A phase II, multicenter, open-label, single-arm study of berzosertib plus topotecan in patients with relapsed platinum-resistant small-cell lung cancer (DDRiver SCLC 250).

Introduction
Small-cell lung cancer (SCLC), the most aggressive subtype of lung cancer, accounts for ∼15% of lung cancer cases1 and is characterized by rapid growth and widespread metastases even at the time of first presentation.2 SCLC is a heterogeneous disease that comprises both highly chemosensitive and chemoresistant clones. As a result, patients with SCLC often respond well to first-line therapy initially. These initial therapies typically include platinum-based chemotherapy combinations with or without programmed death-ligand 1 inhibitors, such as atezolizumab or durvalumab, and local radiotherapy for limited-stage disease.3, 4, 5, 6, 7 However, chemoresistant clones survive and eventually proliferate, leading to disease recurrence and progression.8 Therefore, most patients experience relapse, with a median survival of 10-12 months.9,10
Platinum-free interval (PFI) and performance status (PS), as measured by the Eastern Cooperative Oncology Group (ECOG), are two major prognostic factors to consider in patients with SCLC who have progressed after first-line treatment.11,12 PFI is measured by the elapsed time from the last day of the regimen of a platinum-based therapy until the first day of documented disease progression.13 Patients with a PFI of <90 days are classified as ‘platinum-resistant’.3 The first approved treatment for platinum-resistant SCLC was the topoisomerase 1 (TOP1) inhibitor, topotecan.14 However, it has shown limited efficacy, with response rates ranging from 5% to 17%.15 The subgroup of patients whose disease has a PFI of <30 days are classified as ‘platinum-refractory’ and are characterized by a particularly poor prognosis.16
Topotecan was the only United States Food and Drug Administration (FDA)-approved treatment for relapsed SCLC until 2020, when newer options began to be integrated into its treatment algorithm.14 In Japan, amrubicin, a TOP2 inhibitor, is preferred over topotecan in the second-line treatment of SCLC; however, it is not approved elsewhere globally.17,18 In addition, other therapies might be considered according to clinical aspects of the disease and physicians’ preference, such as irinotecan, paclitaxel, temozolomide, and rechallenge with platinum-based chemotherapy for platinum-sensitive patients.19 More recently, treatment options for relapsed SCLC include the conditionally approved DNA-damaging agent lurbinectedin, and the bispecific T-cell engager immunotherapy tarlatamab.19, 20, 21, 22, 23, 24 Although lurbinectedin received accelerated FDA approval for relapsed SCLC based on a single-arm, open-label, phase II trial,25,26 its combination with doxorubicin failed to improve overall survival (OS), compared with physician’s choice therapy, in the phase III ATLANTIS trial.27 In a phase II trial, tarlatamab showed antitumor activity [objective response rate (ORR) 36%] and positive survival outcomes in patients with relapsed SCLC, thereby receiving accelerated approval from the FDA.24,28 A confirmatory phase III trial has corroborated the preliminary results, demonstrating OS gain with tarlatamab compared with the chemotherapy arm [median OS 13.6 months versus 8.3 months; hazard ratio 0.60, 95% confidence interval (CI) 0.47-0.77, P < 0.001].29
Given that enhanced replication stress is a hallmark of SCLC, these tumors are vulnerable to drugs targeting key proteins involved in the replication stress response.30, 31, 32, 33 The DNA damage response (DDR) pathway, including the apical regulator ataxia telangiectasia and Rad3-related protein kinase (ATR), plays a key role in preventing genomic instability.34, 35, 36, 37, 38, 39, 40 ATR is the primary responder to replication stress and is activated and recruited to stalled replication forks in the presence of single-stranded DNA.37 Therefore, ATR and TOP1 inhibition may be a rational treatment strategy for SCLC.33
Berzosertib is an intravenously administered, highly potent, selective ATR inhibitor.41,42 In chemical genetic screens, the inhibition of ATR and TOP1 were shown to be synergistically cytotoxic in SCLC, and in cell-based assays, berzosertib potentiated the effect of topotecan by trapping TOP1 cleavage complexes on DNA.31 In prior phase I and II studies, this combination demonstrated encouraging antitumor activity and was well tolerated in patients with advanced solid tumors, including platinum-resistant and -sensitive relapsed SCLC.31,43
In the DDRiver SCLC 250 study, we evaluated the combination of berzosertib and topotecan in patients with relapsed platinum-resistant SCLC.

