Chemotherapy in combination with anti-PD-1 agents as adjuvant therapy for high-risk oral mucosal melanoma

Abstract

Background

Adjuvant therapy plays a critical role in the treatment of oral mucosal melanoma (OMM). Anti-programmed cell death-1 (PD-1) agents are recommended as front-line therapy for metastatic melanoma, but their efficacy as adjuvant therapy for high-risk OMM remains unclear.

Patients and methods

A single-center, retrospective cohort study was conducted in 193 nodular-type oral mucosal melanoma (NOMM) patients who received chemotherapy alone or in combination with high-dose interferon-α2b (HDI) or anti-PD-1 agents as adjuvant therapy. Multivariate analysis was performed to identify significant prognostic factors for the 2-year overall survival (OS) and progression-free survival (PFS).

Results

Tumor thickness, ulceration and invasion level were found to be independent prognostic factors for both 2-year OS and PFS, while T-stage was only associated with OS. The 2-year OS and PFS were 43.5% and 10.9% in patients who received only chemotherapy. In comparison, the 2-year OS was improved, albeit not significantly (47.4%; p > 0.05), and PFS was significantly improved (43.6%; p = 0.0028) in patients who received chemotherapy plus HDI; and both 2-year OS (71.0%; p = 0.0118) and PFS (53.6%; p = 0.0001) were significantly improved in patients received chemotherapy plus anti-PD-1. The serious adverse event (SAE) (p < 0.0001) and discontinued treatment due to SAE (p < 0.0001) were significantly lower in patients who received anti-PD-1 than in patients who received HDI.

Conclusions

Invasion level and tumor thickness are independent prognostic factors for NOMM. Chemotherapy plus anti-PD-1 agents seem to be the adjuvant therapy of choice for NOMM, as it is safer and more tolerable than HDI and, more importantly, it can significantly improve the OS and PFS.

Introduction

Oral mucosal melanoma (OMM) is a highly aggressive malignant tumor arising from the oral cavity with a 5-year overall survival (OS) of 6.6–40.0% over the past three decades (Hicks and Flaitz 2000; Lopez-Graniel et al. 1999; Yamada et al. 2017; Perri et al. 2017). Nodular OMM (NOMM) is found to be the most malignant subtype of OMM with a 5-year OS of less than 10%, and most patients would die within two years after diagnosis. NOMM is also the most common subtype of OMM, accounting for 60% of OMM in the clinic. NOMM is associated with a high risk of neck recurrence and distant metastases even following definitive local therapy (Lopez-Graniel et al. 1999; Tanaka et al. 1994; Wu et al. 2014, 2018; Ma et al. 2017). However, no adjuvant therapy is recommended for NOMM in the National Comprehensive Cancer Network (NCCN) Guidelines (Version 1, 2021).

High-dose interferon-α2b (HDI) and chemotherapy are recommended as adjuvant therapy for mucosal melanomas in the Guidelines of the Chinese Society of Clinical Oncology (CSCO) (Version 2021), and a randomized trial has demonstrated that both of them can significantly improve the outcome of mucosal melanomas (Lian et al. 2013). However, as HDI and chemotherapy alone may not be effective for high-risk OMMs, chemotherapy in combination with HDI is suggested for NOMM in the 9th Hospital of Shanghai Jiaotong University. Immune checkpoint agents can improve the outcome of advanced melanoma, and anti-programmed cell death-1 (PD-1) agents have also been used as adjuvant therapy of NOMM since 2016 in our hospital.

NOMM is very aggressive and may invade blood vessels, and thus it has a high risk of recurrence or metastasis. For this reason, postoperative adjuvant therapy is needed in NOMM patients. The main purpose of this study is to determine the optimal adjuvant therapy for NOMM.

Patients and methods

NOMM patients treated in the 9th Hospital of Shanghai Jiaotong University from May 2009 to May 2019 were retrospectively reviewed if the primary oral melanoma was completely controlled and radical resection of involved cervical lymph nodes (CLN) was performed. Patients with Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 or 1, no evidence of distant metastases, and at least one form of adjuvant therapy were eligible for this study. Patients with macular, metastatic or amelanotic melanomas were excluded. Finally, a total of 193 NOMM patients were included in this study.

