Main

B7H3 (CD276) is a member of the B7 family of immune checkpoint molecules, which play a dual role in both immune regulation and tumor progression1. It is overexpressed across a wide spectrum of malignancies, including small cell lung cancer (SCLC)2, nasopharyngeal carcinoma (NPC)3, prostate cancer4, non-small cell lung cancer (NSCLC)5, breast cancer6, and head and neck squamous cell carcinoma7. Its overexpression is frequently associated with a poor prognosis. Beyond its role in immune evasion, B7H3 contributes to tumor growth, metastasis, resistance to therapy and angiogenesis through non-immunological mechanisms8. These characteristics make B7H3 a promising therapeutic target9,10.

Despite its potential as a therapeutic target, the development of effective B7H3-targeting molecules has been challenging. Previous attempts, including B7H3 bispecific antibodies and monoclonal antibodies, have demonstrated limited success in clinical trials11,12. Antibody–drug conjugates (ADCs) represent a cutting-edge approach in oncology by combining the specificity of monoclonal antibodies with cytotoxic agents, which selectively deliver toxic drugs to target antigen-expressing tumor cells13,14. ADCs targeting B7H3, such as MGC018, DS7300a, HS-20093 and MHB088C, have shown promising results in clinical trials, yet there are no approved ADCs that target B7H3 (refs. 15,16,17,18).

YL201 is a novel ADC comprising a human anti-B7H3 monoclonal antibody conjugated to a novel topoisomerase 1 inhibitor via a protease-cleavable linker, with a drug-to-antibody ratio of 8. YL201 was developed using MediLink’s tumor microenvironment activable linker-payload (TMALIN) platform (patent number WO2022170971A1), which is characterized by a highly hydrophilic linker-payload, remarkable stability in circulation and a dual payload release mechanism that includes cleavage both extracellularly in the tumor microenvironment and intracellularly within lysosomes19,20. In a preclinical study, YL201 exhibited potent anti-tumor activity in various CDX/PDX mouse models with SCLC, NSCLC, esophageal cancer and ovarian cancer21. The efficacy and safety data from this preclinical study suggest a favorable benefit–risk balance, supporting further development of YL201 in the clinic.

Here we present the results from this global, multicenter, phase 1 trial, focusing on the safety, tolerability and anti-tumor activity of YL201 across multiple tumor types. We also explored the relationship between B7H3 expression and treatment outcomes, providing insights into the potential of YL201 as a therapeutic option for heavily pretreated patients with advanced solid tumors.

Results

Patients

Between 26 May 2022 and 31 July 2024, a total of 312 patients were enrolled in this study, including 49 patients in the dose-escalation phase (phase 1) and 263 patients in the dose-expansion phase (phase 1b, 177 patients in the 2.4 mg kg−1 group and 86 patients in the 2.0 mg kg−1 group) (Fig. 1 and Extended Data Fig. 1). Enrollment in the 2.0 mg kg−1 dose group is still ongoing. The most common tumor types among the enrolled patients were extensive-stage small cell lung cancer (ES-SCLC, n = 79), NPC (n = 75), NSCLC (n = 68) and esophageal squamous cell carcinoma (ESCC, n = 37). There were also 53 patients with other tumors, including head and neck cancer (n = 20), pancreatic cancer (n = 10), epidermal growth factor receptor (EGFR)-mutated NSCLC (n = 9), sarcoma (n = 5), prostate cancer (n = 2), colorectal cancer (n = 2), breast cancer (n = 1), cervical cancer (n = 1), esophageal adenocarcinoma (n = 1), unknown subtype NSCLC (n = 1) and thymic carcinoma (n = 1). The demographics for patients are summarized in Table 1; for patients in the dose-expansion phase and for different tumor types in the efficacy analysis set, demographics are summarized in Supplementary Tables 1 and 2. Most (77.6%) of the participants in this trial were male, with a median age of 57 years (range, 19–87 years). Brain metastases at baseline were observed in 15.4% of patients. All patients had been treated with prior standard therapy, and 59.9% had at least two prior treatment lines. At the data cutoff (26 September 2024), the median follow-up time was 7.5 months (95% confidence interval (CI): 6.5–7.9).

