Continuous Infusion of Endostar Combined with Chemotherapy in Patients with Advanced or Recurrent Mucosal Melanoma: A Real-World Cohort Study

and The we conducted this study to test continuous intravenous Endostar infusion plus chemotherapy in this in the

China. 6,7 Data on mucosal melanoma remain limited, and most studies are derived from Asian countries due to ethic disparities in prevalence.
Recombinant human endostatin (Endostar) has an additional 9-amino-acid sequence at the N-terminus of the protein and a six-histidine tag. Since 1997, Endostar has been reported to e ciently block angiogenesis and suppress tumor growth, with an acceptable safety pro le in previous clinical studies. [8][9][10][11][12] In patients with advanced non-small-cell lung cancer (NSCLC), Endostar plus standard chemotherapy improved survival and was thus approved for the treatment of NSCLC by the State Food and Drug Administration of China in 2005. 13 In addition, preclinical and clinical investigations have demonstrated the bene t of Endostar in multiple malignancies, including gastrointestinal cancer, colorectal cancer, ovarian cancer, nasopharyngeal carcinoma and extensive-stage small-cell lung cancer (SCLC). [14][15][16][17][18][19][20][21][22] Previously, a phase 2 study evaluated the e cacy and safety of Endostar plus dacarbazine in patients with metastatic melanomas and revealed superior progression-free survival (PFS, 4.5 vs. 1.5 months, p=0.013) and overall survival (OS, 12.0 vs. 8.0 months, p=0.005) in the Endostar group compared with the placebo group. Given the angiogenic nature of melanoma, the considerable survival bene t of Endostar combination treatment and the limited number of patients with mucosal melanoma in that study (7 in the placebo group and 9 in the Endostar group). Endostar-containing treatment strategies are worth further investigation in advanced mucosal melanoma. 23 In the present study, we evaluated the e cacy and safety of Endostar plus polychemotherapy (dacarbazine/temozolomide and cisplatin) in a cohort of patients with advanced or relapsed mucosal melanoma. Endostar is commonly administered at 7.5 mg/m 2 once daily for 14 days, but this real-world cohort study aimed to evaluate the e cacy and safety of continuous Endostar infusion for 168 hours combined with chemotherapy. We also retrospectively analyzed the prognostic factors of this cohort and summarized our experience in treating patients with advanced or recurrent mucosal melanoma in the era of immunotherapy.

Study population
Patients with pathologically con rmed mucosal melanoma treated with Endostar-containing regimens in Fudan University Shanghai Cancer Center between April 2017 and August 2020 were retrospectively enrolled. Patients who were included in the present study also had advanced or recurrent disease, with a baseline Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0-1. The Endostarcontaining regimens were either Endostar plus dacarbazine and cisplatin or Endostar plus temozolomide and cisplatin. Endostar was given by continuous infusion at a dosage of 105 mg/m 2 for 168 hours (Civ 168h) in the present study.

Assessment and statistical analysis
Adverse events (AEs) were assessed by the investigator according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE), version 4.03. Radiologic response was evaluated per the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. ORR was de ned as the proportion of patients achieving a complete response (CR) or partial response (PR). PFS refers to the interval from the rst date of Endostar-containing treatment to date of radiologic progression, initiation of new treatment, patient death, or last follow-up. OS was calculated from the initiation of Endostarcontaining treatment to patient death from any cause or last follow-up. Patients who were alive at the last follow-up were censored. Survival was calculated using the Kaplan-Meier method and compared by the log-rank test. Cox proportional hazards regression was applied to identify potential prognostic indicators. A P value of 0.05 or less (2-sided test) was considered signi cant. All these analyses were conducted with R software, version 3.3.3 (http://www.R-project.org) and SPSS 22.0 (SPSS Inc, Chicago, IL).
This study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center.

