Thank you for your interest in CCO content. As a guest, please complete the following information fields. These data help ensure our continued delivery of impactful education.
Become a member (or login)? Member benefits include accreditation certificates, downloadable slides, and decision support tools.
Considerable advances have been made in the past few years surrounding the management of patients with advanced HCC. This is an area that has had great unmet needs and, I would argue, still has unmet needs today, but great progress has been made in helping our patients live longer. Before 2007, there were no systemic therapies that improved survival for patients with advanced HCC. That year, sorafenib was approved, demonstrating improved survival over placebo. This approval was followed by a 10-year gap in which there was considerable clinical research but no successful drug approvals in this space. That changed in 2017 with the approval of regorafenib in the second-line setting, and since then, we’ve seen a number of new drugs and combinations approved for the treatment of our patients. In this activity, I discuss how to optimize the care of patients with advanced HCC to maximize survival.
Let’s begin our discussion with the case of a man who is 65 years of age and presents to his primary care doctor complaining of fatigue, weight gain, and edema. He has a past medical history of hypertension, diabetes, and thrombocytopenia; he has had a low platelet count for some time that has not been adequately evaluated. On this presentation to his doctor, he has a normal white blood cell count, hemoglobin of 10 g/dL (which is normocytic), and a platelet count of 98 x 109/L. He has elevated total bilirubin (1.4 mg/dL), relatively normal aspartate aminotransferase (AST), slightly elevated alanine aminotransferase (ALT) (55 units/L), and a slightly decreased albumin (3 g/dL).
The patient undergoes an abdominal ultrasound, which reveals a nodular liver, evidence of splenomegaly, and a liver mass. A very common cause of thrombocytopenia is undiagnosed portal hypertension that comes from liver disease. The patient subsequently undergoes a CT scan, which confirms a nodular cirrhotic liver with portal hypertension and shows an 8-cm lesion in the right lobe. This lesion is hypervascular with delayed washout, which is characteristic of liver cancer and removes the need for a biopsy, and the radiologist says this is consistent with LiRADs 5, confirming the diagnosis of liver cancer. The patient has no other sites of disease. His AFP is 12. Hepatitis serologies show that he does not have hepatitis.
On physical examination, he looks well. He is mildly obese with some edema, but otherwise an unremarkable examination.
The first thing we need to think about when we approach a treatment recommendation for this patient is what is his stage. This patient appears to be well compensated. His performance status is good; he is up and doing his normal activities. When we assess his liver function, he has no encephalopathy; he has no ascites; and he has a normal international normalized ratio (INR) and a slightly decreased albumin (which gets him 1 point). In addition, his bilirubin is normal. He is considered Child-Pugh A with a score of 6.
The first thing to think about when approaching a treatment recommendation for this patient is staging. This patient appears to be well compensated. His performance status is good and he is able to perform his normal activities. His liver function would be assessed as Child-Pugh A6, based on the following that factor in when calculating Child-Pugh score: He has no encephalopathy, no ascites, a normal INR, normal bilirubin, and slightly decreased albumin.
Given that he has liver-confined disease, this patient would be considered to be Barcelona Clinic Liver Cancer (BCLC) stage B, which is intermediate stage. The standard of care for intermediate-stage patients is chemoembolization (eg, TACE).
The patient, therefore, undergoes TACE. Although the patient has a large 8-cm tumor, TACE would be appropriate to minimize toxicity and maximize survival. However, at 1 month post TACE, imaging reveals that the patient has developed an expansile tumor thrombus into the right portal vein extending to the confluence of the right and left veins. Subsequent re-evaluation shows that his laboratory studies are stable.
Now we are faced with the question of what to do next for this patient. Has this patient’s stage changed and is different treatment indicated, or do we proceed with another locoregional treatment?
This slide shows the Barcelona staging system and its corresponding data‑driven recommendations as of early 2017. Up to this point, the only approved systemic therapy was sorafenib, and this was approved for patients with unresectable disease. When we return to this staging system again at the end of this activity, you will see that numerous drugs have been added for those patients with advanced disease.
The staging system is divided into 5 different stages from early (stage 0) to very advanced (stage D). Stage D includes those patients who have very decompensated liver disease and are not candidates for cancer treatment, other than perhaps a liver transplant if they would qualify for this treatment.