Materials and methods

Study design and objectives
The DDRiver SCLC 250 (NCT04768296) was a global multicenter single-arm phase II study conducted at 39 sites in China, Europe, Japan, and the United States. The study also included a Japan-only safety run-in part to assess the safety, tolerability, and pharmacokinetic (PK) profile of the combination in Japanese patients as there was no prior experience with berzosertib in Japanese patients before the study. Details of the Japan safety run-in part are provided in the Supplementary Material, available at https://doi.org/10.1016/j.esmoop.2025.105918. In this article, we report results from the final analysis of the main part of the study.
The study included a screening period lasting up to 28 days, a treatment period consisting of 21-day cycles of berzosertib–topotecan combination with regular tumor assessments until disease progression or other criteria for treatment discontinuation were met, an end-of-treatment visit, a safety follow-up period of 30 days after the last treatment dose, and survival follow-up visits every 12 weeks from the time of last treatment dose.
The study was approved by the institutional review board or independent ethics committee of each center and was conducted in accordance with the Declaration of Helsinki, International Conference on Harmonization Good Clinical Practice, local laws, and applicable regulatory requirements. All patients provided written informed consent for participation.

Patient eligibility
Eligible patients were ≥18 years of age and had histologically confirmed SCLC; an ECOG PS ≤2; radiologically confirmed progression after first-line or chemoradiation platinum-based therapy, with or without immunotherapy, for limited or extensive-stage SCLC, with a PFI of <90 days; measurable disease per the Response Evaluation Criteria in Solid Tumors, version 1.1 (RECIST v1.1); and adequate hematologic, renal, and hepatic functions. Patients were excluded if they had clinically relevant uncontrolled intercurrent illness, unstable brain metastases, or other prior relevant malignant disease within the last 3 years.

Study treatment, endpoints, and assessments
Patients received berzosertib at a dose of 210 mg/m2 intravenously on days 2 and 5 of each 21-day cycle and topotecan at a dose of 1.25 mg/m2 intravenously on days 1-5 of each 21-day cycle.
The primary endpoint of the study was the ORR, defined as the proportion of patients with a confirmed complete response (CR) or partial response (PR), as assessed by an Independent Review Committee (IRC) according to RECIST v1.1. Confirmation of the response was required no sooner than 4 weeks after the initial documentation of response. Secondary endpoints included duration of response (DoR), progression-free survival (PFS), OS, safety, and biomarker assessment. DoR and PFS were assessed by the IRC according to RECIST v1.1. Adverse events (AEs) were coded according to the Medical Dictionary for Regulatory Activities (MedDRA) v26.0, and the severity of AEs was graded by the investigator using the National Cancer Institute Common Terminology Criteria for Adverse Events v5.0. The association of tumor markers with treatment responses was evaluated in the biomarker analysis.
The assessments for efficacy were conducted every 6 weeks from the first dose until week 36 and then every 9 weeks until disease progression, start of a new anticancer therapy, patient death, withdrawal of consent, loss to follow-up, or the end of the study. Tumors were assessed via computed tomography (CT) or magnetic resonance imaging (MRI) of the chest, abdomen, and pelvis, with additional anatomic areas investigated based on individual signs or symptoms. Brain imaging by MRI (preferred) or CT (with contrast) was mandatory at baseline, and thereafter as clinically indicated until disease progression or new treatment started.

PK assessment
Sparse PK sampling for population PK and exposure-response analysis were carried out. Berzosertib and its metabolite, M11, and topotecan in plasma were quantified using the validated bioanalytical method.

