Patients with differentiated thyroid cancer generally have good prognosis and undergo life-time follow-up or suppressive treatment. However, recurrence rates are around 20% and more than one third occur after five years of initial treatment . Recurrence rates and clinical-pathological variables associated with thyroid cancer recurrence in low- and middle-income countries are still under-reported . Identification of factors attributed to disease recurrence is very important to provide cost-effective management and surveillance schemes [14, 16]. Although remarkable improvement has been made during the past two decades for treatment and detection of postoperative recurrences, the clinical recommendations have not been implemented in every hospital. Especially in low-resource settings, there are typically only a few centers that have facilities for thyroid scan, radionucleotide ablation iodine (RAI) testing, and systematic monitoring and follow-up [10, 17].
In our patient cohort, recurrence rates were relatively high reaching 29%, 43%, and 68% for patients initially diagnosed in Stage I, Stage II, and Stage III; respectively (Table 2). Gan et al.  reported a recurrence rate of 3% among patients with differentiated thyroid cancers after a median follow-up of 27 months. A recent study found significantly different recurrence rates in patients diagnosed in early and advance stages (7.2% vs 28.2%) . Liu et al. reported a higher loco-regional recurrence in patients with positive lymph nodes (31.5%) in comparison to those with node negative (5.2%) . In comparison to these reports, our study showed relatively higher recurrence rates including in patients initially diagnosed in Stages I and II.
We identified several variables associated with increased risk of recurrences including older age, being male patients, primary thyroid cancer with ETE, and positive neck lymph nodes. In the multivariable analysis, however, only older age and positive lymph nodes were significantly associated with disease recurrence (Table 3). Hollenbeak et al. reported recurrence rates up to 39% in older patients with differentiated thyroid cancer . Thyroid cancers are generally diagnosed in patients younger than 55 years and around 20% are initially detected in individuals above 65 year-old [2, 21]. In our cohort, only 10% patients were diagnosed above 65 years (Table 1); yet, they were associated with higher risks of recurrence (Table 3). Females are more affected by goiter and differentiated thyroid cancer, although the etiology is still not fully addressed. However, males affected by differentiated thyroid cancer are associated with poor prognosis [22, 23]. Other studies, on the other hand, have shown that sex is not an independent risk factor for poor prognosis [24, 25]. Male individuals with thyroid cancer more frequently present in advanced stages, with larger tumor size, positive lymph nodes, and ETE [22, 26], which might also explain the association with higher risks of recurrence. Sequential analysis shows that adverse outcomes attributable to the male sex gradually decrease over time with the improvement of diagnosis and treatment .
Our study showed the significant association of positive lymph nodes with higher recurrence rates. Involvement of neck lymph nodes has been associated with higher rates of thyroid cancer recurrence and mortality [20, 28], although the strength of the association varies widely among studies. With the improvements to detect involvement of lymph nodes, the indication for neck dissection in differentiated thyroid cancer is still debatable, particularly for micrometastasis to the lymph nodes. Rather than simply classify patients into binary variables (positive and negative), Schneider et al. found that the lymph node ratio was also associated with papillary thyroid cancer recurrence . Collinearity should also be considered as patients with positive lymph nodes tend to have other clinicopathological adverse features including larger tumor size, ETE and LVI. Our data also showed that ETE, LVI and multifocal tumors were significantly associated with higher recurrence rates. ETE is considered as tumor extension outside the thyroid capsule with infiltration into surrounding tissues including strep muscles, trachea, larynx, laryngeal nerve, jugular vein and carotid artery. In our cohort, 13% patients had ETE (Table 1), which was parallel with previous studies (9-12%) . According to ATA guidelines, the presence of ETE is considered with intermediate risk factors and extensive ETE is associated with high risk and both are associated with an increased risk of recurrence . LVI is also associated with worse prognosis in differentiated thyroid cancer . Our study found patients with LVI had a higher risk of recurrence, although it was not an independent factor. Wagner et al. have also revealed that LVI is associated with worse RFS especially in the presence of positive lymph node infiltration .
We also found that T4, positive neck lymph nodes, older age, and LVI were associated with shorter RFS (Fig. 1). Palme et al.  also reported that ETE (T4) and advanced stages at initial diagnosis were the most significant independent risk factors of recurrence in patients with well-differentiated thyroid cancers. Using the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database, Banerjee et al.  revealed that tumor size, stage at diagnosis, and receiving radioactive iodine were associated with higher risks of recurrence in patients with differentiated thyroid cancer. In addition to higher frequency of ETE, our cohort also showed much larger tumor size (mean 6.4 cm) in comparison to other studies (mean tumor sizes were 2.1 cm  and 3.3 cm ). Parallel with our findings, Ito et al.  showed that older age was a prominent prognostic factor for overall survival in comparison to positive node and metastatic lesions in patients with papillary thyroid cancer. Another report demonstrated that primary tumor size at diagnosis and cervical nodal status were associated with shorter disease-free survival in patients with differentiated thyroid cancer .
Our study highlights some risk factors of disease recurrence among patients who are diagnosed with relatively larger tumor size (mean 6.4 cm) and advanced stages of differentiated thyroid cancer. The relatively higher rates of recurrence rates particularly in those who are initially diagnosed in Stages II and III within a median follow-up of 57 months might indicate that improvements of diagnosis, clinical management, and surveillance are warranted. In general, differentiated thyroid cancer has very good prognosis and recurrence usually occurs after a period of follow-up of more than 5 years [12, 35]. In our cohort, more than 80% of the patients were diagnosed with tumor size larger than 4 cm (Table 1) indicating that public awareness concerning thyroid cancer might need to be improved. Referral and health care system, assessment, and diagnosis of thyroid lesions should be improved to reduce delayed detection and treatment. Implementation of risk assessment, safe surgery procedures, and expansion of adjuvant RAI also need to be improved to provide more comprehensive treatment for patients with differentiated thyroid cancer in our region. However, there are some limitations of this study including factors associated with the single center, retrospective study design, and duration of follow-up and surveillance. Future multicenter studies with more comprehensive assessment of clinical markers, metabolic comorbidities, as well as social determinants are required to inform and improve the management of differentiated thyroid cancer and to reduce the recurrence rates.