Stage 0 and A are early‑stage patients. Some of these can be cured with a simple resection if they have a tumor that is anatomically resectable and do not have portal hypertension. Others may be transplant eligible and/or may receive ablation or chemoembolization to bridge them to transplant. Details for transplantation are beyond the scope of this talk, but all new patients with HCC deserve to be evaluated at a transplant center before that potentially curative option is ruled out.
Most patients presenting in the clinic do so in the intermediate stage. The case patient was an example: at presentation, he had well-compensated Child-Pugh A liver disease, few symptoms, and underwent chemoembolization.
A number of patients will come to clinic and present with advanced-stage characteristics. These are the patients who benefit most from systemic therapy and for whom these therapies are indicated based on clinical trial data. It is important to note that a patient does not need to have a tumor outside the liver to be considered advanced stage, as is common with other solid tumors. If a patient has extrahepatic spread or metastatic disease to lymph nodes or other solid organs, then they would be considered advanced, but the advanced stage also includes patients with tumor that is contained within the liver but has invaded the liver portal venous vasculature. This is the case with the patient we have been discussing today: He was initially an intermediate patient and received locoregional treatment but progressed after a relatively short period and would now be classified as advanced stage. This patient had a very short intermediate-stage period; other patients undergo repeated TACE over many years. All patients eventually stage migrate from intermediate to advanced stage.
The patients with advanced HCC who derive the greatest benefit from systemic treatment are those with preserved liver function. Patients with elevated bilirubin, ascites, or other stigmata of end-stage liver disease will not benefit as much from systemic treatment, and their most likely cause of death will be cirrhosis and liver failure.
When we look at the complexity of liver cancer management, we see that the best outcomes come with a multidisciplinary team. Management of underlying liver disease is interwoven with management of the malignancy, so several key players should all be involved in a patient’s care and treatment decisions, including hepatology, diagnostic and interventional radiology, pathology, gastrointestinal (GI) services, oncology, and a surgical program that offers transplant and not just surgical resection. Patients can also benefit greatly from a team that includes knowledgeable palliative care personnel, when indicated.
Sorafenib was approved in 2007 based on data from the phase III SHARP study, which assessed sorafenib vs placebo for patients with advanced HCC, with compensated (Child-Pugh A) liver disease and no previous systemic therapy. This study was the first study to show a benefit for systemic treatment in patients with HCC, with sorafenib treatment associated with an OS HR of 0.69 (95% CI: 0.55-0.87; P < .001) or a 31% decrease in the risk of death vs placebo. Median OS in the sorafenib arm was 10.7 months vs 7.9 months in the placebo arm. With these findings, sorafenib became the standard of care for patients with advanced HCC for many years. Sorafenib was the comparator in many phase III studies of novel agents and regimens for HCC, but no treatment (until lenvatinib, discussed below) was superior or noninferior to this agent. What is striking is that sorafenib—a multikinase inhibitor against the VEGF receptor and other kinases—improves survival by slowing progression. The number of patients who have true objective responses is low.
The results from SHARP were confirmed in the ASIA PACIFIC study, which assessed sorafenib vs placebo in a similar patient population but concentrated on those from the Asia-Pacific region. The OS HR with sorafenib was 0.68 (95% CI: 0.50-0.93; P = .014), which was very similar to that observed in the SHARP trial.
No other drug was approved in the frontline setting for advanced HCC until lenvatinib in 2018. This approval was based on REFLECT, an open‑label, randomized, noninferiority phase III study comparing sorafenib 400 mg twice daily with daily lenvatinib as first-line treatment for unresectable HCC. Lenvatinib is dosed by weight; in this study, patients with body weight < 60 kg received 8 mg, whereas those who weighed ≥ 60 kg received 12 mg. These doses are significantly lower than the doses of lenvatinib used in other diseases, such as kidney cancer (18 mg/day) or thyroid cancer (24 mg/day). The primary endpoint of this study was OS.
Like the SHARP study, REFLECT enrolled patients who had Child‑Pugh A liver disease (to minimize the risk of death from liver disease) and advanced liver cancer. Unlike other studies, REFLECT excluded patients with > 50% of their liver occupied by tumor, those who had clear bile duct invasion, and those with portal vein invasion outside the liver. The confluence of the splenic vein and the superior and inferior mesenteric veins come together to make the portal vein, which enters the liver and then branches into the right and left portal veins and smaller branches from there. The case patient we were discussing had a tumor thrombus in the liver that did not extend outside the liver. REFLECT, unlike other studies, excluded patients who had portal vein involvement outside the liver and the reason for that was not necessarily biologic. However, the study originated in Japan and, historically, their standard for patients with main portal vein invasion outside the liver includes intra-arterial chemotherapy.