Fig. 1: Study flowchart and patient disposition.
figure 1

Q3W, 1 dose every 3 weeks per cycle.

Full size image
Table 1 Baseline demographic and clinical characteristics
Full size table

Safety

The treatment-related adverse events (TRAEs) are presented in Table 2. In total, 97.1% of patients experienced TRAEs, with the most common hematological adverse events being leukopenia (66.3%), anemia (64.7%) and neutropenia (61.5%). The most common non-hematological TRAEs included anorexia (35.6%), nausea (26.3%) and hypoalbuminemia (22.8%). Grade 3 or higher TRAEs occurred in 54.5% of patients, with neutropenia (31.7%) being the most frequent. All treatment-emergent adverse events (TEAEs) occurring in more than 10% of patients and those with grade 3 or higher TEAEs are summarized in Supplementary Table 3. Treatment-related severe adverse events (SAEs) were reported in 29.2% of patients, with 5.4% (n = 17) of the patients discontinuing treatment and 17.0% having their dose reduced (Table 2). Thirty-five (11.2%) patients experienced the SAE of leukopenia; the median time from the first dose to the first onset of this SAE was 11.0 days; and the median time from first onset to recovery was 7.5 days. For the lineage of the dynamic changes of the white blood cell counts in patients with the SAE of leukopenia, see Supplementary Fig. 1. Eight deaths (2.6%) were considered related to YL201 (Table 2). Treatment-related interstitial lung diseases (ILDs) were observed in only 1.3% of patients. The incidence of TRAEs leading to treatment interruption was 36.2%, mainly due to hematological adverse events. The incidence of treatment-related infusion reactions was 0.3%. A total of 111 patients (35.6%) remained on treatment at the time of data cutoff. Treatment discontinuations due to adverse events occurred in 17 patients (5.4%), and dose reductions were required in 17.0% of patients. Most of these reductions were due to hematological toxicities, which were manageable with supportive care.

Table 2 Most common TRAEs
Full size table

Efficacy

Among the 287 patients evaluable for efficacy, the objective response rate (ORR) was 40.8% (95% CI: 35.0–46.7), and the disease control rate (DCR) was 83.6% (95% CI: 78.8–87.7). The median progression-free survival (mPFS) was 5.9 months (95% CI: 5.5–7.5), and the median duration of response (mDOR) was 6.3 months (95% CI: 4.7–6.7) (Table 3 and Fig. 2a,b). The median overall survival (OS) was not mature. The efficacy results of the different dose groups and tumor types are summarized in Table 3, Extended Data Table 1 and Extended Data Fig. 2. Among the 48 patients with baseline brain metastases, 21 were evaluable; the intracranial ORR was 28.5% (95% CI: 11.3–52.2); and the DOR was 6.2 months (95% CI: 2.8–not reached) (Extended Data Fig. 3 and Supplementary Fig. 2).

