A Phase 2 Trial of Enhancing Immune Checkpoint Blockade by Stereotactic Radiation and In Situ Virus Gene Therapy in Metastatic Triple-Negative Breast Cancer

Triple-negative breast cancer (TNBC) is an aggressive, difficult-to-treat disease. The median overall survival (OS) of locally advanced unresectable or metastatic TNBC (mTNBC) is 8 to 13 months (1, 2). Compared with other breast cancer types, TNBC has higher expression levels of programmed death-ligand 1 (PD-L1; ref. 3) that makes immune checkpoint blockade (ICB) a potentially appealing treatment option. When checkpoint inhibitors as monotherapy were studied in mTNBC, durable responses were seen in those who responded. However, only 5.2% to 25.9% of patients benefited from monotherapy treatment and the responses were seen mostly in patients with high PD-L1 expression in earlier lines of treatment (4–7). Thus, there is an unmet need of treatment strategies to enhance the antitumor activity of ICB.

Radiotherapy has traditionally been applied for local tumor control, based on its cytocidal activity through direct or indirect DNA damage. Recent studies have shown that radiotherapy can modify the tumor microenvironment to induce a systemic antitumor immune response through proinflammatory cytokines and engagement of the innate and adaptive immune systems (8, 9) leading to abscopal effect and response to ICB in distant non-irradiated sites (10, 11). Abscopal effect was more often observed when fractionated instead of a single-dose radiotherapy was given (12, 13). The advent of stereotactic body radiotherapy (SBRT) allows for hypofractionated treatment with precise delivery of high radiotherapy doses, potentially achieving an ideal balance of tumor ablation and immune activation (14).

Even though abscopal effect has been reported, the overall occurrence rate is low (15). Different strategies have been applied to enhance abscopal effect. Adenovirus-mediated expression of herpes-simplex-virus thymidine-kinase (ADV/HSV-tk) can catalyze the phosphorylation of acyclovir to a toxic form capable of inhibiting DNA synthesis and resulting in selective cytotoxicity to cells expressing HSV-tk (16). ADV/HSV-tk gene transduction followed by acyclovir prodrug ganciclovir or valacyclovir therapy has shown both local and systemic antitumor activity in several cancer models (17–19). It was found to induce a pronounced intratumoral infiltrate of macrophages, CD4⁺ and CD8⁺ T cells, and a cytokine profile similar to Th1 immune response (20). Combing ADV/HSV-tk plus ganciclovir gene therapy and radiotherapy appeared to result in increased CD4⁺ T-cell tumor infiltration, additive killing activities, and enhanced systemic antitumor activity in both prostate cancer and mammary tumor models (21, 22).

Studies have shown the antitumor effect of ICB can be enhanced by both radiotherapy and ADV/HSV-tk plus ganciclovir gene therapy. Combining radiotherapy with ICB demonstrated enhanced systemic CD8⁺ T-cell–dependent antitumor activity in several mouse cancer models (23, 24), and improved responses in solid tumors (25–27). In a study evaluating the combination of ADV/HSV-tk plus ganciclovir gene therapy and PD-L1 checkpoint inhibitor in glioblastoma, the combination therapy upregulated PD-L1 expression and improved long-term survival (28).

Here, we explored the approach of using SBRT and ADV/HSV-tk plus valacyclovir gene therapy before pembrolizumab to enhance pembrolizumab efficacy in patients with mTNBC.

Patients with mTNBC who had relapsed on or were refractory to standard-of-care therapy were eligible for the study. Eligible participants were at least 18 years old, had histologically confirmed TNBC or low estrogen/progesterone receptor breast cancer (estrogen/progesterone receptor <10%, and HER2 negative), an Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2, measurable disease based on RECIST v1.1, a target lesion of at least 1 cm in diameter for SBRT, and a measurable distant metastasis apart from the irradiated site of at least 1 cm to evaluate response. Patients with previously treated brain metastases were eligible for the study provided they were stable (without evidence of progression by imaging for at least 4 weeks before the first dose of trial treatment and any neurologic symptoms have returned to baseline), have no evidence of carcinomatous meningitis, or new/enlarging brain metastases, and have not used steroids for at least 7 days before trial treatment. Other required inclusion criteria included a willingness and ability to provide informed consent for the trial, a life expectancy of at least 6 months, and adequate organ functions.

