (%)Thymoma 6 (19.4) n.a. n.a.Thymic carcinoma 25 (80.6) n.a. n.a.

Histological analysis of NSCLC, n (%)Adenocarcinoma n.a. 16 (48.5) 22 (47.8)Squamous cell carcinoma n.a. 13 (39.4) 21 (45.7)Others n.a. 4 (12.1) 3 (6.5)

Number of lines of previous systemic chemotherapy, n (%)Median (range) 2 (1–5) 2 (1–9) 1 (0–6)0 0 (0.0) 0 (0.0) 6 (13.0)1 12 (38.7) 11 (33.3) 23 (50.0)2 9 (29.0) 6 (18.2) 8 (17.4)�3 10 (32.3) 16 (48.5) 9 (19.6)

Sum of measurable target lesions at baseline, cmMedian (range) 12.4 (1.7–27.0) 7.0 (1.5–15.7) 4.6 (1.3–12.2)

Study drug, n (%)Pembrolizumab 31 (100.0) 16 (48.5) 38 (82.6)Nivolumab 0 (0.0) 17 (51.5) 8 (17.4)

Abbreviations: ECOG, Eastern Cooperative Oncology Group; n.a., not applicable.

Kim et al.

Clin Cancer Res; 25(7) April 1, 2020 Clinical Cancer Research2146

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

(range, 1–21 cycles) and amedian of 4 cycles (range, 2–53 cycles)were administered in patients who received pembrolizumab andnivolumab, respectively. A best response of PR was observed in 9patients (27.3%), SD in 9 patients (27.3%), and PD in 15 patients(45.4%). Twelve patients (36.4%) had DCB. The median follow-up time was 18.5 months (range, 6.5–28.5 months).

In NSCLC cohort 2 (n ¼ 46), the patients received a median 1line of previous systemic chemotherapy and no immune check-point inhibitors. Pembrolizumabwas administered in 38 patientsand nivolumab was administered in 8 patients. A median of 6cycles (range, 1–15) and a median of 9 cycles (range, 2–23) wereadministered in patients who received pembrolizumab and nivo-lumab, respectively. PR was observed in 11 patients (23.9%), SDin 19 patients (41.3%), and PD in 16 patients (34.8%). Nineteenpatients (41.3%) had DCB. The median follow-up time was 10.9months (range, 6.2–14.2 months).

Peripheral blood PD-1þCD8þ T cells show a proliferativeresponse 1 week after anti-PD-1 therapy

A previous study examined the percentage of Ki-67þ prolif-erative cells among peripheral blood PD-1þCD8þ T cells, whichwere previously reported to be enriched for tumor antigen–specific T cells (29), 3 weeks after the first dose of pembrolizu-mab and reported that this value could predict treatment out-comes, but only when combined with the initial tumor bur-den (24).We hypothesized that the proliferation of CD8þ T cellspeaks earlier than 3weeks after administration of the first dose ofanti-PD-1 antibodies. When we examined PBMCs from 9patients with NSCLC, we found that the percentage of prolifer-ative (Ki-67þ) cells among PD-1þCD8þ T cells significantlyincreased in the first 7 days after the first dose and then signif-icantly decreased in the following 2 weeks (3 weeks after the firstdose; Fig. 1A), indicating that the proliferative response peakedat 7 days. On the basis of these data, we decided to evaluate theproliferative response of PD-1þCD8þ T cells at 7 days after thefirst dose and found that the percentage of Ki-67þ cells amongperipheral blood PD-1þCD8þ T cells significantly increasedduring the first 7 days in both discovery cohorts, TET cohortand NSCLC cohort 1 (Fig. 1B and C).

To confirm whether a proliferative response was exhibited bytumor antigen–specific CD8þ T cells, we examined the percentageof Ki-67þ cells in the gate of tumor antigen–specific, peripheralblood CD8þ T cells fromHLA-A�0201(þ) patients exhibiting NY-ESO-1 mRNA upregulation in their tumor tissues. HLA-A�0201-restricted NY-ESO-1 dextramer (HLA-A�0201: NY-ESO-1157) wasused to detect tumor antigen–specific CD8þ T cells, and HLA-A�0201-restricted HCMV pp65 pentamer (HLA-A�0201:pp65495) was used to detect nontumor antigen–specific CD8þ

T cells as a control. We found that the percentages of Ki-67þ cellssignificantly increased inNY-ESO-1–specificCD8þT cells (Fig. 1Dand E) on day 7, but not in HCMV pp65-specific CD8þ T cells(Fig. 1F and G), in both the TET cohort and NSCLC cohort 1.

The first-week proliferative response of PD-1þCD8þ T cellspredicts tumor responses to anti-PD-1 therapy and clinicaloutcomes in patients with TETs

The proliferative response after anti-PD-1 therapy was assessedby the fold-change in the percentage of Ki-67þ cells among PD-1þCD8þ T cells on day 7 (Ki-67D7/D0). In the cohort of patientswith TETs, Ki-67D7/D0 was significantly higher in patients withDCB than in those with no durable benefit (NDB; Fig. 2A).