Table 3 Efficacy in different tumor types
Full size table

All patients with ES-SCLC had prior treatment with platinum-based chemotherapy, and 95.8% of patients also had prior treatment with programmed death 1 (PD-1) or programmed death ligand 1 (PD-L1) antibody. The ORR was 63.9% (95% CI: 51.7–74.9); the DCR was 91.7% (95% CI: 82.7–96.9); the mPFS was 6.3 months (95% CI: 5.6–7.6); and the mDOR was 5.7 months (95% Cl: 3.9–6.4) (Fig. 2c,d). Subgroup analyses of the ES-SCLC cohort are shown in Extended Data Table 2. In the dose-escalation study, five patients with SCLC who received prior treatment with a topoisomerase 1 inhibitor (topotecan or irinotecan) were enrolled, with only one patient responding (ORR, 20%). Of the 70 patients with NPC, all had prior treatment with anti-PD-L1 and platinum-based chemotherapy, and 84.3% (59/70) had two or more lines of prior systemic therapies. Two of the patients with NPC achieved a complete response (CR), with an ORR of 48.6% (95% CI: 36.4–60.8) and a DCR of 92.9% (Fig. 2e). The mPFS was 7.8 months (95% CI: 5.8–12.5), and the mDOR was 8.4 months (95% CI: 5.8–not reached) (Fig. 2f). Subgroup analyses of the NPC cohort are shown in Extended Data Table 3. A total of 64 patients with wild-type NSCLC were evaluated, including 28 with adenocarcinoma, 12 with squamous cell carcinoma and 24 with lymphoepithelioma-like carcinoma (LELC). In total, 90.6% (58/64) of patients with NSCLC had prior treatment with anti-PD-L1 and platinum-based chemotherapy. The ORR values for the patients with lung adenocarcinoma, squamous cell carcinoma and LELC were 28.6%, 8.3% and 54.2%, respectively, and the mDOR was 13.6 months, 2.6 months and 6.7 months, respectively (Fig. 2g,h and Table 3). Typical radiographic responses in patients with ES-SCLC, NPC and LELC are shown in Extended Data Fig. 4.

Fig. 2: Waterfall plot of tumor response from baseline and swimming plot.
figure 2

a,c,e,g, The waterfall plot depicts the best objective responses concerning the tumor size in all patients (a), ES-SCLC (c), NPC (e) and LELC (g). b,d,f,h, Swimmer plots depict responses to treatment and the duration of treatment in all patients (b), ES-SCLC (d), NPC (f) and LELC (h).

Full size image

Pharmacokinetics and recommended expansion dose

A total of 46 patients from the phase 1a dose-escalation study were included in the pharmacokinetics (PK) analysis. YL201 showed nonlinear clearance kinetics at doses from 0.8 mg kg−1 to 2.0 mg kg−1 and linear clearance kinetics at doses higher than 2.0 mg kg−1. The mean terminal-phase half-life ranged from 21.7 h to 37.6 h within the dose range of 1.6–2.8 mg kg−1. The mean maximum concentration (Cmax) showed a dose-proportional increase. The minor unconjugated payload YL0010014 was determined with serum concentrations peaking at approximately 6 h after infusion with a mean half-life ranging from 76.4 h to 97.3 h at doses ranging from 1.6 mg kg−1 to 2.8 mg kg−1. Population PK analysis showed linear clearance kinetics of YL201 from 2.0 mg kg−1 (Supplementary Figs. 3–5 and Supplementary Table 4). Both the mean Cmax and the area under the curve (AUC) of YL0010014 exhibited dose-proportional increases (Extended Data Fig. 5 and Supplementary Table 5). The PK data in cycle 1 and cycle 3 are similar (Supplementary Tables 5 and 6 and Supplementary Fig. 6). The exposure–response analysis for both efficacy and safety indicated a positive correlation between a higher ORR and an increased dose and exposure, whereas the safety profiles exhibited steeper trends (Supplementary Fig. 7). A considerable higher risk of hematological TRAEs was estimated at 2.8 mg kg−1 compared to doses lower than 2.4 mg kg−1, particularly for grade 3 or higher neutropenia (50%, 3/6) and anemia (66.7%, 4/6). Anti-drug antibody (ADA) samples were tested from 215 patients in YL201-INT-101-01 and YL201-CN-101-01 studies. The ADA-positive rate after YL201 treatment was 0.00% (0/215).

