In this study, we found that traditional actuarial DSS showed that patients with OCC had a high survival rate. Low DSS was associated with distant disease. The annual risk curve for mortality from OCC across the study population showed a downward trend, peaking at 3 and 10 years after diagnosis. The patients with distant disease showed a steeper curve than those with regional and localized disease. Conditional DSS showed an increasing trend with survival time. In short, we demonstrated that the improved survival probability depended heavily on the time elapsed after diagnosis. Patients with distant disease showed the most pronounced improvement.

Identifying prognostic factors facilitates the assessment of patient prognosis and the formulation of follow-up plans. Younger age at diagnosis, smaller tumor size, tumors confined to the thyroid, lower tumor stage, and no distant metastases at diagnosis had been shown to be associated with improved survival rates [5, 7, 11,12,13,14]. In a study by Zhou et al. [5] based on the SEER database, it was shown that distant SEER stage was an independent risk factor for death in OCC. In our study, we also found that SEER stage was the most influential prognostic factor among all independent prognostic factors. Therefore, we focused on the impact of different SEER stages on patient prognosis.

Traditional actuarial DSS for patients with OCC can provide information on the cumulative time distribution of carcinoma deaths. In a study of 239 patients [3], the 1-, 5-, 10-, and 20-year DSS rates for the overall population reached 98.0%, 94.6%, 92.5%, and 87.4%, respectively. In other studies [7], 10-year DSS rates ranged from 49.0% to 93.1%. Similarly, in our study, the prognosis of OCC patients was good. However, the DSS of patients with distant disease decreased significantly.

In our previous study, the hazard curve for PTC [15] showed a bimodal distribution with a small peak at 10 years, a trough at 14 years, and a major peak at 20 years after diagnosis in the whole population. In this study, the mortality from OCC hazard curve for the entire study cohort showed a bimodal pattern, with two peaks at 3 and 10 years. The hazard curve for (anaplastic thyroid carcinoma, ATC) in the entire study showed a decreasing trend, with an inflection point at 2.5 years after diagnosis and no significant peak [16]. This discrepancy may be due to the different types of malignancies.

In other studies of nasopharyngeal [17], esophageal [18, 19] and bladder carcinoma [20], the conditional DSS rates increased as the time to diagnosis progressed. Similar results were also observed in our study. For carcinoma studies with relatively good prognosis, such as gastric carcinoma [21], upper tract urothelial carcinoma [22], and bladder carcinoma [20], clinicians and patients were more concerned with 5-year, or even 10-year survival rates. In our research, patients with OCC had a probability of 1-year conditional DSS of more than 99.1% at 1, 2, 3, 5, or even 10 years after diagnosis. Therefore, we estimated the patients' conditional DSS at 5 and 10 years, respectively. The 5- and 10-year conditional DSS increased in OCC patients, with the greatest improvement observed in patients with distant disease. More importantly, the 10-year traditional actuarial DSS for patients with distant disease was only 36.5% compared to 63.9% at 5-year conditional 5 years after diagnosis. This suggests that if patients remain alive 5 years postoperatively, the probability of survival for another 5 years would increase from 36.5 to 63.9%. This dynamic observation of the patient's prognosis will provide a better understanding of the patient's prognosis and formulate more precise treatments to improve the patient's prognosis. This correlation also exists in other malignant tumors. Wang et al. [17] analyzed the survival of 1993 patients with non-metastatic nasopharyngeal carcinoma after intensity-modulated radiation therapy and found that the 5-year conditional DSS improved significantly in surviving patients with advanced-stage disease. Ploussard et al. [20] investigated the changes in the 5-year conditional DSS rates of 8141 patients treated with radical cystectomy for bladder carcinoma and found that improvements of 5- and 10-year conditional DSS were mainly noted in patients with advanced disease. In a study based on 7531 patients with melanoma [23], the same variation in conditional DSS rates was found in patients with different stages. For stage I patients, the 5-year conditional DSS increased from 61.7 to 93.2% if the patient survived for 4 years without recurrence and the stage II–IV patients had the similar trend. This improvement was more pronounced in stage IV patients [23]. Panunzio et al [24] also using the SEER database, the survival probability of patients with adrenocortical carcinoma treated with surgery was found to increase with the duration of the disease-free interval. Patients with stage III disease showed the most significant improvement [24]. The 3-year survival rate for colorectal liver metastasis patients with perineural invasion (PNI) [25] on pathological was 46.4% and the patients without PNI were 68.6%. The 3-year survival rate of patients with PNI was comparable to that of patients without PNI among those who lived 3 years postoperatively. These results suggest that for advanced malignant tumors, the conditional DSS improved significantly with increasing time since diagnosis.

First, the SEER database was limited to the U.S. population and may not be applicable to other populations. Second, due to the limitations of using registry data such as potential issues with coding inaccuracies or missing data, we were not precisely able to capture variables. Third, there is a lack of the duration, dose, or frequency of postoperative adjuvant therapy. Therefore, we were unable to interpret the impact of chemotherapy and radiation therapy on survival. Fourth, recurrence-free survival rates could not be analyzed because of the lack of relevant data. Thus, recurrence-free survival in OCC patients deserves further study. Fifth, there were only 126 cases of OCC patients with distant disease, and information about the number and location of distant metastases was limited; therefore, we were not able to further analyze the impact of the number and location of distant metastases on conditional DSS.