Study Participant Characteristics
The final study cohort consisted of 89 patients with CSCC, whose clinicopathologic characteristics are shown in Table 1, with a median follow-up of 41.3 months (range: 7.7-61.9 months). The median age at diagnosis of CC was 61 years (range: 30-81 years). Staging according to the criteria of the FIGO 2018, the number of patients with stage IB3-II. and III CC were 6 (6.7%), 38 (42.7%) and 45 (50.6%), respectively. In this study, lymph node metastasis was determined by a short diameter of lymph nodes ≥1 cm on MRI, of which 28 patients (31.5%) had lymph node metastasis and 61 patients (68.5%) were lymph node-negative. The median values of pre SCC-Ag and post SCC-Ag were 14.2 ng/mL (range: 5.0-52.2 ng/mL) and 1.7 ng/mL (range: 1.2-2.3 ng/mL) respectively. The plasma samples of 89 patients were evaluated by LC-MS/MS. The median level of omega-3 PUFAs in the whole cohort was as follows: EPA of 58.0 nmol/mL (range :42.5-86.0 nmol/mL); DHA of 301.0 nmol/mL (range: 223.0-364.0 nmol/mL), and DPA of 69.0 nmol/mL (range: 55.0-89.5 nmol/mL). There is no clearly defined threshold for the relationship between fatty acids and cancer prognosis, and we categorized patients into high- and low-level groups using the median value of variables. MRI scans 3 months after the end of CCRT showed that 63 patients (70.8%) achieved CR and 26 patients (29.2%) did not achieve CR. The non-CR group included 24 (27.0%) patients with PR and 2 (2.2%) patients with PD or SD.
Table 1. The clinical characteristics of all patients (n = 89)
Factor
|
|
|
Total (N = 89)
|
|
%
|
|
Age
BMI
<25
≥25
Hypertension history
Yes
No
Diabetes history
Yes
No
Gravidity
0–2
≥3
Unknown
FIGO stage
IB3
II
III
Size
<4 cm
≥ 4 cm
Lymph nodes metastasis
Positive
Negative
Pre SCC-Ag
≤2.5
>2.5
Post SCC-Ag
≤2.5
>2.5
HPV Status
16+
Others
Negative
Unknown
HGB
Eicosapentaenoic acid
Docosahexaenoic acid
Docosapentaenoic acid
|
|
|
61(53-68)
51
38
25
64
11
78
48
28
13
6
38
45
23
66
28
61
14.15(5.03-52.17)
9
80
1.70(1.17-2.28)
74
15
35
6
5
43
122.50±18.30
58.00(42.50-86.00)
301.00(223.00-364.00)
69.00(55.00-89.50)
|
|
57.3
42.7
28.1
71.9
12.4
87.6
53.9
31.5
14.6
6.7
42.7
50.6
25.8
74.2
31.5
68.5
10.1
89.9
83.1
16.9
39.3
6.7
5.6
48.3
|
|
Abbreviations: BMI, body mass index; HGB, hemoglobin; FIGO, International Federation of Gynecology and Obstetrics; SCC-Ag, squamous cell carcinoma antigen.
