Phase I Dose-Escalation Trial of MIW815 (ADU-S100), an Intratumoral STING Agonist, in Patients with Advanced/Metastatic Solid Tumors or Lymphomas

The stimulator of IFN genes (STING) pathway senses intracellular DNA, triggering a number of host defense pathways and resulting in production of type I IFN and other cytokines (1, 2). Under physiologic conditions, the STING pathway is regulated by cyclic GMP–AMP (cGAMP) synthase (cGAS), which binds cytosolic DNA and generates the cyclic dinucleotide, 2′–5′ cGAMP (3). Interaction between cGAMP and STING at the endoplasmic reticulum leads to trafficking through the Golgi to perinuclear vesicles, in complex with TANK-binding kinase 1 (TBK1; 4, 5). The STING–TBK1 complex subsequently interacts with IFN regulatory factor 3 (IRF3), leading to phosphorylation of this transcription factor and nuclear activation of downstream target genes, notably IFNβ (6). Consistent with its role in immune activation, gain-of-function mutations in human TMEM173 (the gene encoding STING) are associated with autoinflammatory diseases, including a syndrome with features similar to lupus and STING-associated vasculopathy with onset in infancy (SAVI; refs. 7, 8).

Cancer immunotherapy with immune checkpoint blockade has changed the therapeutic landscape for many solid tumors, yet not all patients benefit from treatment. Response has been associated with baseline immune infiltration of tumors, suggesting that priming innate immunity might lead to greater antitumor adaptive immunity, manifesting as an increased presence of tumor-infiltrating lymphocytes and stronger treatment responses (9, 10). Several innate immune, nucleic acid–sensing pathways are under investigation as candidates for therapeutic intervention, and the STING pathway is a high priority based on preclinical modeling and human knockout phenotypes (11, 12). STING plays a crucial role in activating antigen-presenting cells (APCs), leading to increased production of proinflammatory cytokines and recruitment and priming of cluster of differentiation 8 (CD8)–positive (CD8⁺) T cells against tumor antigens (13). STING activation–associated TNFα production has also been implicated in STING-mediated tumor cell killing (14).

MIW815 (ADU-S100) is a novel, synthetic, cyclic dinucleotide that activates the STING pathway in vitro and in vivo (2, 15, 16). MIW815 is administered intratumorally (i.t.), which has the potential to provide higher local concentration of dose than systemic administration, inducing type I IFN-associated immunotherapy and activation of local, tumor-resident, APCs. In murine tumor models, i.t. injection of MIW815 resulted in tumor regression in both injected and noninjected lesions (2, 15, 16). These preclinical data suggest that local administration of MIW815 has the potential to generate systemic antitumor immunity and may additionally provide long-lived immunologic memory. Systemic exposure of MIW815 after local injection may also contribute to distant antitumor activity when administered at high doses (16). Mechanistic studies demonstrate that STING-mediated antitumor immunity is associated with a tumor-specific, CD8⁺ T-cell response, as well as an acute proinflammatory cytokine response that can lead to local vascular collapse. At sufficiently high doses, however, MIW815 can induce cell death, resulting in impaired T-cell activation (15, 16).

Here, we describe single-agent, dose-escalation data from a multicenter, open-label, first-in-class, phase I study of MIW815 in patients with advanced/metastatic solid tumors or lymphomas.

This study (ClinicalTrials.gov: NCT02675439) was performed in accordance with the Declaration of Helsinki and the principles of Good Clinical Practice. The protocol was approved by an Institutional Review Board at each investigative site, and all patients provided written informed consent before any study procedures. The study was designed by the sponsor (Novartis Pharma AG and Aduro Biotech); the sponsor collected the data and analyzed them in conjunction with the authors.

