Incidental Thyroid Nodules on Computed Tomography: A Systematic Review and Meta-Analysis Examining Prevalence,Follow-Up,and Risk of Malignancy

Abstract

Background:

The increased utilization of computed tomography (CT) has led to a higher detection rate of thyroid incidentalomas. Currently, there are no widely agreed-upon guidelines for managing these incidentalomas. This study aims to investigate the prevalence, follow-up practices, and malignancy rates of thyroid incidentalomas detected by CT.

Methods:

We conducted a comprehensive search of PubMed, Embase, and Cochrane databases to identify relevant studies published before April 12, 2024 (PROSPERO #42024535501). Studies reporting on the prevalence, follow-up, and risk of malignancy (ROM) of thyroid incidentalomas detected by CT were included. Combined outcomes were analyzed using pooled proportion with a random-effects model. The risk of bias was assessed using the Cochrane risk-of-bias tool for randomized trials (RoB 2) and the Newcastle–Ottawa Scale tool. Subgroup analyses were conducted based on characteristics including size of the incidentaloma, CT area, and age of the study population.

Results:

Thirty-eight studies involving 195,959 patients were included in the prevalence analysis, revealing a prevalence of thyroid incidentalomas on CT of 8.3% (confidence interval [CI], 7.4–9.3). The prevalence was higher in neck CT (16.5%, CI, 11.0–22.1) compared with chest CT (6.6%, CI, 5.3–7.9). Multiple incidentalomas were found in 27.0% (CI, 12.9–41.1) of patients. Of the nodules, 46.3% (CI, 32.3–60.3) were ≥1 cm, and 28.6% (CI, 19.9–37.3) were ≥1.5 cm. Thyroid ultrasounds, biopsies, and surgeries were performed in 34.9% (CI, 26.1–43.7), 28.4% (CI, 19.9–36.9), and 8.2% (CI, 2.1–14.4) of cases, respectively. Additionally, 25 studies with 6272 patients reported a ROM of 3.9% (CI, 3.0–4.9) for thyroid incidentalomas detected on CT. A higher ROM was observed in incidentalomas ≥1 cm (11.7%, CI, 3.9–19.4) and ≥1.5 cm (24.9%, CI, 0–52.7) compared with those <1 cm (0.1%, CI, 0–0.8) and <1.5 cm (0%, CI, 0–0.2).

Conclusions:

Most thyroid incidentalomas identified on CT are benign. Implementing a collaborative protocol between radiologists and thyroid specialists to manage high-risk thyroid incidentalomas can ensure appropriate follow-up and optimal patient care.

Introduction

The use of computed tomography (CT) imaging has dramatically increased over the past few decades.¹ This is driven by the advancements in imaging technology and its widespread application in clinical practice.² The rise in CT imaging has unintentionally led to an increase in incidental findings, such as thyroid nodules or incidentalomas, often discovered during imaging studies conducted for unrelated medical reasons. Depending on the study population and imaging settings, the reported prevalence of thyroid incidentalomas on CT can vary significantly, ranging from 0.5% to 39%.^3,4

Currently, there are no widely agreed-upon guidelines for the management of thyroid incidentalomas found on CT. The American Thyroid Association recommends evaluating all nodules that are 1 cm or larger. Nodules smaller than 1 cm will only require further evaluation if there are clinical symptoms or associated lymphadenopathy.⁵ In contrast, the American College of Radiology (ACR) uses age as an additional risk stratifier and proposes a 1.5 cm threshold for thyroid incidentaloma workup in patients over 35 years old.⁶ However, these recommendations were formed based on individual reports of thyroid incidentalomas with mixed quality, which can limit their generalizability. Furthermore, discrepancies between the guidelines of thyroid specialists and radiologists can lead to inconsistent patient care.

The primary aim of this study was to examine the follow-up and risk of malignancy (ROM) of thyroid incidentalomas on CT by performing a systematic review and meta-analysis, thereby testing the validity of current guidelines and informing clinical decision-making. The secondary aim was to explore the prevalence of thyroid incidentalomas on CT to understand the associated disease burden.