Key exclusion criteria were bone metastasis–only disease; prior treatment with immunotherapy, including anti–PD-1, anti–PD-L1, or anti–PD-L2 agents; prior treatment with gene vector therapy; immunodeficiency or receipt of any form of systemic immunosuppressive therapy (including glucocorticoids) within 7 days before the first dose of trial treatment; active autoimmune disease that required systemic treatment for the preceding 2 years; supplemental oxygen dependence; carcinomatous meningitis; known active hepatitis B or hepatitis C infection; or major surgery within 4 weeks before study enrollment.

Study design and subject consent forms were reviewed and approved by the hospital institutional review board. All patients provided written informed consent. The study was sponsored by Merck & Co., Inc. and was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Guidelines for Good Clinical Practice. Adverse events (AE) were recorded regularly according to the protocol and were subject to mandatory reporting by the investigators. Any serious AEs were reported to Merck within 24 hours.

STOMP was designed as a single-center Phase 2 study evaluating the approach of enhancing the abscopal effect of SBRT with ADV/HSV-tk plus valacyclovir gene therapy in augmenting the efficacy of pembrolizumab monotherapy in patients with mTNBC. ADV/HSV-tk was created by using a backbone Ad5-dl309 with a Rous sarcoma virus promoter that drives the herpes simplex virus thymidine kinase gene insert (Center for Gene and Cell Therapy, Houston Methodist Hospital) as previously reported (29). ADV/HSV-tk [5 × 10¹¹ virus particles (vp)] in a 2-mL total volume was injected intratumorally to a single target lesion per patient on day 0 of the study using imaging guidance along with fiducial placement for SBRT. Valacyclovir was administered orally at a dose of 2 g three times a day for 14 days from days 1 to 15. The injected target lesion was radiated. Metastatic target lesions were treated with SBRT (30 Gy; 6 Gy × 5 fractions) and primary breast/chest wall sites were treated with SBRT, IMRT, or 3DCRT (30 Gy; 6 Gy × 5 fractions, BED10 = 48 Gy or 33.6 Gy; 4.2 Gy × 8 fractions, BED10 = 47.71 Gy) at the treating physician's discretion. SBRT was administered over 2 weeks from days 2 to 16 of the study. Pembrolizumab at a dose of 200 mg was administered intravenously every 21 days after radiotherapy. Pembrolizumab monotherapy was continued until disease progression, development of unacceptable side effects, withdrawal of consent, or up to 24 months (35 cycles) in patients without disease progression.

Peripheral blood for immune correlative analysis was collected on days 1 (baseline), 17, and 38 and two panels of Cytometry by Time of Flight (CyTOF) on paired peripheral blood mononuclear cells were performed. One panel of 35 myeloid markers (M-panel, Supplementary Table S1) and a second panel of 35 T-cell markers (T-panel, Supplementary Table S2) were used on each sample. Barnes-Hut tSNE implementation in Rtsne package was used to plot tSEN. Data were 99th percentile normalized before the analysis, and we used the default tSNE parameters (initial dimensions, 110; perplexity, 30; and theta, 0.5).

In addition, paired core biopsies were performed on days 1 (baseline) and 17. Imaging mass cytometry (IMC) was performed after staining of these patient samples with a 35-marker panel of metal-tagged antibodies and ablation by Hyperion (Fluidigm Inc.; Supplementary Table S3).