This was a phase I, first-in-human, multicenter, open-label, dose-escalation study of MIW815 as a single agent in patients with advanced/metastatic solid tumors and lymphomas. Patients were treated with weekly, i.t. injections of MIW815 (50–6,400 μg) on a 3-weeks-on/1-week-off schedule until they experienced unacceptable toxicity, disease progression as per the immune-related response criteria (irRC) for solid tumors, or confirmed progressive disease in lymphoma, and/or patient/investigator decision. Injection volumes were 0.5 to 1.0 mL, 1.0 to 2.0 mL, and 2.0 to 4.0 mL for lesions of diameter 10 to 25 mm, >25 to 50 mm, and >50 to <100 mm, respectively. The primary objectives were to characterize the safety and tolerability of MIW815 as a single agent and identify a recommended dose for future studies. Secondary objectives included evaluation of the preliminary antitumor activity of single-agent MIW815, characterization of the pharmacokinetic (PK) properties of MIW815, and assessment of the pharmacodynamic effects of MIW815 in injected and distal lesions. Biomarkers of response to MIW815 were investigated as an exploratory objective.

Dose escalation used a Bayesian hierarchical logistic model (BHLRM; ref. 17). Dose escalation of single-agent MIW815 was guided by the escalation with overdose control (EWOC) principle (18) to determine the maximum tolerated dose (MTD) and/or recommended dose for expansion (RDE) based on dose-limiting toxicities (DLTs) observed in cycle 1. Prior to determination of the MTD and/or RDE, a minimum of six evaluable patients were required to have received treatment at that dose level.

Adult patients with advanced/metastatic solid tumors or lymphomas with cutaneous, subcutaneous, and/or nodal lesions that were visible, palpable, or detectable by ultrasound guidance, who progressed despite standard therapy, were intolerant to standard therapy, or for whom effective standard therapy does not exist, were eligible. All patients had an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1. Patients were required to have measurable disease per the Response Evaluation Criteria In Solid Tumors (RECIST) v1.1 or Cheson 2014 criteria (19) and at least two distinct lesions, both accessible for baseline and on-treatment biopsies, with patient consent. The injected lesion was required to measure ≥10 to <100 mm in longest diameter and be accessible for repeated i.t. injection.

Patients were excluded if they had symptomatic central nervous system (CNS) metastases or CNS metastases requiring local CNS-directed therapy, impaired cardiac function or clinically significant cardiac disease, and a history of, or active, autoimmune disease, with the exception of vitiligo or resolved childhood asthma/atopy; patients may have received prior immunotherapy. The following treatments were not permitted: systemic anticancer therapy within 4 weeks of the first dose of study treatment, chronic systemic steroid therapy other than replacement-dose steroids in the setting of adrenal insufficiency, or systemic treatment with any immunosuppressive medication.

Adverse events (AEs) were assessed at every visit according to the Common Terminology Criteria for Adverse Events v4.03.

Tumor response was determined locally according to irRC, RECIST v1.1, and Cheson 2014 criteria for lymphoma, at screening and on day 1 of cycle 3, every 8 weeks up to cycle 11, and then every 12 weeks until disease progression or patient withdrawal. Brain scans were performed if disease was detected at baseline or if clinically indicated. Target lesions were selected according to irRC/RECIST v1.1, or Cheson 2014 criteria; injected lesions could be selected as target lesions, but this was not mandatory.

Pre- and postinjection (±2 minutes) blood samples for MIW815 PK analysis were collected at multiple timepoints on cycle 1, days 1 and 15, and cycle 3, day 1. Plasma samples were assayed for MIW815 concentration using a validated liquid chromatography/tandem mass spectrometry assay with a lower limit of quantification of 50 pg/mL.

Blood samples were also used for pharmacodynamic assessments, including measurement of soluble cytokines, clonal expansion analysis, and cell-based markers. Plasma cytokines, including IFNβ, MIP1-β, IL6, monocyte chemo-attractant protein 1 (MCP-1), and C-X-C motif chemokine ligand 10 (CXCL10), were measured using ELISA. Systemic T-cell clonal expansion was assessed by Adaptive Biotechnologies and run on the hsTCRB v4 assay at deep sequencing resolution.

Newly obtained, pre- and posttreatment paired biopsies from injected and distant lesions were collected at screening and in cycle 2, days 18 to 25, or cycle 3, day 1. Tumor biopsies were used to assess target modulation by IHC methods, including CD8 (percent marker area), programmed death ligand-1 (PD-L1; percent positive tumor score and combined positivity score; Dako PD-L1 IHC 22C3 pharmaDx), cluster of differentiation 68 (CD68; percent marker area), and forkhead box protein P3 (FOXP3; percent marker area). RNA expression analysis was also performed (NanoString; gene lists shown in Supplementary Table S1).