Furthermore, Ki-67D7/D0 had a high predictive value for DCB[AUC 0.86, 95% confidence interval (CI), 0.71–1.00; P ¼0.001; Fig. 2B). An optimal cut-off of 2.8 for Ki-67D7/D0 wasdetermined from the ROC curve. The sensitivity and specificity ofthis cut-off were 90.9% and 75.0%, respectively. The probabilityof DCB was significantly higher in patients with Ki-67D7/D0 � 2.8than in patientswith Ki-67D7/D0 < 2.8, although 33.3%of patientswith Ki67D7/D0 � 2.8 progressed within 6 months and 6.3% ofpatients with Ki67D7/D0 < 2.8 presented with DCB (Fig. 2C).Among patients with Ki-67D7/D0 � 2.8, 5 achieved PR, 9 had SD,and 1 had PD, whereas among patients with Ki-67D7/D0 < 2.8, 1achieved PR, 9 had SD, and 6 had PD as best response (P¼ 0.052;Supplementary Table S1). Furthermore, patients with Ki-67D7/D0

� 2.8 had a significantly prolonged PFS but not OS (Fig. 2D andE). The median PFS was 8.7 months (95% CI, 4.3–13.2 months)in patients with Ki-67 D7/D0 � 2.8 and 3.9 months (95% CI, 1.2–6.6 months) in those with Ki-67D7/D0 < 2.8 (P ¼ 0.027; Fig. 2D).The median OS was 14.8 months (95% CI, 0.0–30.7 months) inpatients with Ki-67 D7/D0 � 2.8 and 15.2 months (95% CI, 0.0–33.0months) in those with Ki-67D7/D0 < 2.8 (P¼ 0.78; Fig. 2E). Amultivariate analysis adjusted for other clinicopathologic factorsrevealed that Ki-67D7/D0 was independently and significantlyassociated with PFS [adjusted HR (aHR) 0.27; 95% CI, 0.09–0.82; P ¼ 0.020), whereas Ki-67D7/D0 was not significantly asso-ciated with OS (Supplementary Table S2).

Ki-67D7/D0 predicts tumor response to anti-PD-1 therapy andclinical outcomes in patients with NSCLC

We then tested Ki-67D7/D0 as a predictive biomarker in acohort of patients with NSCLC (NSCLC cohort 1, n ¼ 33),which has a larger application of anti-PD-1 therapy. Similar tothe TET cohort, Ki-67D7/D0 was significantly higher in patientswith DCB than in those with NDB (Fig. 3A). Ki-67D7/D0

significantly predicted DCB (AUC 0.78, 95% CI, 0.61–0.94;P ¼ 0.009; Fig. 3B) and the same optimal cut-off of 2.8 wasdetermined from the ROC curve. The probability of DCB wassignificantly higher in patients with Ki-67D7/D0 � 2.8 than inpatients with Ki-67D7/D0 < 2.8, though 37.5% of patients withKi67D7/D0 � 2.8 progressed within 6 months and 11.8% ofpatients with Ki67D7/D0 < 2.8 presented DCB (Fig. 3C). Amongpatients with Ki-67D7/D0 � 2.8, 7 achieved PR, 5 had SD, and 4had PD, whereas among patients with Ki-67D7/D0 < 2.8, 2achieved PR, 4 had SD, and 11 had PD as best response (P¼ 0.053; Supplementary Table S3). The median PFS was 6.0months (95% CI, 3.3–8.7 months) for patients with Ki-67D7/D0

� 2.8 and 1.4 months (95% CI, 1.0–1.8 months) for patientswith Ki-67D7/D0 < 2.8 (P ¼ 0.004; Fig. 3D). The median OS was13.8 months (95% CI, 10.0–17.5 months) in patients with Ki-67D7/D0 � 2.8 and was 2.0 months (95% CI, 0.0–4.6 months)in patients with Ki-67D7/D0 < 2.8 (P ¼ 0.001; Fig. 3E).

Next, we validated Ki-67D7/D0 as a predictive biomarker in anindependent cohort of patients with NSCLC (NSCLC cohort 2,n ¼ 46). The percentage of Ki-67þ cells among peripheral bloodPD-1þCD8þ T cells also significantly increased during the first 7days in NSCLC cohort 2 (Fig. 4A). In addition, Ki-67D7/D0 wassignificantly higher in patients with DCB than in those with NDB(Fig. 4B) and Ki-67D7/D0 significantly predicted DCB (AUC 0.81,95% CI, 0.68–0.94; P < 0.001; Fig. 4C). Using the predefined cut-off of 2.8, we stratified the patients with Ki-67D7/D0 � 2.8 andKi-67D7/D0 < 2.8. The probability of DCB was significantly higherin patients with Ki-67D7/D0� 2.8 than in patients with Ki-67D7/D0

