Phase I First-in-Human Study of TRK-950, an IgG1 Antibody Specific to CAPRIN-1, in Patients with Advanced Solid Tumors Open Access

Cytoplasmic activation/proliferation-associated protein-1 (CAPRIN-1) is a highly conserved cytoplasmic phosphoprotein associated with cell proliferation in vertebrates (1, 2). CAPRIN-1 expression decreases when cells cease to divide and increases in all phases of the cell cycle. The lack of CAPRIN-1 in cells can delay progression from the G₁ phase to the S-phase of the cell cycle (2). Previous reports have shown that CAPRIN-1 can selectively bind to mRNA encoding c-Myc and cyclin D2, which are involved in cell proliferation, and can also directly bind to Ras GTPase–activating protein-binding protein 1 to promote mammalian stress particle formation. Overexpression of CAPRIN-1 can lead to an overall inhibition of stress granule formation in cells (3–11). Several studies have also demonstrated that the intracellular expression of CAPRIN-1 is positively correlated with cancer progression and poor prognosis, suggesting that CAPRIN-1 could be a diagnostic biomarker and a suitable cancer therapeutic target (12–19).

We previously demonstrated that CAPRIN-1 is strongly expressed on the cell membrane surface in or on most solid cancers but not in or on normal tissues and that CAPRIN-1 is present in the cytoplasm in both cancer and normal tissues, which is consistent with previous reports. In cancer cells, part of the total CAPRIN-1 is exposed on the cell membrane surface, a phenomenon that is not observed in noncancerous cells. In addition, we found that cancer stem cells derived from patients with pancreatic cancer also strongly expressed CAPRIN-1 on their cell surface and exhibited enhanced in vitro colony-forming activity and in vivo tumorigenicity. Based on these prospective data, we concluded that CAPRIN-1 is a promising target for antibody drugs in solid cancers (20).

TRK-950 is a humanized monoclonal IgG subclass 1 (IgG1) antibody specific to human CAPRIN-1 that is under development for the treatment of patients with CAPRIN-1–expressing solid malignant tumors. TRK-950 binds to human CAPRIN-1 with high affinity and can strongly induce tumor cell death through fragment crystallizable-mediated effector functions, including antibody-dependent cell-mediated cytotoxicity and antibody-dependent cell-mediated phagocytosis (ADCP). Consistent with the antibody-dependent antitumor activity, TRK-950 significantly inhibited tumor growth in mouse xenograft models transplanted with several human cancer cells. Furthermore, a comprehensive series of toxicology studies demonstrated that TRK-950 had a favorable safety profile in cynomolgus monkeys at doses up to 200 mg/kg. The cynomolgus monkeys’s no-observed-adverse-effect level of 200 mg/kg provided a predicted safety margin 67-fold above the clinical starting dose in the initial clinical trial (20).

Based on these promising results in nonclinical studies, we conducted a phase I study to establish a safe dose, determine the pharmacokinetics (PK), and document the preliminary antitumor activity of TRK-950 given intravenously in patients with progressive, locally advanced, or metastatic solid tumors.

This first-in-human phase I study (NCT02990481) was an open-label, multicenter, dose-escalation, nonrandomized study of TRK-950 in patients with progressive locally advanced or metastatic solid tumors. The primary objective of this study was to determine the safety, tolerability, and maximum tolerated dose (MTD) of TRK-950. The secondary objectives were to determine the PK of TRK-950 and observe the preliminary antitumor activity of TRK-950. The exploratory objectives were to identify potential biomarkers that may help predict the response to TRK-950 and to establish the dose of TRK-950 recommended for future phase II studies.

For the dose escalation cohort, a traditional 3 + 3 design was used to evaluate dose-limiting toxicity (DLT) and to identify the MTD. The starting dose was 3 mg/kg administered intravenously over 60 minutes weekly for 3 weeks, followed by a 1-week rest period. Depending on the safety findings, dose escalations could proceed up to 30 mg/kg weekly for 3 weeks, followed by a 1-week rest period. Patients were administered TRK-950 intravenously on days 1, 8, and 15 of a 28-day cycle in the dose escalation cohort. After enrollment in the dose escalation was completed, 14 additional patients who were enrolled in the colorectal cancer cohort were administered TRK-950 intravenously weekly (10 mg/kg on days 1, 8, 15, and 22) or biweekly (30 mg/kg on days 1 and 15) in a 28-day cycle to determine the optimal dose level and schedule and to confirm safety. In parallel, 12 patients were enrolled in the cholangiocarcinoma cohort, in which patients were administered TRK-950 intravenously weekly (10 mg/kg on days 1, 8, 15, and 22) in a 28-day cycle to assess antitumor activity. Patients who successfully completed treatment cycle 1 (28 days) without evidence of significant treatment-related toxicities or clinical evidence of progressive disease continued to receive the treatment. Treatment could continue as long as there was a perceived benefit or until disease progression. The study protocol and other relevant documents were approved by the relevant Institutional Review Boards or ethics committees, and all participants provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki and International Conference on Harmonization Guidelines for Good Clinical Practice.