Factors associated with CR after CCRT
CR after CCRT has been shown to be a reliable surrogate endpoint for survival in patients with LACC. Table 2 shows the relationship between the clinical response of CCRT and clinicopathologic characteristics, and univariate analysis showed that the response to CCRT was significantly correlated with the post SCC-Ag level (P=0.010). The proportion of post SCC-Ag levels that were low (≤2.5 ng/mL) (90.5%) was much higher in the CCRT CR group than the proportion of post SCC-Ag levels that were high (>2.5 ng/mL) (9.5%). In the CCRT non-CR group, the proportion of patients with low levels of post SCC-Ag (≤2.5 ng/m L) (65.4%) was also higher than that of patients with high levels of post SCC-Ag (>2.5 ng/m L) (34.6%), but the difference was not as great as in the effective group (Table 1). It is suggested that high levels of SCC-Ag after CCRT may be a risk factor for poor CCRT outcomes. Similarly, we also found that pretreatment plasma EPA levels were markedly higher in patients in the CCRT CR group than in the non-CR group, and that low levels of EPA were considerably linked with a reduced likelihood of CR (66.0 (48.0-90.0) nmol/mL vs. 48.5 (31.8-69.3) nmol/mL, P=0.018). DHA levels were slightly higher in the CR group than in the non-CR group, however, the results were not statistically different (P=0.246). Similarly, the level of DPA in the two groups was close to the same (P=0.601). There were no differences between the two groups in BMI, age, tumor stage, incidence of positive lymph nodes, tumor size, and HPV type. We included factors significantly associated with response to CCRT in the univariate analysis in a multifactorial binary logistic regression analysis to further identify factors influencing the prediction of CR after CCRT. Multifactorial analysis showed that post SCC-Ag levels >2.5 ng/mL (OR, 4.752; 95% CI, 1.431-15.786, P=0.011) were independently associated with a reduced incidence of CR after CCRT. In contrast, high levels of plasma EPA were independently associated with an increased incidence of CR after CCRT (OR, 0.980; 95% CI, 0.962-0.999, P=0.038) in Table 3.
Table 2. Univariate analysis of clinical variables with response to CCRT.
|
|
|
CR
(n=63)
|
Non-CR
(n=26)
|
P
|
Age
|
|
|
61(53-69)
30(47.6)
33(52.4)
34(54)
29(46)
19(30.2)
44(69.8)
5(7.9)
58(92.1)
35(55.6)
18(28.6)
10(15.9)
6(9.5)
28(44.4)
29(46.0)
16(25.4)
47(74.6)
17(27.0)
46(73.0)
12.95(6.93-40.76)
6(9.5)
57(90.5)
1.60(1.16-2.05)
57(90.5)
6(9.5)
26(41.3)
4(6.3)
4(6.3)
29(46)
123.80±17.40
66.00 (48.00-90.00)
303.00 (246.00-375.00)
69.00 (55.00-91.00)
|
60(52-66)
13(50.0)
13(50.0)
17(65.4)
9(34.6)
6(23.1)
20(76.9)
6(23.1)
20(76.9)
13(50)
10(38.5)
3(11.5)
0(0)
10(38.5)
16(61.5)
7(26.9)
19(73.1)
11(42.3)
15(57.7)
16.58(4.38-54.46)
3(11.5)
23(88.5)
2.00(1.24-2.80)
17(65.4)
9(34.6)
9(34.6)
2(7.7)
1(3.8)
14(53.8)
119.50±20.50
48.50 (31.75-69.25)
289.00 (192.80-354.50)
71.00 (53.30-89.25)
|
|
<60
≥60
BMI
<25
≥25
Hypertension history
Yes
No
Diabetes history
Yes
No
Gravidity
0–2
≥3
Unknown
FIGO stage
IB3
II
III
Size
<4 cm
≥ 4 cm
Lymph nodes metastasis
Positive
Negative
Pre SCC-Ag
≤2.5
>2.5
Post SCC-Ag
≤2.5
>2.5
HPV Status
16+
Others
Negative
Unknown
|
|
|
0.838
0.322
0.499
0.105
0.636
0.173
0.881
0.157
1.000
0.010
0.871
|
HGB
Eicosapentaenoic acid
Docosahexaenoic acid
Docosapentaenoic acid
|
|
|
0.319
0.018
0.246
0.601
|
Abbreviations: BMI, body mass index; HGB, hemoglobin; FIGO, International Federation of Gynecology and Obstet rics; SCC-Ag, squamous cell carcinoma antigen; CR, complete response.
