3.1 Gender differences in terms of overall survival in elderly patients with laryngeal cancer were shown in stage I
There were 20,167 cases of laryngeal squamous cell carcinoma in the SEER database from 2010–2018, excluding 358 cases with unknown cause of death, unclear information on surgical modality and radiotherapy, and 9650 cases aged < 65 years, and finally, 10,159 cases met the inclusion criteria (Fig. 1). K-M analysis showed that there was a significant gender difference in survival prognosis in the laryngeal cancer patients aged ≥ 65 years, with males having significantly better OS than females (P < 0.001) (Fig. 2a). To further explore the gender difference in laryngeal cancer patients, K-M analysis was performed on patients diagnosed with stage I, stage II, stage III, and stage IV, respectively. It was found that the gender difference of elderly laryngeal cancer was mainly manifested in patients diagnosed with stage I (P = 0.024), while the gender differences in patients diagnosed with stage II, stage III, and stage IV were not significant (P = 0.564, P = 0.938, P = 0.240, respectively) (Fig. 2b-2e).
After screening out patients with stage II (N = 1588), stage III (N = 1616), stage IV (N = 2378), and stage unknown (N = 579), there were 3998 patients with stage I remaining. In the guidelines of the National Comprehensive Cancer Network (NCCN), the main treatment methods for early laryngeal cancer are laryngeal preservation surgery and radiation therapy (RT) [18]. To better investigate the epidemiological characteristics of stage I patients and the gender differences between treatment modalities, we further screened out patients treated with chemotherapy or (and) total/radical laryngectomy (N = 168), leaving 3830 patients remaining.
3.2 Prognosis analysis of elderly patients with stage I laryngeal cancer
Variables such as patients' general clinical information and treatment modality were analyzed univariately by the K-M method and Log-rank test, and variables with P < 0.05 were screened into Cox proportional hazards regression model to analyze independent risk factors affecting OS in patients. The variables included in the multifactorial analysis were: gender (P = 0.045), age at diagnosis (P < 0.001), race (P = 0.001), marital status (P < 0.001), tumor size (P = 0.005), tumor site (P < 0.001), pathological grade (P = 0.004) and treatment modality (P < 0.001).
Multifactorial Cox analysis showed that gender (P = 0.008), age at diagnosis (P < 0.001), marital status (P < 0.001), tumor site (P < 0.001), and treatment modality (P < 0.001) were prognostic factors affecting patients' OS. In further stratified analysis: Female and marital status were protective factor, female (vs male, hazard ratio (HR) = 0.797, 95% CI = 0.675–0.941, P = 0.008), married (vs single, HR = 0.702, 95% CI = 0.619–0.795, P < 0.001); Advanced age was a risk factor, oldest old (vs young old, HR = 3.257, 95%CI = 2.768–3.833, P < 0.001), older old (vs young old, HR = 1.699, 95%CI = 1.493–1.934, P < 0.001); Supraglottis and subglottis were risk factors, supraglottis (vs glottis, HR = 2.828, 95% CI = 1.594–5.017, P < 0.001), subglottis (vs glottis, HR = 2.741, 95% CI = 2.290–3.282, P < 0.001); All treatment modalities significantly improved prognosis (vs. none, HR < 0.05, P < 0.001). (Table 1)
Table 1
Univariate analysis and multivariate Cox regression analysis (Forward:LR) of prognostic factors influencing overall survival of elderly patients with stage Ⅰ laryngeal cancer.