Like sorafenib, lenvatinib is a multikinase inhibitor that inhibits VEGF receptors and additional targets, such as FGFR. This latter target is important in liver cancer, as the FGFR family has been shown to be important in angiogenesis resistance and tumor cell growth.
The median OS in the sorafenib group was 12.3 months vs 13.6 months in the lenvatinib group (HR: 0.92; 95% CI: 0.79-1.06), with the OS curves overlapping. For superiority, the upper limit of the CI would have had to have been < 1 with an associated low P value. However, the CI crossed 1 but was below the noninferiority threshold of 1.08; therefore, this study was noninferior for OS but did not meet its superiority endpoint.
The median PFS in sorafenib group was 3.7 months vs 7.4 months in the lenvatinib group (HR: 0.66; 95% CI: 0.57-0.77). The PFS was assessed using modified Response Evaluation Criteria in Solid Tumors (RECIST). Modified RECIST is different from conventional RECIST when applied to liver lesions because of the hypervascular nature of liver cancers. The longest diameter of the enhancing component seen on the arterial phase in the liver is measured. With drugs like lenvatinib and other TKIs, we see loss of enhancement, which appears to be associated with some biologic activity, and response and PFS are assessed by increases or decrease in enhancement in liver tumors, not just the shrinkage of the tumor.
Because REFLECT was an open-label study, endpoints were assessed in 3 ways: modified RECIST by investigator, modified RECIST by blinded independent review, and conventional RECIST by blinded independent review. By independent review using conventional RECIST, the ORR was 18.8% with lenvatinib; with modified RECIST by investigator and independent review, ORRs were 24.1% and 41%, respectively. By contrast, ORRs with sorafenib were 6.5%, 9.2%, and 12.4% by the same assessments. As such, lenvatinib differs from sorafenib as it is able to induce objective responses.
Since the study was noninferior for OS, what about those patients who had a response by modified RECIST? Regardless of the treatment arm, OS was extended for those with a response: median OS was 22 months for responders vs 11.4 months for nonresponders (HR: 0.61; 95% CI: 0.49-0.76).
We do need to be careful with these retrospective data, but it is a very provocative idea: Patients do better if they achieve a response to therapy.
Lenvatinib and sorafenib have some overlap in adverse event (AE) profiles, with GI toxicity, such as anorexia, weight loss, diarrhea, and fatigue observed with both agents. There are some differentiating AEs, however. In REFLECT, lenvatinib more frequently caused treatment-emergent higher-grade hypertension vs sorafenib. This may be due to the increased potency of lenvatinib against the VEGF receptor. Higher rates of proteinuria were also observed with lenvatinib, so patients receiving lenvatinib should be monitored for this AE. Something that was not seen as often with lenvatinib was hand–foot skin reaction, which appeared to be more frequent and higher grade with sorafenib. As we have known for some time, this AE can present a challenge when using this drug.
The recent wave of approvals in HCC included nivolumab, an anti–PD-1 antibody. Given their ability to induce deep and meaningful responses, PD-1 and PD-L1 inhibitors have changed the therapeutic approach for many malignancies. These drugs came to the liver cancer space later because of the concern about AEs.
In phase I CheckMate 040 study, the efficacy of nivolumab was initially assessed by etiology (hepatitis B, hepatitis C, or those without viral hepatitis) in patients who had previous sorafenib and then later in the frontline setting. Nivolumab demonstrated response rates in the range of 14% to 23% across etiologies; strikingly, for patients who responded, the duration of response was 16.6 months. Response did not seem to correlate with etiology or biomarkers, such as PD-L1 expression. Based on this study, nivolumab received accelerated approval in the second-line, post-sorafenib setting for advanced HCC.
CheckMate 459 was the definitive study required for full approval of nivolumab. CheckMate 459 was a phase III trial in which patients with advanced HCC, Child-Pugh A liver function, and no previous systemic therapy were randomized to nivolumab 240 mg every 2 weeks or sorafenib 400 mg twice daily. CheckMate 459 did not have any criteria regarding tumor burden, unlike REFLECT, so the study population was broader. The primary endpoint was OS, with a predefined threshold for statistical significance HR of 0.84 (P = .0419).