Data availability
The dataset used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Patient characteristics and treatment details
Of the 43 patients with advanced or recurrent mucosal melanoma, 17 (39.5%) were males, and 26 (60.5%) were females, with a median age of 61 years (25-74 years). The most common primary site was the gastrointestinal tract (n=22, 51.2%), followed by the sinonasal (n=14, 32.6%) and genitourinary tracts (n=6, 14.0%). A total of 32 (74.4%) patients had undergone surgical resection, of whom 18 had radical surgery (surgical resection margins negative on nal pathology). Before initiation of Endostar-containing treatment, most patients had metastatic disease (n=31, 72.1%), and 12 (27.9%) had locally advanced disease that was deemed unresectable. The majority of patients (n=33, 76.7%) had normal LDH levels, while 9 (20.9%) patients had elevated LDH levels. KIT, BRAF and NRAS mutation status was available for all but four patients. More than half of the patients (25/39, 64.1%) did not have mutations in these genes. The most commonly observed mutation was NRAS (n=9, 23.1%), followed by BRAF (n=3, 7.7%) and CKIT (n=2, 5.1%). Of the three patients with BRAF mutations, one patient harbored the BRAF V600E mutation, and the other two patients had the BRAF G606E and T599del mutations. The baseline patient characteristics are presented in Table 1.
The 43 patients received either Endostar plus temozolomide and cisplatin (n=24, 55.8%) or Endostar plus dacarbazine and cisplatin (n=19, 44.2%); this was the rst-line treatment in most patients (40/43, 93.0%). Three patients had rst-line anti-PD-1 treatment, and 11 patients received chemotherapy (paclitaxel or dacarbazine plus platinum) in the adjuvant setting. The median number of Endostar-containing treatment cycles was 4 (range: 1-6 cycles), and seven (16.3%) patients had chemotherapy dose adjustments due to drug-related toxicities. No patient discontinued Endostar during the treatment period.

E cacy
As of September 30, 2020, the median follow-up interval was 17.6 months (range 1.0-35.6 months). Of the 40 evaluable patients, an objective response was observed in 12 (30.0%) patents, four with CR (10%) and eight with PR (20.0%). Five responders (3 with anorectal primary sites and 2 with nasopharyngeal primary sites) received concurrent or sequential radiotherapy, and two received maintenance anti-PD-1 therapy following Endostar-containing treatment. Overall, disease control was achieved in 31 (77.5%) of the 40 patients. According to the primary site, the ORR was highest in those with disease originating from the sinonasal tract (38.5%, 5/13), followed by the genitourinary tract (28.6%, 2/7) and the gastrointestinal tract (25.0%, 5/20). Of the four patients with a CR, two had primary lesions located in the nasal cavity and paranasal sinuses, and two had primary lesions in the anorectum. All four patients had locally advanced disease and underwent concurrent radiotherapy (the two patients with sinonasal melanoma underwent proton heavy ion radiotherapy).

Univariate and multivariate survival analysis
The NLR and LMR cutoffs were determined using receiver operating characteristic (ROC) curves to predict progression at 4.9 months. Univariate analysis was performed and showed that disease stage (p=0.015), previous surgery (p=0.050), liver metastasis (p=0.013), elevated LDH level (p=0.006), and LMR >3.45

Subsequent treatment
As of Sep 30, 2020, 32 patients had experienced disease progression. Among them, 23 were subsequently treated with anti-PD-1/PD-L1 antibody alone or in combination with VEGFR-predominant tyrosine kinase inhibitors (TKIs); four underwent salvage chemotherapy, two patients harboring BRAF mutations were subsequently treated with vemurafenib, and one harboring a KIT mutation was treated with imatinib.