At the analysis cut-off date (February 25, 2020), 47 patients received treatment with single-agent MIW815 in the phase I, dose-escalation part of the study. Patients discontinued treatment due to disease progression (n = 25; 60%), physician decision (n = 12; 26%), patient/guardian decision (n = 4; 9%), death (n = 2; 4%), or incidence of AEs (n = 1; 2%).

Baseline patient demographics and disease characteristics are shown in Table 1. The median age was 62.0 years (range, 26–80 years) and patients had an ECOG performance status of 0 (n = 12; 26%) or 1 (n = 35; 74%). The majority of patients had received prior therapy (n = 42; 89%) and 30 patients (64%) had received at least three prior regimens; 22 patients (47%) had received prior treatment with immune checkpoint inhibitors. Five patients had a primary diagnosis of lymphoma (11%) and 42 (89%) had solid tumors. Common solid tumor types included melanoma (n = 11; 23%; including five patients with cutaneous melanoma, two with uveal melanoma, and one with other, noncutaneous melanoma), breast cancer (n = 6; 13%; including three with triple-negative breast cancer), colorectal cancer, Merkel cell carcinoma (MCC), and sarcoma (n = 3; 6% each). All patients had metastatic disease; seven patients (15%) had one metastatic site, eight patients (17%) had two, and 32 (68%) had three or more metastatic sites; 31 patients (66%) had visceral metastases. The median tumor size of first-injected lesions at baseline was 41 mm (range, 11–100 mm). Most patients (89%) had one injected lesion, and the median number of target and nontarget lesions was three in both cases (Supplementary Table S2). The median duration of exposure to study treatment was 8 weeks (range, 0.1–85.9 weeks).

The most frequently reported, any-grade AEs (≥15% of all patients), regardless of study-drug relationship, were anemia, fatigue, nausea, pyrexia (n = 12 each; 26%), injection-site pain (n = 9; 19%), and chills (n = 8; 17%; Table 2). Grade 3/4 AEs, regardless of study-drug relationship, were reported in 40% of patients; the most frequent (≥5% of patients) were anemia and hyponatremia (n = 4; 9% each; Table 2). A grade 3 injection site ulcer was the only reported DLT observed during the first cycle of treatment, occurring in one patient (2%) with sarcoma who was treated at the highest dose level tested (6,400 μg). This ulceration resolved without intervention after approximately 4 to 5 weeks. The MTD or RDE were not reached.

There was no clear relationship between dose and toxicity (Table 2). The majority of patients (n = 34; 72%) reported at least one AE suspected to be related to study treatment; the most common (≥10% of patients) were pyrexia (n = 8; 17%), chills, injection-site pain (n = 7; 15%), and headache (n = 6; 13%; Supplementary Table S3). The median time to first suspected treatment-related AE was 3.5 days, with 14 of 34 patients experiencing treatment-related AEs on the first day of dosing. Grade 3/4 AEs suspected to be related to the study drug were reported in five (11%) patients: injection-site reaction, increased lipase (n = 2; 4% each), injection-site ulcer, tumor pain, and increased amylase (n = 1; 2% each). Thirty-three serious AEs (SAEs), regardless of causality, were reported in 19 of 47 (40%) patients. Two patients (4%) experienced at least one SAE suspected to be related to the study drug; grade 4 soft tissue abscess (one patient) and grade 2 chills (one patient). All were resolved/resolving at the data cut-off date. On-treatment deaths (defined as deaths that occurred between treatment start and within 30 days of treatment discontinuation) were reported in four patients due to disease progression (n = 3) and suicide (n = 1).

MIW815 was rapidly absorbed from the tumor injection site into the plasma, with median time to peak plasma concentration (T_max) coinciding with the end of injection. Fast elimination/degradation of MIW815 from the plasma was also observed (Fig. 1). MIW815 had biphasic elimination from the systemic circulation, with the terminal phase observed after 2 hours. The terminal phase half-life was approximately 24 minutes. Mean, steady-state plasma exposures of MIW815 increased dose proportionally for doses between 50 and 6,400 μg; no accumulation was observed in the plasma after weekly dosing. MIW815 had high interindividual variability with a coefficient of variation of more than 86%. A summary of MIW816 PK parameters is shown in Supplementary Table S4.

