A Randomized, Controlled Phase 3 Study of Cabozantinib (XL184) in Combination with Atezolizumab versus Sorafenib in Subjects with Advanced Hepatocellular Carcinoma Who Have Not Received Previous Systemic Anticancer Therapy

Liver cancer is the second most frequent cause of cancer deaths worldwide
(Ferlay et al 2015). Hepatocellular carcinoma (HCC) is the most common form of
primary liver cancer, accounting for approximately 90% of cases. In 2015,
854,000 new liver cancer cases were reported worldwide and 810,000 deaths
occurred (Global Burden of Disease Liver Cancer 2017; EASL 2018). The estimated
incidence and mortality rates of liver cancer in the USA in 2016 were
approximately 42,000 and 30,000 cases, respectively (American Cancer Society
2018); incidence and mortality in the EU in 2012 were 52,000 and 48,000 cases,
respectively (Ferlay et al 2013). Surgical resection and transplantation are
potential curative treatment modalities for HCC. Ablative therapies (eg,
radiofrequency ablation [RFA], microwave ablation [MWA], and percutaneous
ethanol injection [PEI]) are being used for early stage unresectable HCC.
Transarterial chemoembolization (TACE) is used generally for intermediate stage
disease. The current standard of care for first line treatment of advanced
unresectable HCC is sorafenib, which is a small-molecule inhibitor of vascular
endothelial growth factor receptor (VEGFR) and other protein kinases. In a
randomized placebo-controlled Phase 3 study (SHARP), sorafenib improved the
primary endpoint of overall survival (OS) in subjects with advanced HCC
(Child-Pugh A) who had not received prior systemic therapy (Llovet et al 2008).
Median OS was 10.7 months in the sorafenib arm and 7.9 months in the placebo
arm (hazard ratio [HR] 0.69; 95% confidence interval [CI] 0.55, 0.87; p-value
< 0.001). A similar HR was observed (with a shorter duration of OS than in
the SHARP trial) in a corresponding placebo-controlled Phase 3 trial conducted
in an Asian-Pacific population in which infection with hepatitis B virus (HBV)
was the main cause of HCC: median OS was 6.5 months vs 4.2 months (HR 0.68; 95%
CI: 0.50, 0.93; p-value = 0.014) (Cheng et al 2009). Recently, the
VEGFR-targeting tyrosine kinase inhibitor (TKI) lenvatinib was shown to be
non-inferior to sorafenib in a Phase 3 study enrolling subjects with advanced
HCC who had not received prior systemic therapy (Kudo et al 2018) leading to US
and EU approval in this population. In that study, median OS for lenvatinib was
13.6 months compared with 12.3 months for sorafenib (HR 0.92; 95% CI 0.79,
1.06). Cabozantinib is an orally bioavailable small molecule TKI that potently
inhibits VEGFR, MET, AXL, and RET, as well as a number of other receptor
tyrosine kinases (RTKs) that have also been implicated in tumor pathobiology,
including KIT and FLT3. Cabozantinib suppresses MET and VEGFR2 signaling,
rapidly inducing apoptosis of endothelial and tumor cells, resulting in tumor
regression in a variety of xenograft models. Cabozantinib capsules (140 mg) are
approved in the United States for the treatment of patients with progressive,
metastatic medullary thyroid cancer (MTC) and in the European Union for the
treatment of patients with progressive, unresectable locally advanced or
metastatic MTC (Cometriq* US prescribing information [US PI] and European
Medicines Agency Summary of Product Characteristics [EMA SmPC]). Cabozantinib
tablets (60 mg) are approved in the United States, Europe, and other regions
for advanced renal cell carcinoma (RCC; different patient populations depending
on region; Cabometyx* US PI and EMA SmPC). Based on the results from the
studies below, cabozantinib tablets (60 mg) as a single agent have also been
approved in the US and EU for the treatment of HCC in patients who have
previously been treated with sorafenib (Cabometyx US PI and EMA SmPC). The
clinical activity and safety of single agent cabozantinib (60 mg, tablets) in
HCC has been demonstrated in a randomized placebo-controlled Phase 3 study
(CELESTIAL) in subjects who had received prior therapy with sorafenib (subjects
were required to have progressed during or following prior systemic therapy and
up to 2 prior lines of systemic therapy were allowed; Abou- Alfa et al 2018).
The primary endpoint of the study was OS. At the second pre-planned interim
analysis, the prespecified event-driven primary efficacy endpoint analysis of
the 707 subjects enrolled at the data cutoff (470 cabozantinib, 237 placebo)
demonstrated a statistically significant improvement in OS for subjects in the
cabozantinib arm compared with placebo (Intent-to-Treat [ITT] population): the
HR, adjusted for stratification factors, was 0.76 (95% CI 0.63, 0.92;
stratified log-rank p-value = 0.0049; critical p-value to reject the null
hypothesis of equal OS = 0.021). The Kaplan-Meier estimates for median duration
of OS were 10.2 months in the cabozantinib arm vs 8.0 months in the placebo
arm. The secondary endpoint analysis of progression-free survival (PFS) as
determined by the investigator yielded a median duration of PFS of 5.2 months
in the cabozantinib arm and 1.9 months in the placebo arm. The HR, adjusted for
stratification factors, was 0.44 (95% CI 0.36, 0.52, stratified log-rank
p-value < 0.0001). Investigator-determined objective response rate (ORR) was
4% and 0.4% for subjects in the cabozantinib and placebo arms, respectively