Methods

Literature search

This systemic review and meta-analysis was performed according to the outlines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. The review was registered prospectively in the PROSPERO database (CRD42024535501). Relevant studies published from inception to April 12, 2024 were identified through searches from PubMed, Ovid Embase, and Cochrane Library. The search strategy was designed and conducted by an experienced librarian, with input from the study’s principal investigator. The detailed search strategy listing all search terms used and how they are combined is shown elsewhere (Supplementary Data S1).

Inclusion and exclusion criteria

Cohort studies or randomized controlled trials providing information on the prevalence, follow-up management, and malignancy rates of incidentalomas found on CT in patients with no prior history of thyroid disease were considered for inclusion. Studies were excluded if they were not published in English or if the results were mixed with other imaging modalities, making it impossible to isolate CT-detected incidentalomas.

Data extraction

Two reviewers extracted the following data using a standard data collection form: first author, publication year, region, sample size, patient demographics, reasons for undergoing CT scans, exclusion of patients with a history of neck radiation, CT body area, CT reconstruction thickness, use of contrast, mean largest diameter of incidentalomas, and methods used to identify incidentalomas. Since all outcomes in this study were binary variables, the number of events of interest and the total number of observed events were extracted from the manuscripts.

Quality assessment

Two reviewers independently assessed the quality of the included studies using two established tools: Version 2 of the Cochrane risk-of-bias tool for randomized trials (RoB 2) for randomized controlled trials, and the Newcastle–Ottawa Scale (NOS) for cohort studies. For randomized controlled trials, the RoB 2 tool evaluates five aspects of bias: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in the measurement of the outcome, and bias in the selection of the reported result (Supplementary Data S2). For cohort studies, the NOS comprises eight items, each scoring one point, except for the “inter-group comparability” item, which is valued at two points (Supplementary Data S3). However, the “selection of the nonexposed cohort” and “comparability of cohorts on the basis of the design or analysis” items were not scored because all cohort studies included were single-arm studies. The outcome of interest was the malignancy rate of the incidentalomas. Consequently, the following items were not applicable to studies that only reported the prevalence of incidentalomas: “Demonstration that outcome of interest was not present at the start of the study,” “assessment of outcome,” “was follow-up long enough for outcomes to occur,” and “adequacy of follow-up of cohorts.” The maximum possible score for studies reporting malignancy risk was six points. A total score of zero to four, or a follow-up rate of less than 80%, indicated a high risk of bias, while a score of five or six with a follow-up rate of at least 80% signified a low risk of bias.⁷

Statistical analysis

All statistical analyses were performed by using Stata version 18.0 (Stata Corp, College Station, TX, USA). For binary outcomes, we calculated the pooled proportion (PP) and reported the 95% confidence interval (CI). The heterogeneity between studies was assessed using the I ² statistic.⁸ The random-effects model (DerSimonian–Laird method) was employed throughout our study because it best handles the inherent variability in studies by assuming effects are distributed, ensuring robust analysis regardless of heterogeneity. To investigate potential sources of heterogeneity, we conducted subgroup and multivariable meta-regression analyses. These analyses considered several factors: the reason for scan (lung cancer screening, COVID-19 screening, trauma, or others), the body area covered by the CT scan (neck, chest, or both), the use of contrast (yes or no), the average age of the study population (≥60 years or <60 years), the risk of bias and the method of identifying incidentalomas. The latter was distinguished between studies where incidentalomas were identified from previous radiology reports and those where all images were reevaluated by radiologists. When studies reported the risk of malignancy, efforts were made to calculate the malignancy rates separately for incidentalomas at least 1 cm, 1.5 cm, and below, unless the results were mixed for all incidentalomas. Sensitivity analysis is used to assess the robustness of the meta-analysis results, and statistical testing using Egger’s or Begg’s method (the number of studies must be ≥8) is used to evaluate whether there is publication bias in the included literature. For results with significant publication bias, the trim-and-fill method is used to assess the impact of publication bias on the results.

Results

Characteristics of the included studies

A flowchart of the study screening process is presented in Figure 1. The initial database search retrieved 916 papers, and an additional 2 were identified through other sources (references from included papers). After removing 180 duplicates, 738 articles were screened by title and abstract. Following this screening, 67 articles were selected for a thorough full-text review, resulting in the exclusion of 18 articles due to ineligibility (Supplementary Table S1). Ultimately, 49 studies^{3,4,9

–55} were included for systematic evaluation and meta-analysis.