Treatment protocols

Definitive local therapy was administered to all patients upon diagnosis, and 2–3 weeks later dacarbazine and cisplatin were repeated every 3 weeks for 4 cycles. Dacarbazine was administered at a dose of 250 mg/m² for 4 days; and CDDP was administered at a dose of 75 mg/m² with hydration. HDI or Anti-PD-1 was administered for at last a year in the absence of recurrence or metastasis. For HDI, recombinant human IFN-α-2b (Intron A, 15 MU/m² per day) was injected intravenously 5 times a week for 4 weeks, followed by administration of 9MU IFN-α-2b 3 times per week for 48 weeks. For anti-PD-1, pembrolizumab was administered at a dose of 200 mg every 3 weeks for a year; nivolumab or toripalimab was administered at a dose of 3 mg/kg every 3 weeks for a year.

Prognostic factors

The following clinical or pathological factors were evaluated: sex, age, anatomic site, ECOG score, tumor size, tumor thickness, T-stage, CLN status, ulceration and invasion level. Tumor thickness was measured by computed tomography (CT) or magnetic resonance imaging (MRI). Invasion level was defined as described in our previous study. Level I (noninvasive, in situ): all tumor cells are above the basement membrane; level II (micro-invasive, partial in situ): most tumor cells are above the basement membrane and small cell clusters have invaded into the superficial lamina propria; level III (partial invasive): most tumor cells are above the superficial lamina propria and small cell clusters have invaded into the lamina propria; level IV (invasive): most tumor cells have invaded into the lamina propria or deeper.

The PD-L1 expression was detected by immunohistochemistry staining with SP263 antibody in all patients who received anti-PD-1 agents, and PD-L1 positive status was defined as the presence of membrane staining of any intensity in 1% or more of tumor cells or the presence of PD-L1 staining of any intensity in tumor-infiltrating immune cells covering 1% or more of tumor area occupied by tumor cells, associated intratumoral cells, and contiguous peritumoral stroma (Sheng et al. 2019).

Statistical analysis

OS was calculated from the date of adjuvant therapy to the death or the last follow-up on July 1, 2021; Progression free-survival (PFS) was calculated from the date of adjuvant therapy to radiological progression, death or the last follow-up on July 1, 2021. Patients who were still alive at the last follow-up were censored. The 2-year OS and PFS were estimated using the Kaplan–Meier method. All statistical analyses were performed using SAS 9.13. The statistical significance of differences between survival curves was determined by the log-rank test, and multivariate analysis was performed with the Cox proportional hazard model. Adverse events and laboratory abnormalities were collected from patients’ medical records and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (Version 4.0).

Results

Clinical and pathological features

A total of 193 patients (105 males and 88 females; mean age at diagnosis: 53 years; age range: 16–82 years; 134 patients aged < 60 years and 59 patients aged ≥ 60 years) were included in this study. Most lesions occurred in the mucosa covering the maxillary or mandible bone, including the hard palate or upper gingiva (n = 154) and lower gingiva (n = 36); while the rest occurred in the buccal mucosa (n = 3) (Fig. 1). Tumor size ranged from 0.8 to 10 cm, and it was < 4 cm in 62 patients and ≥ 60 cm in 131 patients. Tumor thickness was < 1 cm in 148 patients and ≥ 1 cm in 45 patients. Positive CLN was found in 116 patients, and stage III and IV were found in 169 and 24 patients, respectively. Ulceration and invasion level were reviewed by two pathologists in our hospital. Ulceration was found in 137 (71%) out of 170 patients, 102 (53%) of which were confirmed to have level IV invasion.

Fig. 1.

Nodular oral mucosal melanoma occurred on maxillary gum (A), hard palate (B), mandible gum (C), and buccal (D)

Adjuvant therapy

Chemotherapy was administered in the first stage of adjuvant therapy for high-risk OMM. Of the 193 patients, 150 completed the four-cycle chemotherapy with dacarbazine and CDDP, while 43 did not due to disease progression, intolerability or patients’ refusal. HDI or anti-PD-1 was administered in the second stage of adjuvant therapy. Of the 150 patients who completed chemotherapy, 3 refused to participate in the second-stage adjuvant therapy, 78 received at least one cycle of HDI treatment and 69 patients at least one cycle of anti-PD-1 treatment. The demographics of the cohorts were shown in Table 1.

Table 1.

