Priming With Tremelimumab Leads to Greater Efficacy for Durvalumab Treatment in Advanced HCC

September 16, 2020

“In sum, these data show that T300+D regimen provides the best benefit-to-risk profile across arms,” Robin Kate Kelley, MD, said of Study 22 in hepatocellular carcinoma.

Clinical activity was observed with a single priming dose of tremelimumab in combination with monthly durvalumab (Imfinzi) as treatment of patients with hepatocellular carcinoma (HCC) in the second line and beyond, and the regimen also had an acceptable safety profile, according to findings from a safety run-in portion of Study 22.

The combination had the most favorable efficacy, safety, and tolerability across all treatment arms in the study, Robin Kate Kelley, MD, associate professor, University of California, San Francisco, said in her presentation of these findings during the 2020 International Liver Cancer Association (ILCA) Virtual Annual Conference.

The purpose of this study was to answer whether a single dose of tremelimumab as priming therapy could improve immune-mediated clinical activity in patients with HCC while also minimizing toxicity. In the part 2A portion of the randomized phase 1/2 trial (NCT02519348), patients were randomized to 1 of 3 cohorts, which included durvalumab (n = 40), tremelimumab monotherapy (n = 36), or a combination of tremelimumab and durvalumab (n = 39). Part 2B was designed to add another potential combination regimen with a priming dose of tremelimumab only in a safety run-in phase. The initial combination administered 75 mg of tremelimumab for 4 doses with 1500 mg of durvalumab every 4 weeks (T75+D), but the new dosing schedule assessed in part 2B explored administration at 300 mg of tremelimumab once with the same durvalumab dose (T300+D).

In part 3, patients were randomized to receive either durvalumab alone (n = 64), tremelimumab alone (n = 33), T75+D (n = 45), or T300+D (n = 65). The primary end point of Study 22 was safety, while key secondary end points included overall survival (OS), objective response rate (ORR), and duration of response (DOR).

“The patient demographics were generally similar across arms and representative of an advanced HCC population,” Kelley said during her presentation. “The majority of patients had progressed on or were intolerant of prior sorafenib [Nexavar] and therefore were treated in the second-line context, accounting for 73% of the T300+D arm. Patients who refused prior sorafenib were also eligible and comprised a subset of each arm.”

Adverse events (AEs) appeared tolerable overall across all study arms, in which any-grade AEs and treatment-related AEs (TRAEs), respectively, occurred in 98.6% and 82.4% of patients in the T300+D arm, 94.1% and 60.4% of the durvalumab arm, 97.1% and 84.1% in the tremelimumab arm, and 97.6% and 69.5% in the T75+D arm. Grade 3/4 AEs and TRAEs, respectively, also occurred in 58.1% and 35.1% of patients in the T300+D arm, 55.4% and 17.8% in the durvalumab arm, 66.7% and 43.5% in the tremelimumab arm, and 61.0% and 23.2% in the T75+D arm.

TRAEs resulting in death occurred in 2.7% of patients in the T300+D arm, 3.0% of the durvalumab arm, and 1.2% in the T75+D arm. AEs led to study treatment discontinuation in 12.2% of patients in the T300+D arm, 11.9% of the durvalumab arm, 18.8% of the tremelimumab arm, and 13.4% of the T75+D arm, while TRAEs led to discontinuation in 10.8%, 7.9%, 13.0%, and 6.1%, respectively.

Overall, a number of patients across all arms required systemic steroids for treatment of TRAEs, which included 24.3% of those in the T300+D arm, 9.9% of the durvalumab arm, 26.1% of the tremelimumab arm, and 24.4% of the T75+D arm. TRAEs of grade 3/4 severity required systemic steroids in 10.8% of patients in the T300+D arm, 6.9% in the durvalumab arm, 20.3% in the tremelimumab arm, and 9.8% in the T75+D arm.

In terms of TRAEs occurring in at least 5% of patients, these were generally grade 1/2 in severity while grade 3 events appeared less frequently overall. Grade 3 or higher TRAEs occurred in 28 patients (37.8%) of the T300+D arm, 21 (20.8%) of the durvalumab arm, 30 (43.5%) of the tremelimumab arm, and 20 (24.4%) of the T75+D arm.

In the tremelimumab priming arm, the most common any-grade TRAEs included pruritus (32.4%), rash (32.4%), aspartate aminotransferase (AST) increase (16.2%), alanine aminotransferase (ALT) increase (14.9%), amylase increase (12.2%), and fatigue (10.8%).