PD-L1 expression was measured as the percentage of tumor area involved by PD-L1 positive tumor infiltrating immune cells as assessed by IHC using the Ventana SP142 antibody testing that was validated in the IHC section of our institutional pathology laboratory. PD-L1 expression was also measured as a combined positive score (CPS) using the PD-L1 IHC 22C3 PharmDx kit (Neogenomics performed the staining and results were interpreted by our institutional pathologists) or E1L3N antibody (Cell Signaling Technology, our laboratory performed the staining and results were interpreted by outside pathologists). CPS was defined as the ratio of PD-L1–positive cells out of the total number of tumor cells × 100.

Tumor assessments in non-irradiated sites by CT scans or MRI were performed at screening, every 8 weeks thereafter until completion of the protocol-specified study treatment and/or at the discretion of the treating physician, when clinically indicated, and 30 days after the last dose of pembrolizumab. Tumor response was assessed in these non-irradiated metastatic lesions by two independent radiologists according to RECIST v1.1. AEs were evaluated throughout the trial until 30 days after the end of treatment and were graded according to the National Cancer Institute Common Terminology Criteria for AEs (NCI CTCAE), version 4.03.

The primary end point was clinical benefit rate [CBR; the proportion of patients with complete response (CR), partial response (PR), or stable disease (SD) for ≥24 weeks]. Secondary end points included OS (the time from intratumoral viral injection to death or last date of contact), duration on treatment (DoT), and safety. Exploratory objectives included immune response to treatment, correlative tissue, and blood-based biomarkers that included PD-1 and PD-L1 expression, immune infiltrates, and cytokine expression (IL1, IL2, IL6, IL12, IFNc, TNFα, and GM-CSF). Efficacy and safety were assessed in all patients who received intratumoral ADV/HSV-tk injection.

With the null and alternate hypotheses of 19% CBR and 39% CBR, respectively, an enrollment of 28 patients with TNBC was planned for the target response rate of 39%. Power study revealed that with a sample size of 28 patients if the observed number of responders is 11 (39.3%), the 95% confidence interval (CI) estimate will have a margin of error of 18.1% points (i.e., 21.2% to 57.4%) based on the normal approximation. The Kaplan–Meier method was used to obtain estimates of 95% CIs for mean and median survival times. Pearson correlation analysis was used to assess the association of response with lactate dehydrogenase (LDH), disease-free interval (DFI), PD-L1 expression assessed by Ventana SP142 antibody testing, and CPS as assessed by IHC using the Dako 22C3 PharmDx kit.

The data generated in this study are stored in a secured drive and available upon request from the corresponding author.

From June 7, 2017 to May 7, 2020, a total of 39 patients were screened and 28 patients with histologically confirmed mTNBC were enrolled (Table 1). Median age was 54 years (range, 34 to 78). The majority of patients were heavily pretreated, with a median of two (range, 0–7) previous lines of cytotoxic treatment in the metastatic setting, and 8 patients (25.0%) had received ≥3 prior metastatic lines of chemotherapy. The median duration of response from the last line of treatment was 2 months (range, 1 to 7). Of note, 3 patients (10.7%) had treated brain metastasis at the time of enrollment. The majority of patients were PD-L1 negative; 18 (64.3%) patients had <1% PD-L1 expression by Ventana SP142 antibody testing. Four patients (14.3%) had CPS combined score ≥10 by PD-L1 IHC 22C3 PharmDx kit testing.

At the time of data cutoff on October 22, 2021, patients had received a median of three pembrolizumab doses (range, 0 to 23). Median duration of follow-up was 8.3 months (95% CI, 3.0–10.1 months). One patient (3.6%) remained on treatment at the data cutoff date. Disease progression was the most common reason for treatment discontinuation (89.3%).

