Baseline characteristics of the SEER cohort
In total, 41,270 female and 13,188 male patients (female to male ratio was 3.13) obtained from the SEER database met the inclusion criteria (Table 1). Female patients are typically diagnosed at an earlier age than males on average (47.98 vs. 52.45 years old). On the ethnic level, there are more female cases in the white and black population. Compared with female patients, male patients developed a larger tumor size (17.44 vs. 23.51cm) and had a larger proportion of lymph node metastasis (33.2 vs. 21.0%) and distant metastasis (2.3 vs. 0.9%), consistent with the distribution of combined summary stage (Regional 36.1 vs. 26.1%, Distant 4.0 vs. 1.9%). A rare pathological type, medullary carcinoma, occurred more frequently in male patients than in female patients (2.9 vs. 1.2%). And survival month of male patients was significantly lower than that of female patients (72.39 vs. 76.8 months). Male seems to have a worse prognosis, considering that thyroid cancer caused more specific death (2.9 vs. 1.1%) in them. We found sex differences in whether surgery performed or not (surgery performed Male: Female = 98.7%: 99.4%) which might be attributed to tumor progression and malignancy. Median income was also different in female and male (P<0.001). But the difference between laterality of male and female was non-significant (P=0.376). These findings revealed a large sex disparity in clinical characteristics of thyroid cancer.
Differences in the survival of female and male patients in the SEER cohort
During the time period 1975-2018, the age-adjusted incidence rate of thyroid cancer is 9.2 (4.9 in male, 13.3 in female) per 100,000. From 2010 to 2016, the incidence has been at a stable level. Female were still affected 3 times more than male, although the curves shifted downward during 2017-2018 (Fig. 2a). Compared with female patients in the same period, male diagnosed with thyroid cancer each year seems to have a higher potential of specific death (Fig. 2b). Considering that the proportion of patients died of thyroid cancer among mortality generally rose up (Fig. 2c). Thyroid cancer still threatens human health, especially female due to the higher incidence. As showed in Kaplan-Meier analysis of the SEER cohort, OS declined more sharply for male (Log-rank test, p<0.001) (Fig. 3a). Similar results were observed for CSS (Log-rank test, p<0.001) (Fig. 3b). Namely, female had better OS and CSS prognosis compared with the male. In the univariate and multivariate analyses of OS and CSS in the SEER cohort, there was significant difference between male and female with different tumor size, age or median income subgroups, and the male versus female HRs were > 1.000 (Table S1and S2, Additional File 1).
Additionally, we observed sex differences in nodal ratio. The number of metastatic lymph node number and metastatic lymph node ratio could be prognostic metrics for thyroid cancer[23-25], and can improve predictive ability in a novel N staging system compared with the AJCC system[26]. In this process, 21,825 female and 7,256 males were collected after excluding patients with unknown/not stated regional nodes. The mean number of positive nodes was 2.0 in female, whereas the mean number of positive nodes was 4.0 in male. Nodal Ratio (NR) is defined as the number of positive lymph nodes out of the number of total lymph nodes in the specimen. Mean NR of female and male was 0.192 and 0.297, respectively (Table S3, Additional File 2). Leboulleux et al. demonstrated that the number of positive LNs is related to a significantly higher 10-year risk of recurrence: 3% for <5 LN metastases [27]. Nam et al. demonstrated that NR >0.3 was associated with higher rates of any site and nodal recurrence[28]. Taking into account the results of previous studies, our results on the positive number and positive ratio of cervical lymph nodes in patients with thyroid cancer support the conclusion that male patients have a worse prognosis to a certain extent.Together, these results demonstrated male thyroid cancer associated with a worse prognosis.
Differences in the survival of female and male patients in cohorts after propensity score matching
To eliminate selection bias in observational non-randomized controlled trials, male and female patients diagnosed with thyroid cancer between 2004 and 2018 were matched via PSM. We conducted several matchings including different factors to confirm our conclusion. OS and CSS functions after propensity score matching indicated the sex disparity of prognosis on different levels (Fig. 4).
After 1:1 matching including all the factors, 12779 male and 12779 female patients were collected to form a new post-PSM cohort (Table S4, Additional File 1). Survival months were longer (74.14±49.14 vs 72.68±49.23, P=0.018) in the female. There is more overall and specific death in male (overall death: 10.6 vs. 6.9%, specific death: 2.5 vs. 1.9%, P<0.001, respectively).
Kaplan-Meier curves of OS and CSS for male and female patients in the post-PSM cohort supported the same conclusion as in the SEER cohort (Fig. 5a, b). Compared with female patients, the hazard ratio (HR) for all-cause death in male patients was 1.62 (95%CI=1.514-1.734, P<0.001), and the HR for cancer-specific death was 1.437 (95%CI=1.248-1.654, P<0.001) (Table 2, 3). After adjustment for clinical basic data, these HRs for all-cause and cancer-specific death became 1.632 (95%CI=1.499-1.777, P<0.001) and 1.473 (95% CI=1.245-1.741, P<0.001) (Table 2, 3), respectively.
Factors affecting the prognosis of female and male thyroid cancer
The statistically significant factors affecting OS of female thyroid cancer (Table 4) were race (white versus black: HR= 0.736, 95% CI= 0.636-0.852, P<0.001, other versus black: HR= 0.682, 95%CI= 0.565-0.822, P <0.001), age of diagnosis (<55 versus ≥55 years old: HR= 6.195, 95%CI= 5.612-6.838, P<0.001), median income (≥$55000 versus <$55,000: HR= 0.841, 95% CI= 0.753-0.940, P<0.001), surgery (versus not performed: HR= 0.134, 95% CI= 0.108-0.167, P<0.001), distant in combined summary stage (versus localized: HR= 2.133, 95% CI= 1.671-2.723, P<0.001) T stage, M stage and histology recode (versus medullary: HR= 0.892, 95% CI= 0.853-0.933, P<0.001). The risk of death increases with upgrading of T, M stage. Similar results were found in analysis of OS for male patients (Table 5). As for CSS, the factors for female (Table 6) were tumor size (>1cm versus ≤1cm: HR= 1.578, 95% CI= 1.199-2.077, P<0.001), age of diagnosis (<55 versus ≥55 years old: HR= 7.054, 95%CI= 5.562-8.946, P<0.001), surgery (versus not performed: HR= 0.153, 95% CI= 0.109-0.217, P<0.001), histology recode (versus medullary, HR= 0.750, 95% CI= 0.696-0.809) and the risk of death increased with upgrading of T, N, M and combined summary stage. In male patients, median income (≥$55000 versus <$55,000: HR= 0.739, 95% CI= 0.574-0.953, P= 0.020) had an extra impact compared with female (Table 7).These findings implicated race, age of diagnosis, median income, surgery, tumor size, and histology recode, T, N, M and combined summary stage contributed to the prognosis of thyroid cancer.