Predictive Peripheral Blood Biomarker for Anti-PD-1 Therapy

www.aacrjournals.org Clin Cancer Res; 25(7) April 1, 2020 2147

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

< 2.8 the patients (Fig. 4D). Among patients with Ki-67D7/D0 �2.8, 10 achieved PR, 10 had SD, and 6 had PD, whereas amongpatients with Ki-67D7/D0 < 2.8, 1 achieved PR, 9 had SD, and 10had PD as best response (P ¼ 0.022; Supplementary Table S4).Themedian PFS was 10.9months (95%CI, 4.7–17.1months) forpatients with Ki-67D7/D0 � 2.8 and 2.1 months (95% CI, 2.0–2.3

months) for patients with Ki-67D7/D0 < 2.8 (P ¼ 0.002; Fig. 4E).The median OS was not reached in patients with Ki-67D7/D0 �2.8 and 7.0 months (95% CI not evaluable) for patients withKi-67D7/D0 < 2.8 (P ¼ 0.037; Fig. 4F).

Multivariate analysis of Ki-67D7/D0 adjusted for other clinico-pathologic variables confirmed that Ki-67D7/D0 independently

Figure 1.

Proliferative response of tumorantigen–specific CD8þ T cellsand PD-1þCD8þ T cells inperipheral blood collected earlyafter anti-PD-1 therapy. A, Thepercentage of Ki-67þ cells inperipheral blood PD-1þCD8þ

T cells before anti-PD-1 therapy(day 0) and 7 days and 21 daysafter the first dose for patientswith NSCLC treated withpembrolizumab every 3 weeks(n¼ 9). Representative plotsare shown in the right.Percentage of Ki-67þ cellsamong PD-1þCD8þ T cellsfrom patients with TETs(n¼ 31; B) and NSCLC (n¼ 33;C). Percentage of Ki-67þ cellsamong NY-ESO-1-specific CD8þ

T cells in patients with TETs(n¼ 6; D) and NSCLC (n¼ 7; E).Percentage of Ki-67þ cellsamong HCMV pp65-specificCD8þ T cells from patients withTETs (n¼ 6; F) and NSCLC (n¼7; G). Representative multi-colorflow cytometry plots are shownto the right of each graph.Statistical comparisons wereperformed using theWilcoxonsigned-rank test (A,D–G) orpaired t test (B and C). ns, notsignificant (� , P < 0.05;�� , P < 0.01; ���� , P < 0.0001).

Kim et al.

Clin Cancer Res; 25(7) April 1, 2020 Clinical Cancer Research2148

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

correlated with both PFS (aHR 0.34; 95% CI, 0.19–0.60; P <0.001) and OS (aHR 0.34; 95% CI, 0.17–0.75; P ¼ 0.007;Supplementary Table S5) in the combined cohort of patientswith NSCLC. Performance status also independently correlatedwith OS (aHR 2.98; 95% CI, 1.36–10.12; P¼ 0.013), but was notsignificantly associated with best tumor response (P ¼ 0.15)and DCB (P ¼ 0.23), implying that it was a prognostic factorrather than a predictive biomarker. We further analyzed thepredictability of Ki-67D7/D0 separately in patients with squamouscell carcinoma (n ¼ 34) and patients with nonsquamous histol-ogy (n ¼ 45). Ki67D7/D0 significantly correlated with DCB, PFS,and OS in both squamous and nonsquamous NSCLC (Supple-mentary Table S6).

Next, we investigated which subpopulation within PD-1þCD8þ T cells responded to anti-PD-1 therapy in patients withNSCLC. We focused on CD28þ cells because recent studiesreported that CD28 is a major target of the PD-1–mediatedinhibitory signal and that anti-PD-1–induced T-cell reinvigora-tion depends on CD28 (30, 31). Whenwe analyzed Ki-67D7/D0 inthe CD28�PD-1þCD8þ and CD28þPD-1þCD8þ T-cell popula-tions separately, we found that Ki-67D7/D0was significantly higher

in the CD28þPD-1þCD8þ T-cell population than in theCD28�PD-1þCD8þ T-cell population (Fig. 4G).

Combination of Ki-67D7/D0 and PD-L1 TPS as predictivebiomarkers

Having established that Ki-67D7/D0 may serve as a promisingbiomarker in patients treated with anti-PD-1 therapy, we inves-tigated the association between PD-L1 TPS, the most commonlyused biomarker in practice, and Ki-67D7/D0. PD-L1 TPS wasevaluable in 25 (80.6%) and 68 (86.1%) patients in the TETcohort (n ¼ 31) and the combined NSCLC cohort (n ¼ 79),respectively (Supplementary Table S7). No significant differencein Ki-67D7/D0 was observed according to the level of PD-L1 TPS inpatients with TETs and NSCLC (Fig. 5A and B).

Next, we assessed whether combining PD-L1 TPS with Ki-67D7/D0 improves the predictive power compared with Ki-67D7/D0 as a single biomarker. In the TET cohort, PD-L1TPS did not significantly predict DCB (AUC 0.54, 95% CI,0.31–0.77; P ¼ 0.73) and had lower predictive power thanKi-67D7/D0 (AUC 0.89, 95% CI, 0.76–1.00; P ¼ 0.002) andthe combination of Ki-67D7/D0 and PD-L1 TPS (AUC 0.91,

Figure 2.