The characteristics of patients of by adjuvant therapy

	C alone	C + IFN	C + PD-1	P
Age				0.28
< 60 years	34	57	43
≥ 60 years	12	21	26
Sex				0.94
Men	24	43	38
Women	22	35	31
Location				0.13
Maxilla	34	60	60
Mandible	12	16	8
ECOG score				0.59
0	21	32	25
1	25	46	44
Staging				0.98
III	12	21	19
IV	34	57	50

C Chemotherapy, CLN cervical lymph node, IFN interferon, PD-1 programmed death-1, ECOG Eastern Cooperative Oncology Group

The univariate analysis revealed that sex, age and primary site were not significant prognostic factors for the 2-year OS or PFS, and tumor size, tumor thickness, T-stage, CLN status, ulceration and invasion level were significant prognostic factors for the 2-year OS and PFS. The multivariate analysis revealed that invasion level (Fig. 2a, b), ulceration (Fig. 2c, d) and tumor thickness (Fig. 2e, f) were independent prognostic factors for both OS and PFS, while T-stage was only associated with OS. In conclusion, invasion level was the most significant prognostic factor for NOMM (Table 2).

Fig. 2.

The 2-year overall survival (A) and progression-free survival (B) of patients with different invasion level; The 2-year overall survival (C) and progression-free survival (D) of patients with or without ulceration; The 2-year overall survival (E) and progression-free survival (F) of patients with tumor thickness ≤ 1 cm or > 1 cm

Table 2.

2-year OS and PFS of all patients by Prognostic Variables

Variable	No.	2-year	P (chi-square)		2-year	P (chi-square)
		OS %	U	M	PFS %	U	M
Sex			0.99	–		0.20	-
Men	105	55.2			34.3
Women	88	54.5			45.4
Age			0.67	–		0.62	-
< 60 years	134	56.0			38.1
≥ 60 years	59	52.5			42.4
Primary site			0.28	–		0.10	-
Maxilla	154	57.1			42.2
Mandible	36	47.2			27.8
Buccal	3	33.3			33.3
Tumor size			0.0017 (9.81)	0.1972 (1.66)		0.0017 (9.84)	0.0857 (2.95)
< 4 cm	62	71.0			54.8
≥ 4 cm	131	47.3			32.1
Tumor thickness			< 0.0001 (39.00)	0.0210 (5.32)		< 0.0001 (25.51)	0.0069 (7.30)
≤ 1 cm	148	64.9			44.6
> 1 cm	45	24.4			22.2
Invasion level			< 0.0001 (51.19)	< 0.0001 (47.42)		< 0.0001 (39.19)	< 0.0001 (35.18)
≤ III	91	80.2			56.0
IV	102	32.4			24.5
Ulceration			< 0.0001 (17.59)	0.0089 (6.83)		< 0.0001 (16.04)	0.0054 (7.73)
Absent	56	78.6			41.7
Present	137	45.3			13.5
T-stage			< 0.0001 (28.57)	0.0541 (3.71)		< 0.0001 (16.44)	0.0677 (3.34)
III	169	60.4			42.6
IV	24	16.7			16.2
CLN			0.0002 (14.08)	0.0015 (10.05)		0.0008 (11.33)	0.0240 (5.09)
Positive	116	44.0			30.2
Negative	77	71.4			45.6
Treatment			0.0004 (15.68)			< 0.0001 (29.09)
C alone	46	43.5		–	10.9		–
C + IFN	78	47.4		0.1943 (1.66)	43.6		0.0028 (8.95)
C + PD-1	69	71.0		0.0118 (6.35)	53.6		0.0001 (14.50)

U univariate, M multivariate, C chemotherapy, CLN cervical lymph node, IFN interferon, PD-1 programmed death-1

Of the 193 patients, 46 received only chemotherapy, 78 received chemotherapy plus HDI, and 69 received chemotherapy plus anti-PD-1. The prognosis was best in patients who received chemotherapy plus anti-PD-1 with the 2-year OS and PFS being 71% and 53.6%, followed by patients who received chemotherapy plus HDI with the 2-year OS and PFS being 47.4% and 43.6%, and then by patients received only chemotherapy with the 2-year OS and PFS being 43.5% and 10.9%, respectively. The cox proportional hazard model showed that compared to chemotherapy, chemotherapy plus anti-PD-1 significantly improved both OS and PFS, while chemotherapy plus HDI only improved PFS (Table 2).

