OS was defined as the time from randomization to death from any cause. Participants alive at the time of the analysis were censored at the date when they were last known to be alive as documented by the investigator. Kaplan-Meier method was used to estimate the median. 95% CI for median was computed using the method of Brookmeyer and Crowley.

The ITT population included all randomized participants, whether or not the participant received the assigned treatment.

PFS was defined as the time from randomization to the first occurrence of PD, as determined by the investigator according to response evaluation criteria in solid tumors version 1.1 (RECIST v1.1), or death from any cause (whichever occurred first). PD was defined as at least a 20% increase in the sum of the longest diameters of target lesions, taking as reference the smallest sum of the longest diameters of the target lesions recorded since the treatment started, including screening, or the appearance of one or more new lesions. In addition, the sum of diameters also demonstrated an absolute increase of ≥ 5 millimeters (mm). Participants who were alive and did not experience PD at the time of analysis, were censored on the date of last tumor assessment. Participants with no post-baseline tumor assessment were censored at the date of randomization. Kaplan-Meier method was used to estimate the median. 95% CI for median was computed using the method of Brookmeyer and Crowley.

The ITT population included all randomized participants, whether or not the participant received the assigned treatment.

Confirmed ORR was defined as the percentage of participants with a complete response (CR) or partial response (PR) on two consecutive occasions ≥4 weeks apart, as determined by the investigator according to RECIST v1.1. CR= disappearance of all target lesions. In addition, any pathological lymph nodes must have a reduction in short axis to < 10 mm. PR= at least a 30% decrease in the sum of diameters of all target lesions, taking as reference the baseline sum of diameters. 95% CIs for rates were constructed using the Clopper-Pearson method. Percentages have been rounded off.

The ITT population included all randomized participants, whether or not the participant received the assigned treatment.

DOR for participants with confirmed ORR=time from first documented objective response to PD, as determined by investigator per RECIST v1.1, or death from any cause (whichever occurred first). PD=≥ 20% increase in sum of longest diameters of target lesions, taking as reference smallest sum of longest diameters of target lesions recorded since treatment started, including screening, or appearance of new lesions. In addition, sum of diameters also demonstrated an absolute increase of ≥ 5 mm. Confirmed ORR=percentage of participants with a CR or PR on two consecutive occasions ≥4 weeks apart, as determined by investigator per RECIST v1.1. CR= disappearance of all target lesions. PR=≥ 30% decrease in sum of diameters of all target lesions. Participants who had not progressed and who did not die at the time of analysis were censored at the time of last tumor assessment date. Kaplan-Meier method was used to estimate median. 95% CI for median was computed using Brookmeyer and Crowley method.

Participants in the ITT population who had a confirmed objective response (CR or PR) as determined by the investigator per RECIST v1.1 were included in the analysis.

TTCD analyses was performed for patient-reported PF (items 1 to 5) of European Organisation for Research and Treatment of Cancer Quality of Life-Core 30 Questionnaire (EORTC QLQ-C30). The PF was measured on 4-point scale (1='Not at all' to 4='Very much'). TTCD for PF was defined as time from date of randomization to first confirmed clinically meaningful decrease from baseline in PF score held for at least 2 consecutive assessments or initial clinically meaningful decrease from baseline followed by death from any cause within 21 days or until next tumor assessment, whichever occurs first. A score change of ≥ of 10-point on EORTC QLQ-C30 PF scale was determined as being clinically meaningful. Scores were averaged, transformed to 0-100 scale; where higher score represented high/healthy level of functioning. Kaplan-Meier method was used to estimate the median. 95% CI for median was computed using the method of Brookmeyer and Crowley.

The ITT population included all randomized participants, whether or not the participant received the assigned treatment. Participants without a confirmed deterioration at the time of analysis were censored at the last time they were known to have not deteriorated.