Biomarker assessment
Common SCLC oncogenes, DDR genes, tumor suppressors and oncogenes, and ATR candidate biomarkers along with selected individual genes, RNA expression signatures, and protein markers in tumor tissue were chosen as the response biomarkers. Formalin-fixed paraffin-embedded tumor tissue samples were analyzed using whole-exome sequencing and RNA sequencing via next-generation sequencing (NGS) at Personalis (Fremont, CA), and analysis of cell-free DNA (cfDNA) from blood samples was carried out at Guardant Health, Inc. (Redwood City, CA).
Truncating loss-of-function mutations as well as the homozygous loss of a gene locus were used to define the loss-of-function status for a given biomarker gene. The associations of the biomarkers’ loss-of-function status and disease stabilization [PR + stable disease (SD)] as clinical response were assessed using Fisher’s exact tests. The association to time-to-event data (PFS, OS) were assessed with log-rank tests. Differences in the expression of gene sets in the responder and non-responder groups were assessed at the level of single-sample gene set enrichment analysis scores as implemented in the GSVA R package.44 Individual gene expression differences as well as differences in expression levels of gene sets were assessed for their association to clinical response using rank sum tests.
Four gene expression signatures were used to assess the association of high replicative stress and clinical response through the Biocarta pathway (ATR-BRCA pathway), Pathway Interaction Database (ATR pathway), Reactome (activation of ATR in response to replication stress),45,46 and a pathway defined by Jo et al.47

Statistical analysis
The study had a total planned sample size of 80 patients to be recruited in two stages (stage 1 and stage 2). An interim analysis for futility with a potential to stop the study was to be conducted after the first 40 patients had completed their second on-treatment tumor assessment (or dropped out/died prematurely). If an objective response was observed in five or fewer patients in stage 1, the study had to be stopped for futility—this allowed the study to be stopped with a probability of 79% in case the true ORR was only 10%. Stage 2 was to continue if at least 6 of the first 40 patients had a confirmed objective response.
Continuous variables were summarized with descriptive statistics (n, mean, median, standard deviation, min, max, Q1, Q3) and categorical variables with frequencies and percentages. The ORR was assessed together with a two-sided Clopper–Pearson 95% CI. DoR, PFS, and OS were analyzed using the Kaplan–Meier method. All treated patients were included in both efficacy and safety analysis sets.

Results
At the planned interim analysis for futility (data cut-off: 14 April 2022), the efficacy threshold was not reached. Consequently, the Safety Monitoring Committee recommended stopping the study and the total enrollment target of 80 patients was not met. Here we report the results of the final analyses (data cut-off: 10 July 2023).

Patient characteristics and disposition
Between 22 March 2021 and 12 June 2023, 73 patients were enrolled and treated in the study (Figure 1). The median age of the patients was 62 years (range 44-78 years), with 37% aged ≥65 years, and most of the patients (78.1%) were male. At baseline, 53.4% of the patients had liver metastases, 15.1% had brain metastases, and 69.9% had an ECOG PS of 1. In alignment with the study eligibility requirements, all patients received prior platinum-based therapy, with 25 (34.2%) patients being platinum-refractory (PFI <30 days). The majority of patients (61.6%) received immunotherapy as prior therapy. Table 1 summarizes the baseline characteristics of the patients. In terms of exposure, the median number of cycles administered for both berzosertib and topotecan was 3, with a range from 1 to 19 cycles.

Efficacy
Four patients achieved confirmed responses, all of which were PRs, resulting in an ORR of 5.5% (95% CI 1.5% to 13.4%) (Supplementary Figure S1, available at https://doi.org/10.1016/j.esmoop.2025.105918). The disease control rate, defined as the proportion of patients who achieved a confirmed CR, PR, or SD, was 31.5% (95% CI 21.1% to 43.4%), including the 4 patients with a PR and 19 patients with SD (n = 23). The four patients with PR had DoR of 2.1, 2.6, 2.8, and 8.5 months. The median PFS was 2.2 months (95% CI 1.5-2.7 months) (Figure 2A). The 6-month PFS rate was 10.0% (95% CI 4.2% to 18.9%). There were no major differences in the median PFS time among the subgroups analyzed, namely, age (<65 years versus ≥65 years), gender, race, ethnic origin, country (North America, Europe, Asia), ECOG PS status (≤1 versus >1), disease stage at initial diagnosis (limited versus extensive), prior immunotherapy, presence or absence of brain/liver metastases, and number of organs with metastatic disease (<3 versus ≥3). The median duration of follow-up for OS was 15.4 months, and the median OS was 6.4 months (95% CI 4.2-7.6 months) (Figure 2B). The 12-month OS rate was 17.7% (95% CI 9.7% to 27.5%).