Of the 28 patients enrolled, 6 patients (21.4%) had clinical benefit, including 2 with CR (7.1%), 1 with PR (3.6%) and 3 with SD (10.7%; Table 2). The median OS was 6.6 months (95% CI, 3.0 to 10.1 months) in the overall population, and 14.7 months (95% CI, 6.5 to 47.5 months) in the patients who had the clinical benefit (Fig. 1A). The median DoT was 2.2 months (95% CI, 1.4 to 2.9 months) in the overall population, and 9.6 months (95% CI, 4.8 to 23 months) in the patients who had clinical benefit (Fig. 1B). In the overall population, 57.1% were alive at 6 months and 19.5% at 12 months; whereas in the responders, 100% were alive at 6 months and 66.7% at 12 months (Fig. 1C).

Two patients who had clinical response were alive at the data cutoff date and 1 patient remained on treatment (Table 3). Patient 2 had biopsy-proven recurrent metastatic disease to the lungs and liver, with a 1.2-cm liver lesion diagnosed on PET scan that decreased in size on treatment. She then achieved CR but discontinued pembrolizumab after 18 doses due to Grade 3 pneumonitis. She remains in CR and has not received any systemic treatment for over 47 months. Patient 5 had recurrent biopsy-proven metastatic disease in the liver and had SD on treatment for over 22 months.

We explored the potential predictive values of LDH, DFI, and CPS. Clinical benefit was moderately associated with PD-L1 expression (r = 0.46), but not with LDH, DFI, or CPS.

Treatment-related AEs of any cause were reported in 22 patients (78.6%; Table 4). The most common AEs were any-grade constitutional (35.7%), hypothyroidism (21.4%), diarrhea (14.3%), nausea (10.7%), dyspnea on exertion (10.7%), alkaline phosphatase elevation (10.7%), and hyperthyroidism (10.7%). Grade 3 or higher treatment-related AEs were reported in 10 patients (35.7%). Immune-related AEs at any grade occurred in 13 patients (46.4%). The most common immune-related AEs were hypothyroidism (21.4%) and hyperthyroidism (10.7%). Grade 3 or higher immune-related AEs were reported in 2 patients (7.1%), including 1 pneumonitis (3.6%), and 1 transaminitis (3.6%). Two patients (7.1%) discontinued treatment due to Grade 3 pneumonitis and Grade 2 transaminitis. There were no treatment-related deaths.

To investigate the association between response and the immune populations in peripheral blood, CyTOF was performed on 25 patients (5 responders and 20 non-responders) on days 1 and 17 and on 20 patients (4 responders and 16 non-responders) on day 38. Cell populations were clustered, identified (Supplementary Table S4), and plotted on tSNE plots (Fig. 2A). The log₂ fold change between Days 17 or 38 and day 1 of responders and non-responders was calculated and illustrated as heatmaps (Fig. 2B).

When the cell populations were compared between responders and non-responders at the same time points (Fig. 2C), significantly more T cells [effector memory (Tem) CD4 T cells and Tem CD8 cells, unconventional T cells, and natural killer T (NKT) cells] were seen in the non-responders on day 1. At day 17, these T-cell populations increased in the responders (Fig. 2B and C), with Tem CD8 and the activated CD8 T-cell (PD-1⁺ CD8) showing the most profound increase. In the non-responders, a statistically significant increase of myeloid cells was observed in 14/20 (70%) patients. In those patients who had increased myeloid cells, 12/14 patients (85.7%) had a decrease of T cells. On day 38, there was no continuous increase of CD8 T cells in responders except for 1 patient (Fig. 2B). In comparison, continuous increase of myeloid cells was seen in the non-responders (10/16, 62.5%).

The tumor microenvironment was further assessed on paired tumor biopsies by IMC. A detailed description of IMC was previously described (30). In responders, tumors contained myofibroblasts (SMA⁺) on day 1 and became infiltrated with immune cells on day 17 after treatment. In contrast, non-responder tumor cells were cohesive with few myofibroblasts present on day 1 and limited immune cell infiltration was observed on day 17 after treatment (Supplementary Fig. S1A). In addition, more CD4 and CD8 T cells were found in responders with few Ki67⁺ tumor cells, whereas more CD68⁺ macrophages were found in the non-responders with more Ki67⁺ tumor cells. The cell populations from 3 responders and 14 non-responders were then clustered and identified by unsupervised algorithms (Supplementary Fig. S1B) and the log₂ fold change between days 1 and 17 was plotted in a heatmap (Supplementary Fig. S1C). Consistent with CyTOF findings, there was a significant increase of M2-like monocytes (CD163⁺CD14⁺) in non-responders (P = 0.0012).