Variable | Mean survival (95%CI a) months | Univariate analysis | Multivariate analysis |
Log-rank(χ2) | P-value | HRb (95% CI) | P-value |
Gender | 4.01 | 0.045 | | 0.008 |
Male | 74.89(73.33–76.45) | | | Reference | |
Female | 70.95(67.13–74.78) | | | 0.797(0.675–0.941) | 0.008 |
Age at diagnosis, years | 202.56 | < 0.001 | | < 0.001 |
65–74 | 81.57(79.70-83.43) | | | Reference | |
75–84 | 70.53(68.08–72.98) | | | 1.699(1.493–1.934) | < 0.001 |
≥ 85 | 51.31(47.21–55.41) | | | 3.257(2.768–3.833) | < 0.001 |
Race | | 13.78 | 0.001 | | 0.001 |
White | 74.06(72.48–75.63) | | | Reference | |
Black | 71.61(67.30-75.91) | | | 1.132(0.950–1.348) | 0.166 |
Otherc | 86.75(80.59–92.91) | | | 0.698(0.538–0.907) | 0.007 |
Marital status | 64.27 | < 0.001 | | < 0.001 |
Singled | 66.42(63.84-69.00) | | | Reference | |
Married | 78.38(76.57–80.19) | | | 0.702(0.619–0.795) | < 0.001 |
Unknown | 77.27(71.80-82.74) | | | 0.698(0.538–0.907) | 0.007 |
Tumor Size, cm | 12.82 | 0.005 | | 0.017 |
< 3 | 75.10(72.15–78.06) | | | Reference | |
3–5 | 55.95(38.01–73.89) | | | 2.077(1.231–3.505) | 0.006 |
> 5 | 87.23(72.82-101.64) | | | 0.487(0.121–1.960) | 0.311 |
Unknown | 74.18(72.51–75.85) | | | 1.140(0.991–1.312) | 0.066 |
Tumor Site | | 148.81 | < 0.001 | | < 0.001 |
Glottis | 77.04(75.54–78.55) | | | Reference | |
Subglottis | 46.12(26.19–66.05) | | | 2.828(1.594–5.017) | < 0.001 |
Supraglottis | 52.89(47.73–58.05) | | | 2.741(2.290–3.282) | < 0.001 |
Other | 57.44(48.15–66.72) | | | 1.797(1.340–2.411) | < 0.001 |
Grade | 13.08 | 0.004 | | 0.124 |
Glottis | 75.19(71.99–78.40) | | | Reference | |
Subglottis | 73.99(71.81–76.17) | | | NA | NA |
Supraglottis | 67.19(62.28–72.10) | | | NA | NA |
Unknown | 76.71(72.85–75.74) | | | NA | NA |
Therapy | | 111.06 | < 0.001 | | < 0.001 |
None | 50.74(44.50-56.98) | | | Reference | |
RTe | 73.67(71.60-75.75) | | | 0.455(0.375–0.553) | < 0.001 |
OPSf | 79.01(75.91–82.11) | | | 0.379(0.303–0.474) | < 0.001 |
OPS + RT | 77.43(74.64–80.23) | | | 0.400(0.322–0.495) | < 0.001 |
a 95%CI: 95% Confidence Interval; b HR: Hazard Ratio; c Other: American Indian/AK Native, Asian/Pacific Islander and other races; dSingle: Never married, Domestic Partner, Widowed, Separated or Divorced; eRT: Radiation Therapy; fOPS: Organ Preservation Surgery |
3.3 Gender differences among elderly patients with stage I laryngeal cancer
The stage I population included 3266 male patients (85.3%) and 564 female patients (14.7%). In the K-M analysis, the average survival time (AST) of OS was 74 months for male patients compared with 70 months for females. Similarly, the AST was 94 months for male patients compared with 90 months for female patients in cancer specific survival (CSS) comparison. The differences were statistically significant (P < 0.05). We analyzed the composition of male and female patients using chi-square tests and found statistically significant differences in the variables of age at diagnosis, marital status, tumor site and lymph node dissection (LND). Compared to male patients, females were diagnosed at an older age (age of diagnosis ≥ 85 years: 15.1% vs. 11.2%), with more singles (56.7% vs. 29.8%), more supraglottic types (21.8% vs. 6.4%) and more likely to undergo LND (2.0% vs. 0.6%).