In the phase 1a cohort, two dose-limiting toxicities (DLTs) (one with grade 4 thrombocytopenia accompanied by grade 2 hemorrhage and the other with grade 4 neutropenia lasting for >7 days) were observed in three patients enrolled at 3.0 mg kg−1, and one DLT (grade 4 febrile neutropenia) was observed in six patients enrolled at 2.8 mg kg−1. According to the 3 + 3 rule, the maximum tolerated dose (MTD) was determined to be 2.8 mg kg−1. Based on the safety data and PK data collected from 49 patients in phase 1, a relatively high rate of grade 3 or higher TRAEs (83.3%) was observed after multiple doses at 2.8 mg kg−1, with a positive relationship established between YL201 exposure and the incidence of grade 3 or higher neutropenia. Comprehensive assessment of the safety and PK analyses suggested 2.0 mg kg−1 and 2.4 mg kg−1 as the recommended expansion dose.

In all patients, for the 2.4 mg kg−1 and 2.0 mg kg−1 dose groups, the incidence of grade 3 or higher TRAEs was 58.5% and 45.8%, respectively, and the ORR was 37.5% and 48.3%, respectively. In patients with ES-SCLC, for the 2.4 mg kg−1 and 2.0 mg kg−1 dose groups, the incidence of grade 3 or higher TRAEs was 71.1% and 50.0%, respectively; the ORR was 63.9% and 67.7%, respectively; and the mPFS was 5.7 months and 7.6 months, respectively. In patients with NPC, for the 2.4 mg kg−1 and 2.0 mg kg−1 dose groups, the incidence of grade 3 or higher TRAEs was 56.8% and 48.6%, respectively; the ORR was 47.1% and 48.6%, respectively; and the mPFS was 9.6 months and 7.8 months, respectively. Detailed efficacy data for the different tumor types in the 2.0 mg kg−1 and 2.4 mg kg−1 dose groups are shown in Supplementary Table 7.

Biomarkers

A total of 152 patients (including 48 with NPC, 14 with LELC, 12 with lung adenocarcinoma, 46 with SCLC and 32 with ESCC) had pretreatment specimens available to explore the relationship between B7H3 membrane expression and activity. Nearly all tumors expressed detectable B7H3 by immunohistochemistry (Fig. 3a). The median baseline B7H3 H-scores in the patients with NPC, LELC, lung adenocarcinoma, SCLC and ESCC were 137.5, 152.5, 130, 82.5 and 142.5, respectively. However, no significant correlation was observed between B7H3 expression and the clinical response across different tumor types (Fig. 3 and Extended Data Fig. 6). These findings suggest that B7H3 expression alone may not be a sufficient predictive biomarker for YL201 efficacy.

Fig. 3: Expression of B7-H3 in different tumor types and the relationship between B7-H3 expression and response.
figure 3

a, Expression of B7H3 in different tumor types. b–f, The relationship between B7-H3 expression and response in all patients (b), ES-SCLC (c), NPC (d), ESCC (e) and LELC (f). The median, quartile, and maximum and minimum values are shown. Comparisons were performed using Kruskal–Wallis one-way ANOVA tests. NE, not evaluated; cBOR, confirmed best overall response.

Source data

Full size image

A total of 223 patients (including 68 with ES-SCLC, 61 with NPC, 33 with ESCC, 24 with LELC, 27 with lung adenocarcinoma and 10 with lung squamous cell carcinoma) had pretreatment blood samples available to explore the relationship between the soluble B7H3 (sB7H3) concentration and response. The median concentration of sB7H3 was 16.6 ng ml−1. No significant correlation was observed between the sB7H3 concentration and the clinical response across different tumor types (Extended Data Fig. 7).

Discussion

To our knowledge, we report here the largest cohort to date for evaluation of a B7H3-targeted ADC in patients with advanced solid tumors; in particular, this is the first report on their use in Epstein–Barr virus (EBV)-associated tumors, such as NPC and LELC. YL201 demonstrated a manageable safety profile and a promising anti-tumor activity, particularly in patients with ES-SCLC, NPC or LELC. These findings provide strong confidence for the further development of YL201 in randomized trials.

