Avelumab in patients with previously treated metastatic adrenocortical carcinoma: phase 1b results from the JAVELIN solid tumor trial

Adrenocortical carcinoma (ACC) is a rare endocrine tumor with an incidence of approximately 0.7 cases/1 million population per year [1]. Genomic studies indicate that mutations in defined driver genes, including zinc and ring finger 3 (ZNRF3), telomerase reverse transcriptase (TERT), protein kinase cAMP-dependent type I regulatory subunit alpha (PRKAR1A), ribosomal protein L22 (RPL22), telomeric repeat binding factor 2 (TERF2), cyclin E1 (CCNE1), and neurofibromin 1 (NF1), may have a major role in the etiology of ACC [2], and that ACC has a moderate overall mutation burden compared with other tumor types [3]. Patients with ACC have a poor prognosis, with an estimated 5-year survival rate for metastatic ACC (mACC) of < 20% [4–6]. Therapeutic options for mACC are limited and include surgery; platinum agents, mitotane, streptozocin, and other systemic therapies; and locoregional radiotherapy [7]. Although cytotoxic treatments are widely administered, they are associated with limited efficacy and high toxicity, and targeted agents have not shown clinically meaningful activity in this disease [8]. Thus, there is an urgent need for new treatment options.

Cancer cells can exploit the programmed death 1 (PD-1)/programmed death ligand 1 (PD-L1) immune checkpoint pathway to promote an immunosuppressive environment and allow tumor growth. Specifically, PD-L1 expressed on tumor cells binds to PD-1 expressed on activated T cells, resulting in downregulation of antitumor immune responses [9, 10]. Monoclonal antibodies targeting the PD-1/PD-L1 axis have shown clinical activity in multiple tumor types [11]. PD-L1 is expressed in ACC tissues throughout various stages of disease, providing a rationale for assessing anti–PD-1/PD-L1 antibodies in mACC [12].

Avelumab is a human IgG1 monoclonal antibody with a wild-type Fc region that was designed to specifically bind and block PD-L1 [13]. Unlike other anti–PD-L1/PD-1 antibodies in clinical practice, avelumab is unique in that preclinical models show that it can induce antibody-dependent cellular cytotoxicity of tumor cells [14, 15]. Thus, avelumab may engage both the adaptive and innate immune systems. Avelumab treatment has been associated with durable responses in several tumor types [16–18], and avelumab has been approved in various countries for the treatment of patients with metastatic Merkel cell carcinoma and in the United States and Canada for treatment of locally advanced or metastatic urothelial carcinoma that progressed during or after platinum-containing chemotherapy [19]. Here we report results from a phase 1b cohort of patients with previously treated mACC who received avelumab.

Of 50 patients enrolled, 3 had confirmed PRs, and the ORR was 6.0% (95% CI, 1.3% to 16.5%) (Table 2). Of the 3 responding patients, 2 were receiving concomitant mitotane (both had prior progression while receiving mitotane), and 2 had PD-L1+ tumors (based on 1% or 5% expression cutoffs). In patients who had received 1 (n = 13), 2 (n = 18), or ≥ 3 (n = 19) prior lines of systemic therapy since diagnosis, ORR was 15.4%, 5.6%, and 0%, respectively. Response was ongoing in 1 patient, who had a duration of response of 19.4 months at data cutoff (Fig. 1); this was a man aged 63 years with a PD-L1− tumor who had received prior treatment with cisplatin, etoposide, and mitotane and had target lesions in the liver and lymph node (Additional file 1: Figure S1). The patient received avelumab with concurrent mitotane, had a PR documented at the second assessment (week 13), and remained on treatment without progression throughout 22.0 months of follow-up. Twenty-one patients had a best response of stable disease, resulting in a disease control rate of 48.0%, which was the same in patients with or without concomitant mitotane. Using immune-related response criteria, ORR was 6.0% (95% CI, 1.3% to 16.5%), and the disease control rate was 58.0% (Additional file 1: Table S1). Of 48 patients evaluable for change in tumor size, 16 (33.3%) had a reduction of any level vs baseline, including a reduction of ≥30% in 4 patients (8.0%) (Fig. 2a). One patient had an initial assessment showing progressive disease (~ 100% increase vs baseline and new lesion at week 7) but later assessments showed a decrease vs baseline (nadir − 74%; Fig. 2b).