Patients eligible for the dose escalation cohort were male or female patients ≥18 years of age with histologically confirmed locally advanced or metastatic solid carcinomas who were likely to overexpress CAPRIN-1 and had already received or been considered for all therapies known to confer clinical benefit, with tumor progression after receiving standard/approved chemotherapy or in which there was no approved therapy or that were not amenable to curative treatment. Patients were to have measurable disease per RECIST version 1.1 (primary or metastases), a Karnofsky performance status ≥ 70%, and a life expectancy of at least 3 months. Patients eligible for the colorectal cancer cohort (10 and 30 mg/kg) were histologically confirmed to have locally advanced or metastatic colorectal cancer. The patients eligible for the cholangiocarcinoma cohort (10 mg/kg) had histologically confirmed cholangiocarcinoma. Key exclusion criteria included New York Heart Association Class III or IV; cardiac disease; myocardial infarction within the past 6 months; unstable arrhythmia; evidence of ischemia on ECG; uncontrolled bacterial, viral, or fungal infections; known active infection with human immunodeficiency virus; hepatitis B; hepatitis C; symptomatic brain metastases; serious nonmalignant disease; and treatment with radiotherapy, surgery, chemotherapy, immunotherapy, or investigational therapy within 4 weeks prior to study entry. The representativeness of the study population is included in Supplementary Table S1.

Safety assessments performed during the study included monitoring of adverse events (AE), physical examination, vital signs, 12-lead electrocardiogram, and laboratory tests (hematology, serum chemistry, and urinalysis). Safety data were collected from the date of written informed consent until 28 days after the administration of the last dose of the study drug. All AEs that occurred after the first dose of the study drug were considered treatment-emergent AEs (TEAE). All AEs were classified using the MedDRA version 19.1 classification system. The severity of toxicity was graded according to the NCI Common Terminology Criteria for Adverse Events version 4.03. A DLT was defined as one of the following toxicities: grade 4 neutropenia lasting ≥5 days, grade 3 or 4 neutropenia with fever and/or infection, grade 4 thrombocytopenia (or grade 3 with bleeding), grade 3 or 4 treatment-related nonhematologic toxicity, and a dosing delay greater than 2 weeks due to TEAEs or related severe laboratory abnormalities. The MTD was the highest dose level that did not meet the toxic dose level definition. The toxic dose level was defined as the lowest dose level at which a DLT was experienced in ≥2 out of a maximum of six patients.

Serum was collected for PK analysis as described in Supplementary Table S2. The serum concentrations of TRK-950 were measured using a validated immunoassay method. The serum concentrations of TRK-950 from the first dose in cycle 1 were analyzed by noncompartmental analysis using Phoenix WinNonlin (version 8.3, Certara, 2020) to determine the following PK parameters: time to maximum serum concentration, maximum observed serum concentration (C_max), area under the serum concentration–time curve (AUC) from time zero to the time of the last quantifiable concentration (AUC_last), AUC extrapolated from time zero to infinity (AUC_inf), terminal elimination half-life, total body clearance, and apparent steady-state volume of distribution.

Dose proportionality was assessed within the dose escalation cohort using plots of mean C_max, mean AUC_last, and mean AUC_inf by dose and an empirical model relating the log of the PK parameter linearly to the log of the dose (the “power model”).

Tumor response was determined according to RECIST 1.1 criteria using objective measurements of target lesions, assessment of nontarget lesions, and identification of new lesions. Radiographic tumor burden assessment (CT/MRI) was performed at screening to establish a baseline and was reassessed at the end of every other cycle until cycle 6. After cycle 6 was completed, the tumor burden was assessed every three cycles.