We then performed neighborhood analysis to assess the spatial relationship between the identified cell populations (Supplementary Fig. S2). Comparing before and after treatment in responders, the epithelial cells or cancer cells (cluster 3) were separated from most immune populations (cluster 17–24) at baseline, but became surrounded by these immune cells after treatment; CD8 T cells (cluster 23) were surrounded by immunosuppressive CD15⁺ neutrophils (cluster 27) at baseline, but were freed from these neutrophils after treatment. Comparing responders and non-responders, the non-responders already had some immune populations surrounding epithelial cells or cancer cells (cluster 3) at baseline, which is consistent with our previous observation that non-responders had more CD4 T cells at baseline (Fig. 1C). These results demonstrate distinct spatial patterns of tumor microenvironment and dynamic change with treatment in responders and non-responders.

In this Phase 2 study to evaluate the efficacy of enhancing pembrolizumab monotherapy by engaging innate and adaptive immunity through SBRT and intratumoral ADV/HSV-tk gene therapy in patients with mTNBC who had relapsed on or were refractory to standard-of-care therapy, the treatment demonstrated an acceptable safety profile. The incidence of grade 3 or higher treatment-related AEs was 35.7% and immune-related AEs was 7.1%. The two grade 3 or higher immune-related AEs were managed with glucocorticoids and discontinuation of treatment. There were no treatment-related deaths. The safety profile of this trial was consistent with that seen in previous studies (5–7, 31).

ICB monotherapy in mTNBC yields low ORR of 7.6% to 10% in second line and beyond metastatic setting (5, 6), 2.6% to 5% in patients with low or negative PD-L1 (6, 7), and nearly no response in patients with liver metastases. Primary and acquired immune resistance involving antigen-presenting machinery and T-cell dysfunction play important roles in the low response rate to ICB in mTNBC (32). Both radiotherapy and ADV/HSV-tk plus ganciclovir gene therapy can induce tumor antigen release, increase T-cell infiltration and activation and PD-L1 expression (23, 24, 28); adding ICB to either modality can potentially overcome PD-L1 inhibitory signals and enhance antitumor response. Indeed, many animal studies of combining radiotherapy or ADV/HSV-tk plus ganciclovir gene therapy with ICB have shown enhanced antitumor effect (12, 23, 24). However, combining either SBRT or virus therapy with ICB only yielded moderate response in early-phase clinical trials. In TONIC trial that explored different immune induction strategies with chemotherapy agents or radiotherapy in mTNBC, the radiotherapy arm had an ORR of only 8% (33). In contrast with an animal study showing radiotherapy can induce T-cell infiltration and activation (25), biopsy after treatment did not show change of tumor-infiltrating lymphocytes or CD8 cells in TONIC trial. Even though two Phase 2 studies combining SBRT with ICB in treated mTNBC (34, 35) showed an ORR of 17.6% to 22%, the median duration of response was only 2.5 to 5 months and no response was observed in patients with liver metastases. Similarly, oncolytic virus type I herpes-simplex-virus Talimogene laherparepvec (TVEC, IMLYGIC) in combination with ICB in mTNBC yielded only one PR in a Phase 1 study (36).