PD-L1 expression

We also explored the relationship between PD-L1 expression and the therapeutic efficacy of anti-PD-1. It was found that 23 patients were PD-L1 negative and 46 patients were PD-L1 positive (Table 2). PD-L1 positive patients had higher 2-year OS (73.9% v.s. 65.2%) and PFS (60.9% v.s. 39.1%) than PD-L1 negative patients, but the difference was not statistically significant (Table 3).

Table 3.

2-year OS and PFS by PD-L1 expression

PD-L1 expression		NO. of patient	2-year OS %	p	2-year PFS %	p
Results				0.3763		0.0557
Negative (23)	0%	6	65.2		39.1
	< 1%	17
Positive(46)	1–5%	6	73.9		60.9
	6–10%	15
	11–25%	9
	26–49%	5
	≥ 50%	11

Adverse events and discontinued treatment

The three most common adverse events were decreased appetite (181, 94%), gastrointestinal disorders (169, 88%) and anaemia (116, 60%) in patients who received only chemotherapy, and serious adverse events (grade 3–4) occurred in 93 (48%) patients, 35 (18%) of which discontinued chemotherapy. The three most common adverse events were fatigue (55, 71%), fever (51, 65%) and hepatotoxicity (49, 63%) in patients who received chemotherapy plus HDI, and serious adverse events (grade 3–4) occurred in 46 (59%) patients, 25 (32%) of which discontinued HDI treatment. The three most common adverse events were anaemia (38, 55%), fatigue (34, 49%), and diarrhea (31, 45%) in patients who received chemotherapy plus anti-PD-1, and serious adverse events (grade 3–4) occurred in 11 (16%) patients, only one of which (1%) discontinued anti-PD-1 treatment because of diarrhea (grade 4). It was concluded that anti-PD-1 was safer and more tolerable than HDI (p < 0.0001) (Table 4).

Table 4.

Adverse events (AE) and discontinued treatment due to serious AE (SAE)

Group (no.)	AE		No. of discontinued treatment due to SAE (%)
	Top three, (no.%)	No. of grade 3–4(%)
Chemotherapy (193)		93 (48%)	35 (18%)
	Decreased appetite (181, 94%)
	Gastrointestinal disorders (169,88%)
	Anaemia (116,60%)
IFN (78)		46 (59%)	25 (32%)
	Fatigue (55, 71%)
	Fever (51,65%)
	Hepatotoxicity (49, 63%)
Anti-PD-1(69)		11 (16%)	1(1%)
	Anaemia (38, 55%)
	Fatigue (34, 49%)
	Diarrhea (31, 45%)
P (Chi-square) IFN vs Anti-PD-1		< 0.0001(28.56)	< 0.0001(23.55)

Form of first relapse	Oral recurrence no. (%)	Neck recurrence no. (%)	Distant metastasis no. (%)
Chemotherapy alone (n = 46)	29 (63%)	7 (17%)	5 (11%)
Chemotherapy + IFN (n = 78)	26 (33%)	9(12%)	9 (12%)
Chemotherapy + PD-1 (n = 69)	21 (30%)	8(12%)	2 (3%)
P (IFN vs Anti-PD-1)	0.71	0.99	0.047

Recurrence and distant metastasis

The 2-year PFS of the 193 patients was 39%, and 117 (61%) patients had the first relapse within two years, including oral recurrence in 76 patients, neck recurrence in 24 patients, and distant metastasis in 16 patients. New melanin plaque was the most common form of oral recurrence, which was mostly found at the border of the original surgical area (Fig. 3a) and occasionally in the contralateral area (Fig. 3b). Compared with chemotherapy, chemotherapy plus HDI or anti-PD-1 significantly reduced the oral recurrence rate and improved the PFS (Table 4). Lung was the most common site of distant metastasis, which occurred in 14 patients. Compared with HDI, PD-1 treatment significantly reduced the distant metastasis rate (p = 0.047).

Fig. 3.