Fold-change in the percentage of Ki-67þ cells among PD-1þCD8þ T cells 7 days after the first dose of anti-PD-1 antibodies (Ki-67D7/D0) and treatment outcomesin the TET cohort. A, The Ki-67D7/D0 values were compared between patients with TETs with NDB (n¼ 20) and those with DCB (n¼ 11). The box plot representsthe interquartile range (IQR)with the horizontal line indicating the median. Whiskers extend to a maximum of 1.5� IQR beyond the box. The representative multi-color flow cytometry plots to the right show Ki-67 detection in a patient with NDB and a patient with DCB. B, The ROC curve and optimal cut-off point (2.8) forpredicting patients with DCB. C, Percentage of DCB in patients with Ki-67D7/D0� 2.8 and Ki-67D7/D0 < 2.8. Kaplan–Meier curves of PFS (D) and OS (E) in patientswith Ki-67D7/D0� 2.8 or Ki-67D7/D0 < 2.8. Statistical comparisons were performed using the Student t test (A), Pearson x2 test (C), or log-rank test (D and E;�� , P < 0.01; ��� , P < 0.001).

Predictive Peripheral Blood Biomarker for Anti-PD-1 Therapy

www.aacrjournals.org Clin Cancer Res; 25(7) April 1, 2020 2149

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

95% CI, 0.80–1.00; P ¼ 0.001; Fig. 5C). In addition, thecombination of Ki-67D7/D0 and PD-L1 did not significantlyimprove the predictive power compared with Ki-67D7/D0

alone (P ¼ 0.32; Fig. 5C). In the NSCLC cohort, PD-L1significantly predicted DCB (AUC 0.64, 95% CI, 0.51–0.77;P ¼ 0.049), but had lower predictive power than Ki-67D7/D0

(AUC 0.81, 95% CI, 0.70–0.91; P < 0.001) and combinationof Ki-67D7/D0 and PD-L1 TPS (AUC 0.82, 95% CI, 0.72–0.93;P < 0.001; Fig. 5D). No significant difference was foundbetween the AUC values using Ki-67D7/D0 as a single bio-marker and using the combination of Ki-67D7/D0 and PD-L1TPS (P ¼ 0.57; Fig. 5D).

DiscussionImmune checkpoint inhibitors that block the PD-1/PD-L1

interaction have shown promising results in patients withadvanced cancers (3–8). Although patients with a tumor responsecommonly experience long-term tumor control, only a smallfraction of patients experience clinical benefit with these inhibi-tors. Unfortunately, there is currently no robust biomarker thatpredicts the treatment outcome in response to anti-PD-1 therapy.In this study, we identified and validated a blood-based biomark-

er that predicted the response to anti-PD-1 therapy in 3 indepen-dent cohorts of 2 distinct types of cancer, TETs and NSCLC. Wefound that the fold-change in the percentage of Ki-67þ cellsamong peripheral blood PD-1þCD8þ T cells 7 days after the firstdose, a marker we termed Ki-67D7/D0, predicted tumor responseand survival.

Peripheral blood CD8þ T cells are composed of T cellsspecific to various antigens. In this study, we showed thatanti-PD-1 therapy increases the expression of Ki-67 in tumorantigen–specific T cells, but not in irrelevant virus-specific Tcells, using MHC I multimers. However, detecting tumor-spe-cific CD8þ T cells using MHC I multimers is limited to only afew HLA types and tumor antigens. In a previous study, tumor-specific CD8þ T cells were identified among peripheral bloodPD-1þCD8þ T cells of patients with cancer, but not in PD-1�CD8þ T cells (29). Thus, the proliferative response in PD-1þCD8þ T cells in patients with cancer may reflect the prolif-erative response of tumor-specific CD8þ T cells, even withoutMHC I multimers.

Considering the mode of action of anti-PD-1 blocking anti-bodies, it stands to reason that PD-1þCD8þ exhausted T cellsmustbe reinvigorated and recover their functions for patients to achievelong-term tumor control and show improved survival, regardless

Figure 3.

Ki-67D7/D0 predicts treatment outcomes in NSCLC cohort 1. A, The Ki-67D7/D0 value was compared between patients with NSCLC with NDB (n¼ 21) and thosewith DCB (n¼ 12). The box plot represents the IQR with the horizontal line indicating the median. Whiskers extend to a maximum of 1.5� IQR beyond the box.The representative multi-color flow cytometry plots to the right show Ki-67 detection in a patient with NDB and a patient with DCB. B, The ROC curve andoptimal cut-off point (2.8) for predicting patients with DCB. C, Percentage of DCB in patients with Ki-67D7/D0� 2.8 and with Ki-67D7/D0 < 2.8. Kaplan–Meiercurves of PFS (D) and OS (E) in patients with Ki-67D7/D0� 2.8 or Ki-67D7/D0 < 2.8. Statistical comparisons were performed using the Student t test (A), Pearson x2

test (C), or log-rank test (D and E; �� , P < 0.01).