(unstratified Fisher exact test p-value = 0.0059); all were partial responses
(PRs). In addition, there was a high rate of stable disease (SD) in the
cabozantinib arm relative to placebo (60% vs 33%). Adverse events reported for
>= 20% of subjects in the cabozantinib arm by decreasing frequency were
diarrhea, decreased appetite, palmar-plantar erythrodysesthesia (PPE), fatigue,
nausea, hypertension, vomiting, aspartate aminotransferase (AST) increased, and
asthenia. Grade 3 or 4 adverse events (AEs) regardless of causality were
reported for 68% of subjects in the cabozantinib arm and 36% in the placebo
arm. Grade 3 or 4 AEs reported for >= 5% of subjects in the cabozantinib arm by
decreasing frequency were PPE, hypertension, AST increased, fatigue, diarrhea,
asthenia, and decreased appetite. The results of this study formed the basis
for regulatory applications to the US FDA and EMA to approve cabozantinib for
treatment of patients with advanced HCC who have received prior therapy.
Earlier clinical evaluation of cabozantinib in HCC was conducted in a Phase 2
study that included both previously-treated and treatment-nai*ve subjects with
advanced HCC (n=41; Kelley et al 2017). Progression-free survival from first
dose throughout the study was estimated for all HCC subjects using a piecewise
method; median PFS was 5.2 months. Tumor regression appeared independent of
prior sorafenib exposure. Atezolizumab is a humanized immunoglobulin (Ig) G1
monoclonal antibody which potently and selectively inhibits binding of
programmed death ligand 1 (PD-L1) on tumor cells and tumor infiltrating immune
cells in the tumor microenvironment (McDermott et al 2016). Through this
interaction, atezolizumab interrupts the negative regulatory effects of PD-L1
on T-cell proliferation and function that result from PD-L1 binding to
programmed death receptor 1 (PD-1) and B7.1 (CD80) expressed on T lymphocytes
and other immune cells. The result is an increase in the susceptibility of
tumor cells to T-cell-mediated immune response, an effect that has been
demonstrated in clinical activity across several tumor types. Atezolizumab
injection, for intravenous (IV) use (1200 mg once every 3 weeks [q3w]), has
been approved in the United States and the European Union for the treatment of
patients with advanced urothelial carcinoma (UC) after prior platinum
containing chemotherapy or in a subset patients who are considered
cisplatin-ineligible (different patient populations are indicated depending on
region; Rosenberg et al 2016, Balar et al 2017). Atezolizumab is also approved
for patients with locally advanced or metastatic non-small cell lung cancer
(NSCLC) after prior chemotherapy (Fehrenbacher et al 2016; Tecentriq* US PI and
EMA SmPC). Recently, atezolizumab was also granted accelerated approval in the
US for treatment in combination with paclitaxel proteinbound for adult patients
with unresectable locally advanced or metastatic triple negative breast cancer
(TNBC) whose tumors express PD-L1 (Schmid et al 2018) and was also approved for
first-line treatment in combination with carboplatin and etoposide in adult
patients with extensive-stage small cell lung cancer (ES-SCLC; Horn et al 2018,
Tecentriq US PI). Treatment with atezolizumab is generally well-tolerated but
can be associated with immune-related adverse events (irAEs). The clinical
activity and safety of atezolizumab has been evaluated in subjects with
advanced HCC either as single agent or in combination therapy. In a Phase 1b
study of atezolizumab (1200 mg q3w) in combination with the anti-VEGF targeting
antibody bevacizumab, 103 subjects with advanced HCC nai*ve to systemic therapy
had been enrolled at the data cutoff of 26 July 2018 (NCT02715531; Pishvaian et
al 2018). Among 73 efficacy-evaluable subjects, the median survival follow-up
was 7.2 months. The ORR by independent radiology facility (IRF) was 27% (with 4
complete responses [CRs]) per Response Evaluation Criteria in Solid Tumors
version 1.1 (RECIST 1.1) and was 34% (with 8 CRs) per modified RECIST
(mRECIST); ORR by Investigator per RECIST 1.1 was 32% with 1 CR. Confirmed
responses were reported across the patient population regardless of HCC
etiology, geographic region, baseline alpha-fetoprotein (AFP) levels, or
extrahepatic spread of tumor. The investigator-assessed median PFS per RECIST
1.1 was 14.9 months, and the IRF-assessed median PFS per RECIST 1.1 was 7.5
months. Median estimates for duration of response (DOR) and OS were not yet
reached at the data cutoff of 26 July 2018. Among the 103 safety evaluable
subjects, treatment-related Grade 3 or 4 AEs were reported in 28 subjects
(27%), most commonly hypertension (n = 10 [10%]). Five (5) Grade 5 AEs were
observed, 2 of which were assessed as treatment related (one sepsis, one
pneumonitis). A total of 19 subjects (18%) experienced treatment-related
serious adverse events (SAEs). Adverse events of special interest (AESIs) of
any grade for atezolizumab were reported for 54% of subjects, and AESIs of any
grade for bevacizumab were reported for 47% of subjects. Immune-related AESIs
for atezolizumab of >= Grade 3 requiring corticosteroid treatment included
pneumonitis (2 subjects), autoimmune encephalitis, drug-induced liver injury
(DILI), colitis, AST increased, γ-glutamyltranspeptidase (GGT) increased,
diabetes mellitus, and pancreatitis (1 subject each). The high response rate