TTCD analyses was performed for GHS and quality of life (QoL) (items 29 and 30) of EORTC QLQ-C30. GHS/ QoL items were scored on a 7-point scale that ranges from "very poor" to "excellent." TTCD for GHS/QoL was defined as the time from the date of randomization to the first confirmed clinically meaningful decrease from baseline in GHS/QoL score held for at least two consecutive assessments or an initial clinically meaningful decrease from baseline followed by death from any cause within 21 days or until the next tumor assessment, whichever occurs first. A score change of ≥ 10-point on the EORTC QLQ-C30 GHS/QoL scale was determined as being clinically meaningful. Scores were averaged, transformed to 0-100 scale; where higher score for GHS/QoL= better health-related QoL. Kaplan-Meier method was used to estimate the median. 95% CI for median was computed using the method of Brookmeyer and Crowley.

Participants without a confirmed deterioration at the time of analysis were censored at the last time they were known to have not deteriorated. The ITT population included all randomized participants, whether or not the participant received the assigned treatment. Only responders were analysed in this endpoint.

PFS rates were defined as the percentage of participants alive and without PD as assessed by the investigator according to RECIST v1.1 at 6 months and 1 year after randomization. PD was defined as at least a 20% increase in the sum of diameters of target lesions, taking as reference the smallest sum of diameters at prior timepoints, including baseline. In addition, the sum of diameters also demonstrated an absolute increase of ≥ 5 mm. Kaplan-Meier method was used to estimate the median. 95% CI for median was computed using the method of Brookmeyer and Crowley. Percentages have been rounded off.

The ITT population included all randomized participants, whether or not the participant received the assigned treatment.

OS rates were defined as the percentage of participants who were alive at 1 and 2 years. Participants alive at the time of the analysis were censored at the date when they were last known to be alive as documented by the investigator. Kaplan-Meier method was used to estimate the median. 95% CI for median was computed using the method of Brookmeyer and Crowley. Percentages have been rounded off.

The ITT population is defined as all randomized participants, whether or not the participant received the assigned treatment. Participants alive at the time of the analysis were censored at the date when they were last known to be alive as documented by the investigator. At the time of the analysis, there were no participants with 24 months or more of survival follow-up, therefore, survival rate at the 2-year timepoint was not estimable.

An AE was defined as any untoward medical occurrence in a clinical investigation patient administered a pharmaceutical product, regardless of causal attribution. An AE can therefore be any of the following: unfavorable and unintended sign (including an abnormal laboratory finding), symptom, or disease temporally associated with the use of a medicinal product, whether or not considered related to the medicinal product; any new disease or exacerbation of an existing disease; recurrence of an intermittent medical condition not present at baseline; any deterioration in a laboratory value or other clinical test; AEs related to a protocol-mandated intervention. AEs were assessed using National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0 (NCI CTCAE, v5.0). Percentages have been rounded off.

Safety-evaluable population included all randomized participants who had received any amount of study drug, with participants grouped according to the actual treatment received.

Pharmacokinetic (PK)-evaluable population for atezolizumab included all participants who had received any dose of atezolizumab and who had evaluable PK samples. Number analyzed is the number of participants with data available for analysis at the specified timepoint. Different participants may have contributed data for each timepoint.

PK-evaluable population included all participants who had received any dose of atezolizumab and who had evaluable PK samples.

PK-evaluable population included all participants who had received any dose of cabozantinib and who had evaluable PK samples. Number analyzed is the number of participants with data available for analysis at the specified timepoint. Different participants may have contributed data for each timepoint.

Cmax was not collected for this outcome measure because PK of cabozantinib was well characterized through the cabozantinib development for mono- therapy. An established population PK model for cabozantinib is available for the PK data from NCT04471428 (study GO41892). The PK data collected in the current study is sufficient for the population PK model to characterize the exposure of cabozantinib in this study.

Participants were considered to be ADA positive if they were ADA negative or had missing data at baseline but developed an ADA response following study drug exposure (treatment-induced ADA response), or if they were ADA positive at baseline and the titer of one or more post-baseline samples was at least 0.60 titer unit greater than the titer of the baseline sample (treatment-enhanced ADA response). The total number of participants who developed ADAs to atezolizumab was determined by summing the ADA-positive participants across all timepoints.

Safety-evaluable population included all randomized participants who had received any amount of study drug, with participants grouped according to the actual treatment received. Overall number analyzed is the number of participants with data available for analysis.