Unfortunately, this was a negative study, which surprised many in the field. The median OS of 16.4 months with nivolumab was the longest observed in a phase III study at that time; for sorafenib, the median OS of 14.7 months was the longest survival observed with sorafenib in any phase III study. The difference was close to significant (HR: 0.85; 95% CI: 0.72-1.02; P = .0752) but did not reach the predefined threshold.
The median PFS was also not significantly improved with nivolumab vs sorafenib (3.7 vs 3.8 months). However, looking at benchmarks over time, we do see a separation of the curves. For example, at 24 months, 14% of patients treated with nivolumab were progression free whereas only 6% of patients receiving sorafenib were progression free. This relates to some of the challenges seen with PD-1/PD-L1 inhibitors in cancer trials, where medians may not always capture this tail of the curve. The OS curve suggests that there may be a subgroup of patients who derive a greater benefit than the general population. It behooves us to try to identify a biomarker or clinical characteristics that will allow the identification of these patients.
The ORR with nivolumab was 15%, while that with sorafenib was 7%. Despite this, the disease control rate, including stable disease, was very similar between arms.
Because CheckMate 459 was a frontline study, OS is vulnerable to posttreatment progression, as it is not possible to control what treatments patients receive after their frontline treatment. In the sorafenib arm, 20% of patients received subsequent immunotherapy and 11% of patients received other investigational agents, the majority being immunotherapies. So at least 20%, and perhaps as many as 30%, of patients were treated with immunotherapy after progression on sorafenib, which could complicate the OS analysis.
Nivolumab was very well tolerated in this study and no new toxicities with sorafenib were reported. As shown in the graphic on the slide, there are several AEs that occur frequently and could affect the quality of life for patients treated with sorafenib—such as diarrhea and hand–foot skin reaction—that do not generally appear for patients treated with nivolumab.
It was hypothesized that one way to increase response rates associated with checkpoint inhibitors would be to combine VEGF and immune checkpoint inhibition. As we will discuss in a moment, the combination of the PD-L1 and VEGF inhibitors atezolizumab and bevacizumab was recently shown to be highly efficacious for treating advanced HCC.
The biology behind combining these types of agents evolved over time.[11-13] We and others assessed the potential of the VEGF antibody bevacizumab in preclinical and clinical studies around the time of the development of sorafenib but did not observe particularly high response rates with this agent. Studies of VEGF inhibitors were driven by the fact that HCC is a very vascular cancer and that VEGF seemed to be important in prognosis and possibly pathology with liver cancer. However, our understanding of how VEGF/VEGFR signaling functions has moved beyond altering the vasculature and providing oxygen to the tumor. When VEGF or VEGFR is targeted, the inflammatory milieu around a tumor is altered to promote APC maturation and trafficking and T-cell trafficking into the tumor bed, all of which can stimulate an immune response against a cancer cell. This immune response was postulated to be accelerated with atezolizumab, which targets PD-L1.
This was borne out in the phase III IMbrave150 study, which evaluated atezolizumab plus bevacizumab vs sorafenib in an open-label, randomized trial. Similar to other frontline studies, it enrolled patients with locally advanced, intermediate-stage or advanced-stage C HCC, with Child-Pugh A liver function, good performance status, and no previous systemic treatment. This study had coprimary endpoints of OS and PFS as assessed by independent review. Atezolizumab and bevacizumab were given intravenously every 3 weeks, atezolizumab at a flat dose of 1200 mg and bevacizumab at 15 mg/kg; sorafenib was given at 400 mg twice daily.
The baseline characteristics for this study represent a very typical study population, with the majority of patients being male and between 60 and 70 years of age. Most patients came from outside of Asia, and > 80% of patients in each arm were staged as BCLC stage C. Regarding the patients who were staged as BCLC stage B (~ 15%), recall that although some intermediate-stage patients get repeated TACE and do not develop vascular invasion or extrahepatic spread, others develop contraindications to repeated TACE, such as rapid recurrence after TACE, HCC symptoms, or liver function decline.
This study was otherwise well balanced. It was a fairly high-risk population as > 75% of patients had macrovascular invasion and/or extrahepatic spread.