Recurrent oral melanin plaque occurred at the border of the original surgical area (A), and in the contralateral gum (B)

Discussion

OMM is a highly aggressive neoplasm with a poor long-term outcome. NOMM is the middle or late stage of oral melanoma and it often has a quick vertical growth phase like the nodular cutaneous melanoma (CM)(Susok et al. 2021; Robsahm et al. 2021; Allais et al. 2021). As expected, the outcome of NOMM is even worse with a 5-year OS of < 20%. In this study, invasion level and tumor thickness are found to be independent prognostic factors of NOMM, both of which are related to the vertical distance and thus could better characterize the growth of NOMM. In line with this, Breslow thickness has also been demonstrated to be the most important prognostic factor of CM (Gershenwald et al. 2017).

Mucosal melanoma is the second most common subtype of malignant melanomas in Asians(Xu et al. 2019; Cui et al. 2018). There is substantial evidence that adjuvant treatment is effective in improving the outcome of high-risk melanoma, including OMM (Lian et al. 2013; Hillner 1998; Kirkwood et al. 1996; Wang et al. 2015; Flaherty et al. 2014). Previous randomized trials have also suggested that HDI can significantly improve the relapse-free survival rather than the overall survival of patients with high-risk melanoma (Kirkwood et al. 2001). A randomized trial showed that chemotherapy with temozolomide and cisplatin was more effective than HDI in improving relapse-free survival in patients with resected mucosal melanoma (Lian et al. 2013). Thus, chemotherapy is often considered the treatment of choice for all OMMs. However, it should be noted that chemotherapy alone may not be enough for high-risk OMMs, as relapse or metastatic diseases occur in nearly 90% of NOMM patients within 2 years (Wu et al. 2014). Chemotherapy and HDI are recommended by CSCO as adjuvant therapy for resected mucosal melanoma, and thus they are combined for the treatment of NOMM in this study. The results suggest that chemotherapy in combination with HDI could significantly improve the PFS of NOMM compared to chemotherapy alone. However, it may result in higher toxicity, and nearly 60% of patients have grade 3–4 AEs during the treatment, and more than 30% of patients discontinued treatment because of AEs.

Anti-PD-1 agents are safe and tolerable and thus recommended as front-line therapy for metastatic melanoma by NCCN and CSCO guidelines (Koelblinger et al. 2021; Tang et al. 2019; Hamid et al. 2013; Yun et al. 2016; Ramelyte et al. 2017). However, they are seldom used as adjuvant therapy for OMM. The chemotherapy in combination with HDI is more effective but also more toxic (Flaherty et al. 2014). Therefore, there is an urgent need to find a safer alternative to HDI. In this study, chemotherapy in combination with anti-PD-1 is administered as adjuvant therapy for high-risk OMM, and the results show it can improve the OS and PFS of NOMM, and it is more tolerable than HDI and only one patient discontinued treatment because of toxicity. It is further found that HDI improves the PFS mainly by reducing the oral recurrence rate, while anti-PD-1 agents could reduce distant metastasis and oral recurrence rate. However, both of them could not significantly reduce neck recurrence.

Based on these results there seems to be an improved outcome when combining Chemo to PD-1 compared to the other adjuvant treatments. Further prospective randomized studies are required to confirm this benefit. Invasion level and tumor thickness are independent prognostic factors of NOMM, which indicate that vertical phase growth is an advanced stage of OMM.

Limitations of the retrospective study

The Charlson Index is a tool used to assess probable mortality for patients with multiple serious illnesses, which may have an impact on the patient’s survival. Since Charlson Index was not recorded in the original medical record, it is regrettable that the Charlson Index was not included in the study. As a retrospective study, this study has many other limitations, including selective bias and memory bias.

Abbreviations

OMM

Oral mucosal melanoma

OS

Overall survival

NOMM

Nodular oral mucosal melanoma

CSCO

Chinese Society of Clinical Oncology

NCCN

National Comprehensive Cancer Network

HDI

High-dose interferon-α2b

PD-1

Programmed cell death-1

CLN

Cervical lymph nodes

ECOG

Eastern Cooperative Oncology Group

PS

Performance status

CT

Computed tomography

MRI

Magnetic resonance imaging

PFS

Progression free-survival

CM

Cutaneous melanoma

Funding

This work was supported by Fundamental research program funding of Ninth People's Hospital affiliated with Shanghai Jiao Tong University School of Medicine [No. JYZZ073].

Declarations

Conflict of interest

The authors have declared no conflicts of interest.

Ethical approval

This research study was conducted retrospectively from data obtained for clinical purposes.

Informed consent

Because of the retrospective nature of this study, consent was not obtained.