CAPRIN-1 expression was examined by the IHC method using a rabbit anti-human CAPRIN-1 monoclonal antibody and was measured for all patients in all cohorts at screening using tumor tissue biopsied from the primary tumor or metastatic lesions at a centralized pathologic laboratory. For the dose escalation cohort, if archival tissue was obtained more than 1 year prior to screening, a new biopsy was required to obtain tissue. For the colorectal cancer and cholangiocarcinoma cohorts, if archival tissue was available from a biopsy taken <18 months prior to screening, it was collected. If archival tissue was obtained >18 months prior to screening, tumor tissue for CAPRIN-1 determination was obtained with a pretreatment biopsy. The proportion of cancer cells that positively stained for CAPRIN-1 on the cell membrane was categorized as follows: no reactivity (0), faint/barely perceptible reactivity (1+), weak to moderate reactivity (2+), and strong reactivity (3+). The category was determined by the intensity of membrane staining and the percentage of stained cells, as shown in Supplementary Table S3.

Multi-immunofluorescence analysis of tumor-infiltrating macrophages was performed on formalin-fixed, paraffin-embedded tumor tissue samples obtained from each participant in the colorectal cancer and cholangiocarcinoma cohorts. After antigen retrieval (36 minutes at 95°C, pH 8.4), 4-µm-thick sections of tumor specimens obtained from core-needle biopsies at the screening and C1D22 time points were stained with mouse monoclonal anti-human CD68 (IgG2b, clone 298807, R&D Systems) and mouse monoclonal anti-human CD163 (IgG1, clone 10D6, Leica Biosystems) and then incubated with AF647 goat anti-mouse IgG2b and AF488 goat anti-mouse IgG1 (both from Invitrogen). The sections were then counterstained with DAPI (ref D1306, Invitrogen). Macrophage populations, such as CD68⁺/CD163⁺ cells and CD68⁺/CD163⁻ cells, were manually counted on digital images from all available biopsies, and the data were expressed as the number of cells per mm².

The sample size for the dose escalation cohort was based on a 3 + 3 design, whereas that for the other cohorts was determined without statistical considerations. Descriptive statistics were used to evaluate the safety, PK, and other measurements. The 95% confidence intervals for the disease control rate were calculated using the Pearson–Clopper method. Safety analysis was performed on the population that received at least one dose of the study drug. Analysis of antitumor activity was performed on the population that received at least one dose of the study drug and had at least one post-baseline tumor assessment. All analyses were performed using the SAS statistical software (version 9.4; SAS Inc.).

All clinical trial data presented in the article are not publicly available, as they could compromise patient privacy or consent, but are available upon reasonable request by contacting the corresponding author.

The patient demographics and baseline characteristics are summarized in Table 1, and the CONSORT diagram is shown in Supplementary Fig. S1. The total number of patients enrolled who received at least one dose of the study medication was 36: 10 patients in the dose escalation cohort, 14 patients in the colorectal cancer cohort, and 12 patients in the cholangiocarcinoma cohort. The median age of the patients was 62.5 years (range, 36–78 years). There were a variety of anatomic locations of the individual patients’ primary cancers, with the largest numbers being reported for colorectal cancer and cholangiocarcinoma (15 and 13 patients, respectively). Two patients each had breast, pancreatic, and gastric cancer. Most patients had stage IV disease at the time of study enrollment. All patients had received at least one prior chemotherapy session, and the majority had a history of three or more chemotherapy treatments. Among 14 patients in the colorectal cancer cohort, all were treated with 5-fluorouracil. Additionally, 12 patients received irinotecan, 13 patients received oxaliplatin, 11 patients received bevacizumab, and five patients were treated with an anti-EGFR antibody, including cetuximab. Among 12 patients in the cholangiocarcinoma cohort, all were treated with gemcitabine and cisplatin, and six patients were treated with fluoropyrimidine as a single agent in the adjuvant setting or in combination with irinotecan and/or oxaliplatin. Most patients had either a 3+ or 2+ CAPRIN-1 semiquantitative expression score [15 (41.7%) and 10 (27.8%) patients, respectively].