Table 3. Multivariate analysis of clinical variables with response to CCRT.
|
|
β
|
|
OR (95% CI)
|
|
P
|
Post SCC-Ag
Eicosapentaenoic acid
|
1.559
-0.020
|
|
4.752 (1.431-15.786)
0.980 (0.962-0.999)
|
|
0.011
0.038
|
Abbreviations: SCC-Ag, squamous cell carcinoma antigen; OR, odds ratio; 95% CI, 95% confidence interval.
Survival analysis
Of the 89 patients, 13 (29.5%) of 44 patients with an EPA level <58.0 nmol/mL were diagnosed with locally recurrent or metastatic disease, and 4 (8.9%) of 45 patients with an EPA level ≥58.0 nmol/ were diagnosed with locally recurrent or metastatic disease (P=0.010). The PFS of the low EPA group at 1 and 3 years was 88.6% and 69.7%, respectively, and the PFS of the high EPA group at 1 and 3 years was 97.8% and 90.6%, respectively. The PFS is shown in Figure 1A. Patients with EPA levels ≥58.0 nmol/mL have significantly better PFS than patients with EPA levels <58.0 nmol/mL. In terms of OS, 8 (18.2%) of 44 patients with EPA levels <58.0 nmol/mL died, and 2 (4.4%) of 45 patients with EPA levels ≥58.0 nmol/mL died (P=0.042). OS was 95.5% and 86.0% at 1 and 3 years in the low EPA group and 100% and 95.3% at 1 and 3 years in the high EPA group, respectively. The OS data are shown in Figure 1B. Patients with EPA levels <58.0 nmol/mL had relatively poor OS. Similarly, the difference in 3-year PFS and OS between patients with post SCC-Ag levels ≤2.5 ng/mL and >2.5 ng/mL was statistically significant in Figure 1 C and D.
Figure 1. Kaplan–Meier curve for progression-free survival and overall survival regarding low vs high C20:5 levels (1A) (P=0.010); (1B) (P=0.042). Kaplan–Meier curve for progression-free survival and overall survival regarding low vs high Post SCC-Ag levels (1C) (P<0.001); (1D) (P=0.002). Abbreviations: Post SCC-Ag, posttreatment squamous cell carcinoma antigen; C20:5, eicosapentaenoic acid, EPA.
Tables 4 and 5 summarize the univariate and multivariate HR and 95% CI for PFS and OS (HPV status was not included in the survival analysis because of high missingness). Univariate analysis showed that pretreatment plasma EPA ≥58.0 nmol/mL level was considerably connected with improvement in PFS (HR, 0.255; 95% CI, 0.083-0.784, P=0.017), and other clinicopathologic variables that were dramatically linked with improvement in PFS included achievement of CR after CCRT (HR, 0.238; 95% CI, 0.090-0.625, P= 0.004). Lymph node metastasis, later FIGO staging, and post SCC-Ag level >2.5 ng/mL were all significantly linked with decreased PFS, (all P<0.05). In the Multivariate analysis, pretreatment plasma EPA level ≥58.0 nmol/mL remained an independent prognostic factor for improved PFS (HR, 0.249; 95% CI, 0.079-0.780, P=0.017). Other independent factors associated with reduced PFS included lymph node metastasis (HR, 4.850; 95% CI, 1.700-13.832, P=0.003), post SCC-Ag level >2.5 ng/mL (HR, 2.996; 95% CI, 1.098-8.171, P=0.003).
Univariate analysis showed that achieving CR after CCRT was significantly associated with improved OS (HR, 0.233; 95% CI, 0.065-0.827, P=0.024), whereas no significant correlation with OS was observed with pretreatment EPA level ≥58.0 nmol/mL (HR, 0.229; 95% CI, 0.049-1.080, P=0.063). Similarly, we also observed that lymph node metastasis (HR, 6.917; 95% CI, 1.777-26.924, P=0.005), post SCC-Ag level >2.5 ng/mL (HR, 5.532; 95% CI, 1.593-19.211, P=0.007) were significantly associated with decreased OS. Through multivariate analysis, only lymph node metastasis was an independent predictor of OS reduction (HR, 6.405; 95% CI, 1.584-25.908, P=0.009), and similarly multifactorial analysis showed that pretreatment EPA ≥58.0 nmol/mL level was not an independent predictor of OS (HR, 0.216; 95% CI, 0.045-1.030, P=0.054).