The duration of treatment with MIW815 and response per patient are shown in Supplementary Fig. S1. Across all dose levels, one patient experienced a systemic confirmed partial response (PR) per RECIST v1.1, for an overall response rate (ORR) of 2.1%. A reduction in lesion size was observed for the majority of injected lesions (Fig. 2). Stable disease was reported as the best response in 18 patients (38%); no complete responses were observed during dose escalation (Supplementary Table S5). The disease control rate, defined as the proportion of patients experiencing a response or stable disease, was 40% (Supplementary Table S5).

The patient with a PR had a diagnosis of MCC and had received prior chemotherapy but no prior immunotherapy. This patient received MIW815 at the 100-μg dose level and achieved a PR after four cycles of treatment, which was confirmed after six cycles, before progressing. Pre- and on-treatment images of injected lesions are shown in Supplementary Fig. S2. Unconfirmed PRs were observed in two patients (Supplementary Fig. S2); one immunotherapy-experienced patient (prior pembrolizumab) with parotid cancer who completed eight cycles of MIW815 treatment (800-μg dose level) and one patient with myxofibrosarcoma who had received prior therapy with an investigational checkpoint inhibitor who completed eight cycles of MIW815 treatment (6,400-μg dose level). Prolonged stable disease was observed in an immunotherapy-naïve patient with a collecting duct carcinoma of the kidney who completed 21 cycles of treatment (800-μg dose level). This patient previously progressed after 266 days on treatment with an investigational agent. Prolonged stable disease was also observed in a patient with estrogen receptor–positive breast cancer who was on study for 10 cycles (3,200-μg dose level). The prior two regimens for this patient were capecitabine (307 days) and eribulin (305 days).

Due to limitations in data collection, it was not possible to report the percentage change from baseline of noninjected lesions; the best percentage change from baseline in all target lesions and in injected lesions is shown in Fig. 2A and B, respectively. Lesion size was stable or decreased in the majority of evaluable injected lesions (33/35; 94.3%; Fig. 2B). The numbers of target, nontarget, and injected lesions are summarized in Supplementary Table S2.

Tumor biopsies from injected lesions and from noninjected lesions were obtained at screening and on-treatment (during cycle 2, between days 18 and 25, or cycle 3, day 1). At baseline, 28% of patients (11/39 evaluable) had tumor cells positive [tumor proportion score (TPS)] for PD-L1 expression in injected lesions (≥1% staining by IHC) and 28% (10/36) had PD-L1 TPS-positive, noninjected lesions. A subset of patient samples was also analyzed using the combined positivity score (CPS) method; 80% of patients (12/15 evaluable) had biopsies positive for PD-L1 CPS expression in injected lesions (≥1% IHC) and 79% of patients (11/14 evaluable) in noninjected lesions. Baseline CD8⁺ T-cell infiltration was seen in 39% (15/38 evaluable patients) of injected lesions (≥1% staining by IHC) and 43% (15/35) of noninjected lesions (Fig. 3). Where patients had baseline biopsies available for both injected and noninjected lesions, there was good agreement between injected and noninjected lesions: 88% of patients (29/33) had overlapping results for PD-L1 TPS expression (≥1% vs. <1%), and 81% of patients (25/31) had overlapping results for CD8 positivity (≥1% vs. <1%). There was no evidence of consistent increase in PD-L1 TPS or CD8 positivity on treatment, in injected or noninjected lesions (Fig. 3). There was no correlation between response to MIW815 treatment in target lesions and baseline, or on-treatment changes in PD-L1 or CD8 expression (Fig. 3); this was also true for response to treatment seen in injected lesions. Similarly, baseline or on-treatment expression of CD68 (CD68⁺ macrophage infiltration) and FOXP3 [FOXP3/cluster of differentiation 3 (CD3)⁺ T-cell infiltration] were not associated with response (Fig. 3).