TEAEs reported at CTCAE all grades in >10% or more patients and at CTCAE grade 3 or higher and treatment-related TEAEs are summarized in Table 2. No DLTs were observed during dose escalation, and no MTD was determined in this study. After the administration of TRK-950, all 36 patients (100.0%) had at least one TEAE, and 10 patients (27.8%) had at least one treatment-related TEAE. The TEAEs that the investigators attributed as at least possibly related to the study drug (n = 10) were generally mild to moderate, with the exception of one event (2.8%: fatigue) that was considered severe (grade 3) in the cholangiocarcinoma cohort. Fifteen (41.7%) patients had at least one serious adverse event (SAE) out of a total of 21 SAEs. One patient (12.5%) who received 30 mg/kg in the colorectal cancer cohort had an SAE of cerebral thrombosis that was considered to be related to the study drug (grade 2). Notably, in the analysis of patients reporting TEAEs, there was no noticeable pattern either overall or by individual preferred term in the frequency of patients reporting TEAEs during the individual dosing days. For each preferred term, the most commonly reported TEAEs (in ≥20% of patients) were abdominal pain (18 patients, 50.0%), followed by fatigue (14 patients, 38.9%), constipation (11 patients, 30.6%), back pain and nausea (10 patients each, 27.8%), and decreased appetite (nine patients, 25.0%). Treatment-related TEAEs (≥2 patients) included fatigue and nausea (two patients, 5.6%).

Among the 30 patients evaluable for tumor burden assessment in any cohort, five (16.7%) patients had a best overall response of stable disease, including patients with colorectal cancer (three patients), cholangiocarcinoma (one patient), and appendiceal cancer (one patient; Table 3). Of the five patients, three had positive CAPRIN-1 expression (3+). The other two patients could not be adequately evaluated due to missing specimens and lack of tumor cells (Fig. 1A; Supplementary Fig. S2A and S2B). All patients in the cholangiocarcinoma cohort had progressive disease as their best response. However, in the colorectal cancer cohort, two patients had stable disease. One patient had received three prior lines of treatment, including an experimental agent, whereas the other patient had received 12 prior treatments, also including experimental agents. There was no obvious bias in the metastatic sites or tumor burden in these patients; both patients had relatively large sums of target lesions, measuring 360 and 147 mm, respectively. In addition, one patient with cholangiocarcinoma in the dose escalation cohort, treated with 10 mg/kg weekly for 3 weeks followed by a 1-week rest period, remained in the study with stable disease for eight cycles until this patient experienced progression of nontarget lesions in the lung (Fig. 1A). During the treatment period with TRK-950, the tumor marker CA19-9 ranged from 4 to 27 U/mL, and carcinoembryonic antigen ranged from 1.6 to 3.4 ng/mL, with both marker values remaining within the reference range (Supplementary Table S4). In addition, this patient had high intensity (3+) and percentage (70%) of CAPRIN-1 expression (Fig. 1B). Furthermore, several of the pulmonary metastases in CT scans through the four imaging time points showed signs of cavitation, potentially indicative of antitumor activity (Fig. 1C–F).

The PK profile of TRK-950 was assessed from serum concentrations after the first intravenous administration at doses of 3, 10, and 30 mg/kg, administered weekly or biweekly during cycle 1 (Supplementary Fig. S3). The PK parameters based on the noncompartmental analysis are presented in Table 4. The total body clearance, apparent steady-state volume of distribution, and terminal elimination half-life were estimated to be 0.362 mL/hour/kg, 64.1 mL/kg, and 125.5 hours (5.2 days), respectively. C_max, AUC_last, and AUC_inf increased in a dose-dependent manner, and a dose proportionality analysis with a power model revealed that C_max and AUC_last were dose-proportional, whereas AUC_inf was not (Supplementary Table S5).

To characterize the immunologic events associated with the treatment of TRK-950, we evaluated whether TRK-950 promoted an increase in the number of tumor-infiltrating immune cells. In particular, we focused on macrophage infiltration as the primary pharmacologic action of TRK-950, which is considered ADCP, based on the results of nonclinical studies. Matched pretreatment and on-treatment tumor biopsies were available for eight patients in the colorectal cancer cohort and were analyzed by immunofluorescence. All biopsies performed for the analyses were obtained from metastatic sites: the liver in seven cases and the adrenal gland in one case. Analysis of CD68/CD163 staining revealed that the density of macrophages (total CD68⁺ cells) significantly increased after treatment in four patients, decreased in one patient, and remained stable in the other patients (Fig. 2A). This increase was mostly in CD163⁻ macrophages (M1) in two patients and both CD163⁻ and CD163⁺ subsets in the other patients (Fig. 2B and C).

