A total of 21 patients were screened and 14 were enrolled; eight (including two patients who were not evaluable for DLTs) and six patients were assigned to cohort 1 (7-day regimen) and cohort 2 (14-day regimen), respectively (Fig. 1). A summary of baseline characteristics and demographics is given in Table 1. All patients were Chinese; median age was 56.0 years (range 39‒74 years); 9/14 patients were female (64.3%). Most patients had lung cancer (n = 9; adenocarcinoma [n = 6], squamous cell carcinoma [n = 2], small cell lung cancer [n = 1]). One patient in cohort 1 had stage III disease, and all other patients had stage IV disease. Patients were not selected for PD-L1 status, and all patients had received at least three prior lines of systemic anticancer treatment; 12/14 patients (85.7%) had received prior immunotherapy. Of the patients who had received prior immunotherapy, 6/12 (50.0%) had primary resistance, 2/12 (16.7%) developed secondary resistance to prior anti-PD-(L)1 therapy, and one patient had early relapse; status was not known for 3/12 (25.0%) patients.

Study flow. *One patient did not receive durvalumab as a result of the COVID-19 pandemic. AE adverse event, BID twice daily

At data cutoff (20 October 2023), seven patients (50.0%) had completed the study (Fig. 1). Of the remaining patients, two had died (14.3%, one patient in each cohort) and five patients were ongoing in the study (35.7%). At data cutoff, two patients were still receiving treatment in cohort 1 and six patients had discontinued (AEs, n = 2; progressive disease, n = 3; investigator decision, n = 1). In cohort 2, one patient was still receiving both treatments, and five patients had discontinued (progressive disease, n = 4; investigator decision, n = 1).

The median duration of ceralasertib treatment was 5.2 months (range 0.2‒12.2 months) for cohort 1 and 2.3 months (range 1.4‒6.9 months) for cohort 2; median duration of treatment with durvalumab was 7.5 months (range 0.9‒12.1 months) for cohort 1 and 1.9 months (range 0.9–6.9 months) for cohort 2. The median relative dose intensity of ceralasertib and durvalumab is given in Table S1 in the electronic supplementary material.

In total, most patients (13/14; 92.9%) had at least one AE; grade ≥ 3 AEs occurred in 7/14 patients (50.0%) (Table 2). Twelve patients (six in each cohort) were evaluable for DLTs; no DLTs were observed and no AEs leading to death were reported. One patient in cohort 1 discontinued both study treatments due to an AE of ascites, which was not considered to be related to either study drug. Dose interruptions due to AEs occurred in eight patients (cohort 1, n = 5; cohort 2, n = 3); ceralasertib treatment was interrupted as a result of AEs in five patients and durvalumab treatment was interrupted due to AEs in three patients. The most common AEs across both cohorts were anemia (11/14, 78.6%), platelet count decreased (8/14, 57.1%), white blood cell count decreased (7/14, 50.0%), and hyperglycemia (6/14, 42.9%) (Table 3). Individual patient plots for hematological laboratory analyses are given in Figs. S1 and S2 in the electronic supplementary material. The most common grade ≥ 3 AEs occurring in ≥ 10% of patients across both cohorts were anemia (5/14, 35.7%), lymphocyte count decreased (2/14, 14.3%), thrombocytopenia (2/14, 14.3%), and white blood cell count decreased (2/14, 14.3%) (Table 3). In total, 6/14 patients experienced SAEs (42.9%, n = 3 in each cohort); a summary of treatment-related SAEs is given in Table S2 in the electronic supplementary material. One patient in cohort 1 had an immune-mediated AE of hypothyroidism (grade 1).

During the ceralasertib monotherapy lead-in (i.e., cycle 0), 11/14 patients (78.6%) experienced an AE; grade ≥ 3 AEs occurred in 1/14 patients (7.1%). The most common AEs were nausea (4/14, 28.6%) and vomiting (4/14, 28.6%) (Table S3 in the electronic supplementary material).

Following single-dose administration of , concentrations were detectable for 11 h, i.e., the last sampling timepoint within the dose interval following the first dose (Figs. S3a, b in the electronic supplementary material). was detectable in all patients following multiple-dose administration, including up to 72 h post-dose (Figs. S3c, d in the electronic supplementary material). Additional PK results are given in the Supplementary Results.

At data cutoff of 20 October 2023, 13 patients were evaluable for antitumor activity, and three patients had a confirmed partial response (PR); ORR was 23.1% (3/13, 80% CI 8.8‒44.4%). Among the seven response-evaluable patients in cohort 1 (one patient was excluded due to the absence of measurable disease at baseline), two patients had insufficient exposure to the investigational drugs; one patient had no durvalumab exposure during the study and one patient had insufficient ceralasertib exposure (i.e., less than 75%) in cycle 0. The two patients with PRs were in cohort 1 (2/7 patients. ORR, 28.6%; 80% CI 7.9‒59.6%); one patient had lung adenocarcinoma and the other patient had squamous cell carcinoma (SCC) of the cervix uteri. The third patient with PR had lung squamous cell carcinoma and was in cohort 2 (1/6 patients. ORR 16.7%; 80% CI 1.7‒51.0%). The responder with cervical cancer had acquired resistance and the other two responders had primary resistance to prior PD-(L)1 immunotherapies. Four patients (30.8%) had stable disease for ≥ 6 months.

In cohort 1, DoR was 5.7 months in the patient with SCC of the cervix uteri, and 9.4 months in the patient with lung adenocarcinoma. In cohort 2, DoR was 2.7 months for the patient with lung squamous carcinoma. Median PFS in cohort 1 was 5.6 months (95% CI 2.0‒not calculable [NC]); in cohort 2, median PFS was 2.7 months (95% CI 0.8‒NC).

The best percentage change in target lesion size and duration of response is shown in Figs. 2 and S4 (in the electronic supplementary material). Of the seven patients who were evaluable for response in cohort 1, one patient with SCC of the cervix uteri had a reduction in the size of the target lesion of 30% and a second patient with lung adenocarcinoma had a reduction of 80%. Of the six evaluable patients in cohort 2, one patient with lung squamous carcinoma had a reduction in the size of the target lesion of 50%.

Duration of response (a) and percentage change from baseline in (b) cohort 1 and (c) cohort 2. *Patients received insufficient dose exposure and were not DLT evaluable. DLT dose-limiting toxicity, NSCLC non-small cell lung cancer, SCC squamous cell carcinoma, SCLC small cell lung cancer