This section focuses on the efficacy of durvalumab in combination with etoposide plus either carboplatin or cisplatin in adults with untreated ES-SCLC, as evaluated in the open-label (unmasked to investigators and patients, but not to sponsor) multicentre, phase III CASPIAN trial (Fig. 1) .
Supplementary file1 Video (MP4 8106 kb)
CASPIAN enrolled patients (n = 805) aged ≥ 18 years (≥ 20 years in Japan) with treatment-naïve histologically or cytologically documented ES-SCLC, a WHO performance status score of 0 or 1, measurable disease according to Response Evaluation Criteria in Solid Tumours (RECIST) [version 1.1], suitability for first-line platinum-based chemotherapy, life expectancy of ≥ 12 weeks from the study start, and adequate organ and marrow function. Patients with active or previous autoimmune or inflammatory disorders; uncontrolled, concurrent illness or active infections; or a history of radiotherapy to the chest or planned consolidation chest radiotherapy were excluded .
Randomization was stratified according to planned platinum (carboplatin or cisplatin) . Eligible patients were randomized in a 1:1:1 ratio (in blocks of six) to receive durvalumab (1500 mg) plus chemotherapy [which consisted of etoposide 80–100 mg/m2 (administered on days 1–3 of each 21-day cycle) plus the investigator’s choice of either carboplatin area under the curve 5–6 mg/mL/min or cisplatin 75–80 mg/m2 (administered on day 1 of each cycle)], durvalumab plus tremelimumab (75 mg) plus chemotherapy, or chemotherapy alone (up to 6 cycles every 3 weeks plus optional prophylactic cranial irradiation after discontinuation of chemotherapy). Patients treated with immunotherapy received four cycles of platinum-etoposide plus durvalumab, with or without tremelimumab, every 3 weeks followed by maintenance durvalumab every 4 weeks. All drugs were administered intravenously. Treatment continued until unacceptable toxicity, other discontinuation criteria were met, or disease progression, although continuation was permitted after disease progression provided there was evidence of clinical benefit .
The two primary endpoints were OS, assessed in the intent-to-treat population, for durvalumab plus chemotherapy versus chemotherapy alone and for durvalumab plus tremelimumab plus chemotherapy versus chemotherapy alone . Secondary endpoints included PFS, objective response rate (ORR; unconfirmed) [both investigator-assessed according to RECIST version 1.1], OS rate at 18 months, PFS rates at 6 months and 12 months, and safety. A hierarchical multiple-testing procedure with a gatekeeping strategy was used across the primary OS analyses and secondary PFS analyses (i.e. PFS was only tested if both OS primary analyses achieved significance), and the study was considered positive if either of the OS analysis results achieved statistical significance .
At the time of the planned interim analysis (data cut-off 11 March 2019), OS with durvalumab plus chemotherapy versus chemotherapy alone met the prespecified threshold for statistical significance (Table 1); therefore, the independent data monitoring committee recommended unmasking these groups to the sponsor and the interim analysis was considered to be the final result in terms of formal statistical testing for this comparison, despite ongoing follow-up . Discussion in this section will be limited to the approved combination of durvalumab plus chemotherapy, with no further discussion of the findings for the combination of durvalumab plus tremelimumab plus chemotherapy, which did not meet the predefined statistical significance threshold at the time of either the interim analysis or the updated (final) analysis [11, 12].
At baseline, demographics and disease characteristics were generally similar between the durvalumab plus chemotherapy (n = 268) and chemotherapy (n = 269) treatment groups . Across both groups, the majority of patients were male (70%), White (84%), aged < 65 years (60%; median age 63 years), and had stage IV disease (90%) and a WHO performance status of 1 (65%). A total of 10% of patients had brain or CNS metastases and 39% of patients had liver metastases .
At the data cut-off for the interim analysis, the median duration of follow-up for OS was 14.2 months and 336 deaths had occurred (155 and 181 events in the durvalumab plus chemotherapy and chemotherapy alone groups; 62.6% maturity) . The addition of durvalumab to chemotherapy with platinum-etoposide in the first-line treatment of ES-SCLC significantly extended OS compared with chemotherapy alone with a hazard ratio (HR) of 0.73 (95% CI 0.59–0.91; p = 0.0047); the median OS was 2.7 months longer in the durvalumab plus chemotherapy group than in the chemotherapy alone group (Table 1) . An OS benefit of durvalumab plus chemotherapy over chemotherapy alone (i.e. HR < 1) was also observed across all prespecified subgroups based on planned platinum, age, sex, WHO performance status, smoking status, brain or CNS metastases, disease stage at diagnosis, race and region. The 12- and 18-month OS rates with durvalumab plus chemotherapy were 54% and 34% (vs. 40% and 25% with chemotherapy alone) .