The patient who had a PR to treatment had an increase in CD8 positivity on treatment in the injected lesion (0.4% to 6.3%) and no change in CD8 positivity in noninjected lesions (4.2% to 5.5%; Supplementary Fig. S3A). No PD-L1 tumor cell expression was detected at baseline or on treatment; an increase in PD-L1 CPS was seen on treatment in the injected lesion (10% to 50%) and a decrease seen in noninjected lesions (35% to 25%). An increase was also seen in CD68 positivity on treatment in the injected (2.0% to 7.0%) and noninjected lesions (5.4% to 10.9%). For FOXP3, levels were low at baseline and on treatment—injected lesion (0.1% to 0.7%), noninjected lesion (0.6% to 0.5%). Of the two patients with unconfirmed PRs, one did not have biomarker data available, and one did not have detectable PD-L1 or CD8. The patient with collecting duct carcinoma and prolonged stable disease did have biomarker data available, and increases in CD8 positivity, PD-L1 tumor cell expression, CD68, and FOXP3 staining were observed in the injected lesion following MIW815 treatment (Supplementary Fig. S3B).

RNA expression analysis of pre- and posttreatment tumor biopsies revealed limited changes, but did show a nonsignificant increase in natural killer cell lineage markers in some patients (Supplementary Fig. S4), consistent with the mechanism of action of MIW815 and immunotherapy-mediated antitumor effects (gene lists shown in Supplementary Table S1; refs. 20, 21).

Plasma cytokines, including IFNβ, MIP-1β, IL6, MCP-1, and CXCL10 were increased at 6 hours postinjection compared with predose levels (Fig. 4), consistent with the mechanism of action of MIW815. Plasma levels of IFNβ showed dose-dependent increases with increasing MIW815 plasma concentrations (Fig. 4). Systemic T-cell clonal expansion was measured as a proxy for antigen-specificity; systemic clonal expansion of T cells was observed at cycle 1, day 8 and cycle 2, day 1 compared with pretreatment levels (P = 0.02; Fig. 5). Individual patient data are shown in Supplementary Fig. S5.

In this phase I, dose-escalation clinical trial, MIW815 was well tolerated in patients with advanced solid tumors and lymphomas. The most common AEs suspected to be related to treatment were pyrexia, chills, injection-site pain, and headache, consistent with what has been seen for other agents associated with type I IFN agonism (22–24). The incidence of treatment-related AEs was generally low, which may have been related to limited systemic exposure resulting from i.t. administration and rapid PK clearance, although it may also suggest that higher-potency STING agonism could be necessary to achieve therapeutic effects. No MTD was reached, and due to a lack of clear association between dose and STING-associated pharmacodynamics, and between dose and response to treatment, an RDE was not determined.

Clinical response to immune checkpoint blockade has been associated with tumor mutational burden (TMB), neoantigenicity, and the presence of a T-cell–inflamed tumor microenvironment (TME; refs. 9, 10, 25, 26). Despite the lack of a spontaneous antitumor immune response, noninflamed tumors can contain potentially immunogenic antigens, highlighting a role for therapies that stimulate de novo responses in the TME (27). In this context, agonism of innate pattern recognition receptors (PRR), such as STING, has emerged as a mechanistically promising approach to convert tumors from noninflamed to inflamed, for use in combination with immune checkpoint blockade (12, 28).

In our study, pharmacodynamic, immune-monitoring assessments revealed biological activity following treatment with MIW815, including increased expression of proinflammatory cytokines such as IFNβ, and systemic clonal expansion of T cells. With the exception of IFNβ, these effects were not clearly associated with dose and, therefore, did not support selection of a single, biologically active dose. Systemic CXCL10, recently reported to increase in response to treatment with another STING agonist, MK-1454 (22), did not show dose-dependent changes in our study. Assessment of gene expression changes in pre- and posttreatment biopsies revealed a trend towards increased immune activation; however, this was not significant across the trial. Postdose biopsies were taken over 2 weeks after treatment, which may have been beyond the window to see the early pharmacodynamic changes expected based on the mechanism of action for MIW815. It is also possible that MIW815 levels did not sufficiently impact the STING pathway.