Discussion And Perspectives
Mucosal melanoma is a distinct and rare subtype of melanoma that presents in unique and often concealed anatomic locations, which results in more advanced stage at diagnosis. Local recurrence and metastatic spread commonly occur after surgical resection, with recurrence rates as high as 50-90%. 24 Thus, the survival rate for this population is dismal overall. Due to the rarity, to date, no consensus guidelines have been formulated on the optimal systemic therapy for mucosal melanoma.
Dacarbazine and temozolomide are longstanding cytotoxic agents used in the treatment of advancedstage melanoma. The ORRs in patients with metastatic melanoma are 12.1% and 13.1% with temozolomide and dacarbazine monotherapy, respectively. 25 Despite the increase in ORR with polychemotherapy, no signi cant improvement in OS has been detected. 26 Both agents are widely applied in patients with mucosal melanoma. However, limited e cacy data are available regarding these chemotherapeutic agents in mucosal melanoma patients. Although some evidence suggests that the response rate of mucosal melanoma is equivalent to that of cutaneous melanoma, worse outcomes have been reported in patients with mucosal melanoma treated with similar dacarbazine-based regimens. 24,27 Previously, the e cacy and safety of the Endostar plus dacarbazine regimen was explored in a phase 2 randomized study in patients with metastatic melanoma. A response was observed in 5 of the 56 (8.9%) patients receiving Endostar plus dacarbazine treatment. The PFS and OS durations of this group were 4.5 and 12.0 months, 23 respectively, which were slightly lower than those of the present cohort (PFS, 4.9 months; OS, 15.3 months). Noticeably, different from this phase 2 study, which included all treated metastatic melanoma patients (16 [14.5%] of which were mucosal melanoma patients), our study applied combination chemotherapy and focused on patients with advanced or recurrent mucosal melanoma, a population with a devastating prognosis and limited treatment options. The results of our study indicate the acceptable safety pro le of the Endostar plus chemotherapy combination and highlight the promising activity of this combination. Different from the conventional administration route, we used continuous Endostar infusion in addition to chemotherapy. Indeed, Endostar delivered by continuous intravenous injection has been explored previously. In a combined report from two phase 2 trials, Ma HL et al. described the favorable e cacy and safety pro le of continuous intravenous Endostar (7.5 mg/m 2 /24 hours, 120 hours) in unresectable stage III NSCLC. 28 Continuous Endostar infusion (7.5 mg/m 2 /day or 15 mg/m 2 /day 1-14) has also been investigated in patients with metastatic acral melanoma. The preliminary results of this study were presented at ASCO 2015, demonstrating that dacarbazine in combination with continuous Endostar infusion was well tolerated and e cacious in the 20 patients involved, with an expected median PFS of approximately 6 months. 29 Further detailed data from this study are still awaited. Likewise, the real-world data of the present study showed the potential bene t of combining continuous Endostar infusion with chemotherapy in patients with advanced or recurrent mucosal melanoma. Notably, despite the median PFS of 4.9 months for the entire cohort, patients (n=8) receiving concurrent/sequential radiotherapy or anti-PD-1 maintenance treatment achieved a median PFS of 12.1 months (95% CI 4.0 -20.0), which provides clues for further study.
Recently, the use of ICIs has been well established in melanoma, with pembrolizumab and nivolumab approved as standard rst-line therapies in advanced cutaneous melanoma. However, the poor e cacy of ICIs has been reported in mucosal melanoma, 3 with an even more disappointing ORR in Chinese patients (ORR 0-13.3%). 30,31 More recently, toripalimab in combination with a multitarget tyrosine kinase inhibitor has demonstrated increased e cacy compared to that of current treatments. With an ORR of 48.3% and a median PFS of 7.5 months in mucosal melanoma in a phase I trial, toripalimab plus axitinib has achieved orphan-drug designation by US Food and Drug Administration. 32 Nevertheless, a similar combination, toripalimab plus vorolanib (CM082), exhibited an ORR of 22.2% in the preliminary results of a phase II trial. 33 Thus, more data are warranted to identify the optimal rst-line treatment regimen for mucosal melanoma. LMR, derived from routine blood tests, has been suggested to re ect the host immune system and correlate with clinical outcomes in various cancers. [34][35][36] In melanoma, LMR has been indicated as a prognostic predictor of advanced melanoma treated with chemotherapy or checkpoint inhibitors. [37][38][39] Few studies have focused on the correlation between LMR and the survival of patients with mucosal melanoma. 40 Our study suggests that a high LMR is a strong indicator for favorable PFS and OS in patients with advanced and recurrent mucosal melanoma. Future studies on incorporating the LMR into the prognostication process are warranted to better stratify patient risk. Additionally, speci c therapies targeting systemic in ammation may be an effective option in patient subsets with a low LMR, which warrants further exploration. In a previous cohort, the association of mutations and survival was investigated among patients with sinonasal mucosal melanoma, 41 and no impact of mutation was found on patient survival. Noticeably, in the present study, the multivariate analysis results suggested that mutations were associated with favorable OS but not PFS, which could be explained by the subsequent use of targeted therapies, such as imatinib and vemurafenib. Such individual examples were observed in this cohort. One patient harboring the CKIT mutation progressed after two cycles of Endostar-containing treatment and was given 8 cycles of imatinib plus nab-paclitaxel/carboplatin. He achieved a PR following this treatment and was alive at the cutoff of this study.
There are several limitations to this study. In addition to the retrospective nature of the study, the missing detailed information regarding the initial surgery, along with anti-PD-1 maintenance and radiotherapy, may have introduced di culties and bias in the data analysis. Additionally, the number of patients included was small, and the follow-up period was relatively short. Nevertheless, the size of the study provided reasonable con dence in the ndings despite the above limitations.
In summary, our study demonstrated the promising activity and acceptable safety pro le of continuous Endostar infusion in combination with chemotherapy in the treatment of advanced or recurrent mucosal melanoma. These real-world data con rmed that a high LMR was correlated with favorable PFS and OS in this patient population. The detailed mechanisms of Endostar for the treatment of mucosal melanoma, its impact on the immune system and its use in combination with immunotherapy are worth investigations in the near future.  Table 2 Adverse events of any grade reported in at least 5% of patients and all reports of adverse events of grade 3-4.  Figure 1 Kaplan-Meier curve of progression-free survival (A) and overall survival (B) of the entire cohort.