Combining both SBRT and ADV/HSV-tk plus valacyclovir gene therapy with ICB may result in complementary mechanisms of DNA damage, alter and potentiate tumor-specific antigenicity, and elicit a stronger immune response (22, 37). By combining three modalities, a CBR of 21.4% (2 CR, 1 PR, and 3 SD), a median DoT of 9.6 months and two long-term responders were observed. Although cross-trial comparisons could be confounded by different trial designs, patient populations, and other factors, trimodality combination strategy improved ORR and DoT numerically compared with SBRT and ICB combination strategy (34, 35). Response observed was not associated with the line of treatment delivered in metastatic setting or PD-L1% or CPS. Among patients who had received chemotherapy in metastatic setting, 5/20 (25%) responded to the combination treatment, and 2/18 (11.1%) patients with negative PD-L1 had clinical benefit. Out of the 6 responders, 3 had CPS ≥10 and 2 patients had CPS <10, consistent with the observation that CPS is not predictive of response beyond first-line metastatic setting (38). Taken together, these data suggest the enhanced antitumor effect of SBRT, ADV/HSV-tk plus valacyclovir gene therapy and ICB combination therapy in pretreated patients and in patients with low or negative PD-L1.

We observed some atypical patterns of response in our patients, in particular in patients with liver metastases. Of the 6 responders in our study, 2 patients had liver metastases (Table 3). Patient 2 had biopsy-proven lung metastases and one liver lesion on PET scan that decreased in size with treatment. She then achieved CR but discontinued pembrolizumab after 18 doses due to Grade 3 pneumonitis. She remains in complete remission without any systemic treatment for over 47 months. Patient 5 had biopsy-proven relapsed metastatic disease in the liver and had remained on treatment for over 22 months with stable disease. Visceral metastases, especially liver metastases, have been a poor prognostic factor in mTNBC (39, 40). Liver metastases have the lowest PD-L1 level among metastatic sites (41). Historically, ICB monotherapy, and ICB and SBRT combination therapy had very low response rates in those patients (5, 34, 36). In contrast, previous studies of ADV/HSV-tk plus valacyclovir gene therapy demonstrated significant regression of liver metastases in breast cancer and lung cancer mouse models (42, 43). In the Phase 1 study of combining oncolytic type I herpes simplex virus TVEC (IMLYGIC) with ICB in patients with mTNBC and liver metastases, 1 out of 7 patients had a PR (36). The durable responses in these 2 patients in our study further suggest the role of ADV/HSV-tk plus valacyclovir gene therapy in re-priming T cells and potential added synergistic effect of ADV/HSV-tk plus valacyclovir gene therapy and SBRT in enhancing ICB in these with liver metastases, an observation that warrants further investigation.

In non-responders, immune profile analysis revealed high baseline T cells; however, there was no increase of these T cells after treatment suggesting that these T cells may be tolerant to the tumor and are not proliferating. In contrast, in responders, the increase of T cells suggests that these T cells are newly proliferating in response to the treatment and can antigen-specifically target tumor cells. In addition, neighborhood analysis suggested that these immune cells were actively engaged with cancer cells for tumor killing. Myeloid cells in the tumor can exhibit suppressive activity on T cells. In responders, CD8 T cells were distant from CD15⁺ neutrophils after treatment, indicating that the immune populations were freed from immunosuppressive neutrophils. In contrast, myeloid cells increased in non-responders after treatment. These findings further support that myeloid cells play an important role in immunosuppression and treatment response. Six non-responders exhibited immune profiles similar to that of responders with increased Tem CD8 cells and PD-L1⁺ CD8 T cells on D17. Among these 6 patients, 3 had increased PD-L1⁺ monocytes at D17 and 2 exhibited delayed increase of myeloid cells on D38. Moderate upregulation of Tregs and exhausted CD4 T cells were also seen in one of these 6 non-responders. These observations are consistent with previously reported mechanisms of ICB resistance (44–46).