Kim et al.

Clin Cancer Res; 25(7) April 1, 2020 Clinical Cancer Research2150

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

of the type of tumor. Indeed, the proliferative response of tumor-infiltrating CD8þ T cells after anti-PD-1 therapy in patients withmelanoma has been shown to be associated with tumorresponse (18). In our study, the proliferative response of PD-1þCD8þ T cells after anti-PD-1 therapy was important for thetumor response in 2 distinct types of tumors, TETs andNSCLC. Ki-67D7/D0 exhibited high predictive ability for NDB, and 89%of thepatients with Ki-67D7/D0 < 2.8 progressed within 6 months. Thisindicates that tumor control may hardly be achieved without thereinvigoration of PD-1þCD8þ T cells early after anti-PD-1 ther-

apy. This finding also implies that Ki-67D7/D0 may be used toscreen for patients who are least likely to respond to anti-PD-1therapy early after treatment initiation. However, there is stillroom for improvement of this new biomarker due to the falsevalues. Thirty-seven percent of patients with Ki67D7/D0 � 2.8progressed within 6 months, whereas 11% of patients withKi67D7/D0 < 2.8 presented DCB. Combining other biomarkerswith Ki-67D7/D0 may improve the predictive ability of Ki-67D7/D0

in discriminating the responders and nonresponders. By identi-fying the nonresponders in advance, patients may have an

Figure 4.

Validation of Ki-67D7/D0 in NSCLC cohort 2. A, Percentage of Ki-67þ cells among PD-1þCD8þ T cells at day 0 and day 7 in NSCLC cohort 2 (n¼ 46). B, The Ki-67D7/D0 value was compared between patients with NDB (n¼ 27) and those with DCB (n¼ 19). The box plot represents the IQR with the horizontal line indicatingthe median. Whiskers extend to a maximum of 1.5� IQR beyond the box. C, The ROC curve to predict DCB after treatment.D, Percentage of DCB in patients withKi-67D7/D0� 2.8 and with Ki-67D7/D0 < 2.8. Kaplan–Meier curves of PFS (E) and OS (F) in patients with Ki-67D7/D0� 2.8 or Ki-67D7/D0 < 2.8. G, Ki-67D7/D0 inCD28�PD-1þCD8þ T cells and CD28þPD-1þCD8þ T cells in patients with NSCLC (n¼ 79). The box plot represents the IQR with the horizontal line indicating themedian. Whiskers extend to a maximum of 1.5� IQR beyond the box. Representative multi-color flow cytometry plots on the right show Ki-67 expression in theCD28�PD-1þCD8þ and CD28þPD-1þCD8þ T cell gates. Statistical comparisons were performed using paired t test (A, G), Student t test (B), Pearson x2 test (D),or the log-rank test (E and F; ��, P < 0.01; ��� , P < 0.001; ���� , P < 0.0001).

Predictive Peripheral Blood Biomarker for Anti-PD-1 Therapy

www.aacrjournals.org Clin Cancer Res; 25(7) April 1, 2020 2151

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

opportunity to try other treatment options or combinations ratherthan blindly taking anti-PD-1 monotherapy.

Previous studies have investigated other biomarkers, includinghigh levels of PD-L1 expression in tumor tissues, abundant T-cellinfiltration into tumors, high mutational load, and mismatchrepair deficiency (14, 16–22). However, none of these biomarkerswere sufficiently robust for precise predictions. One explanationis that the treatment response to immune checkpoint blockadeis influenced by multiple factors rather than by a single fac-tor (33, 34). Ki-67D7/D0 is a dynamic immunologic responsemarker that may incorporate multiple factors that influenceT-cell proliferation, such as the neoantigen load, the adequacyof antigen presentation and costimulation, cytokine regulation,and the efficiency of treatment in terms of blocking the PD-1/PD-L1 interaction.

We also found that CD28þPD-1þCD8þ T cells preferentiallyproliferated in response to anti-PD-1 therapy compared withCD28�PD-1þCD8þ T cells, which is in line with a recent studyreporting CD28 dependency of the effect of PD-1 blockade onexhausted PD-1þCD8þ T cells (30, 31). Considering these data,further studies are needed to investigate therapies that can rein-vigorate CD28�PD-1þCD8þ T cells to overcome unresponsive-ness to anti-PD-1 therapy.

We further investigated the association between tumor tissuePD-L1 expression andKi-67D7/D0. Ki-67D7/D0 did not significantlydiffer between patients with high PD-L1 expression and low PD-L1 expression. As mentioned above, Ki-67D7/D0 could be influ-enced by multiple factors; therefore, PD-L1 itself may not show asignificant correlation with Ki-67D7/D0. We further investigated

whether combining Ki-67D7/D0 and PD-L1 TPS could betterpredict treatment outcomes. However, the addition of PD-L1 TPSto Ki-67D7/D0 was not able to further improve the predictivepower. Nevertheless, due to the small number of patients includ-ed in the analyses, future studies are needed to confirm thesefindings.