Because bevacizumab has been associated with bleeding, patients were required to have an endoscopy within 6 months of starting the regimen. Approximately 25% of patients in each arm had known varices at baseline and some of them required treatment prior to coming on study.
The coprimary endpoint of median OS was 13.2 months with sorafenib and the median OS had not yet been reached with atezolizumab plus bevacizumab. The survival curves separated early and remained separated; this translated into an HR of 0.58, or a 42% decrease in the risk of death (95% CI: 0.42-0.79; P = .0001). This is actually a greater benefit for survival than we saw with sorafenib vs placebo.
Median PFS (as measured by conventional RECIST criteria) was also improved from 4.3 months with sorafenib to 6.8 months with atezolizumab plus bevacizumab (HR: 0.59 [or a 41% decrease in the risk of progression]; 95% CI: 0.47-0.76; P < .0001).
Why was this study positive when the CheckMate 459 study of single-agent nivolumab was not? It is likely related to the increased activity of this combination compared with the single-agent immune checkpoint inhibitor. Specifically, by independent review RECIST, the confirmed ORR was 27.3% with atezolizumab plus bevacizumab compared with 11.9% with sorafenib. The ORR with atezolizumab plus bevacizumab included 5.5% of patients who had CRs and 22% of patients who had PRs. An increase in the proportion of patients with stable disease and a 73.6% disease control rate were observed with the combination treatment vs a 55.0% disease control rate with sorafenib. These data were consistent when assessed by modified RECIST.
Overall, there were no differences in the rates of all-case AEs or grade 3/4 AEs between the two arms. Grade 5 events were 5.8% with sorafenib vs 4.6% with atezolizumab plus bevacizumab. Serious AEs occurred in 38% of patients who were treated with atezolizumab plus bevacizumab arm vs 31% who were treated with sorafenib. AEs leading to protocol‑required withdrawals occurred in 15% of patients in the atezolizumab plus bevacizumab group vs 10% of patients in the sorafenib group. Overall, this safety profile seems to be very consistent with what we know about both drugs.
This slide shows all‑cause AEs that occurred with ≥ 10% frequency in either arm and > 5% difference between arms. In light colors are those AEs of all grades and in darker colors are grade 3/4 events. Diarrhea, hand–foot skin reaction, and anorexia were more frequent with sorafenib, whereas more frequent and higher-grade hypertension and proteinuria occurred with atezolizumab plus bevacizumab.
Looking at treatment-related AEs occurring at > 10% in any arm, a very similar pattern emerges. There was a slight increase in all-grade and grade 3/4 elevations in ALT and AST with atezolizumab plus bevacizumab vs sorafenib. One thing to keep in mind is that bleeding events were slightly increased with atezolizumab plus bevacizumab vs sorafenib, but those were mostly driven by low‑grade bleeding events such as epistaxis. Some bleeding deaths occurred with atezolizumab plus bevacizumab; these were due to upper GI or variceal bleeding. These were single‑digit events that occurred in 3 patients treated with atezolizumab plus bevacizumab.
Finally, quality-of-life data also favored atezolizumab plus bevacizumab. The median time to deterioration in quality of life with sorafenib was 3.5 months compared with 11.2 months with atezolizumab plus bevacizumab.
Overall, the safety profile and efficacy data from this study clearly support the use of atezolizumab plus bevacizumab as frontline therapy for most patients who are candidates for systemic treatment.
Now, let’s return to our case. The patient is a 65-year-old man who presented with fatigue, weight gain, edema. He had a history of hypertension, diabetes, and thrombocytopenia. Laboratory testing showed the following: white blood cells 4500/µL, hemoglobin 10 g/dL, platelets 98 x 109/L, total bilirubin 1.4 mg/dL, AST 45 U/L, ALT 55 U/L, and albumin 3.3 g/dL. Abdominal ultrasound revealed a nodular liver, splenomegaly, and a liver mass. The patient subsequently underwent a CT scan, which confirmed a nodular cirrhotic liver with portal hypertension and showed an 8-cm lesion in the right lobe. This lesion was hypervascular with delayed washout, which is characteristic of liver cancer and removes the need for a biopsy, and the radiologist says this is consistent with LiRADs 5, confirming the diagnosis of liver cancer. The patient has no other sites of disease. His AFP is 12.