observed suggested that the combination of atezolizumab with bevacizumab has
synergistic activity in advanced HCC and compared favorably to early
single-agent atezolizumab data in treatment-nai*ve HCC. Limited information is
currently available for atezolizumab as a single agent in patients with
advanced HCC. A total of 12 subjects were enrolled in two Phase 1 studies. In
these two studies, few responses were observed: one study with 5 subjects had
no responses (NCT01375842), and one study with 7 subjects had two confirmed
responses per investigator assessment (NCT02825940). Targets of cabozantinib
are also implicated in promoting tumor immune suppression including TYRO3, MER,
and AXL (TAM family kinases). Preclinical studies (Kwilas et al 2014, Lu et al
2017) and clinical observations on circulating immune suppressive cells and
immune effector cells in cancer patients (Apolo et al 2014) suggest that
cabozantinib promotes an immunepermissive environment that may present an
opportunity for synergistic effects from combined treatment with immune
checkpoint inhibitors (ICIs). A Phase 1b study (NCT03170960) is currently
evaluating the combination of cabozantinib with atezolizumab in multiple tumor
cohorts. In the dose escalation stage of the study, cabozantinib dose levels of
40 mg and 60 mg once daily (qd) were evaluated in 6 subjects each (atezolizumab
was administered at 1200 mg IV q3w for all subjects). At both cabozantinib dose
levels in the dose-limiting toxicity (DLT) evaluation period of the dose
escalation stage of the study there were no DLTs or SAEs (n=12 subjects total).
The majority of AEs were of Grade 1 or 2 including irAEs. Grade 3 AEs included
five events of hypertension, two events each of diarrhea and hypophosphatemia,
and one event each of pulmonary embolism, hyperglycemia, GGT increased, AST
increased, ALT increased, lymphocyte count decreased, lipase increased,
muscular weakness, nephritis, and myositis (verbatim term). No Grade 4 or 5 AEs
were reported. Among 10 subjects with clear cell RCC enrolled in the dose
escalation stage, the investigator-assessed confirmed ORR was 70% with 1 CR and
6 PRs. Cabozantinib 40 mg in combination with atezolizumab 1200 mg was selected
as the recommended dose for the Expansion Stage cohorts because of its
favorable safety profile over a prolonged time on study treatment with minimal
dose reductions and encouraging preliminary efficacy, which was deemed to
optimize the benefit/risk of the combination. The study is ongoing and
currently enrolling expansion cohorts in multiple solid tumor types including
subjects with advanced HCC who have not received prior systemic anticancer
therapy. As of 05 February 2019, 157 additional subjects had been enrolled in
the expansion cohorts evaluating cabozantinib (40 mg, qd) + atezolizumab (1200
mg IV q3w); information on the 141 subjects evaluable for safety as of 29
January 2019 is provided in Section 1.4. Targeting the VEGF signaling pathway
with small molecule TKIs has improved the clinical outcome of patients with
advanced HCC; however, the OS has been modest with a median OS between 10.7 and
13.6 months (SHARP trial, Llovet et al 2008; REFLECT trial, Kudo et al 2018).
More recently ICI therapies are being evaluated as potential new treatment
strategy in HCC. Both cabozantinib and atezolizumab have shown encouraging
clinical activity in advanced HCC. Based on the potential synergistic effects
the combination of cabozantinib with atezolizumab appears to be a promising
treatment opportunity for subjects with advanced HCC. Therefore, further
evaluation of cabozantinib in combination with atezolizumab in subjects with
previously untreated advanced HCC is warranted. This Phase 3 study evaluates
the safety and efficacy of cabozantinib in combination with atezolizumab
(approximately 370 subjects) versus the standard of care sorafenib
(approximately 185 subjects) in subjects with advanced HCC who have not
received previous systemic anticancer therapy. A single-agent cabozantinib arm
(approximately 185 subjects) will be enrolled in which subjects will receive
single-agent cabozantinib in order to determine its contribution to the overall
safety and efficacy of the combination with atezolizumab in this patient
population. It is planned to include up to 148 subjects from mainland China in
this study to assess safety and efficacy in the China subpopulation. In the
event enrollment in mainland China is incomplete by the time the global
enrollment phase has been completed, a Mainland China Extension Phase will be
implemented to complete enrollment and facilitate subpopulation analyses as
needed. The global population will include all subjects enrolled during the
global enrollment phase (including subjects enrolled at sites in mainland China
during that phase), and the China subpopulation will include all subjects
enrolled at sites in China (ie, during both the global enrollment phase and the
Mainland China Extension Phase). XL184-312 Protocol Amendment 2.0 introduced
considerations and study-related measures necessary due to the COVID-19
pandemic. XL184-312 Protocol Amendment 3.0 expanded the COVID-19-related
guidance to include instructions for managing subjects who become infected on
study, considerations for administration of COVID-19 vaccines, and confirmation
that the COVID-19 accommodations are temporary and will be repealed back to
standard study conduct when conditions allow.