Given that themode-of-actionof PD-1blocking antibodies andPD-L1 blocking antibodies are similar, further investigation of thepredictive power of Ki-67D7/D0 in patients treated with atezoli-zumab or durvalumab may be an interesting subject of research.In addition, combination therapy of anti-PD-1/PD-L1 with othertreatment modalities, such as chemotherapy and chemora-diotherapy, are being actively investigated (35, 36). In thesestudies, the clinical benefit of anti-PD-1/PD-L1 therapy wasobserved regardless of PD-L1 status. Thus, new biomarkers areneeded in these treatment settings and Ki-67D7/D0 may be a goodcandidate for a predictive biomarker.

In this study, Ki-67D7/D0 predicted treatment outcome in 2distinct types of tumors, TETs andNSCLC. Considering themode-of-action of PD-1 blocking antibodies, measuring the extent ofPD-1þCD8þ T-cell proliferation may be applicable to broadertypes of cancer. However, Ki-67D7/D0 significantly predictedOS inpatients with NSCLC, but not in the patients with TETs. Thisimplies that the predictive power of Ki-67D7/D0 may vary accord-ing to the type of cancer or histology and should be confirmed infuture studies.

This study has some limitations. The sample size was small,patients with different types of histology were included, andfollow-up time was short, which may limit the statistical power

Figure 5.

Combination of Ki-67D7/D0 and PD-L1 TPS aspredictive biomarkers. A, Ki-67D7/D0 was comparedbetween patients with TETs with low PD-L1 TPS (<50%, n¼ 12) and high PD-L1 TPS (� 50%, n¼ 13). B,Ki-67D7/D0 was compared between patients withNSCLC with low PD-L1 TPS (< 50%, n¼ 27) and highPD-L1 TPS (� 50%, n¼ 41). The box plot representsthe IQR with the horizontal line indicating the median.Whiskers extend to a maximum of 1.5� IQR beyondthe box. The ROC curve of Ki-67D7/D0 (blue solid line),PD-L1 TPS (black solid line), and the combination ofKi-67D7/D0 and PD-L1 TPS (green solid line) to predictDCB in patients with TETs (C) and NSCLC (D).Statistical comparisons were performed using Studentt test (A and B), or Delong test (C and D). ns, notsignificant.

Kim et al.

Clin Cancer Res; 25(7) April 1, 2020 Clinical Cancer Research2152

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

and interpretation of our data. Thus, our results must be furthervalidated in larger cohorts with longer follow-up. In addition, weonly tested 2 types of cancer and further validation should beperformed in other types of cancer that are approved for pem-brolizumab or nivolumab treatment, such as malignant melano-ma, head and neck squamous cell carcinoma, urothelial cancer,and hepatocellular carcinoma.

In conclusion, this preliminary study suggests an association ofthe fold-change in the percentage of Ki-67þ cells among PD-1þCD8þ T cells 7 days after the first dose of anti-PD-1 antibodies(Ki-67D7/D0) with treatment outcomes in patients treated withanti-PD-1 therapy. Further prospective studies withmore patientswith various types of cancer will be needed to validate Ki-67D7/D0

as a useful biomarker in patients with solid tumors treated withanti-PD-1 therapy.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: K.H. Kim, M.-J. Ahn, E.-C. ShinDevelopment of methodology: K.H. Kim, S.-H. Park, M.-J. Ahn

Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): K.H. Kim, J. Cho, B.M. Ku, J.-M. Sun, S.-H. Lee,J.S. Ahn, J. Cheon, Y.J. Min, K. Park, vAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K.H. Kim, J. Cho, S.-H. Park, K. Park, M.-J. AhnWriting, review, and/or revision of the manuscript: K.H. Kim, B.M. Ku,S.-H. Lee, J.S. Ahn, K. Park, M.-J. Ahn, E.-C. ShinAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): K.H. Kim, J. Koh, M.-J. AhnStudy supervision: M.-J. Ahn, E.-C. Shin

AcknowledgmentsThis study was supported by a National Research Foundation (NRF) grant

(NRF-2017R1A2A1A17069782), the Collaborative Genome Program for Fos-teringNew Post-Genome Industry of theNRF (NRF-2017M3C9A6044632), theGlobal PhD Fellowship of the NRF (NRF-2017H1A2A1046131), and by theKorea Advanced Institute of Science and Technology (KAIST) Future SystemsHealthcare Project, which is funded by the Ministry of Science and ICT (MSIT).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received May 10, 2018; revised August 23, 2018; accepted January 7, 2019;published first January 15, 2019.

References1. Pardoll DM. The blockade of immune checkpoints in cancer immuno-

therapy. Nat Rev Cancer 2012;12:252–64.2. Sharma P, Allison JP. Immune checkpoint targeting in cancer therapy:

toward combination strategies with curative potential. Cell 2015;161:205–14.