To reduce data bias and confounding factors, we performed propensity score matching (PSM) on general epidemiological and clinicopathological data except for "treatment modality". After matching with a propensity score of 1:2 and a caliper value of 0.02, the distribution of propensity scores (Fig. 3a) and univariate SD scatter plots (Fig. 3b) indicated a good matching effect. After matching, there was no longer a significant difference in the composition ratio various variables for male and female patients. Meanwhile, the survival difference between male and female patients disappeared. (Table 2)
Table 2
The portion and distribution of clinical characteristics in elderly patients with stage Ⅰ laryngeal cancer before and after propensity score matching.
Variable | Before PSMa(n = 3830) | After PSM(n = 1542) |
Male (n = 3266, 85.3%) | Female (n = 564, 14.7%) | P-value | Male (n = 1011, 65.6%) | Female (n = 531, 34.4%) | P-value |
Age of diagnosis, years | 74.73 ± 7.11 | 75.43 ± 7.60 | 0.032 | 75.07 ± 7.51 | 75.16 ± 7.48 | 0.833 |
Age of diagnosis, years | | 0.025 | | | 0.714 |
65–74 | 1774(54.3%) | 301(53.4%) | | 549(54.3%) | 294(55.4%) | |
75–84 | 1126(34.5%) | 178(31.6%) | 330(3.6%) | 163(30.7%) |
≥ 85 | 366(11.2%) | 85(15.1%) | 132(13.1%) | 74(13.9%) |
Race | | | 0.563 | | | 0.240 |
White | 2749(84.2%) | 468(83.0%) | | 850(84.1%) | 441(83.1%) | |
Black | 384(11.8%) | 75(13.3%) | 114(11.3%) | 72(13.6%) | |
Otherb | 133(4.1%) | 21(3.7%) | 47(4.6%) | 18(3.4%) | |
Marital status | | | < 0.001 | | | 0.075 |
Singlec | 974(29.8%) | 320(56.7%) | | 501(49.6%) | 293(55.2%) | |
Married | 2092(64.1%) | 208(36.9%) | | 447(44.2%) | 203(38.2%) | |
Unknown | 200(6.1%) | 36(6.4%) | | 63(6.2%) | 35(6.6%) | |
Site | < 0.001 | | | 0.172 |
Glottis | 2966(90.8%) | 416(73.8%) | | 825(81.6%) | 416(78.3%) | |
Subgiottis | 18(0.6%) | 5(0.9%) | | 4(0.4%) | 0(0.0%) | |
Supraglottis | 208(6.4%) | 123(21.8%) | | 157(15.5%) | 101(19.0%) | |
Other | 74(2.3%) | 20(3.5%) | | 25(2.5%) | 14(2.6%) | |
Grade | | | 0.176 | | | 0.634 |
Grade 1 | 659(20.2%) | 104(18.4%) | | 190(19.7%) | 100(18.8%) | |
Grade 2 | 1422(43.5%) | 266(47.2%) | | 467(46.2%) | 249(46.9%) | |
Grade 3 | 281(8.6%) | 56(9.9%) | | 85(8.4%) | 54(10.2%) | |
Unknown | 904(27.7%) | 138(24.5%) | | 260(25.7%) | 128(24.1%) | |
Tumor Size, cm | | | 0.964 | | | 0.964 |
< 3 cm | 800(24.5%) | 138(24.5%) | | 221(21.9%) | 124(23.4%) | |
3–5 cm | 33(1.0%) | 6(1.1%) | | 9(0.9%) | 5(0.9%) | |
> 5 cm | 10(0.3%) | 2(0.4%) | | 4(0.4%) | 2(0.4%) | |
Unknown | 2423(74.2%) | 418(74.1%) | | 777(76.9%) | 400(75.3%) | |
Therapy | | | 0.451 | | | 0.761 |
None | 200(6.1%) | 42(7.0%) | | 70(6.9%) | 33(6.2%) | |
RTd | 1595(48.8%) | 279(46.8%) | | 490(48.5%) | 248(46.7%) | |
OPSe | 708(21.7%) | 226(37.9%) | | 221(21.9%) | 127(23.9%) | |
OPS + RT | 763(23.4%) | 37(6.2%) | | 230(22.7%) | 123(23.2%) | |