There are several limitations to this study: (i) As a single-arm study, patients were not randomized and the exact contribution of single modality to the enhanced antitumor effect is unknown. Given that oncolytic virus alone (47) or in combination with ICB (36) had a low response rate of 0%–12.5%, and ICB and SBRT combination only improved response with short duration of response in pre-treated patients without liver metastases (34), the results from our study suggest that all three treatment modalities contributed to the observed enhanced antitumor effect. However, larger randomized studies are needed to elucidate the exact role of each modality and their additive effects on tumor microenvironment response. (ii) Initial trial design defined PD-L1 expression by IHC using the Ventana SP142 antibody. CPS score as assessed by the PD-L1 IHC 22C3 PharmDx kit was later added as another measurement after FDA approval of pembrolizumab in combination with chemotherapy in untreated mTNBC based on CPS. This led to the unavailability of CPS in almost one third of patients. (iii) Although the correlative biomarker analyses suggest that an increase of CD8 T cells and myeloid cells may serve as useful predictive markers for responders and non-responders, the results need to be interpreted with caution in view of the small sample size of this study. Larger studies are needed to validate these initial findings.

Even though the study did not meet its statistical significance, the durable responses in heavily pretreated patients with mTNBC suggest that the efficacy of ICB monotherapy may be enhanced by harnessing the abscopal effect of SBRT and intratumoral ADV/HSV-tk plus valacyclovir gene therapy, even in those with low PD-L1 level and liver metastases. Early detection of increased effector and effector memory CD8 T cells and myeloid cells correlates with response and non-response, respectively.

P. Niravath reports grants from Merck during the conduct of the study. N. Gupta reports grants and non-financial support from Siemens Healthineers, as well as personal fees and non-financial support from GE Healthcare outside the submitted work. J.A. Mejia reports grants from Merck Research Labs during the conduct of the study, as well as other support from Merck Research Labs outside the submitted work. E.H. Bernicker reports personal fees from Blueprint Medicine, as well as grants from Genentech outside the submitted work. J.C. Chang is the sole inventor on patent application no. 10420838 entitled “Methods for treating cancer using iNOS-inhibitory compositions” held by Houston Methodist Hospital. No disclosures were reported by the other authors.

K. Sun: Conceptualization, resources, data curation, supervision, methodology, writing–original draft, writing–review and editing. Y. Xu: Resources, data curation, software, visualization, methodology, writing–original draft, writing–review and editing. L. Zhang: Resources, data curation, software, visualization, methodology, writing–review and editing. P. Niravath: Resources, investigation, project administration, writing–review and editing. J. Darcourt: Resources, investigation, project administration, writing–review and editing. T. Patel: Resources, investigation, project administration, writing–review and editing. B.S. Teh: Resources, investigation, project administration, writing–review and editing. A.M. Farach: Resources, investigation, project administration, writing–review and editing. C. Guerrero: Resources, methodology, writing–review and editing. S. Mathur: Resources, data curation, formal analysis, writing–review and editing. M.A. Sultenfuss: Resources, investigation, project administration, writing–review and editing. N. Gupta: Resources, investigation, project administration, writing–review and editing. M.R. Schwartz: Resources, investigation, project administration, writing–review and editing. S.L. Haley: Resources, investigation, project administration, writing–review and editing. S. Nair: Resources, investigation, project administration, writing–review and editing. X. Li: Resources, project administration, writing–review and editing. T.T.A. Nguyen: Resources, project administration, writing–review and editing. J.D. Butner: Data curation, formal analysis, writing–review and editing. J. Ensor: Resources, formal analysis, project administration, writing–review and editing. J.A. Mejia: Resources, supervision, funding acquisition, investigation, visualization, project administration, writing–review and editing. Z. Mei: Resources. E.B. Butler: Conceptualization, resources, investigation, project administration, writing–review and editing. S.-h. Chen: Conceptualization, resources, data curation, software, visualization, methodology, project administration, writing–review and editing. E.H. Bernicker: Conceptualization, resources, data curation, supervision, funding acquisition, validation, investigation, methodology, project administration, writing–review and editing. J.C. Chang: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing.

The clinical trial was supported by Merck Sharp & Dohme Corp. The authors would like to acknowledge Dr. Mary and Ron Neal for their generous financial support; Drs. H. Heslop and M. Brenner for their input in this research.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.