Safety and tolerability
All patients were reported to have experienced at least one treatment-emergent AE (TEAE) (Supplementary Table S1, available at https://doi.org/10.1016/j.esmoop.2025.105918). The most common grade ≥3 TEAEs were anemia (26.0%), thrombocytopenia, and neutropenia (20.5% each), while the most common grade ≥4 TEAEs were neutropenia (12.3%) and thrombocytopenia (9.6%). Treatment-related TEAEs were observed in 67 patients (91.8%), with grade ≥3 treatment-related TEAEs being reported in 45 patients (61.6%) (Table 2). Serious TEAEs were reported in 31 patients (42.5%) (Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2025.105918). Treatment-related TEAEs led to a berzosertib dose reduction in 12 patients (16.4%) and a topotecan dose reduction in 20 patients (27.4%). Treatment-related TEAEs led to permanent discontinuation of berzosertib in four patients (5.5%) and topotecan in five patients (6.8%). TEAEs led to death in six patients (8.2%), and these included disease progression, multiple organ dysfunction syndrome, coronavirus disease 2019 (COVID-19), pneumococcal pneumonia, sepsis, dyspnea, and respiratory failure. One patient (1.4%) experienced a treatment-related death, in association with multiple related fatal TEAEs (sepsis, pneumococcal pneumonia, and COVID-19).

PK analysis
PK parameters were calculated only in Japan safety run-in part (Supplementary Figure S2, available at https://doi.org/10.1016/j.esmoop.2025.105918), and the results are provided in the Supplementary Material, available at https://doi.org/10.1016/j.esmoop.2025.105918. Since samples were collected sparsely from the main study, no further analysis was carried out.

Biomarker analyses
Forty tumor samples were available for biomarker analyses, from which 38 were archival samples and 2 were from new biopsies. For the biomarker assessments, 9 of the 40 available samples had usable messenger RNA expression data. The association of gene expression levels with clinical response was therefore limited to nine patients of whom four had achieved SD. None of the assessed samples were from patients who had achieved a response. Tumor NGS data could be generated for 17 patients and cfDNA could be analyzed from 30 blood samples. Fifteen tumor biopsy samples could be analyzed for their protein levels by immunofluorescence. The mutation frequencies observed for TP53 and RB1 in blood samples were in the expected range for the setting48: 97% for TP53, 53% for RB1, and 50% for TP53 + RB1. The frequencies in the tumor samples were not representative given the smaller sample size. Overall, due to the limited number of samples with differential biomarker status and a low rate of responses, no firm conclusions for predictive or prognostic power of individual biomarkers could be drawn. Still, regarding the RNA expression, we note that all four patients with SD who were included in the biomarker analysis showed a high expression of either ASCL1 or NEUROD1, two well-known neuroendocrine-specific transcription factors. No signs of higher replication stress in those four patients have been found based on analysis of four replication stress gene expression signatures.