OPS | | | 0.647 | | | 0.191 |
None | 1795(55.0%) | 304(53.9%) | | 560(55.4%) | 281(52.9%) | |
Yes | 1471(45.0%) | 260(46.1%) | | 451(44.6%) | 250(47.1%) | |
LNDf | | | 0.003 | | | 0.114 |
None | 3242(99.3%) | 551(97.7%) | | 1001(99.0%) | 519(97.7%) | |
Yes | 18(0.6%) | 11(2.0%) | | 7(0.7%) | 10(1.9%) | |
Unknown | 6(0.2%) | 2(0.4%) | | 3(0.3%) | 2(0.4%) | |
RT | | | 0.076 | | | 0.310 |
None/unknown | 908(27.8%) | 174(30.9%) | | 291(28.8%) | 160(30.1%) | |
Yes | 2358(72.2%) | 390(69.1%) | | 720(71.2%) | 371(69.9%) | |
OSg, ASTh, Months (95%CIi) | 74.89 (73.33, 76.45) | 70.95 (67.13, 74.78) | 0.045 | 71.55 (68.70, 74.40) | 72.25 (68.35, 76.15) | 0.726 |
CSSj, AST, months (95%CI) | 94.48 (93.24, 95.73) | 90.38 (87.05, 93.70) | 0.005 | 91.86 (89.44, 94.27) | 91.43 (88.08, 94.78) | 0.780 |
aPSM: Propensity Score Matching; bOther: American Indian/Alaska Native, Asian/Pacific Islander and other races; cSingle: Never married, Domestic Partner, Widowed, Separated or Divorced; dRT: Radiation Therapy; eOPS: Organ Preservation Surgery; fLND: Lymph Node Dissection; gOS: Overall Survival; hAST: Average Survival Time; i95%CI: 95% Confidence Interval; jCSS: Cancer Specific Survival |
3.4 Gender differences in survival of elderly patients with stage I laryngeal cancer under different treatment modalities
In the univariate analysis of the survival prognosis of male and female patients with different treatment modalities, we found that in RT group, OPS group and OPS + RT group, male patients had better OS and CSS than females, among which the gender difference of OS in the RT group was the most obvious, but the difference was not statistically significant (Fig. 4). The prognosis of the OPS group was comparable to that of the OPS + RT group (OS: 75.17 months vs. 75.32 months; CSS: 91.87 months vs. 90.51 months), but both better than that of the RT group (OS: 69.30 months; CSS: 90.38 months) in elderly women. (Table 3)
Table 3
Survival comparison between different treatments in elderly patients with stage I laryngeal cancer
Variable | Overall Survival, ASTa, months (95%CI)b | P-value | Cancer Specific Survival, AST, months (95%CI) | P-value |
Male | Female | Male | Female |
Therapy | | | | | | |
None | 51.05(44.19–57.91) | 48.90(33.90–63.90) | 0.733 | 75.64(68.36–82.93) | 67.84(51.90-83.78) | 0.204 |
RTc | 74.45(72.21–76.69) | 69.30(63.76–74.85) | 0.063 | 94.22(92.41–96.03) | 90.38(85.51–95.25) | 0.077 |
OPSd | 79.56(76.18–82.94) | 75.17(68.37–81.97) | 0.701 | 96.46(93.91–99.01) | 91.87(86.75–96.99) | 0.940 |
OPS + RT | 77.85(74.87–80.82) | 75.32(67.40-83.24) | 0.439 | 97.32(95.10-99.54) | 90.51(83.59–97.44) | 0.022 |
aAST: Average Survival Time; b95%CI: 95% Confidence Interval; cRT: Radiation Therapy; cRT: Radiation Therapy; dOPS: Organ Preservation Surgery |