Plasma PK analyses demonstrated that MIW815 was rapidly absorbed from the injection site into systemic circulation and had a short half-life. It was recognized during the trial that analyses of injection-site tissue PKs may have provided more relevant information on drug exposure based on the mechanism of action; however, these local i.t. exposure data are not available for correlation with plasma PKs, as they required repeated tissue sampling in patients. Collection of data surrounding drug delivery, such as injection technique, tumor tissue pressure, injection-site drug residence, and the time course of tumor tissue exposure, may be important in the investigation of i.t.-injected small-molecule immune agonists in the future.

The ORR seen for single-tumor injection with MIW815 was lower than expected based on the potent antitumor activity observed in preclinical mouse models (2, 15, 16). This may be explained in part by the fact that this patient population was heavily pretreated and heterogenous, with more than 15 different tumor types. This variation in tumor types could have manifested in inconsistent drug exposure as a consequence of differing i.t. pressure upon injection. Similarly, the primary immunobiology of the TME in distinct tumor types, such as endothelium, fibroblasts, and immune-suppressive myeloid cells, as well as genomic retention of the STING pathway, may have affected the immunomodulatory and clinical outcomes. We also note that in the absence of i.t. PK data, we are unable to confirm that the local injected MIW815 concentration was equivalent to that required to show the activity in preclinical models. Limited systemic exposure of MIW815 was observed, but not correlated with clinical activity or toxicity. In murine models of i.t. STING agonism, systemic PK has not been associated with therapeutic effect and we therefore do not feel that the systemic PK profile strongly informs our results here. To date, the greatest success with i.t. therapies has been observed in melanoma, where a discrete population of patients have disease amenable to injection, and immunobiology associated with high neoantigenicity and a T-cell–inflamed TME (10, 23, 24, 29).

In the context of a single tumor-site injection for patients with advanced solid tumors, preliminary signs of clinical activity were observed, including a partial response in a patient with MCC and unconfirmed partial responses in two patients with parotid cancer and myxofibrosarcoma, two tumor types not known to be responsive to immunotherapy. As most patients had three or more target lesions and only one injected lesion, clinical activity was unlikely to have been driven solely by the shrinkage of injected lesions; the patient with a confirmed response had four target lesions. It is also noteworthy that lesion size was stable or decreased in almost all injected lesions (94%), which compares favorably with what has been seen with i.t. injection of the immunotherapeutic agent, talimogene laherparepvec (TVEC; ref. 29). However, there were important differences between this study and the OPTiM phase III TVEC trial (29). For example, in this trial, patients received MIW815 per irRC guidelines, with only short-term treatment permitted beyond progression, and changing injected lesions was not prioritized, focusing only on a single injected lesion. Preliminary data from i.t. injection and radiation studies suggest that treatment of more than one lesion may enhance systemic immunity and improve clinical outcomes (24, 30).

There is a strong rationale that activation of the STING pathway combined with checkpoint inhibition may trigger a more effective antitumor immune response, given the combined effects on priming via APC activation and IFNβ production, as well as effector function via release of inhibitory feedback mechanisms in the TME associated with IFNγ (13). This hypothesis is supported by preclinical studies and emerging clinical data, which show greater activity with combined treatment compared with STING agonist monotherapy (15, 16, 22).

There are many other STING agonists in various stages of development (11, 12, 31). These include intratumorally or intramuscularly administered agents, intravesical therapy, systemically delivered agents, and an orally bioavailable agent, many of which are being investigated in combination with checkpoint inhibitors (32–37). Whether systemically administered agents will be able to achieve a therapeutic window in humans without inducing immune-related toxicities remains to be seen. There is preclinical evidence to suggest that high-dose or systemic administration of STING agonists may mediate antitumor activity through a more ablative, less durable mechanism than low-dose, local administration (16). To our knowledge, this phase I study represents the first investigation of STING agonism in human patients to treat cancer. Continued investigation into the regulation of the STING pathway, pharmacology of STING agonists, and optimized routes of delivery to the TME will be important for further development of these therapies.

Novartis will not provide access to patient-level data, if there is a reasonable likelihood that individual patients could be reidentified. Phase I studies, by their nature, present a high risk of patient reidentification; therefore, individual patient results for phase I studies cannot be shared. In addition, clinical data, in some cases, have been collected subject to contractual or consent provisions that prohibit transfer to third parties. Such restrictions may preclude granting access under these provisions. Where codevelopment agreements or other legal restrictions prevent companies from sharing particular data, companies will work with qualified requestors to provide summary information where possible.

