1. Baseline profiles
A total of 56 patients with advanced melanoma who did not respond to the initial immune therapy and received anti-PD-1 antibody combined with IFN-α1b plus anlotinib as the second-line option were included. Demographic characteristics and disease-related profiles are presented in Table 1. The cohort consisted of 29 males(51.8%)and 27 females(48.2). The median age was 58.4 years old (32.1–80.6 yo). The leading mutated gene was cKIT (n = 12/56, 21.4%), followed by NRAS (n = 8/56,14.3%), while BRAF mutations were found in only 4 cases(7.1%). Half of the patients were negative for the above 3 gene mutations, and 4 (7.1%) patients’s mutation data were unavailable. IV M1c accounted for the greatest proportion of the cohort with 30.4% (17/56), IIIC-IIID and IV M1b accounted for the same proportion of 26.8% (15/56), IV M1a was 12.5% (7/56), and IV M1d was the least with only 3.6% (2/56). The acral subtype made up nearly a half of the enrolled cases (n = 27/56, 48.2%), the cutaneous subtype came second (n = 16/56, 28.6%) and the mucosal subtype was seen in 12/56 (21.4%). 26 patients (46.4%) had elevated LDH.
Table 1
|
n
|
%
|
Sex
|
|
|
Male
|
29
|
51.8
|
Female
|
27
|
48.2
|
Age
|
|
|
≤ 60 years old
|
30
|
53.6
|
>60 years old
|
26
|
46.4
|
median years
|
58.4
|
|
range
|
32.1–80.6
|
|
Stage
|
|
|
IIIC-IIID
|
15
|
26.8
|
IV M1a
|
7
|
12.5
|
IV M1b
|
15
|
26.8
|
IV M1c
|
17
|
30.4
|
IV M1d
|
2
|
3.6
|
LDH
|
|
|
Normal(≤ ULN)
|
30
|
53.6
|
Elevated(>ULN)
|
26
|
46.4
|
ECOG PS
|
|
|
0
|
46
|
82.1
|
≥ 1
|
10
|
17.9
|
Primary site
|
|
|
Cutaneous
|
16
|
28.6
|
Acral
|
27
|
48.2
|
Mucosal
|
12
|
21.4
|
Unkown
|
1
|
1.8
|
Front line treatment
|
|
|
PD-1 /IFN
|
3
|
5.4
|
PD-1 + IFN
|
53
|
94.6
|
Front-line treatment response
|
|
|
PD
|
54
|
94.6
|
SD
|
2
|
3.6
|
Gene mutation
|
|
|
Wild type
|
28
|
50
|
BRAF
|
4
|
7.1
|
NRAS
|
8
|
14.3
|
cKIT
|
12
|
21.4
|
Unkown
|
4
|
7.1
|
PD-1 Monoantibody type
|
|
|
Pembrolizumab
|
15
|
26.8
|
Toripalimab
|
39
|
69.6
|
Sintilimab
|
2
|
3.6
|
IFN dose
|
|
|
600µg
|
44
|
78.6
|
300µg
|
12
|
21.4
|
2. Efficacy
As of the data cutoff on 20 Sep 2023, the median follow-up time was 13.6 months (95% CI 10.7–16.5 months). The median overall survival (mOS) was 17.5 months (95% CI 12.4–22.6 months) (Fig. 1A)and the median progression-free survival (mPFS) was 2.8 months(95% CI 1.4–4.3 months) (Fig. 1B). The 6-month PFS rate was 21.4% (95% CI 10.7%-32.1%). The 6-month and 1-year OS rate was 80.4%(95% CI 70.0%-90.8%) and 35.7%(95% CI 23.2%-48.2%), respectively. By RECIST V1.1 assessment, only 1 (1.8%) patient achieved a complete response and 4 (7.1%) patients achieved a partial response, with an ORR of 8.9% and a DCR of 46.4% (Table 2). The patient with CR had acral melanoma, and of the 4 PR patients, 2 had acral melanoma, 1 had cutaneous melanoma, and 1 had mucosal melanoma. Patients with stage IIIC-IIID showed a significantly longer mOS than patients with M1c-M1d (29.9 months vs 10.87 months, p = 0.002, Fig. 2A). Unexpectedly, the mOS of cutaneous melanoma did not differ significantly compared to that of either acral or mucosal subtypes(p = 0.251 and 0.365, Fig. 2B). Also, the mOS in patients younger than 60 years old (29.9 months, 95% CI 7.196–52.604 months) was more than twice as the number in patients aged 60 years old or over (12.4 months, 95% CI 7.175–17.685 months), though the difference was not statistically significant (p = 0.146) (Fig. 2C). Univariable Cox regression identified stage was the only factor associated with worse overall survival (IV M1a-M1b: HR = 0.2, p = 0.009; IV M1c-M1d: HR = 0.4, p = 0.056) (Supplemental table 1) while no significant association was found between PFS and other variables of interest (Supplemental table 2).