The primary objective of this study is to evaluate the efficacy of cabozantinib
in combination with atezolizumab versus sorafenib in subjects with advanced HCC
who have not received previous systemic anticancer therapy. An secondary
objective is to evaluate the activity of single-agent cabozantinib compared
with sorafenib in this patient population.

This is a multicenter, randomized, open-label, controlled Phase 3 trial of
cabozantinib in combination with atezolizumab versus sorafenib in subjects with
advanced HCC who have not received previous systemic anticancer therapy. The
multiple primary efficacy endpoints are PFS
and OS for the experimental arm (cabozantinib + atezolizumab) vs the control
arm (sorafenib).
Additionally there will be a third arm to evaluate the safety and clinical
activity of single-agent cabozantinib. Approximately 740 eligible subjects with
advanced HCC were planned to be randomized in a 2:1:1 ratio at approximately
250 sites in this trial in the global enrollment phase of the study. However,
enrollment commenced under a randomization ratio of 6:3:1 per the original
protocol design, and there was a dynamic transition in randomization allocation
over time. As a result, the needed enrollment of 185 subjects in the
single-agent cabozantinib arm was not expected to be reached at the planned
total global enrollment phase sample size of 740 subjects. Therefore, to ensure
complete enrollment in the single-agent cabozantinib arm, the total global
enrollment phase was extended to accrue a total of approximately 840 subjects.