In this phase I study, TRK-950 monotherapy demonstrated a favorable safety profile up to the maximum dose of 30 mg/kg weekly for 3 weeks, followed by a 1-week rest period, and was well tolerated in the patient population with advanced/metastatic solid tumors. In addition, no DLTs were observed, and the MTD was not reached. There was no significant dose dependence for the frequency or severity of TEAEs. Furthermore, the most common AEs reported were abdominal pain, fatigue, constipation, back pain, nausea, and decreased appetite, mainly grade 1/2 in severity, and no particular pattern was observed in the frequency of TEAEs reported across different dose levels among patients who reported TEAEs. These favorable safety profile data suggest that AEs are not derived from specific organ or cell characteristics and are consistent with the safety profile of TRK-950 in cynomolgus monkeys. Although CAPRIN-1 expression was observed at a low level in the cytoplasm of some normal tissues, CAPRIN-1 was barely expressed on the cell membranes of many normal tissues (20). This notable characteristic suggests that the safety profile shown in this study is consistent with the lack of expression of the binding site for TRK-950 in many normal tissues.

The PK profile of TRK-950 is similar to that of other IgG1 therapeutic antibodies (21–23). The distribution volume of TRK-950 was approximately 1.5 times the total plasma volume in humans (3 L for a 70 kg body weight; ref. 24). It has been suggested that TRK-950, which shows high affinity for CAPRIN-1, a protein strongly expressed on the surface of cancer cells, is mainly distributed in the circulating blood and tumor tissue. In the dose range of 3 to 30 mg/kg and serum concentration range of approximately 20 to 850 μg/mL, no steep decline in serum concentration was observed, indicating that a nonlinear PK profile due to target-mediated drug disposition clearance was not present (25, 26). This suggests that TRK-950 fully saturated its binding to CAPRIN-1 within this dose range.

Although no objective responses were observed in the dose escalation cohort, the best response of stable disease was observed in two cases of colorectal cancer and one case of cholangiocarcinoma. Among these three cases, CAPRIN-1 expression could not be evaluated in one case, but the remaining two cases exhibited high expression (3+). We previously demonstrated that the level of CAPRIN-1 expression was crucial for the antitumor effect of TRK-950 (20). Therefore, we selected these two cancer types, which showed the best response, to explore the correlation between TRK-950 efficacy and CAPRIN-1 expression, as well as the characterization of the immunologic events associated with the treatment of TRK-950.

One of the objectives of this study was to establish a recommended dose of TRK-950 for future clinical trials. Because no DLTs were observed at dose levels up to 30 mg/kg and the MTD was not reached, dose-related safety concerns could not be factored into the dose-finding estimates. Additionally, the safety and tolerability of TRK-950 monotherapy at doses ranging from 3 to 30 mg/kg were favorable, with dose-proportional exposure. The PK profile suggested complete binding with CAPRIN-1 at 3 to 30 mg/kg, and the dose–exposure relationship also demonstrated proportionality. Thus, the dose range of 3 to 30 mg/kg was supported as the recommended dose of TRK-950 for future clinical trials. Furthermore, PK/pharmacodynamic analysis (unpublished data) utilizing nonclinical data (20) and the PK profile of this study revealed that the minimally active dose and saturated active dose were determined to be 10 mg/kg, which led to the selection of 10 mg/kg.

Although the assessment of the antitumor activity of TRK-950 was not the primary objective of this first-in-human study and the enrolled patients were heavily pretreated, disease control was achieved in five patients, including one patient with cholangiocarcinoma who was treated for approximately 8 months. Interestingly, 70% of the tumor cells in this patient were identified as CAPRIN-1 at a strong intensity (3+). Owing to the limited number of patients, no clear conclusions could be drawn about the association between CAPRIN-1 expression and the antitumor activity of TRK-950. Further investigation into these relationships is required.

In our previous nonclinical studies, we demonstrated that the mechanism of action of TRK-950 involved ADCP and antibody-dependent cell-mediated cytotoxicity activity. In addition, we analyzed tumor-infiltrating immune cells in a mouse xenograft model when TRK-950 was administered and found that macrophages, which are important for ADCP activity, were the main infiltrating cells (20). In this study, we analyzed changes in tumor-infiltrating immune cells in the tumor tissues of patients before and after TRK-950 administration. Although this analysis was performed on a small subset of patients with colorectal cancer, we observed that the number of macrophages was clearly increased after TRK-950 treatment in four of eight patients with colorectal cancer. This increase was mostly associated with CD163⁻ macrophages (M1) in two patients and both CD163⁻ and CD163⁺ subsets in two patients. Although it should be noted that these patients did not respond to TRK-950 treatment, the results support the data obtained in nonclinical studies, as they suggest that TRK-950 treatment may increase macrophage infiltration into the tumor in some patients.

