Purpose: Understanding the method of thyroid cancer detection has potential implications on interpreting incidence rates, the diagnosis and management of thyroid cancer. We conducted a systematic review of studies reporting methods of thyroid cancer detection to estimate the frequency of incidentally found cancers and classify triggers of incidental thyroid cancer diagnosis. Methods: We searched multiple bibliographic databases from inception to June 2020. A pair of reviewers, working independently and in duplicate selected studies for inclusion, extracted data, and evaluated each trial’s risk of bias. Studies enrolling patients older than 18 years with thyroid cancer confirmed histologically were included. Results: In total, 17 cohorts and 1 cross-sectional study, conducted between 1991 and 2018, enrolling 4668 patients with thyroid cancer were included: 88% had papillary thyroid cancer and 23% had papillary thyroid microcarcinoma. The proportion of patients with non-incidental and incidental thyroid cancer was similar: 49% [95% confidence interval (CI): 40–58%]. Subgroup analysis showed that most patients with incidental thyroid cancers had tumor size <10 mm (76%; 95% CI: 56–92%), age >45 (61%; 95% CI: 56–67%), and were detected through imaging (35%; 95% CI: 26–45%), of which ultrasound was the most common modality (27%; 95% CI: 14–43%). The heterogeneity for all the effect sizes was large and significant. Conclusions: About half of thyroid cancers were found incidentally through the use of imaging studies, in particular neck ultrasound. These incidentally found cancers were mostly small papillary thyroid cancer. These results highlight opportunities for interventions aimed at reducing drivers of overdiagnosis.