Table 4. Univariate and multivariate analyses of prognostic factors for PFS among patients with CC
|
|
Univariate
|
|
|
Multivariate
|
|
Characteristics
|
|
HR (95% CI)
|
|
P
|
HR (95% CI)
|
P
|
Age
(<60 vs. ≥60)
BMI
(≥25 vs. <25)
Hypertension history
(yes vs. no)
Diabetes history
(yes vs. no)
Lymph nodes metastasis
(yes vs. no)
FIGO stage
IB3
II
III
Pre SCC-Ag
(>2.5 vs.≤2.5)
Post SCC-Ag
(>2.5 vs.≤2.5)
Size
(<4 vs. ≥4)
CR achieved
(yes vs. no)
HGB
(≥110 vs.<110)
Eicosapentaenoic acid
(≥58 vs.<58)
Docosapentaenoic acid
(≥69vs.<69)
Docosahexaenoic acid
(≥301vs.<301)
|
0.829(0.319-2.145)
0.525(0.185-1.490)
0.520(0.149-1.810)
0.391(0.052-2.949)
5.378(1.982-14.594)
3.508(1.213-10.144)
1.957(0.259-14.760)
5.296(2.036-13.776)
0.814(0.287-2.312)
0.238(0.090-0.625)
0.631(0.222-1.794)
0.255(0.083-0.784)
0.463(0.171-1.254)
0.621(0.236-1.631)
|
|
0.697
0.226
0.520
0.362
0.001
0.021
0.515
0.001
0.700
0.004
0.388
0.017
0.130
0.333
|
4.850(1.700-13.832)
2.996(1.098-8.171)
0.249(0.079-0.780)
|
0.003
0.032
0.017
|
Abbreviations: HR, hazard ratio; 95% CI, 95% confidence interval; BMI, body mass index; HGB, hemoglobin; FIGO, International Federation of Gynecology and Obstetrics; SCC-Ag, squamous cell carcinoma antigen.
Table 5. Univariate and multivariate analyses of prognostic factors for OS among patients with CC
|
|
Univariate
|
|
|
Multivariate
|
|
Characteristics
|
|
HR (95% CI)
|
|
P
|
HR (95% CI)
|
P
|
Age
(<60 vs. ≥60)
BMI
(≥25 vs. <25)
Hypertension history
(yes vs. no)
Diabetes history
(yes vs. no)
Lymph nodes metastasis
(yes vs. no)
FIGO stage
Ib3
II
III
Post SCC-Ag
(>2.5 vs.≤2.5)
Size
(<4 vs. ≥4)
CR achieved
(yes vs. no)
HGB
(≥110 vs.<110)
Eicosapentaenoic acid
(≥58 vs.<58)
Docosapentaenoic acid
(≥69vs.<69)
Docosahexaenoic acid
(≥301vs.<301)
|
1.454(0.409-5.161)
0.336(0.071-1.581)
0.270(0.034-2.132)
0.801(0.101-6.354)
6.917(1.777-26.924)
3.967(0.899-17.499)
5.532(1.593-19.211)
0.537(0.151-1.905)
0.233(0.065-0.827)
0.625(0.161-2.422)
0.229(0.049-1.080)
0.521(0.146-1.854)
0.592(0.166-2.106)
|
0.563
0.167
0.214
0.834
0.005
0.069
0.007
0.335
0.024
0.497
0.063
0.314
0.418
|
6.405(1.584-25.908)
3.414(0.951-12.258)
0.216(0.045-1.030)
|
0.009
0.060
0.054
|
Abbreviations: HR, hazard ratio; 95% CI, 95% confidence interval; BMI, body mass index; HGB, hemoglobin; FIGO, International Federation of Gynecology and Obstetrics; SCC-Ag, squamous cell carcinoma antigen.