The favorable safety profile of TRK-950 suggests that it may be a suitable candidate for radiopharmaceutical development and may be safely used in combination with a standard treatment that has strong cytotoxicity and activates immune cells. Cytotoxic drugs such as cisplatin and doxorubicin have been reported to have a direct activating effect on antigen-presenting cells (27). In addition, bevacizumab (an IgG1 antibody targeting VEGF-A) has been shown to normalize the vascular system and promote the maturation of antigen-presenting cells (28). Furthermore, we showed that TRK-950 had synergistic antitumor effects in combination with chemotherapeutics, including gemcitabine, cisplatin, and the monoclonal antibody bevacizumab, suggesting that the combination of TRK-950 with standard therapeutic agents that have a direct activating effect on antigen-presenting cells and standard therapeutic agents with an inhibitory effect on angiogenesis would be expected to have a synergistic effect on each other (unpublished data).

In conclusion, this study showed that TRK-950 has a favorable safety profile and can be safely administered to patients with advanced solid tumors. In addition, preliminary evidence of potential antitumor activity in patients with high CAPRIN-1 expression in tumors was observed, thus providing initial clinical experience and a rationale for the further development of TRK-950 in patients with advanced solid tumors expressing CAPRIN-1. TRK-950 is currently in a Ph-Ib study in combination with standard-of-care treatments (chemotherapy, immune checkpoint inhibitors, and targeted agents) for various cancers, including advanced gastric, ovarian, and renal cancer (NCT03872947), and a phase II clinical study in patients with advanced gastric cancer (NCT06038578).

P.A. Cassier reports grants and other from Toray Industries during the conduct of the study as well as other from AbbVie, Adlai Nortye, Alligator, Astellas, Boehringer Ingelheim, C4 Therapeutics, Dragonfly, Daiichi Sankyo, Ellipses, Greywolf, GSK, Incyte, Ipsen, iTeos, Kinnate, Kazia, Loxo/Eli Lilly, Molecular Partners, OnKure, Regeneron, Relay, Roche/Genentech, SOTIO, Tango, Toray Industries, Transgene, and Arcus; grants, nonfinancial support, and other from Novartis; personal fees and other from MabQuest, OSE Immunotherapeutics, Bristol Myers Squibb, MSD, and AstraZeneca; and personal fees from Servier outside the submitted work. M.J. Borad reports grants from Toray Industries during the conduct of the study. B. Dubois reports grants from Toray Industries during the conduct of the study. C. Caux reports grants and other from Toray Industries during the conduct of the study. F. Okano reports that he has patents WO2010/016526 and WO2013/018886 issued and he is a senior director of Toray Industries. D.D. Von Hoff reports other from Toray Industries during the conduct of the study. J.-Y. Blay reports grants from Toray Industries during the conduct of the study. No disclosures were reported by the other author.

P.A. Cassier: Investigation, writing–original draft, writing–review and editing. M.J. Borad: Investigation, writing–review and editing. S. Sharma: Investigation, writing–review and editing. B. Dubois: Investigation, writing–original draft, writing–review and editing. C. Caux: Investigation, writing–original draft, writing–review and editing. F. Okano: Conceptualization, data curation, supervision, writing–original draft, writing–review and editing. D.D. Von Hoff: Conceptualization, supervision, writing–original draft, writing–review and editing. J.-Y. Blay: Conceptualization, supervision, writing–original draft, writing–review and editing.

This study was funded and supported by Toray Industries, Inc. (Tokyo, Japan). We would like to thank Justine Berthet and Sarah Barrin-Ghamry for multi-immunofluorescence staining and analysis and the Rhône Alpes Auvergne region (Grant IRICE RRA18 010792 01 – 10365) for the initial funding of the Lyon Immunotherapy for Cancer Laboratory multi-immunofluorescence platform. We also thank the patients, their families, investigators, and research staff at all the study sites. We would also like to thank the TD2 and Biotrial teams for their support with study operations, data analysis, and quality checks.