At the time of the interim analysis, 226 and 233 patients in the durvalumab plus chemotherapy and chemotherapy alone groups had disease progression or died . Median PFS was similar across the durvalumab plus chemotherapy and chemotherapy alone groups, but the risk of progression or death was 22% lower with durvalumab plus chemotherapy than with chemotherapy alone (Table 1) . PFS could not be formally tested for significance within the multiple-testing procedure at the interim analysis because of the hierarchical design of CASPIAN. The 6- and 12-month PFS rates in the durvalumab plus chemotherapy group were 45% and 18% (vs. 46% and 5% in the chemotherapy alone group) .
A higher proportion of patients in the durvalumab plus chemotherapy group had an unconfirmed objective response compared with the chemotherapy alone group (Table 1) . Amongst patients who had a confirmatory scan no sooner than 4 weeks after the initial response, the confirmed ORR was 68% and 58% in the durvalumab plus chemotherapy and chemotherapy alone groups; the median duration of confirmed response was 5.1 months in each group. In patients with a confirmed response in the durvalumab plus chemotherapy and chemotherapy alone groups, the estimated percentages of patients remaining in response at 6 months were 39% and 34%, and at 12 months were 23% and 6% .
In a preplanned subgroup analysis of patients recruited in Japan (n = 34; median follow-up 12.5 months), the addition of durvalumab to chemotherapy (n = 18) compared with chemotherapy alone (n = 16) was associated with an HR for OS of 0.77 [(95% CI 0.26–2.26); median OS not reached vs. 15.2 months] and an estimated OS rate at 12 months of 72.2% versus 60.2% . In the durvalumab plus chemotherapy and chemotherapy alone groups, median PFS was 4.5 and 4.7 months (HR 0.90; 95% CI 0.43–1.89) and the confirmed ORR was 89% and 69% (odds ratio 3.64; 95% CI 0.65–28.75) .
The addition of durvalumab to platinum-etoposide in CASPIAN improved survival outcomes without detrimentally impacting patient-reported outcomes (PROs), as assessed using the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (QLQ-C30) [version 3] and its lung cancer module, Quality of Life Questionnaire-Lung Cancer 13 (QLQ-LC13) . For the key prespecified disease-related symptoms [cough, dyspnoea and chest pain (as measured by QLQ-LC13) and fatigue and loss of appetite (QLQ-C30)], numerical reductions in the symptom burden over 12 months or until disease progression were observed in both the durvalumab plus chemotherapy and chemotherapy alone groups. A between-group difference (nominal p = 0.009) in favour of durvalumab plus chemotherapy was evident for the improvement of appetite loss from baseline. The time to deterioration for global health status/quality of life (QoL), all functioning scales (cognitive, emotional, physical, role and social) and all QLQ-C30 and QLQ-LC13 symptom scales was longer (i.e. HR < 1) in the durvalumab plus chemotherapy group than in the chemotherapy alone group .
At the data cut-off for the updated analysis (27 January 2020), the median duration of follow-up for OS was 25.1 months and 441 deaths had occurred across the two treatment groups (210 and 231 events in the durvalumab plus chemotherapy and chemotherapy alone groups; 82.1% maturity) . The updated analysis of OS with 11 additional months of follow-up indicated a sustained OS improvement with durvalumab plus chemotherapy versus chemotherapy alone that was consistent with the results of the interim analysis (Table 1) . An OS benefit with durvalumab plus chemotherapy versus chemotherapy alone (i.e. HR < 1) was consistently observed across all prespecified subgroups , as well as post hoc subgroups defined by the extent of disease (thoracic only or any extra-thoracic disease  and presence or absence of liver metastases ) at baseline. The 24-month OS rate was higher in the durvalumab plus chemotherapy group (22.2%) than in the chemotherapy alone group (14.4%) .
At the time of the updated analysis, median PFS was similar in the durvalumab plus chemotherapy and chemotherapy alone groups (Table 1); the 24-month PFS rate was higher in the durvalumab plus chemotherapy group (11.0%) than in the chemotherapy alone group (2.9%) . Consistent with the interim analysis, the unconfirmed ORR was higher with durvalumab plus chemotherapy than with chemotherapy alone (Table 1) and the confirmed ORR was 68% and 58%, respectively. The estimated proportion of patients in the durvalumab plus chemotherapy group remaining in response at 24 months was 13.5% (vs. 3.9% in the chemotherapy alone group) .