After completion of the global enrollment phase, additional subjects (up to
148) may be enrolled in a Mainland China Extension Phase at sites in mainland
China for evaluation in a China subpopulation. If initiated, subjects recruited
in the Mainland China Extension Phase will be randomized according to the same
2:1:1 scheme as subjects in the global enrollment phase. The global population
will include all subjects enrolled during the global enrollment phase
(including subjects enrolled at sites in mainland China during that phase), and
the China subpopulation will include all subjects enrolled at sites in China
(ie, during both the global enrollment phase and the Mainland China Extension
Phase).
The sample size for the global study may be increased up to an additional 25%
if a review of the accumulating data suggests that the COVID-19 pandemic has
caused the rate of study dropout or non-compliance to increase to a degree that
the ability to adequately evaluate study endpoints may be undermined. The
Mainland China Extension Phase will not be expanded beyond 148 subjects.

The trial consists of the
following phases:
Pre-Treatment Period: Potential subjects will be screened to determine if they
meet the required eligibility criteria. Qualifying screening assessments must
be performed within 28 days before randomization unless otherwise specified.

Treatment Period: Subjects who meet all study eligibility criteria will be
randomly assigned in a 2:1:1 manner to receive study treatment as follows:
Experimental arm (at least 370 subjects):
cabozantinib (40 mg oral, qd) +
atezolizumab (1200 mg infusion, q3w)

Control arm (at least 185 subjects):
sorafenib (400 mg, twice a day [bid])
Single-Agent Cabozantinib Arm (approximately 185 subjects):
cabozantinib (60 mg qd)
Randomization will be stratified by the following factors established at
screening:
• Disease etiology (HBV [with or without hepatitis C virus {HCV}], HCV [without
HBV], or Other)
• Region (Asia, Other)
• Presence of extrahepatic disease and/or macrovascular invasion (Yes, No)
Subjects will receive study treatment as long as they continue to experience
clinical benefit in the opinion of the Investigator or until there is
unacceptable toxicity, the need for subsequent systemic anticancer treatment,
or any other reasons for the treatment discontinuation listed in the
protocol. Treatment may continue after radiographic progression per RECIST 1.1
according to the criteria outlined in Section 5.7.6.3. Subjects on the
experimental arm (cabozantinib and atezolizumab) are allowed to discontinue one
component of the study treatment but continue to
receive the other. Escalation of cabozantinib from 40 mg qd to 60 mg qd in the
experimental arm is allowed after Sponsor approval for subjects who are
tolerating the 40 mg cabozantinib dose level well and have been treated on this
dose level for at least 4 weeks. In general, subjects who
develop clinically relevant adverse events (eg, Grade 3 or 4 AEs) are not
allowed to dose escalate cabozantinib from 40 qd to 60 mg qd. Crossover from
the control to experimental therapy will not be allowed unless study
transitions to a Crossover Phase (see below).