3. Robert C, Schachter J, Long GV, Arance A, Grob JJ, Mortier L, et al.Pembrolizumab versus Ipilimumab in advanced melanoma. N EnglJ Med 2015;372:2521–32.

4. Weber JS, D'Angelo SP, Minor D, Hodi FS, Gutzmer R, Neyns B, et al.Nivolumab versus chemotherapy in patients with advanced melanomawho progressed after anti-CTLA-4 treatment (CheckMate 037): a rando-mised, controlled, open-label, phase 3 trial. Lancet Oncol 2015;16:375–84.

5. Brahmer J, Reckamp KL, Baas P, Crino L, Eberhardt WE, Poddubskaya E,et al. Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung cancer. N Engl J Med 2015;373:123–35.

6. Borghaei H, Paz-Ares L, Horn L, Spigel DR, Steins M, Ready NE, et al.Nivolumab versus docetaxel in advanced nonsquamous non-small-celllung cancer. N Engl J Med 2015;373:1627–39.

7. Reck M, Rodriguez-Abreu D, Robinson AG, Hui R, Csoszi T, Fulop A, et al.Pembrolizumab versus chemotherapy for PD-L1-positive non-small-celllung cancer. N Engl J Med 2016;375:1823–33.

8. Seiwert TY, Burtness B, Mehra R, Weiss J, Berger R, Eder JP, et al. Safetyand clinical activity of pembrolizumab for treatment of recurrent ormetastatic squamous cell carcinoma of the head and neck (KEYNOTE-012): an open-label, multicentre, phase 1b trial. Lancet Oncol 2016;17:956–65.

9. Balar AV, Castellano D, O'Donnell PH, Grivas P, Vuky J, Powles T, et al.First-line pembrolizumab in cisplatin-ineligible patients with locallyadvanced and unresectable or metastatic urothelial cancer (KEYNOTE-052): a multicentre, single-arm, phase 2 study. Lancet Oncol 2017;18:1483–92.

10. Muro K, Chung HC, Shankaran V, Geva R, Catenacci D, Gupta S, et al.Pembrolizumab for patients with PD-L1-positive advanced gastric cancer(KEYNOTE-012): a multicentre, open-label, phase 1b trial. Lancet Oncol2016;17:717–26.

11. Doi T, Piha-Paul SA, Jalal SI, Saraf S, Lunceford J, KoshijiM, et al. Safety andantitumor activity of the anti-programmed death-1 antibody pembrolizu-mab in patients with advanced esophageal carcinoma. J Clin Oncol 2018;36:61–7.

12. Nishino M, Ramaiya NH, Hatabu H, Hodi FS. Monitoring immune-checkpoint blockade: response evaluation and biomarker development.Nat Rev Clin Oncol 2017;14:655–68.

13. Topalian SL, Taube JM, Anders RA, Pardoll DM. Mechanism-driven bio-markers to guide immune checkpoint blockade in cancer therapy.Nat Rev Cancer 2016;16:275–87.

14. Garon EB, Rizvi NA, Hui R, Leighl N, Balmanoukian AS, Eder JP, et al.Pembrolizumab for the treatment of non-small-cell lung cancer.N Engl J Med 2015;372:2018–28.

15. Gettinger SN,HornL,GandhiL,SpigelDR,AntoniaSJ, RizviNA, etal.Overallsurvival and long-term safety of nivolumab (Anti-Programmed Death 1Antibody, BMS-936558, ONO-4538) in patients with previously treatedadvanced non-small-cell lung cancer. J Clin Oncol 2015;33:2004–12.

16. Rizvi NA, Hellmann MD, Snyder A, Kvistborg P, Makarov V, Havel JJ, et al.Cancer immunology. Mutational landscape determines sensitivity to PD-1blockade in non-small cell lung cancer. Science 2015;348:124–8.

17. Le DT, Uram JN,WangH, Bartlett BR, Kemberling H, Eyring AD, et al. PD-1blockade in tumors with mismatch-repair deficiency. N Engl J Med 2015;372:2509–20.

18. Tumeh PC, Harview CL, Yearley JH, Shintaku IP, Taylor EJ, Robert L, et al.PD-1 blockade induces responses by inhibiting adaptive immune resis-tance. Nature 2014;515:568–71.

19. Daud AI, Loo K, Pauli ML, Sanchez-Rodriguez R, Sandoval PM, Taravati K,et al. Tumor immune profiling predicts response to anti-PD-1 therapy inhuman melanoma. J Clin Invest 2016;126:3447–52.

20. Hugo W, Zaretsky JM, Sun L, Song C, Moreno BH, Hu-Lieskovan S, et al.Genomic and transcriptomic features of response to Anti-PD-1 therapy inmetastatic melanoma. Cell 2017;168:542.

21. Weide B, Martens A, Hassel JC, Berking C, Postow MA, Bisschop K, et al.Baseline biomarkers for outcome of melanoma patients treated withpembrolizumab. Clin Cancer Res 2016;22:5487–96.

