Baseline antibody profiles predict toxicity in melanoma patients treated with immune checkpoint inhibitors

Table 1 Baseline patient characteristics

		Anti-CTLA-4	Anti-PD-1	Combination	Fisher’s test
		(n = 37)	(n = 27)	(n = 11)	p value
		No. (%)	No. (%)	No. (%)	p value
Gender	Female	12 (32)	11 (41)	4 (36)	0.848
	Male	25 (68)	16 (59)	7 (64)
Age at treatment initiation	Mean (SD)	66.2 (13)	69.9 (14)	59.9 (13)	1
	Median	67.4	71	61.1
ECOG PS (pretreatment)	0	28 (76)	19 (70)	7 (64)	0.736
	> 1	9 (24)	8 (30)	4 (36)
LDH (pretreatment)	Normal	31 (91)	19 (70)	9 (82)	0.114
	Elevated	0 (9)	8 (30)	2 (18)
	Unknown	3	0	0
Response to treatment	POD	22 (59)	10 (37)	4 (36)	0.01
	SD	10 (27)	5 (19)	1 (9)
	PR	5 (14)	8 (30)	2 (18)
	CR	0	2 (7)	4 (36)
	UNC	0	2 (7)	0
Toxicity	None	8 (22)	4 (15)	0
	Mild	20 (54)	15 (55)	4 (36)	0.16
	Severe	9 (24)	8 (30)	7 (64)
GI toxicity	Mild	9 (23.1)	12 (42.9)	3 (27.3)	0.08
	Severe	6 (15.4)	3 (10.7)	6 (54.5)
Skin toxicity	Mild	15 (38.5)	17 (60.7)	5 (45.5)	0.43
	Severe	0	1 (3.6)	1 (9.1)
Endocrine toxicity	Mild	5 (12.8)	11 (39.2)	4 (36.4)	0.71
	Severe	0	1 (3.6)	1 (9.1)
Required treatment termination	Yes	4 (11)	3 (11)	6 (54)	0.006
	No	33 (89)	24 (89)	5 (46)

Summary of clinical features from 75 melanoma patients treated with anti-CTLA-4 (n = 37), anti-PD-1 (n = 27), or anti-CTLA-4 and anti-PD-1 (n = 11). LDH lactate dehydrogenase, POD progression of disease, SD stable disease, PR partial response, CR complete response, UNC unclassified. Fisher’s exact test was used to examine the significance of the association between patient characteristics and treatment type. Two anti-CTLA-4 patients were sampled twice (11-311, in 2011 and 2013; 12-071, in 2012 and 2013), and one anti-PD-1 patient was sampled twice (13-185, in 2015 and 2016)