The patient underwent TACE; however, 1 month after TACE, repeat imaging shows some response but new expansile tumor thrombus in the right portal vein extending to the confluence.
There is ongoing interest in trying to improve outcomes for patients with advanced HCC by combining immune checkpoint inhibitors with VEGF- or VEGFR-targeted agents, and several regimens are now in phase III clinical trials for the first-line treatment of patients with advanced HCC.
Data recently became available for one such combination, in the form of an open-label phase Ib study of lenvatinib plus the PD-1 inhibitor pembrolizumab for patients with unresectable HCC and no prior systemic therapy. This study assessed this combination initially in a small safety population and then in an expansion population. Note that pembrolizumab has been assessed in clinical trials as monotherapy for patients previously treated with sorafenib; these data will be discussed later in the activity.
In KEYNOTE-524, the ORR with lenvatinib plus pembrolizumab was 36%. As seen in the waterfall plot, the majority of patients had some response. The disease control rate was 88%.
This slide shows OS according to best response to lenvatinib plus pembrolizumab. For patients who had a CR or PR, the median OS has not been reached; that’s a very flat curve out to approximately 22 months with for these patients. For those patients who had stable disease (~ 52%), median OS was 22.0 months, and in those patients whose best response was progressive disease, there was a fairly short median OS of 2.3 months.
These data have been now moved forward into a phase III study, the LEAP-002 study, which has completed accrual. This study randomized patients with advanced HCC, Child-Pugh A liver function, and no previous systemic therapy to lenvatinib plus or minus pembrolizumab.
Also ongoing is the phase III COSMIC-312 study, in which patients with advanced HCC, Child-Pugh A liver function, and no previous systemic therapy are being randomized to atezolizumab plus cabozantinib, cabozantinib alone, or sorafenib. Cabozantinib is a multikinase inhibitor that inhibits the VEGF receptor but uniquely also targets c-MET (or hepatocyte growth factor receptor) and AXL. We will have a closer look at cabozantinib during our discussion of second-line therapy, but COSMIC-312 represents an effort to move cabozantinib into the frontline setting in combination with atezolizumab, building on data with PD-1/PD-L1 inhibitor and TKI combinations.
Another combination strategy that is being evaluated in the frontline setting is the combination of nivolumab and ipilimumab. The phase I/II CheckMate 040 study included an assessment of 3 dosing schemes of nivolumab plus ipilimumab for patients with advanced HCC and previous sorafenib treatment. Treatment with nivolumab plus ipilimumab in any of these dosing schemes was associated with an ORR> 30%.
Of interest, patients in arm A (nivolumab 1 mg/kg + ipilimumab 3 mg/kg every 3 weeks) had a median OS of 22.8 months compared with approximately 13.0 months for the other 2 arms. However, the combination of nivolumab plus ipilimumab is associated with greater toxicity than single-agent nivolumab.
Despite this, the provocative response rates observed with nivolumab plus ipilimumab did lead to accelerated approval of this combination for second-line, post-sorafenib treatment of advanced HCC.
A definitive phase III study of nivolumab and ipilimumab is ongoing in the frontline setting. In CheckMate 9DW, patients with advanced HCC, no previous systemic therapy, and Child-Pugh A liver function will be randomized to nivolumab plus ipilimumab or sorafenib or lenvatinib.
The final immune checkpoint inhibitor combination in advanced clinical development is durvalumab plus tremelimumab. Durvalumab is a PD-L1 inhibitor and tremelimumab is a CTLA-4 antibody (like ipilimumab) that works in the immune checkpoint pathway in the priming phase of T-cells.
A phase I/II trial was conducted in which patients with unresectable HCC with progression on or intolerance to sorafenib were treated with 2 different doses of tremelimumab plus durvalumab or tremelimumab or durvalumab monotherapy. The combination of durvalumab plus the higher dose of tremelimumab was associated with the longest median OS, at almost 19 months. This was based on a response rate of 24%, and this combination had an acceptable toxicity profile.
Durvalumab plus tremelimumab is now being evaluated in the HIMALAYA study, which has completed accrual. In this phase III study, patients with advanced HCC, no previous systemic therapy, and Child-Pugh A liver function are randomized to one of 2 regimens of durvalumab plus tremelimumab, durvalumab alone, or to sorafenib.
Before continuing, let’s return to a question from earlier in the activity.