Crossover Phase: The study may transition to a Crossover Phase if the analysis
of the multiple primary endpoint of OS for the global ITT population (ie, not
including subjects enrolled in the Mainland China Extension Phase) shows
statistically significant and clinically meaningful evidence of improvement.

The Crossover Phase will only be implemented upon decision by the Sponsor and
following any required discussion with regulatory authorities following review
of the data. Crossover may be implemented independently and at different points
in time for study sites and subjects in mainland China compared to other sites
and subjects in the global study. If the decision is made to enter the
Crossover Phase, study sites will have 8 weeks to determine eligibility and
begin administration of crossover treatment (cabozantinib + atezolizumab
combination) to eligible subjects randomized to the control arm (sorafenib) or
the single-agent cabozantinib arm; subsequently no further crossover will be
allowed.
• Subjects randomized to the sorafenib control arm or the single-agent
cabozantinib arm will have the option to cross over to receive the cabozantinib
+ atezolizumab combination if they meet predefined eligibility criteria.
• Subjects randomized to the cabozantinib + atezolizumab experimental arm who
are still receiving study treatment and subjects randomized to the sorafenib
control arm or the singleagent cabozantinib arm who are still receiving study
treatment and do not cross over to the combination treatment (cabozantinib +
atezolizumab) may continue on their originally assigned study treatment until a
criterion for protocol-defined discontinuation has been met.
• Subjects randomized to the cabozantinib + atezolizumab experimental arm who
are in the Post-Treatment Period and subjects randomized to the sorafenib
control arm or the single-agent cabozantinib arm who do not cross over to
cabozantinib + atezolizumab and are in the Post-Treatment Period will continue
with post treatment assessments.

The study is expected to have completed enrollment in the global enrollment
phase at the time of transitioning to the Crossover Phase for those subjects,
but accrual of subjects in the Mainland China Extension Phase may still be
ongoing at that point. In the Crossover Phase safety assessments and efficacy
assessments will be performed per the schedule of assessments in Appendix B;
PK, biomarker, health-related quality of life (HRQOL), and healthcare resource
utilization assessments will be discontinued.
Post-Treatment Period: A first Post-Treatment Follow-up visit (FU-1) for safety
evaluation (including subjects in the Maintenance Phase [below]) is to occur 30
(+14) days after the date of the decision to permanently discontinue study
treatment (defined as the later of the date of the decision to permanently
discontinue study treatment or the date of the last dose of study treatment). A
second Post-Treatment Follow-up visit (FU-2) for safety evaluation will be
conducted approximately 100 days (±14 days) after the date of the decision to
permanently discontinue study treatment. Further details for follow-up and data
collection requirements for AEs, SAEs, and AESIs are summarized in Appendix L.
Radiographic tumor and HRQOL assessments are to continue, regardless of whether
study treatment is given, reduced, held, or discontinued until a
protocol-defined criterion for ending radiographic assessments is met (see
Section 5.7.6.2). Consequently these assessments may be
required in the Post-Treatment Follow-up Period for some subjects.
In addition, subjects are to be contacted every 12 weeks (± 14 days) after FU-2
to assess survival status and document receipt of subsequent anticancer
therapy. This follow-up will continue until the subject expires or the Sponsor
decides to discontinue collection of these data; however, these assessments are
not required in the Maintenance Phase (below). Every effort must be made to
perform these evaluations unless consent for non-interventional study
assessment is withdrawn.

Mainland China Extension Phase: After completion of the global enrollment
phase, additional subjects (up to 148) may be enrolled in a Mainland China
Extension Phase at sites in mainland China for evaluation in a China
subpopulation. The Pre-Treatment, Treatment, and Post-Treatment Periods of the
study will be conducted in the same manner in the Mainland China Extension
Phase as for subjects who were enrolled in the global enrollment phase.

Study Completion: The study will be considered complete if the null hypothesis
is rejected for the primary endpoint of OS (experimental vs control arm) in
either of the planned interim analyses or if the final planned analysis for OS
has been conducted (irrespective of the results) and any required supportive
analyses for China are completed.