F. Meric-Bernstam reports grants from Novartis during the conduct of the study; personal fees from AbbVie, Aduro BioTech Inc., Alkermes, AstraZeneca, DebioPharm, eFFECTOR Therapeutics, F. Hoffman-La Roche Ltd., Genentech Inc., IBM Watson, Infinity Pharmaceuticals, Jackson Laboratory, Kolon Life Science, OrigiMed, PACT Pharma, Parexel International, Pfizer Inc., Samsung Bioepis, Seattle Genetics Inc., Tyra Biosciences, Xencor, Zymeworks, Black Diamond, Eisai, Immunomedics, Inflection Biosciences, Karyopharm Therapeutics, Mersana Therapeutics, OnCusp Therapeutics, Puma Biotechnology Inc., Seattle Genetics, Silverback Therapeutics, Spectrum Pharmaceuticals, Tallac Therapeutics, and Zentalis; and grants from Aileron Therapeutics, Inc., AstraZeneca, Bayer Healthcare Pharmaceutical, Calithera Biosciences Inc., Curis Inc., CytomX Therapeutics Inc., Daiichi Sankyo Co. Ltd., Debiopharm International, eFFECTOR Therapeutics, Genentech Inc., Guardant Health Inc., Klus Pharma, Takeda Pharmaceutical, Puma Biotechnology Inc., and Taiho Pharmaceutical Co. outside the submitted work. R.F. Sweis reports grants from Novartis during the conduct of the study; grants and personal fees from Astellas, AstraZeneca, BMS, Eisai, Mirati, Pfizer; grants from CytomX, Eli Lilly and Company, Genentech/Roche, Immunocore, Moderna; personal fees from Janssen, Seattle Genetics; and grants from QED Therapeutitcs outside the submitted work; in addition, R.F. Sweis has a patent for Neoantigens in Cancer pending. F.S. Hodi reports other support from Novartis during the conduct of the study; personal fees from Merck, Novartis, EMD Serono, Aduro, Apricity, Surface, Pionyr, Checkpoint Therapeutics, Surface, Compass, Bicara, Iovance, Genentech, Gossamer, Sanofi, 7 Hills Pharma, Bioentre, Trillium, Catalym, and Immunocore; and personal fees from Amgen outside the submitted work; in addition, F. Hodi has a patent for MICA Related Disorders pending, licensed, and with royalties paid; a patent for Tumor Antigens and Uses Thereof issued; a patent for Angiopoiten-2 Biomarkers Predictive of Anti-Immune Checkpoint Response pending; a patent for Compositions and Methods for Identification, Assessment, Prevention, and Treatment of Melanoma Using PD-L1 Isoforms pending; a patent for Therapeutic Peptides issued; a patent for Vaccine Compositions and Methods for Restoring NKG2D Pathway Function Against Cancers pending, licensed, and with royalties paid; a patent for Antibodies That Bind to MHC Class I Polypeptide-Related Sequence A pending, licensed, and with royalties paid; and a patent for Anti-galectin antibody biomarkers predictive of anti-immune checkpoint and anti-angiogenesis responses pending. W.A. Messersmith reports grants from Novartis during the conduct of the study. R.H.I. Andtbacka reports other support from Novartis during the conduct of the study and personal fees from Seven and Eight Biopharmaceuticals Inc. outside the submitted work. M. Ingham reports grants from Aduro during the conduct of the study; grants and personal fees from Apexigen; personal fees from Xencor, and Daiichi Sankyo; grants from Mirati Therapeutics; and personal fees from Roche/Genetech outside the submitted work. N. Lewis is an employee of Novartis and has stock/stock options. Xinhui Chen reports personal fees from Novartis during the conduct of the study and personal fees from Novartis outside the submitted work. J. Wu is a full-time employee of Novartis and owns stock shares of Novartis. T.W. Dubensky reports other support from Novartis during the conduct of the study, and patents for US 9,549,944, US 9,695,212, and US 9,724,408 issued to Novartis. S.M. McWhirter reports personal fees from and was an employee of Chinook Therapeutics during manuscript preparation; is a current employee of Lycia Therapeutics outside the submitted work; in addition, S.M. McWhirter is an inventor on issued patents US11040053B2, US10906930B2 and published patents WO2020049534A1, WO2018200812A1, and WO2021086889A1. T. Müller reports other support from Aduro Biotech during the conduct of the study and other support from Aduro Biotech outside the submitted work. J.J. Luke reports support from the Data Safety Monitoring Board, AbbVie, Immutep; is part of the Scientific Advisory Board of 7 Hills, Fstar, Inzen, RefleXion, Xilio (stock), Actym, Alphamab Oncology, Arch Oncology, Kanaph, Mavu, Onc.AI, Pyxis, and Tempest; consultancy with compensation from AbbVie, Alnylam, Avillion, Bayer, Bristol Myers Squibb, Checkmate, Codiak, Crown, Day One, Eisai, EMD Serono, Flame, Genentech, Gilead, HotSpot, Kadmon, KSQ, Janssen, Ikena, Immunocore, Incyte, Macrogenics, Merck, Mersana, Nektar, Novartis, Pfizer, Regeneron, Ribon, Rubius, Silicon, Synlogic, Synthekine, TRex, Werewolf, and Xencor; recieves research support from (all to institution for clinical trials unless noted) AbbVie, Agios (IIT), Astellas, AstraZeneca, Bristol Myers Squibb (IIT & industry), Corvus, Day One, EMD Serono, Fstar, Genmab, Ikena, Immatics, Incyte, Kadmon, KAHR, Macrogenics, Merck, Moderna, Nektar, Next Cure, Numab, Pfizer (IIT & industry), Replimmune, Rubius, Scholar Rock, Synlogic, Takeda, Trishula, Tizona, and Xencor; in addition, J.J. Luke has patents (both provisional) for No. 15/612,657 (Cancer Immunotherapy) and PCT/US18/36052 (Microbiome Biomarkers for Anti-PD-1/PD-L1 Responsiveness: Diagnostic, Prognostic and Therapeutic Uses Thereof). No disclosures were reported by the other authors.