The safety profile of YL201, characterized predominantly by hematological toxicities such as leukopenia, anemia and neutropenia, aligns with the expected profile of most ADCs using topoisomerase inhibitors as the payload16,17,22. The rates of severe (grade ≥3) treatment-related hematological toxicities of YL201 ranged from 13.8% to 31.7%, which are similar to those of other B7H3 ADCs15,16,17. Despite these toxicities, the treatment discontinuation rate of YL201 due to TRAEs (5.4%) was relatively low, and the dose reduction rate during the treatment course was also acceptable (17.0%). In this study, the incidence of treatment-related ILD was low (1%), suggesting a low risk of pulmonary toxicity for YL201. In addition, the incidence of infusion reactions was relatively low (0.3%, 1/312)23. Overall, YL201 showed a manageable safety profile with low frequencies of dose modifications and discontinuations because of adverse events.

The observed efficacy of YL201 in patients with ES-SCLC is particularly noteworthy, with an ORR of 63.9%, which is superior to the response rates reported for standard second-line treatments such as lurbinectedin, tarlatamab and topotecan, where the ORRs typically range from 21.9% to 40%24,25,26. Additionally, the mPFS of 6.3 months and the mDOR of 5.7 months observed in this study suggest that YL201 could provide a meaningful clinical benefit in a heavily pretreated ES-SCLC population. This efficacy was observed irrespective of the B7H3 expression in tumor tissue or soluble B7H3 in the peripheral blood, indicating that the activity of YL201 may not be strictly dependent on the extent of target expression. This finding is consistent with previous observations for other B7H3-targeting ADCs and highlights the potential for broader applicability across tumor types with variable B7H3 expression.

In patients with NPC, where treatment options are limited after failure of platinum-based chemotherapy and PD-1 inhibitors, YL201 demonstrated an ORR of 48.6%. This is particularly significant given the unsatisfied efficacy of patients with recurrent/metastatic NPC after progression on standard therapies. The established standard treatments, including docetaxel or PD-1 inhibitors for immune checkpoint inhibitor (ICI)-naive/treated recurrent or metastatic NPC, exhibit ORRs ranging from 23.5% to 36.7%27,28. A recent study reported that EGFR/human epidermal growth factor receptor (HER) 3 ADC demonstrated an ORR of 37.8% in pretreated patients with recurrent/metastatic NPC22. The presence of EBV and its role in upregulating B7H3 expression in patients with NPC might provide a plausible mechanistic basis for the observed efficacy29, suggesting that B7H3-targeting strategies might be particularly effective in EBV-associated tumors.

The results in patients with LELC are also encouraging, with an ORR of 54.2% and an mDOR of 6.7 months. LELC is a subtype of NSCLC with distinct clinical and pathological characteristics, which are closely associated with EBV infection30. The lack of effective targeted therapies for LELC makes YL201 a potentially valuable therapeutic option for these patients.

The recommended doses may be different among tumor types31,32,33. For example, trastuzumab deruxtecan has been approved for breast cancer, stomach cancer and lung cancer, but the approved doses are different31,32,33. In breast cancer and HER2-mutant metastatic NSCLC, dose–response projections suggest increased toxicities with 6.4 mg kg−1 versus 5.4 mg kg−1, and the recommended dose is 5.4 mg kg−1 (refs. 31,33,34). In gastric cancer, the approved dose is 6.4 mg kg−1 due to the controllable adverse events32. In our study, given the tolerable safety profile and promising efficacy in the 2.0 mg kg−1 dose group for ES-SCLC, the recommended dose used in the ongoing phase 3 clinical trial for ES-SCLC is 2.0 mg kg−1 (NCT06612151); for NPC, a better activity and a similar safety were found in the 2.4 mg kg−1 dose group, which was applied as the recommended dose in the ongoing phase 3 clinical trial (NCT06629597).