Response	With Mitotane (n = 25)	Without Mitotane (n = 25)	Overall (N = 50)
Confirmed best overall response, n (%)
Complete response	0	0	0
Partial response	2 (8.0)	1 (4.0)	3 (6.0)
Stable disease	10 (40.0)	11 (44.0)	21 (42.0)
Progressive disease	10 (40.0)	13 (52.0)	23 (46.0)
Not evaluable	3 (6.0)	0	3 (6.0)
ORR (95% CI), %	8.0 (1.0–26.0)	4.0 (0.1–20.4)	6.0 (1.3–16.5)
Disease control rate, %	48.0	48.0	48.0

Median PFS was 2.6 months (95% CI, 1.4 to 4.0), and PFS rates at 6 months and 1 year were 20.9% (95% CI, 10.6% to 33.5%) and 8.7% (95% CI, 2.6% to 19.6%), respectively (Fig. 3a). Using immune-related criteria, median PFS was 3.8 months (95% CI, 2.4 to 5.5). Median OS was 10.6 months (95% CI, 7.4 to 15.0), and the 1-year OS rate was 43.4% (95% CI, 27.9% to 57.9%) (Fig. 3b).

PD-L1 expression was evaluable in tumor samples from 42 patients (84%). In patients with PD-L1+ (n = 12) or PD-L1– (n = 30) tumors based on a 5% PD-L1 expression cutoff, ORR (95% CI) was 16.7% (2.1% to 48.4%) vs 3.3% (0.1% to 17.2%) (P = .192), median PFS was 5.5 (1.3 to 8.2) vs 1.7 (1.4 to 4.0) months (HR, 0.66 [95% CI, 0.32 to 1.39]), and median OS was 14.4 (7.4 to 14.4) vs 10.6 (7.3 to not evaluable) months (HR, 0.91 [95% CI, 0.33 to 2.49]) (Fig. 3c and Fig. 3d). Findings were similar based on analyses of PD-L1 expression using a 1% cutoff (Additional file 1: Table S2 and Figure S2). Analyses of higher tumor cell PD-L1 cutoffs and PD-L1 expression on tumor-infiltrating immune cells were not informative because few patients had PD-L1+ tumors (0 to 5 patients).

In this single-arm phase 1b study, which to our knowledge is the largest prospective trial of a checkpoint inhibitor in mACC, avelumab showed antitumor activity with an acceptable safety profile in a platinum-treated population. Three patients (6.0%) had an objective response, including patients with PD-L1+ and PD-L1− tumors, and all of whom had received only 1 (n = 2) or 2 (n = 1) prior lines of treatment. This suggests that the activity of avelumab might be greatest in patients with limited pretreatment, although the small patient numbers in this study prevent any definitive conclusion. Reasons for improved response in patients with less pretreatment may include a smaller tumor burden, decreased proportion of treatment-resistant cells within the tumor, and reduced immunosuppression associated with multiple prior lines of chemotherapy. Although the ORR and median PFS were modest in this heavily pretreated population, the disease control rate was 48.0%, median OS was 10.6 months, and the 1-year OS rate was 43.0%. No association was seen between concomitant mitotane treatment and clinical activity of avelumab, although the absence of detailed patient data relating to ongoing mitotane treatment, including drug levels, is a limitation of the study.

Apart from the current study, other data reported in ACC with anti–PD-L1/PD-1 agents are preliminary findings from phase 2 studies of nivolumab and pembrolizumab in patients with previously treated advanced ACC. Of 7 patients who received nivolumab, 5 had a best response of disease progression and 2 were awaiting evaluation [22]. Of 11 patients who received pembrolizumab, 2 had a PR, 1 achieved stable disease, and the 6-month PFS rate was 27% [23]. In addition, in the phase 1a study of avelumab in patients with various advanced malignancies, a PR occurred in a patient with ACC [13].