Discussion
The combination of berzosertib and topotecan did not reach the predefined efficacy threshold at the interim analysis for futility, which evaluated the first 40 patients treated. In the final analysis that included 73 patients, the ORR was 5.5%. Overall, the combination of berzosertib plus topotecan was not associated with any new safety concerns, and the safety profile was consistent with the known risk profile of the respective monotherapies.
Despite the small sample size in the biomarker assessments and it being an exploratory analysis in this study, we did observe limited clinical benefit in patients whose tumors expressed high levels of neuroendocrine-specific transcription factors. This was consistent with a previous study in a similar patient population.31 High neuroendocrine differentiation, which is characterized by enhanced replication stress, might therefore be a biomarker of tumors more likely to respond to ATR and TOP1 inhibition.
Previously, in the phase I portion of a phase I-II study of berzosertib–topotecan combination, encouraging clinical activity was observed in three patients with platinum-resistant, relapsed SCLC who derived a durable clinical benefit (one PR and two prolonged SD; DoR of 7, 6, and 10 months, respectively).43 In the subsequent single-arm phase II portion of the same study in patients with relapsed SCLC, the berzosertib–topotecan combination demonstrated a confirmed ORR of 36% (9 of 25 patients). Responses were reported in patients with both platinum-sensitive [ORR 60.0% (3/5 patients; 95% CI 14.7% to 94.7%)] and platinum-resistant disease [ORR 30.0% (6/20 patients; 95% CI 11.9% to 54.3%)].31 In a recent randomized phase II study that included 60 patients with platinum-resistant (<90 days PFI) or -sensitive, relapsed SCLC, berzosertib–topotecan combination did not increase PFS as compared with topotecan, but was associated with a significantly prolonged OS.49 Differences in the proportion and assessment of patients with platinum-resistant or -sensitive disease, the number of prior therapies, and the study setting and selection of patients (study with few specialized cancer centers versus a multicenter global study) may explain some of the differences in outcomes observed between these studies and the DDRiver SCLC 250 study. The proportion of patients with platinum-refractory disease (34.2%) may have also contributed to the worse outcomes observed in our study. Additionally, the observed differences could simply reflect the role of chance in small studies that was not similarly reproduced in our study.
A review of clinical trials and a meta-analysis evaluating topotecan monotherapy in patients with relapsed SCLC reported ORRs ranging from 5% to 27%, median PFS between 2.6 and 4.3 months, and OS durations ranging from 3.7 to 12.5 months.15,50 In the platinum-resistant setting, where data are scarcer and not homogenously reported, the ORR is as low as 3% and consistently below 10%, with a median PFS of 2.8 months and an OS of 5.7 months.15,18,50,51 In our study, the confirmed ORR was 5.5%, median PFS was 2.2 months, and median OS duration was 6.4 months in patients with platinum-resistant relapsed SCLC. Taken together, these data suggest that our results were comparable with historical data for topotecan monotherapy, further indicating that compared with topotecan monotherapy, berzosertib–topotecan combination therapy offers no improvement in efficacy.
SCLC remains a disease with a dismal prognosis, with few effective therapeutic options, particularly in the relapsed setting, stressing the need for further research and novel therapies. Despite the lack of OS improvement in the ATLANTIS study,27 additional confirmatory clinical studies with lurbinectedin in combination with agents other than doxorubicin are ongoing.52, 53, 54 A phase III study of lurbinectedin in combination with atezolizumab in the first-line maintenance setting has completed enrollment and results have not been published yet,52 while another phase III study in combination with irinotecan in relapsed SCLC is ongoing.53,54 Berzosertib showed synergy with lurbinectedin by abrogating the S-phase arrest induced by lurbinectedin and forcing cell cycle progression, leading to mitotic catastrophe in multiple SCLC cell line, organoid, and in vivo models.20 This combination is currently being assessed in a phase I/II study in patients with SCLC and high-grade neuroendocrine cancers (NCT04802174).55
A limitation of berzosertib is its intravenous administration, which adds complexity in clinical practice and might not yield an ideal exposure for a more sustained ATR inhibition that is potentially required to stall tumor growth. Tuvusertib is a novel, orally administered, potent ATR inhibitor that is currently in phase II clinical development.56,57 Given that tuvusertib achieves the desired exposure for continuous inhibition of ATR over the dosing intervals, with a tolerable safety profile,56 its development has been prioritized over berzosertib. In conclusion, while the combination of berzosertib and topotecan failed to yield the desired results in our study, other ATR inhibitors and/or DDR inhibitors that can potentiate the effects of chemotherapy, radiotherapy, and immunotherapy are being actively explored in both preclinical and clinical settings.58, 59, 60, 61, 62