FIG. 1.

Schema of the literature screening.

Table 1 shows the characteristics of the 49 included studies, encompassing a total of 235,765 patients who underwent CT imaging. Of these, 47 were cohort studies and 2 were randomized controlled trials. The mean age of patients was 56 ± 12 years, and the mean percentage of female patients was 54.9 ± 20.2%. Nine (18.4%) studies excluded patients with a history of neck radiation. The most common indications for CT scans were miscellaneous (40.8%), trauma (16.3%), lung cancer screening (10.2%), and COVID-19 screening (8.2%). Twenty-two (44.9%) studies involved chest CT alone, 13 (26.5%) involved neck CT alone, and the remainder either involved both or did not specify. The reconstruction thickness of CT imaging ranged from 0.75 to 5 mm, and contrast was used in 22 (44.9%) of the studies. The largest mean diameter of the incidentalomas ranged from 0.6 to 3 cm. The majority (59.2%) of the studies identified incidentalomas through text extraction from previous radiology reports, while the remaining 40.8% employed radiologist reevaluation of the original imaging.

Table 1.

Characteristics of the 49 Included Studies

Author, year	Region	Study design	Sample size	Age (year)	Sex (% female)	Patients with neck radiation excluded	Reason for scan	CT area	Reconstruction thickness	Contrast	Nodule size, cm (SD)	Identification of incidentaloma
Ahmed, 2012	USA	RCS	2510	59	48.7	Yes	Miscellaneous	Chest	3–5 mm	Yes	0.9 (0.6)	Radiologist reevaluation
Andersen, 2020	Denmark	RCT	503	64	57.1	NR	Miscellaneous	Chest	2 mm	Yes	NR	Radiologist reevaluation
Atalay, 2011	USA	RCS	335	56	60	NR	Pulmonary embolism	Chest	2 mm	Yes	NR	Radiologist reevaluation
Auladell, 2015	USA	RCS	93	39	67.7	NR	Tuberous sclerosis	Chest	0.74–5 mm	Partial	0.6 (0.5)	Radiologist reevaluation
Baez, 2015	USA	RCS	2525	13	51.4	NR	Miscellaneous	Chest	1.25 mm	Yes	1.2 (0.8)	Extracted from report
Bahl, 2018	USA	RCT	663	61	28.9	NR	Lung cancer screening	Chest	NR	No	2.1 (1.2)	Extracted from report
Baugh, 2014	USA	RCS	555	41	65.2	NR	Trauma	Both	NR	NR	NR	Extracted from report
Cansu, 2017	Turkey	RCS	26	60	36.2	NR	Cancer staging	NR	NR	NR	NR	Extracted from report
Chaikhoutdinov, 2014	USA	RCS	149	59	68.7	Yes	Miscellaneous	NR	NR	NR	NR	Extracted from report
Chan, 2021	Taiwan	RCS	1010	55	28.3	NR	Liver transplant	Chest	NR	NR	1.5 (0.5)	Extracted from report
Chen, 2020	China	RCS	1698	55	45.1	NR	Miscellaneous	Neck	NR	Yes	NR	Extracted from report
Crockett, 2015	Ireland	RCS	302	61	43	NR	Miscellaneous	Neck	1 mm	Yes	NR	Extracted from report
Doyle, 2018	USA	RCS	200	68	50	NR	Stroke	Both	NR	Partial	NR	Extracted from report
Drake, 2022	USA	RCS	1460	70	5.1	NR	Miscellaneous	Chest	NR	Partial	NR	Extracted from report
Drake, 2019	USA	RCS	30339	66	3.5	NR	Miscellaneous	Chest	NR	Partial	NR	Extracted from report
Ergun, 2013	Turkey	RCS	357	45	40.3	NR	Trauma	Neck	3 mm	No	NR	Radiologist reevaluation
Farra, 2017	USA	RCS	28	51	82	NR	Miscellaneous	NR	NR	NR	NR	Extracted from report