Despite these promising results, our study has limitations. This is an ongoing single-arm phase 1 study, and the efficacy analyses are preliminary and should be interpreted with caution. Future endeavors will report the survival benefits and long-term toxicity of YL201. Additionally, the correlation between B7H3 expression and the treatment response remains inconclusive, necessitating further biomarker-driven studies to better identify patients who are most likely to benefit from YL201 therapy. Future research should focus on elucidating the mechanisms underlying the observed efficacy of YL201, including the role of B7H3 expression, the impact of tumor microenvironment activation and potential combination strategies with other immuno-oncological agents.

In conclusion, YL201 has shown a favorable safety profile and encouraging clinical anti-tumor activity in patients with heavily pretreated advanced solid tumors, particularly in those with ES-SCLC, NPC or LELC. Phase 3 clinical trials for SCLC and NPC have already been initiated (NCT06612151 and NCT06629597).

Methods

Study design and participants

This was an open-label, multicenter, phase 1/1b clinical trial evaluating YL201 monotherapy in patients with locally advanced or metastatic solid tumors. The dose-escalation part (phase 1) of this study was conducted globally (ClinicalTrials.gov identifier: NCT05434234, version INT101); the dose expansion part (phase 1b) was conducted in China (ClinicalTrials.gov identifier: NCT06057922, version CN101).

The primary objectives of the phase 1 study were to assess the safety and to determine the MTD and the recommended expansion dose of YL201. For the phase 1b study, the primary objectives were to further evaluate the safety and the primary efficacy of YL201 in selected patients with advanced solid tumors; the secondary objectives were to assess the efficacy in patients with selected advanced solid tumors, PK profiles and immunogenicity of YL201.

Inclusion and ethics

Eligible patients were those older than 18 years (phase 1) or 18–75 years (phase 1b) with histologically confirmed advanced solid tumors for which no standard options were available. Inclusion criteria included a life expectancy of at least 3 months, at least one measurable lesion based on Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1, an Eastern Cooperative Oncology Group (ECOG) performance status of 0–1 and adequate organ and bone marrow function. Exclusion criteria included prior treatment with B7H3-targeting agents, concurrent participation in other clinical trials, recent major surgery (not diagnostic surgery), recent allogeneic hematopoietic stem cell transplantation or solid organ transplantation, recent systemic steroids or other immunosuppressive therapy, recent live vaccinations and uncontrolled or clinically significant cardiovascular disease. The full inclusion and exclusion criteria are listed in the protocol part (pages 30–33) of Supplementary Information.

The study protocol and all amendments were approved by the institutional review boards at all participating sites (details are shown in Supplementary Table 8). Written informed consent was obtained from all patients before any study procedure. This trial was conducted according to the Declaration of Helsinki and Good Clinical Practices. The full study protocol is provided as a part of Supplementary Information.

Procedures

The study drug (YL201) was supplied by the sponsor (MediLink Therapeutics) in single-use glass vials (200 mg per vial). The drug was intravenously administered once every 3 weeks. Phase 1a followed a standard 3 + 3 design. The initial dose was 0.8 mg kg−1, which was determined based on one-sixth of the human equivalent dose of the highest non-severely toxic dose in monkeys (15 mg kg−1). The dose levels included 0.8 mg kg−1, 1.6 mg kg−1, 2.4 mg kg−1 and 3 mg kg−1. Patients were backfilled at doses of 2.0 mg kg−1, 2.4 mg kg−1 and 2.8 mg kg−1, with the number at each dose level varying from six to 15, as determined by the Safety Review Committee’s analysis of emerging data.