Maintenance Phase/Treatment after Study Completion: The purpose of the
Maintenance Phase is to continue to provide long-term access to study drug(s)
to subjects who are deriving clinical benefit even after evaluation of the
study objectives has been completed (Study Completion, see
above). When sufficient data have been collected to adequately evaluate all
study endpoints, subjects who continue study treatment may enter the study
Maintenance Phase. Upon initiation of the Maintenance Phase, the Sponsor
considers the safety and efficacy profile of the experimental
treatment regimen within this study to have been sufficiently established and
data analyses required for regulatory purposes to have been completed. If a
Crossover Phase has been implemented, the Maintenance Phase may not begin
before the Week 9 Day 1 (W9D1) visit has
elapsed in the Crossover Phase for the last subject randomized to sorafenib or
single-agent cabozantinib who crossed over to receive cabozantinib +
atezolizumab. The Sponsor is to notify the sites if or when the study will
enter the Maintenance Phase or if an alternative post-Study
Completion option will be implemented (Section 6.3).
In the Maintenance Phase, subjects will continue to receive study treatment
until they meet the protocol-required criteria for treatment discontinuation.
Subjects are to undergo periodic safety assessments (including local laboratory
tests) and tumor assessments. The nature and frequency of these assessments
during the Maintenance phase are to be performed per institutional standard of
care and guidance from the Sponsor. It is the Investigator*s responsibility to
ensure that subject visits occur frequently enough and adequate assessments are
performed to ensure subject safety.
In order to continue to collect important safety information for subjects
enrolled in the study during the Maintenance Phase, reporting of SAEs; certain
AEs (including irAEs and other AESIs [whether serious or not], and AEs leading
to dose modifications or treatment discontinuation);
and other reportable events (DILI, pregnancy, and medication errors with
sequelae) is to continue per protocol requirements specific to the Maintenance
Phase.
The study clinical database will be closed upon initiation of the Maintenance
Phase. Important safety information (noted above) collected in the Maintenance
Phase will be captured in the safety database. Only data collected prior to
implementation of Maintenance Phase will be
reported in a clinical study report.
End of Trial: End of trial is defined as the last scheduled visit or scheduled
procedure for the last
subject (including Maintenance Phase assessments).

Subjects in the experimental arm will take oral study medication (40 mg of
cabozantinib: 2 tablets containing 20 mg each of cabozantinib) once daily.
Atezolizumab will be administered IV every 3 weeks at 1200 mg dose.
Dose reduction levels of cabozantinib will be allowed in the experimental arm
(20 mg qd and 20 mg every other day [qod]). Dose reductions for atezolizumab
will not be allowed and AEs will be managed by dose delays. Escalation of
cabozantinib from 40 mg qd to 60 mg qd is allowed after Sponsor approval for
subjects who are tolerating the 40 mg cabozantinib dose level well and have
been treated on this dose level for at least 4 weeks. In general, subjects who
develop clinically relevant adverse events (eg, Grade 3 or 4 AEs) are not
allowed to dose escalate cabozantinib from 40 qd to 60 mg qd.
Subjects in the control arm will take sorafenib orally (2 tablets containing
200 mg each of sorafenib) twice daily. Dose interruptions and reductions will
be allowed as per local prescribing information.
Subjects in the single-agent cabozantinib arm will take 60 mg of cabozantinib
qd (1 tablet of 60 mg of cabozantinib). Two dose reduction levels of
cabozantinib will be allowed in this arm (40 mg daily and 20 mg daily).
Subjects may be allowed to re-escalate following a dose reduction according to
the guidance provided in Section 6.6.1.
Subjects will continue study treatment as long as they continue to experience
clinical benefit in the opinion of the Investigator or until unacceptable
toxicity, the need for alternative anticancer treatment, or other reasons for
treatment discontinuation. Continuation of one component of the combination
study in the experimental arm (cabozantinib and atezolizumab) while
discontinuing the other will be allowed as per protocol defined guidelines.

Please see the schedule of events in the protocol, on page 157-161 for a
detailed overview of visits, tests and examinations.
The risks associated with this study are described in the informed consent
form, chapter 6 and appendix D.