F. Meric-Bernstam: Resources, data curation, formal analysis, supervision, funding acquisition, investigation, writing–original draft, writing–review and editing. R.F. Sweis: Conceptualization, data curation, formal analysis, supervision, investigation, methodology, writing–original draft, writing–review and editing. F.S. Hodi: Data curation, formal analysis, supervision, investigation, writing–original draft, writing–review and editing. W.A. Messersmith: Supervision, investigation, methodology, project administration, writing–review and editing. R.H.I. Andtbacka: Investigation, methodology, writing–review and editing. M. Ingham: Investigation, writing–review and editing. N. Lewis: Conceptualization, resources, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. X. Chen: Data curation, formal analysis, investigation, visualization, methodology, writing–review and editing. M. Pelletier: Formal analysis, investigation, visualization, methodology, writing–review and editing. X. Chen: Formal analysis, validation, visualization, methodology, writing–original draft, writing–review and editing. J. Wu: Data curation, formal analysis, validation, visualization, methodology, writing–review and editing. T.W. Dubensky: Conceptualization, funding acquisition, investigation, methodology. S.M. McWhirter: Conceptualization, methodology, writing–original draft, project administration, writing–review and editing. T. Müller: Resources, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. N. Nair: Formal analysis, methodology, writing–review and editing. J.J. Luke: Investigation, visualization, writing–original draft, writing–review and editing.

We thank the patients and their caregivers who participated in this study, as well as the research teams including Krystle Luna (MD Anderson Cancer Center). The authors would also like to thank Sinead Dolan, Yan Ji, Saero Park, Craig Talluto, Damian Trujillo, and Anthony Desbien for their valuable contributions to this study. Medical writing assistance was provided by Laura Hilditch, PhD, and Jenny Winstanley, PhD of Articulate Science Ltd., funded by Novartis Pharmaceuticals Corporation.

The study was funded by Novartis Pharmaceuticals Corporation and Aduro Biotech and partially funded by an NIH/NCATS Center for Clinical and Translational Sciences award (12713238) and NCI MD Anderson Cancer Center Support Grant (P30CA016672).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.