PFS was defined as time from randomization until disease progression (PD) or death from any cause, whichever occurs first for all participants. Progression was based on investigator assessment of disease response per Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.

PD: radiologic detection of ≥ 20% increase in the sum of diameters of target lesions and ≥ 5 mm absolute increase, or unequivocal progression of non-index lesions, or the presence of new lesions. 95% CIs were calculated using the Brookmeyer and Crowley method.

The EORTC QLQ-C30 was developed to assess quality of life in cancer participants across tumor types. It was a self-reported, 30-item generic instrument that assessed five functional scales (physical, role, emotional, cognitive, social); nine disease symptom items (fatigue, nausea and vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea, financial difficulties); and a global health status (GHS)/quality of life (QOL) scale. Questionnaire score ranged from 0 to 100. For the functional scales, higher scores indicated high/healthy level of functioning; for the disease symptom scales, higher scores indicated greater symptom burden; and for the GHS/QoL scale, higher scores represented a high QoL. A negative change from baseline indicated reduction in functioning, improvement in symptom burden and reduction in health and quality of life.

The EORTC QLQ-LC13 was a disease-specific supplement to the EORTC QLQ-C30. It included multi-item and single-item measures of lung cancer symptoms (coughing, hemoptysis, dyspnea, pain) and treatment side effects (hair loss, neuropathy, sore mouth, dysphagia). Scores ranged from 0 to 100, with higher scores indicating a greater degree of symptom severity. A negative change from baseline indicated improvement in symptom burden or side effect severity.

ORR was defined as the percentage of participants with a confirmed best overall response of complete response (CR) or partial response (PR) based on the RECIST v1.1.

CR: complete disappearance of all lesions and pathologic lymph nodes reduced in short axis to < 10 mm. PR: decrease in tumor burden of ≥ 30% relative to baseline, or complete disappearance of all index lesions with the presence of non-index lesions. 95% CIs were calculated using the Brookmeyer and Crowley method.

DCR was defined as the percentage of participants documented to have a CR, PR or stable disease (SD) per RECIST v1.1.

CR: complete disappearance of all lesions and pathologic lymph nodes reduced in short axis to < 10 mm. PR: decrease in tumor burden of ≥ 30% relative to baseline, or complete disappearance of all index lesions with the presence of non-index lesions. SD: Index lesions not meeting the criteria for CR, PR, or PD or the persistence of one or more non-index lesions. Exact 95% CIs were calculated using the Clopper-Pearson method.

DOR was defined as the time from the first documentation of OR until the first documentation of PD or death due to any cause, whichever occurred first determined by the investigator per RECIST v1.1. Only participants who had achieved OR will be evaluated for DOR.

PD: radiologic detection of ≥ 20% increase in in the sum of diameters of target lesions and ≥ 5 mm absolute increase, or unequivocal progression of non-index lesions, or the presence of new lesions.

PFS was defined as time from randomization until PD or death from any cause, whichever occurs first for all participants. Progression was based on investigator assessment of disease response per Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.

PD: radiologic detection of ≥ 20% increase in in the sum of diameters of target lesions and ≥ 5 mm absolute increase, or unequivocal progression of non-index lesions, or the presence of new lesions. 95% CIs were calculated using the Brookmeyer and Crowley method.

OS was defined as time from randomization until death from any cause. Median overall survival was estimated using Kaplan-Meier method. 95% CIs were calculated using the Brookmeyer and Crowley method.

Incidence of treatment emergent adverse events including grade ≥ 3 treatment emergent adverse events, serious treatment emergent adverse events, treatment emergent adverse events leading to treatment discontinuation, fatal treatment emergent adverse events, and treatment-related treatment emergent adverse events.

Blood samples were collected for measurement of serum concentrations of tarlatamab.

The BPI-SF (Pain) was a valid and reliable instrument, which was a commonly used tool to capture severity of pain and its impact on daily functioning. The BPI-SF measured intensity of pain at four time points (worst, least, average, and right now), pain relief, and seven interference items (general activity, mood, walking ability, normal work, relations with other people, sleep, and enjoyment of life). Scoring for pain ranged from 0 (no pain) to 10 (worst pain), and for pain interference from 0 (no interference) to 10 (complete interference). A negative change from baseline indicated a reduction in pain severity or interference.

PGIS was used to measure a patient's perceived severity of symptoms. PGIS scale asked respondents to describe the severity of their symptoms, overall status, etc, over the past week. The 4-level PGIS response options were: 1: none, 2: mild, 3: moderate, and 4: severe. Higher scores indicated severe condition. A negative change from baseline indicated an improvement in severity.

PGIC measures were used in the measurement of within-patient meaningful change (response) thresholds for disease symptoms/impacts of interest. The PGIC scale asked respondents to rate the overall change in symptoms, overall status, etc., since initiating treatment with the medication. The 5-level PGIC response options were: 1: much better, 2: a little better, 3: about the same, 4: a little worse, and 5: much worse. Higher scores indicated worsening condition. A negative change from baseline indicated an improvement in symptoms.

The EQ5D-5L questionnaire was a standardized instrument used as a measure of health outcome developed by the EuroQol group. It was comprised of a 5-dimension health status measure and a VAS. The 5-dimension health status measure evaluated: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression based on a 5-level scale: 1: no problems, 2: slight problems, 3: moderate problems, 4: severe problems, 5: an extreme problem. Higher scores (closer to 5) indicated worse health status. The VAS recorded the participant's self-rated health on a vertical VAS where the endpoints were labelled 'Worst imaginable health state' and 'Best imaginable health state'. VAS ranged from 0 to 100. Higher scores indicated better perceived health.

The EQ5D-5L questionnaire was a standardized instrument used as a measure of health outcome developed by the EuroQol group. It was comprised of a 5-dimension health status measure and a VAS. The 5-dimension health status measure evaluated: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression based on a 5-level scale: 1: no problems, 2: slight problems, 3: moderate problems, 4: severe problems, 5: an extreme problem. Higher scores (closer to 5) indicated worse health status. The VAS recorded the participant's self-rated health on a vertical VAS where the endpoints were labelled 'Worst imaginable health state' and 'Best imaginable health state'. VAS ranged from 0 to 100. Higher scores indicated better Perceived health. A negative change from baseline indicated a reduction in better perceived health.

PRO-CTCAE was a 78-item library used to measure patient-reported symptomatic adverse events. Users selected items that were most relevant to the disease, treatment profile, and fit for purpose to document patients' experience. Each symptom was rated by up to 3 attributes: presence/frequency, severity, and/or interference of the adverse event. Based on the safety profile of tarlatamab and SOC therapy, the study included the following symptoms: shivering or shaking chills, anxiety, constipation, taste changes, vomiting, headache, concentration, rash, palpitations, arm or leg swelling, nausea, dizziness, bruising, fatigue, and decreased appetite, which were deemed relevant for the site of cancer as well as cancer treatment. Responses were typically scored on a 5-point ordinal scale, with values ranging from 0 to 4. Higher scores indicated greater symptom burden.

FACT-G was a 27-item questionnaire designed to measure four domains of health-related quality of life in cancer patients: physical, social, emotional, and functional well-being. The GP5 of the FACT-G was a single item: "I am bothered by side effects of treatment", rated on a 5-point Likert scale from 0: not at all to 4: very much. Higher scores indicated greater impact of side effects. A negative change from baseline indicated improvement in impact of side effects.

OS was defined as time from randomization until death from any cause. Median overall survival was estimated using Kaplan-Meier method. 95% confidence intervals (CIs) were calculated using the Brookmeyer and Crowley method.

The ITT analysis set included all randomized participants.