Helvaci, 2023	Turkey	RCS	35113	48	42.3	Yes	COVID-19	Chest	NR	No	NR	Extracted from report
Jantarapootirat, 2023	Thailand	RCS	1326	49	55.3	No	COVID-19	Chest	1 mm	Partial	0.6 (0.3)	Extracted from report
Jin, 2009	USA	RCS	150	61	74	NR	Miscellaneous	NR	NR	NR	NR	Extracted from report
Khasawneh, 2019	Japan	RCS	1135	56	51.4	NR	Miscellaneous	Neck	NR	Yes	NR	Radiologist reevaluation
Kroeker, 2013	Canada	RCS	729	50	71	NR	Miscellaneous	NR	NR	NR	2 (1.3)	Extracted from report
Lee, 2015	USA	RCS	259	61	66	Yes	Miscellaneous	Both	1–5 mm	No	2.6 (1.9)	Radiologist reevaluation
Lee, 2013	Korea	RCS	1941	58	61.9	NR	Lung cancer screening	Chest	5 mm	No	0.9 (0.7)	Extracted from report
Lehnert, 2014	USA	RCS	315	64	58	Yes	Trauma	Neck	NR	NR	1.5 (0.2)	Extracted from report
Lin, 2021	USA	RCS	563	61	72.7	Yes	Stroke	Neck	NR	NR	1.5 (0.6)	Radiologist reevaluation
Loomans-Kropp, 2019	USA	RCS	53454	61	41	Yes	Lung cancer screening	Chest	NR	NR	NR	Extracted from report
Morgan, 2017	USA	RCS	320	64	38.1	NR	Lung cancer screening	Chest	NR	Yes	NR	Extracted from report
Nguyen, 2013	USA	RCS	84720	40	77	NR	Miscellaneous	Neck	NR	NR	1.7 (0.9)	Radiologist reevaluation
Niedermeier, 2022	Germany	RCS	2871	82	64.2	NR	Trauma	Both	NR	NR	NR	Extracted from report
O’Leary, 2018	Ireland	RCS	100	NR	NR	NR	Gynecological cancer	Chest	NR	NR	NR	Radiologist reevaluation
Park, 2014	Korea	RCS	776	48	99.7	Yes	Breast cancer	Neck	5 mm	No	NR	Radiologist reevaluation
Park, 2017	Korea	RCS	61827	69	60.6	Yes	Miscellaneous	Chest	NR	Yes	1 (0.7)	Extracted from report
Rona, 2020	Turkey	RCS	338	57	50.3	NR	COVID-19	Chest	5 mm	No	NR	Radiologist reevaluation
Rositzka, 2020	Germany	RCS	1175	73	50.3	NR	Stroke	Neck	0.8 mm	Yes	NR	Radiologist reevaluation
Rothberger, 2019	USA	RCS	460	51	78	NR	Thyroid nodule	Neck	NR	NR	NR	Extracted from report
Sahiner, 2019	Turkey	RCS	155	77	63.2	NR	Valve replacement	Chest	0.75 mm	Yes	NR	Radiologist reevaluation
Seah, 2016	UK	RCS	104	38	NR	NR	Trauma	Both	NR	NR	NR	Extracted from report
Seo, 2023	Korea	RCS	2329	61	98	NR	Lung cancer screening	Chest	3–5 mm	No	1.3 (0.8)	Radiologist reevaluation
Shahir, 2015	USA	RCS	95	41	48.8	NR	Pulmonary embolism	Chest	NR	Yes	NR	Radiologist reevaluation
Short, 2022	USA	RCS	21583	NR	59.4	NR	Miscellaneous	Chest	NR	No	2.3 (1)	Extracted from report
Shetty, 2006	USA	RCS	230	65	69	NR	Miscellaneous	Both	2.5–5 mm	partial	2.2 (0.1)	Radiologist reevaluation
Sich, 2018	USA	RCS	674	NR	NR	NR	Trauma	Both	NR	NR	NR	Extracted from report
Sierink, 2013	Netherlands	PCS	321	44	28.3	NR	Trauma	Both	1–5 mm	Yes	NR	Extracted from report
Smith, 2020	USA	PCS	876	57	57.8	NR	Trauma	Both	NR	No	NR	Extracted from report
Willems, 2023	Netherlands	RCS	1992	71	41.6	NR	COVID-19	Chest	NR	NR	NR	Extracted from report
Yoon, 2008	Korea	RCS	734	50	47.6	No	Miscellaneous	Neck	1.5 mm	Yes	1.3 (0.1)	Radiologist reevaluation
Youserm, 1997	USA	RCS	231	56	42	No	Miscellaneous	Neck	3–5 mm	Yes	NR	Radiologist reevaluation
Zhang, 2020	China	RCS	81	NR	74.4	No	Miscellaneous	Neck	2 mm	Yes	3 (2)	Radiologist reevaluation