The DLT observation period was the first treatment cycle (days 1–21) for each dose level. DLTs were assessed during the first 21 days and were defined as TEAEs as follows: (1) hematological toxicities, including grade 4 neutropenia lasting more than 7 days; (2) grade 3 or higher febrile neutropenia; (3) grade 4 lymphocyte count decrease lasting for 14 or more days; (4) grade 4 anemia; (5) grade 3 or higher thrombocytopenia lasting more than 7 days; and (6) any grade 3 or higher non-hematological toxicities, except for alopecia, nausea and vomiting, that could be controlled with adequate treatment as well as grade 3 fatigue lasting less than 7 days. Detailed information of the study design and the DLT definition are listed in the protocol part (pages 36–40) of Supplementary Information. If two or more DLTs occurred within a cohort, that dose level was considered to be above the MTD, and the previous lower dose level was considered the MTD if one or fewer in six patients had a DLT. The recommended expansion dose was selected according to the safety and PK profile.

The dose-expansion stage was carried out in two recommended expansion doses (2.4 mg kg−1 and 2.0 mg kg−1). Patients with selected advanced solid tumors were assigned to the following expansion cohorts: patients with locally advanced or metastatic NSCLC without actionable genomic alterations; patients with ES-SCLC; patients with recurrent or metastatic NPC; and patients with locally advanced or metastatic ESCC. In each cohort, we adopted a sequential allocation method for patient recruitment, where consecutive patients were initially enrolled at 2.4 mg kg−1, and then the 2.0 mg kg−1 dose level was opened.

All patients received YL201 treatment until disease progression, intolerable toxicity or withdrawal of consent.

Assessments

Tumor response was evaluated by local investigators according to RECIST version 1.1 as complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Radiographic tumor assessments (via computed tomography and magnetic resonance imaging) were performed within 28 days of treatment initiation and once every 6 weeks during the first four assessments and every 12 weeks thereafter until disease progression, start of new anti-cancer therapy, death or withdrawal of consent, whichever occurred first.

Safety assessments, including laboratory tests, were assessed throughout the study, extending until 42 days after the last dose or initiation of other anti-tumor treatment. Adverse events in the study were defined and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 5.0. Safety endpoints included TEAEs and TEAEs of special interest (ILD/pneumonitis and infusion-related reactions). All events of ILD/pneumonitis, regardless of severity or seriousness, were followed until resolution, even after drug discontinuation. Follow-up for survival was done every 3 months after treatment discontinuation by assessment of medical records and telephone contact.

According to the predefined protocol guidelines, YL201 treatment would be permanent if there were grade 3 or higher infusion-related reactions, grade 2 or higher ILD/pneumonitis or grade 4 or higher hepatic organ toxicities. Specific criteria for dose interruption of YL201 are listed in the protocol. The investigator evaluated which toxicities were attributable to YL201 and adjusted the dose of YL201 as recommended in the protocol, with a maximum of two dose reductions permitted. All data were collected in an internet-based, controlled, electronic data capture system (Medidata Rave) for further analysis.

Blood samples for PK analysis were collected at pre-infusion; at end of infusion (EOI); at 3 h and 6 h after EOI on cycle 1 day 1 (C1D1); at 24 h after EOI on C1D2; at 168 h after EOI on C1D8; at 336 h after EOI on C1D15; at pre-infusion and EOI on C2D1; at pre-infusion, EOI and 6 h after EOI on C3D1; at 168 h after EOI on C3D8; and at pre-infusion and EOI on C4D1 and every 2 cycles. For phase 1, samples were collected from all patients at cycles 1 and 3. For the dose-expansion phase, samples were collected from 20 patients at cycles 1 and 3 and from the other patients at pre-infusion and EOI on day 1 of each cycle. The serum concentrations of YL201-ADC and unconjugated payload (YL0010014) were determined using a validated ELISA and high-performance liquid chromatography with tandem mass spectrometry, respectively. Exposure–response analyses were also conducted (Supplementary Methods).

Expression of B7H3 in tumor tissue collected before the first dose of YL201 was determined by immunohistochemistry. B7H3 immunohistochemistry was performed centrally on formalin-fixed, paraffin-embedded tissue using anti-B7H3 antibody (Abcam, ab227670) and a Bond Polymer Refine Detection Kit (Leica, DS9800), and interpretation of B7H3 staining was carried out by a qualified pathologist. B7H3 membrane expression on tumor cells was quantified by H-scores. The H-score (range, 0–300) was calculated using the following formula: H-score = (1 × % cells with weak intensity staining) + (2 × % cells with moderate intensity staining) + (3 × % cells with strong intensity staining).

