The addition of durvalumab to tremelimumab showed modest activity in patients with advanced neuroendocrine tumors of gastroenteropancreatic and lung origins.
Jaume Capdevila, MD, PhD
The addition of durvalumab (Imfinzi) to tremelimumab showed modest activity in patients with advanced neuroendocrine tumors (NETs) of gastroenteropancreatic and lung origins, according to results from the DUNE trial (GETNE 1601; NCT03095274) presented during the 2020 ESMO Virtual Congress.1
The multicohort study included patients whose disease stemmed from typical/atypical lung carcinoid (cohort 1), grade 1/2 gastrointestinal (cohort 2), grade 1/2 pancreatic (cohort 3), and grade 3 gastroenteropancreatic origins (cohort 4).
Results showed that at a median follow-up of 10.8 months, the clinical benefit rate (CBR), defined as complete response (CR) plus partial response (PR) and stable disease (SD), was determined to be 7.4% in cohort 1 (n = 27), 32.3% in cohort 2 (n = 31), and 25.0% in cohort 3 (n = 32). Moreover, the 9-month overall survival (OS) in cohort 4, which was one of the primary end points of the trial, was 36.1% (n = 33; 95% CI, 22.9-57) with the combination.
“The CBR at 9 months was not met at the prespecified 50% [threshold for futility],” Jaume Capdevila, MD, PhD an oncologist within the GI and Endocrine Tumor Unit with Vall d'Hebron University Hospital and the Developmental Therapeutics Unit with Vall d'Hebron Institute of Oncology, said in a presentation during the 2020 ESMO Virtual Congress. However, he added, “Cohort 4 really met the prespecified threshold for OS of 9 months, [with] 36% of patients being alive at 9 months.”
The rationale to utilize immune checkpoint blockade in neuroendocrine neoplasms (NENs) has been limited, as these neoplasms typically have low tumor mutational burden, PD-L1 expression, and lymphocyte infiltration. To date, PD-1 inhibitors such as pembrolizumab (Keytruda) and spartalizumab have demonstrated a limited overall response rate (ORR) of less than 5%, although they have yielded more intriguing data in thoracic well-differentiated NETs, noted Capdevila. For example, spartalizumab has been shown to result in an ORR of 20% in this patient population.2,3 Recently, the combination of nivolumab (Opdivo) plus ipilimumab (Yervoy), showed encouraging activity in this patient population.4
“In trying to revert this intrinsic resistance in this tumor [type], the combination of anti–PD-L1 and anti–CTLA-4 may play a role in [treating] these patients,” Capdevila explained. “It’s important to know that, recently, nivolumab and ipilimumab reported some promising activity in high grade NENs with an ORR of 44%. However, no response was shown in low or intermediate grade–disease.
To participate in the DUNE trial, patients needed to have advanced NENs of gastroenteropancreatic or lung origin following progression on standard therapies. Patients in cohort 1 had received previous treatment with somatostatin analogues and/or targeted therapies or chemotherapy; cohort 2 received prior somatostatin analogues and targeted therapies such as everolimus (Afinitor) or radionucleotides; cohort 3 received previous chemotherapy, somatostatin analogues, and targeted therapies (2-4 systemic treatment lines); and cohort 4 had received first-line treatment with platinum-based chemotherapy.
Participants received 4 cycles of durvalumab at 1500 mg in combination with 75 mg of tremelimumab followed by an additional 9 cycles of durvalumab monotherapy, given at a dose of 1500 mg.
“After 1 year of therapy, for patients who achieved SD, PR, or CR, treatment was stopped and follow-up was started,” Capdevila noted. “In the case of disease progression during this follow-up period, patients were allowed to start the same treatment.”
Primary efficacy end points of the trial included CBR for cohorts 1-3 per RECIST v1.1 criteria and the 9-month OS rate in cohort 4. Key secondary end points of the trial included safety, PFS, OS, ORR, duration of response (DOR), and biomarker analyses.
“With a unilateral alpha level of 5% and 80% power, we expected to increase the CBR from 30% to 50% in cohorts 1, 2, and 3,” explained Capdevila said. “For that, we decided to estimate 28 patients per cohort. In cohort 4, we estimated to increase the OS rate of 9 months from 13% to 23%. For that reason, we estimated to include 30 patients in this cohort.”
The median ages of patients on the trial were 61 (n =27; 100%), 62 (n = 31; 100%), 66 (n = 32; 100%), and 55 (n = 33; 100%) in cohorts 1-4, respectively. The majority of patients were male in most cohorts (cohort 1, 66.7%; cohort 2, 58.1%; cohort 4, 66.7%), with the exception of cohort 3, which included more female patients (56.3%).
Additional results from the trial showed that in terms of best change from baseline in target lesions per cohort, cohorts 1 and 2 had an ORR of 0%. Cohorts 3 and 4 achieved higher ORRs of 6.9% and 7.2%, respectively, with 1 CR in cohort 4.
When it came to treatment response according to PD-L1 expression, cohort 1 experienced an ORR of 7.4% per irRECIST criteria. In those with PD-L1 positivity, the ORR was 16.6%; in those without PD-L1 expression the ORR was 0%. In cohort 2, the ORR was 0% in those with or without PD-L1 expression. In cohort 3, the ORR was 6.3%; those with PD-L1 positivity experienced an ORR of 25%. In cohort 4, the ORR was 9.1% with an ORR of 0% in those with PD-L1 positivity and 7.7% in those who were negative for PD-L1.
“As a general message, the number of patients with PD-L1–positive tumors was very low. In cohort 1, only 6 patients with PD-L1–positive tumors were included; there were only 4 patients in cohort 3 with pancreatic origins. These number should be taken with caution,” Capdevila emphasized. “Even with that, we can see a trend toward a better ORR in most patients with PD-L1–positive tumors, mainly in [cohort 1].”
Investigators additionally examined progression-free survival (PFS) in each cohort and found a median PFS of 5.3 months in cohort 1 (95% CI, 4.52-6.06), 8 months in cohort 2 (95% CI, 4.92-11.15), 8.1 months in cohort 3 (95% CI, 3.80-12.46), and 2.5 in cohort 4 (95% CI, 2.15-2.75).
The most commonly reported adverse effects (AEs) for patients who were treated with durvalumab plus tremelimumab included fatigue (all grades, 43.1%; n = 53), diarrhea (31.7%; n = 39), pruritus (23.6%; n = 29), nausea (13.8%; n = 17), skin and subcutaneous tissue disorders (8.9%; n = 11), and hypothyroidism (9.8%; n = 12). The only grade 4 (0.8%; n = 1) and grade 5 (0.8%; n = 1) toxicities reported were diarrhea.
Although objective responses were infrequent using both RECIST v1.1 criteria and irRECIST criteria, durvalumab plus tremelimumab met the predefined threshold of 9-month OS and should be further explored in this patient population, noted Capdevila.
“No new safety concerns were observed in this population of advanced NENs,” Capdevila concluded. “As a final thought, the CBR based on prior studies may not reflect the clinical behavior of this advanced, heavily pretreated population.”