“Skin and subcutaneous disorders of pruritus and rash were the most common TRAEs and had the highest rates in the tremelimumab arm, but again, these were predominantly grade 1/2. Diarrhea was uncommon overall with low rates of grade 3 or higher,” Kelley said. “Rates were highest for tremelimumab monotherapy, which had grade 3 or higher diarrhea in 8.7%. Grade 3 or higher AST/ALT elevations occurred in around 12% or fewer in each arm.”

Immune-mediated AEs, occurring in at least 2% of patients, of any grade and grades 3/4, respectively, were experienced by 23 (31.1%) and 9 (12.2%) patients in the T300+D arm, 16 (15.8%) and 5 (5.0%) in the durvalumab arm, 17 (24.6%) and 11 (15.9%) in the tremelimumab arm, and 22 (26.8%) and 10 (12.2%) in the T75+D arm.

Diarrhea was the most common immune-mediated AE of any grade in the tremelimumab priming arm (6.8%), followed by hypothyroidism (5.4%), colitis (4.1%), hyperthyroidism (4.1%), increased AST (4.1%), and rash (4.1%).

The median OS was longest for patients in the T300+D arm, which was 18.7 months (95% CI, 10.8-27.3). The median OS was 13.6 months (95% CI, 8.7-17.6) with durvalumab alone, 15.1 months (95% CI, 11.3-20.5) for tremelimumab alone, and 11.3 months (95% CI, 8.4-15.0). The 18-month OS rate was 52.0% with T300+D (95% CI, 38.9%-63.6%), 35.3% (95% CI, 25.0%-45.8%) with durvalumab alone, 45.7% (95% CI, 32.8%-57.7%) with tremelimumab, and 34.7% (95% CI, 24.4%-45.2%) with T75+D.

The ORR was the best at 24.0% (95% CI, 14.9%-35.3%) with T300+D, in which 1 patient (1.3%) achieved a complete response (CR), 17 (22.7%) had partial responses (PRs), and 16 (21.3%) reached stable disease (SD). The disease control rate (DCR) in this arm was 45.3%, but the median DOR was not reached. The median time to response was 1.86 months, and the median PFS was 2.17 months (95% CI, 1.91-5.42). Additionally, 6 patients (8.0%) achieved SD >6 months.

Among the monotherapy arms with durvalumab versus tremelimumab, the ORR was 10.6% (95% CI, 5.4%-18.1%) versus 7.2% (95% CI, 2.4%-16.1%), respectively, in which PRs were achieved by 11 (10.6%) versus 5 (7.2%) and SD in 28 (26.9%) versus 29 (42.0%). The DCR with durvalumab was 37.5% and 49.3% for tremelimumab, and the median DOR was 11.2 months and 24.0 months, respectively. The median time to response was 3.65 months with durvalumab and 1.81 months with tremelimumab. The median PFS with durvalumab was 2.07 (95% CI, 1.84-2.83) while the median with tremelimumab was 2.69 (95% CI, 1.87-5.29). Overall, 4 (3.8%) and 10 (14.5%) of patients in the durvalumab and tremelimumab arms, respectively, achieved SD >6 months.

T75+D had an ORR of 9.5% (95% CI, 4.2%-17.9%), with 2 patients (2.4%) achieving a CR, 6 (7.1%) achieving a PR, and 23 (27.4%) achieving SD. The DCR was 36.9%, and the median DOR was 13.2 months. The median time to response was 2.86 months, and the median PFS was 1.87 months (95% CI, 1.77-2.43). Four patients (4.8%) in this arm achieved SD >6 months.

“In sum, these data show that T300+D regimen provides the best benefit-to-risk profile across arms,” Kelley concluded. “Furthermore, the unique association of the T300+D regimen with proliferative CD8-positive T cells provides a strong biologic rationale with the observed treatment response that warrants further study.”

The T300+D regimen and durvalumab alone are under evaluation currently against sorafenib in the ongoing phase 3 HUMALAYA study (NCT03298451) as frontline treatment of patients with HCC.


Kelley RK, Kudo M, Harris W, et al. The novel regimen of tremelimumab in combination with durvalumab provides a favorable safety profile and clinical activity for patients with advanced hepatocellular carcinoma. Presented at: 2020 ILCA Virtual Conference; September 11-13, 2020; Virtual. Abstract O-22.