The 2017 United States Preventive Services Task Force Recommendation for Thyroid Cancer Screening Is No Longer the Gold Standard
Article information
In 2017, the United States Preventive Services Task Force (USPSTF) published recommendation against screening for thyroid cancer (D recommendation) [1]. They concluded that screening for thyroid cancer had at least moderate harms outweighing its benefits based on a review of the evidence [2]. The USPSTF is an independent, volunteer panel of national experts in prevention, primary care, and evidence-based medicine. To ensure the independence of their recommendation, the panel did not contain any experts in endocrinology who cared for patients with thyroid cancer in practice; therefore, this recommendation has been regarded as “a balanced gold standard.” However, the systematic reviews that supplied evidence for this recommendation were neither well-balanced nor adequately designed, undermining their claims. They had five key questions (KQs): KQ1, the benefit of thyroid cancer screening to reduce thyroid-specific morbidity or mortality; KQ2, the performance of screening tests, palpation or ultrasonography (US); KQ3, the harms of thyroid cancer screening; KQ4, the benefit of treating screen-detected thyroid cancer; and KQ5, the harms of treating screen-detected thyroid cancer. Regarding the benefit of screening, they found no studies directly comparing screened versus unscreened thyroid cancer patients. For the two KQs about screening tests, they confirmed that palpation was not sensitive enough to detect thyroid cancer, and US-guided fine-needle aspiration biopsy (FNAB) had no serious harm. For the benefit of treatment, they reviewed Japanese observational studies of screen-detected low-risk papillary thyroid cancers equal to or less than 1 cm in diameter, called papillary thyroid microcarcinomas (PTMCs) [3-5]. These Japanese studies were originally designed to prove the non-inferiority of observation with or without delayed surgery compared to immediate surgery for the treatment of low-risk PTMC patients and succeeded in showing no significant differences in recurrence and mortality between two groups. After these studies, active surveillance was accepted as one of treatment options for low-risk PTMCs. However, those studies also clearly showed that at least some PTMCs (over 40% in young patients during 15 years of observation) did progress [5], confirming that not all PTMCs are the same as latent thyroid cancers detected on autopsy [6]. KQ1 to KQ4 did not provided sufficient evidence to support or oppose screening for thyroid cancer. The crucial evidence against screening was the harm of treatment, which was the most unbalanced and inadequately designed part of the USPSTF recommendation. The panel excluded studies including patients with metastatic disease or anaplastic cancers to approximate screen-detected cancers, while for the harms of treatment of screen-detected thyroid cancer, they in cluded 36 studies on surgical harms, as well as 16 studies on the harms of radioactive iodine therapy (RAIT), which were conducted almost exclusively among patients who underwent total thyroidectomy (TT). They reviewed studies about the incidence of hypocalcemia from permanent hypoparathyroidism and vocal cord palsy after TT with or without lymph node dissection, finding rates of 2%–6% and 1%%–62%, respectively. As complications of RAIT, they examined the risk of second primary cancers (SPCs) and salivary gland sequelae. To summarize the studies included as evidence for the USPSTF recommendation, SPCs increased after a high dose (more than 3.7 GBq, 100 mCi) of RAIT [7-10] or a higher cumulative dose (more than 37 GBq, 1,000 mCi) [11], usually prescribed for patients with high risk of persistent, recurrent, or metastatic disease. About 90% of screendetected thyroid cancers are papillary thyroid cancers, and most of them are PTMCs. During the past decade, hemithyroidectomy, not TT, has been the treatment of choice for these PTMCs [12], and permanent hypoparathyroidism is therefore naturally excluded from the list of surgical complications. In fact, a recent systematic review and meta-analysis on the complication rates of surgery for the treatment of PTMC showed that the rates of permanent hypoparathyroidism after TT and hemithyroidectomy were 1.8% and 0%, respectively [13]. RAIT is not necessary after hemithyroidectomy. If it is performed, we usually use 1.1 GBq (30 mCi) after TT for low-risk PTC [14].
In response to clinicians’ concerns that the USPSTF recommendation would discourage neck examinations, the panel explained that the recommendation only applied to “asymptomatic” adults, not to people with symptoms such as hoarseness, pain, and difficulty in swallowing, or with signs of lumps, swelling, or asymmetry of the neck. As we all know, patients presenting with these symptoms or signs have far advanced cancers requiring TT with neck dissection and repeated high doses of RAIT, are at higher risks for vocal cord palsy, permanent hypoparathyroidism, and later SPCs, and finally may die of cancer. Should we delay thyroid US until thyroid cancer becomes symptomatic and incurable?
Notably, in this issue of Endocrinology and Metabolism, two papers demonstrating the true benefit of thyroid cancer screening are published. In one of these papers [15], the authors conducted an analysis of a nationwide cohort study and evaluated the impact of US screening on the outcomes of thyroid cancer by directly comparing screen-detected versus symptomatic, clinically detected cancer patients. The patients with thyroid-specific symptoms had larger cancers with more advanced T stages, and showed higher thyroid cancer-specific mortality as well as all-cause mortality. In the second of these studies [16], a systematic review and meta-analysis of 12 studies showed that screen-detected incidental thyroid cancers had better clinical staging (smaller tumor size and lower rates of aggressive histology, lymph node metastasis, and distant metastasis), resulting in about a 60% lower rate of recurrence and 70% lower rate of thyroid cancer-specific mortality when compared with symptomatic non-incidental cancer patients.
The USPSTF did not endorse the performance of US-guided FNAB for the detection of thyroid cancer for the reason that the sensitivity was overestimated because patients with negative US findings were not followed. They also argued that there were many unnecessary FNAB tests. However, a recent meta-analysis re-confirmed that FNAB is an integral part of thyroid cancer diagnosis and its accuracy has not changed over time [17]. Furthermore, Joo et al. [18] evaluated the performance of US-based risk stratification systems, including the 2016 Korean Thyroid Imaging Reporting and Data System (K-TIRADS) and 2021 K-TIRADS, and showed that the sensitivity and specificity were 96% versus 76% and 21% versus 50%, respectively. The unnecessary FNAB rate of the 2021 K-TIRADS was 50%, much lower than that of the 2016 K-TIRADS (79%), showing that the 2021 K-TIRADS has stricter criteria for FNAB of thyroid nodules.
“Screening may have the potential for the early detection of malignant thyroid nodules, which could make treatment more effective with less harm than if administered later”—just as described in the introduction of the 2017 USPSTF recommendation. We, endocrinologists, do not think every adult should undergo US to detect thyroid cancer, and do think that at least some PTMCs are over-diagnosed. However, there is no rationale against thyroid cancer screening if the true benefit of screening is evident, especially in the present situation where US is used like a stethoscope at the bedside. Instead, we should make efforts to differentiate whether or not small thyroid cancers will progress and propose guidelines for adequate treatment (not over-treatment) with minimal harm for patients with screen-detected thyroid cancer.
Notes
CONFLICTS OF INTEREST
No potential conflict of interest relevant to this article was reported.