Objective: We conducted this systematic review to support the U.S. Preventive Services Task Force in updating its recommendation on screening for thyroid cancer. Our review addresses the following Key Questions (KQs): 1) Compared with not screening, does screening adults for thyroid cancer lead to a reduced risk of thyroid-specific mortality or morbidity, reduced all-cause mortality, and/or improved quality of life? 2) What are the test performance characteristics of screening tests for detecting malignant thyroid nodules in adults? 3) What are the harms of screening for thyroid cancer in adults? 4) Does treatment of screen-detected thyroid cancer reduce thyroid-specific mortality or morbidity, reduce all-cause mortality, and/or improve quality of life? 5) What are the harms of treating screen-detected thyroid cancer?
Data sources: We searched MEDLINE, PubMed, and the Cochrane Central Register of Controlled Trials to locate relevant studies for all KQs. We searched for articles published from January 1966 to January 2016.
Study selection: We reviewed 10,424 abstracts and 707 articles against specified inclusion criteria. Eligible studies included those written in English and conducted in asymptomatic adult populations at general risk or with a prior personal history of radiation exposure.
Data analysis: We conducted dual independent critical appraisal of all included studies and extracted study details and outcomes from fair- or good-quality studies. We synthesized results by KQ and type of screening test (i.e., palpation or ultrasound). We used primarily qualitative synthesis. We used random-effects meta-analyses to pool surgical harms. We also summarized the overall strength of evidence for each KQ.
Results: We found no studies that met our inclusion criteria for KQ 1. Ten fair-quality studies were included for KQ 2. In two studies, neck palpation was not sensitive to detect thyroid nodules. Two methodologically limited studies that used selected sonographic features demonstrated that screening with ultrasound can be specific for detecting thyroid malignancy; one of these studies suggested that using a combination of high-risk sonographic features, such as microcalcification or irregular shape, can optimize both sensitivity and specificity. Three fair-quality studies met our inclusion criteria for KQ 3, none of which suggested any serious harms from screening or ultrasound-guided fine-needle aspiration. However, we found no screening studies that directly examined the risk of overdiagnosis. Two studies met our inclusion criteria for KQ 4, but neither was designed to determine if earlier or immediate treatment versus delayed or no surgical treatment improves the outcomes of patients with well-differentiated thyroid cancer. Fifty-two studies were included for KQ 5. Based on 36 studies, permanent surgical harms, hypoparathyroidism, and recurrent laryngeal nerve palsy are not uncommon. Best estimates of permanent hypoparathyroidism are from 2 to 6 events per 100 thyroidectomies and are more variable with lymph node dissection. The rate of recurrent laryngeal nerve palsy is estimated at 1 or 2 events per 100 surgeries. Based on 16 studies, treatment of differentiated thyroid cancer with radioactive iodine (RAI) treatment is associated with a small increase in second primary malignancies; RAI treatment is also associated with increased permanent adverse effects on the salivary gland, such as dry mouth.
Limitations: The vast majority of studies that evaluated the diagnostic accuracy of ultrasound to detect thyroid tumors are not in screening populations. High statistical heterogeneity for surgical harms of hypoparathyroidism could not be explained by known clinical heterogeneity across studies. Differences in study designs and variable reporting on radiation doses limits our understanding of the magnitude and precision around the excess risk for second primary malignancies due to RAI.
Conclusions: Although ultrasound of the neck using high-risk sonographic characteristics plus followup cytology from fine-needle aspiration can reasonably identify thyroid cancer, it is unclear if population-based or targeted screening can decrease mortality or improve important patient health outcomes. More importantly, screening results in the identification indolent thyroid cancer, and treatment of these cases of overdiagnosed cancer can pose real patient harms.