Table 2
Objective response to treatment assessed per RECIST v1.1, by investigators
|
N
|
%
|
CR
|
1
|
1.8
|
PR
|
4
|
7.1
|
SD
|
21
|
37.5
|
PD
|
30
|
53.6
|
CR + PR
|
5
|
8.9
|
CR + PR + SD
|
26
|
46.40
|
3. Safety profile
56 enrolled patients all reported adverse events (AEs) of any grade, and 16 (28.6%) reported Grade 3 or 4 AEs. No treatment-related deaths were observed. 18 adverse events occurred in more than 10% of the patients, including fever and fatigue in 57.1% (n = 32), rash (n = 29, 51.8%), vitiligo(n = 26, 46.4%), decreased appetite (n = 22, 39.3%), hypothyroidism (n = 20, 35.7%), alanine/aspartate aminotransferase increased (n = 18, 32.1%), oral ulcer (n = 18, 32.2%), high blood pressure (n = 11, 19.6%), myalgia (n = 8, 14.3%), pruritus (n = 8, 14.3%), vomiting (n = 8, 14.3%), weight loss (n = 8, 14.3%), WBC count decrease (n = 8, 14.3%), hyperthyroidism(n = 8, 14.3%), joint pain(n = 7, 12.5%), hemorrhage(n = 6, 10.8%) and hand-foot syndrome(n = 6, 10.8%)(Table 3). These adverse events were introduced to regression models to examine the association with efficacy. We found patients who developed oral ulcers had a better ORR (OR = 12.5, 95% CI 1.5-106.2, p = 0.020) compared to those without oral ulcers (Supplementary Table 3). Multivariate Cox regression analysis identified rash as the only factor favorable to better PFS (HR = 2.1, 95% CI 1.1–4.3, p = 0.036) and OS(HR = 2.6, 95% CI 1.0-6.4, p = 0.044) (Supplementary Tables 4 and 5).
Table 3
Summary of Treatment-related irAEs
|
Any grade
|
≥Grade 3
|
|
N
|
%
|
N
|
%
|
Any
|
56
|
100
|
16
|
28.6
|
Fatigue
|
32
|
57.2
|
3
|
5.4
|
Fever
|
32
|
57.1
|
0
|
0
|
Rash
|
29
|
51.8
|
1
|
1.8
|
Vitiligo
|
26
|
46.4
|
0
|
0
|
Decreased appetite
|
22
|
39.3
|
3
|
5.4
|
Hypothyroidism
|
20
|
35.7
|
0
|
0
|
Oral ulcer
|
18
|
32.2
|
1
|
1.8
|
Alanine/aspartate aminotrans-ferase increased
|
18
|
32.1
|
0
|
0
|
High blood pressure
|
11
|
19.6
|
0
|
0
|
Vomiting
|
8
|
14.3
|
2
|
3.6
|
Pruritus
|
8
|
14.3
|
0
|
0
|
Myalgia
|
8
|
14.3
|
0
|
0
|
Weight loss
|
8
|
14.3
|
0
|
0
|
WBC count decrease
|
8
|
14.3
|
0
|
0
|
Hyperthyroidism
|
8
|
14.3
|
0
|
0
|
Arthralgia
|
7
|
12.5
|
2
|
3.6
|
Hand-foot syndrome
|
6
|
10.8
|
2
|
3.6
|
Hemorrhage
|
6
|
10.8
|
0
|
0
|
Diarrhea
|
4
|
7.1
|
0
|
0
|
Nausea
|
4
|
7.1
|
0
|
0
|
Lose hair or feathers
|
4
|
7.1
|
0
|
0
|
Dullness of the oral senses
|
3
|
5.4
|
1
|
1.8
|
Glucose increase
|
3
|
5.4
|
0
|
0
|
PLT decrease
|
3
|
5.4
|
0
|
0
|
Hypoadrenia
|
2
|
3.6
|
0
|
0
|
Triglycerides increased
|
2
|
3.6
|
0
|
0
|
Guillain-Barre syndrome
|
2
|
3.6
|
0
|
0
|
Other mucosal adverse reactions
|
2
|
1.8
|
0
|
0
|
Headache
|
2
|
3.6
|
0
|
0
|
Autoimmune pneumonia
|
2
|
3.6
|
0
|
0
|
Myasthenia gravis
|
1
|
1.8
|
1
|
1.8
|
Pituitary inflammation
|
1
|
1.8
|
0
|
0
|
Elevated bilirubin
|
1
|
1.8
|
0
|
0
|
Cough
|
1
|
1.8
|
0
|
0
|
Shortness of breath
|
1
|
1.8
|
0
|
0
|
Hiss
|
1
|
1.8
|
0
|
0
|
Blister
|
1
|
1.8
|
0
|
0
|
Electrocardiographic abnormality
|
1
|
1.8
|
0
|
0
|
Nephritis
|
1
|
1.8
|
0
|
0
|
Reduced hemoglobin
|
1
|
1.8
|
0
|
0
|
Decreased albumin
|
1
|
1.8
|
0
|
0
|
4. T-cell exhaustion markers
To evaluate the impact of T-cell exhaustion on the efficacy, we obtained 4 tumor samples from the individuals with treatment responses (1CR, 3PR) and 4 matched samples from those without treatment responses (1SD, 3PD). All samples were taken before the triple combination treatment was given. mIHC examination indicated a higher expression of TIM3 in patients without ORR. On the other hand, the expression of PD-1, LAG-3, and CD8 were similar between the two groups(Fig. 3 ). Statistical calculation was not performed due to the small sample size.