The manageable safety profile of avelumab seen in patients with mACC was consistent with experience in other tumor types [16–18]. Patients receiving concomitant mitotane had a higher rate of grade 3 TRAEs than those not receiving mitotane (24.0% vs 8.0%), particularly liver enzyme elevations. This reflects the known toxicity profile of mitotane, which includes hepatic, gastrointestinal, neurological, and hematologic AEs [24]. However, our study showed that the tolerability of avelumab and mitotane in combination is acceptable.

Current treatment options for patients with mACC are highly limited. In the first-line setting, response rates with mitotane monotherapy are estimated to be approximately 10% to 30%, although data from prospective trials are lacking [7]. In a randomized phase 3 study of mitotane combined with either etoposide, doxorubicin, and cisplatin or streptozocin in patients with unresectable ACC without prior treatment (except mitotane), the ORR was 23.2% vs 9.2% (P < .001), the disease-control rate was 58.3% vs 31.4% (P < .001), median PFS was 5.0 vs 2.1 months (P < .001), median OS was 14.8 vs 12.0 months (P = .07), and serious AEs occurred in 58.1% vs 41.6% of patients [25]. In a phase 2 trial of gemcitabine plus metronomic fluoropyrimidine as second-/third-line treatment in patients with advanced ACC who were receiving ongoing mitotane treatment (n = 28), the ORR was 7.1%, disease control rate was 46.4%, median time to progression was 5.3 months, and median OS was 9.8 months; grade 3/4 AEs were leukopenia (21.4%), thrombocytopenia (3.5%), and mucositis (3.5%) [26]. Thus, the results from our study indicate that avelumab has comparable clinical activity and may be better tolerated than existing treatment options for this hard-to-treat cancer. A randomized phase 2 study in non-small-cell lung cancer has shown that combining an anti–PD-1 antibody with platinum-based chemotherapy increased the ORR and prolonged PFS vs chemotherapy alone [27]. This suggests that studies in ACC of avelumab in combination with chemotherapy or as maintenance therapy after first-line induction chemotherapy are warranted.

Targeted molecular therapies have been explored in patients with previously treated advanced ACC, but efficacy has been poor or nonexistent. In a phase 3 trial of linsitinib (insulin-like growth factor receptor 1 and insulin receptor inhibitor) vs placebo in 139 patients with previously treated ACC, of whom 55% had received prior cisplatin, ORR was 3.3% vs 0%, disease control rate was 32.2% vs 34.7%, median PFS was 44 vs 46 days (P = .30), and median OS was 323 vs 356 days (P = .77) [28]. Treatment with erlotinib (tyrosine kinase inhibitor [TKI] of epidermal growth factor receptor) plus gemcitabine in 10 patients resulted in tumor-size reduction in 1 patient, whereas 8 patients had progressive disease at first evaluation [29]. In other studies, no patient treated with combinations of bevacizumab (anti-VEGF antibody) plus capecitabine [30] or sorafenib (multi-TKI) plus paclitaxel [31] had an objective response or stable disease. In a study of axitinib (VEGF receptor inhibitor), none of 13 patients had an objective response, 4 patients showed decreased tumor growth, and median PFS was 5.5 months [32]. Among 35 evaluable patients who received sunitinib (multi-TKI), 14.2% had stable disease as best response and median PFS was 2.8 months [33]. In a study assessing cixutumumab (anti-insulin growth factor 1 receptor antibody) plus temsirolimus (mTOR inhibitor), 11 of 26 patients (42.3%) had a best response of stable disease [34]. Thus, the activity of immunotherapy may be more promising than seen to date in studies of targeted therapies.

In patients with mACC previously treated with chemotherapy, avelumab had moderate clinical activity either as monotherapy or with concurrent mitotane (50% of patients), particularly in those with limited pretreatment or PD-L1+ tumors, and with an acceptable safety profile. These findings provide a rationale for studies of avelumab-based combination treatment in patients with advanced ACC.