Outcomes

For phase 1, the primary endpoint was DLTs during the first treatment cycle; the secondary endpoints were to assess the adverse events during treatment, PK parameters, ORR, DCR, DOR, PFS, best tumor response and OS. The levels of sB7H3 in the peripheral blood before and after treatment were exploratory endpoints.

For phase 1b, the primary endpoints were to assess the adverse events during treatment and the ORR; the secondary endpoints were DCR, DOR, PFS and best tumor response. The secondary endpoints included PK parameters, metabolite profile of YL201 and immunogenicity. Exploratory endpoints included the expression level of B7H3 in tumor tissues and the tumor response.

The ORR was defined as the sum of CR and PR. DCR was defined as the percentage of patients who had achieved CR, PR and SD. The DOR was defined as the time from the first documentation of confirmed response (CR or PR) to the first documentation of PD or death, whichever occurred first. PFS was defined as the time from the start of treatment until documentation of PD or death, whichever occurred first. OS was defined as the time from the start of treatment until death. We also supplemented intracranial ORR and DOR in patients with brain metastases as post hoc analyses.

Single-dose and multiple-dose PK parameters of the YL201 ADC (YL201-ADC), YL201 total antibody (YL201-TAb) and unconjugated payload YL0010014 included AUC, Cmax, clearance (CL) rate, volume of distribution (Vd) and half-life time (t1/2).

Statistical analysis

In phase 1, the sample size was based on the number of DLT cases observed at each dose level. Using a standard 3 + 3 design and six planned dose levels, the maximum sample size of DLT-evaluable patients (not including backfilling patients) was estimated to be 36. In phase 1b (dose-expansion study), the sample size was determined according to the FDA guidance for ‘expansion cohorts’35. The sample size for each dose level in each cohort was expected to be 30–45. The safety analysis set included all patients who received at least one dose of YL201. The efficacy analysis set included all patients who received at least one dose of YL201 and had baseline and at least one post-baseline efficacy assessment. The PK concentration analysis set included all patients who received at least one dose of YL201 and had at least one post-dose concentration measurement above the lower limit of quantitation for YL201-ADC, YL201-TAb or YL0010014, without major protocol deviations or events (for example, incomplete dosing or disallowed concomitant medications) affecting PK. The PK parameter analysis included all patients who had at least one PK parameter of interest for YL201-ADC, YL201-TAb or YL0010014, without major protocol deviations or events (for example, incomplete dosing or disallowed concomitant medications) affecting PK. The immunogenicity analysis set included all patients who received at least one dose of YL201 and had at least one ADA assessment. The biomarker analysis set included all patients who received at least one dose of YL201 and had baseline assessment for biomarkers.

Descriptive statistics are provided for selected demographic, safety, PK, efficacy and biomarker data by dose level/cohort within each part and time, as appropriate. The chi-squared test was used for categorical variables. Comparisons of marker levels were performed using Kruskal–Wallis tests followed by Dunn’s correction for multiple comparisons. Time-to-event variables, including DOR and PFS, were summarized descriptively using the Kaplan–Meier method to estimate the median event times and two-sided 95% CIs. The PK analysis was performed using actual sample times and non-compartmental methods. The concentration data were used to perform population PK modeling.

All statistical analyses, except for PK parameters, were performed using SAS software version 9.4 or higher (SAS Institute). The PK parameters were estimated by non-compartmental methods with Phoenix WinNonlin software version 8.2 (Certara) and populational pharmacokinetic (PopPK) methods with NONMEM software version 7.5 (ICON Development Solutions). A logistic regression model was performed for exposure–response analysis using R (version 4.4.1) software.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.