Relationship Between the EPA Level and Other Clinical Characteristics
The clinicopathological features of the two groups were compared, as shown in Table 6. However, there was no difference between the high and low EPA groups in age, BMI, diabetes, hypertension, lymph node metastasis, FIGO stage, pre SCC-Ag, post SCC-Ag and tumor size (all P>0.05).
Table 6. Relationships between clinicopathological data and the different levels of pretreatment C20:5 in CC.
|
|
C20:5<58
|
|
|
C20:5≥58
|
|
Characteristics
|
|
N=44
|
|
|
N=45
|
P
|
Age
<60
≥60
BMI
<25
≥25
Hypertension history
yes
no
Diabetes history
yes
no
Lymph nodes metastasis
yes
no
FIGO stage
IB3
II
III
Pre SCC-Ag
≤2.5
>2.5
Post SCC-Ag
≤2.5
>2.5
Size
<4
≥4
HGB
<110
≥110
|
|
24(54.5)
20(45.5)
25(56.8)
19(43.2)
13(29.5)
31(70.5)
6(13.6)
38(86.4)
14(31.8)
30(68.2)
2(4.5)
17(38.6)
25(56.8)
3(6.8)
41(93.2)
35(79.5)
9(20.5)
13(29.5)
31(70.5)
13(29.5)
31(70.5)
|
|
|
19(24.2)
26(57.8)
26(57.8)
19(42.2)
12(26.7)
33(73.3)
5(11.1)
40(88.9)
14(31.1)
31(68.9)
4(8.9)
21(46.7)
20(44.4)
6(13.3)
39(86.7)
39(86.7)
6(13.3)
13(28.9)
32(71.1)
7(15.6)
38(84.4)
|
0.245
0.927
0.763
0.717
0.943
0.439
0.504
0.370
0.946
0.114
|
Abbreviations: BMI, body mass index; HGB, hemoglobin; FIGO, International Federation of Gynecology and Obstetrics; SCC-Ag, squamous cell carcinoma antigen; C20:5, eicosapentaenoic acid, EPA.
Correlation of SCC-Ag with plasma EPA
To confirm the relationship between SCC-Ag and plasma EPA, we performed Spearman rank correlation analysis. The results showed that pre SCC-Ag and EPA were negatively correlated (r=-0.305, P=0.004), post SCC-Ag and EPA were weakly negatively correlated (r=-0.251, P=0.018) in Figure 2A and 2B.
Figure 2. Correlation of plasma C20:5 and Post- SCC-Ag in 89 patients with CC. The plasma C20:5 was negatively correlated with Pre SCC-Ag (r = −0.305, p=0.004) (2A). The plasma C20:5 was negatively correlated with Post SCC-Ag (r = −0.251, p=0.018) (2B). Abbreviations: SCC-Ag, squamous cell carcinoma antigen; C20:5, eicosapentaenoic acid, EPA; Pre SCC-Ag, pretreatment squamous cell carcinoma antigen; Post SCC-Ag, posttreatment squamous cell carcinoma antigen.
Prognostic Significance of Combining plasma EPA and post-SCC Ag
The prognostic significance of combining EPA and post SCC-Ag for the entire cohort is shown in Figure 3. There was a significant difference in 3-year PFS rates among the four groups (P<0.001). The pretreatment plasma EPA ≥58.0 nmol/mL and post SCC-Ag ≤2.5 ng/mL groups had significantly higher survival probabilities than the other three groups. In contrast, patients with pretreatment plasma EPA <58.0 nmol/mL and post SCC-Ag>2.5 ng/mL had the lowest survival rate among the four groups in Figure 3.
Figure 3. Abbreviations: SCC-Ag, squamous cell carcinoma antigen; C20:5, eicosapentaenoic acid, EPA.