22. Krieg C,NowickaM,Guglietta S, Schindler S,Hartmann FJ,Weber LM, et al.High-dimensional single-cell analysis predicts response to anti-PD-1immunotherapy. Nat Med 2018;24:144–53.

23. Lesterhuis WJ, Bosco A, Millward MJ, Small M, Nowak AK, Lake RA.Dynamic versus static biomarkers in cancer immune checkpoint blockade:unravelling complexity. Nat Rev Drug Discov 2017;16:264–72.

24. Huang AC, Postow MA, Orlowski RJ, Mick R, Bengsch B, Manne S, et al. T-cell invigoration to tumour burden ratio associated with anti-PD-1response. Nature 2017;545:60–5.

www.aacrjournals.org Clin Cancer Res; 25(7) April 1, 2020 2153

Predictive Peripheral Blood Biomarker for Anti-PD-1 Therapy

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

25. Kamphorst AO, Pillai RN, Yang S, Nasti TH, Akondy RS, Wieland A,et al. Proliferation of PD-1þ CD8 T cells in peripheral blood after PD-1-targeted therapy in lung cancer patients. Proc Natl Acad Sci U S A 2017;114:4993–8.

26. Sanmamed MF, Perez-Gracia JL, Schalper KA, Fusco JP, Gonzalez A,Rodriguez-Ruiz ME, et al. Changes in serum interleukin-8 (IL-8) levelsreflect and predict response to anti-PD-1 treatment inmelanoma and non-small-cell lung cancer patients. Ann Oncol 2017;28:1988–95.

27. Chen PL, Roh W, Reuben A, Cooper ZA, Spencer CN, Prieto PA, et al.Analysis of immune signatures in longitudinal tumor samples yieldsinsight into biomarkers of response and mechanisms of resistance toimmune checkpoint blockade. Cancer Discov 2016;6:827–37.

28. Cho J, Kim HS, Ku BM, Choi YL, Cristescu R, Han J, et al. Pembroli-zumab for patients with refractory or relapsed thymic epithelial tumor:an open-label phase ii trial. J Clin Oncol 2018 Jun 15 [Epub ahead ofprint].

29. Gros A, Parkhurst MR, Tran E, Pasetto A, Robbins PF, Ilyas S, et al.Prospective identification of neoantigen-specific lymphocytes in theperipheral blood of melanoma patients. Nat Med 2016;22:433–8.

30. Hui E, Cheung J, Zhu J, Su X, Taylor MJ, Wallweber HA, et al. T cellcostimulatory receptor CD28 is a primary target for PD-1-mediated inhi-bition. Science 2017;355:1428–33.

31. Kamphorst AO, Wieland A, Nasti T, Yang S, Zhang R, Barber DL, et al.Rescue of exhausted CD8 T cells by PD-1-targeted therapies is CD28-dependent. Science 2017;355:1423–7.

32. Giaccone G, Kim C, Thompson J, McGuire C, Kallakury B, Chahine JJ, et al.Pembrolizumab in patients with thymic carcinoma: a single-arm, single-centre, phase 2 study. Lancet Oncol 2018;19:347–55.

33. BlankCU,Haanen JB, Ribas A, Schumacher TN.CANCER IMMUNOLOGY.The "cancer immunogram". Science 2016;352:658–60.

34. ChenDS,Mellman I. Elements of cancer immunity and the cancer-immuneset point. Nature 2017;541:321–30.

35. Gandhi L, Rodriguez-Abreu D, Gadgeel S, Esteban E, Felip E, De Angelis F,et al. Pembrolizumab plus chemotherapy inmetastatic non-small-cell lungcancer. N Engl J Med 2018;378:2078–92.

36. Antonia SJ, Villegas A, Daniel D, Vicente D, Murakami S, Hui R, et al.Durvalumab after chemoradiotherapy in stage III non-small-cell lungcancer. N Engl J Med 2017;377:1919–29.

Clin Cancer Res; 25(7) April 1, 2020 Clinical Cancer Research2154

Kim et al.

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449

2019;25:2144-2154. Published OnlineFirst January 15, 2019.Clin Cancer Res   Kyung Hwan Kim, Jinhyun Cho, Bo Mi Ku, et al.   Tumors

T Cells Predicts the Response to Anti-PD-1 Therapy in Solid+CD8+The First-week Proliferative Response of Peripheral Blood PD-1

  Updated version

  10.1158/1078-0432.CCR-18-1449doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://clincancerres.aacrjournals.org/content/suppl/2019/01/15/1078-0432.CCR-18-1449.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://clincancerres.aacrjournals.org/content/25/7/2144.full#ref-list-1

This article cites 35 articles, 8 of which you can access for free at:

  Citing articles

  http://clincancerres.aacrjournals.org/content/25/7/2144.full#related-urls

This article has been cited by 6 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.org

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/25/7/2144To request permission to re-use all or part of this article, use this link

on November 9, 2020. © 2019 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 15, 2019; DOI: 10.1158/1078-0432.CCR-18-1449