CT, computed tomography; NR, not reported; PCS, prospective cohort study; RCS, retrospective cohort study; RCT, randomized controlled trial; SD, standard deviation.

Quality assessment

Of the two randomized controlled trials included,^10,14 one¹⁰ reported only the prevalence of incidentalomas, while the other¹⁴ also provided follow-up and malignancy rate data. The outcome of the latter was comprehensively documented and considered adequate, leading to its classification as having a low risk of bias (Supplementary Fig. S1). Among the 47 cohort studies included in our review,^{3,4,9,11
–13,15

–55} 24 studies that reported malignancy data yielded scores ranging from 4 to 6.^{11
–13,16
–18,20,22,25,28,30
–32,34,35,40,42,45,47,48,52–55} Of these, 17 were found to have a low risk of bias,^{12,13,16
–18,20,25,28,30,31,35,42,45,47,48,54,55} while 7 were associated with a high risk of bias^{11,22,32,34,40,52,53} (Supplementary Tables S2 and S3).

Prevalence of thyroid incidentalomas on CT

Thirty-eight studies^{3,4,9

–15,18

–21,23,24,26,27,29,32,34

–46,49

–54} reported the prevalence of thyroid incidentalomas based on CT imaging of 195,959 patients. The pooled prevalence of thyroid incidentalomas on CT was 8.3% (CI, 7.4–9.3) (I ² = 99.3%, p < 0.001) (Fig. 2A). Neither Egger’s test (p = 0.686) nor Begg’s test (p = 0.372) indicated publication bias.

FIG. 2.

Pooled prevalence of thyroid incidentalomas on CT: overall (A), subgroup analysis by means of identifying incidentaloma (B), body area that CT covered (C), use of contrast (D), and reason for CT (E). Square represents the weight of each study, and diamond represents the effect size of each study. CI, confidence interval; CT, computed tomography; ES, effect size.

Subgroup analysis revealed a higher prevalence of incidentalomas in studies where incidentalomas were identified by radiologist reevaluation (PP 13.4%, CI, 9.1–17.7) compared with studies where incidentalomas were identified from reports (PP 5.2%, CI, 4.1–6.2) (Fig. 2B). The p-value for meta-regression was 0.301. Additionally, a higher prevalence of incidentalomas was observed in neck CTs (PP 16.5%, CI, 11.0–22.1) compared with chest CTs (PP 6.6%, CI, 5.3–7.9) (Fig. 2C). The p-value for meta-regression was 0.028. Incidentalomas were more prevalent in contrast-enhanced CTs (PP 11.5%, CI, 9.6–13.4) compared with noncontrast CTs (PP 5.5%, CI, 4.1–7.0) (Fig. 2D), although the p-value for meta-regression was 0.999. Furthermore, incidentalomas were less common in patients scanned for lung cancer screening (PP 4.5%, CI, 2.1–6.9) compared with those scanned for other reasons (PP 10.4%, CI, 8.9–11.9) (Fig. 2E), with a p-value for meta-regression of 0.687.

A leave-one-out sensitivity analysis was conducted for the pooled prevalence, demonstrating that the combined results were not significantly influenced by any single study (Supplementary Table S4). The funnel plot is shown in Supplementary Figure S2.

Characteristics of thyroid incidentalomas on CT

Eleven studies^{9,12,14,29,32,33,37,45,47,52,53} reported the imaging characteristics of 1994 patients who had thyroid incidentalomas detected on CT. The percentage of patients with multiple incidentalomas was 27.0% (CI 12.9–41.1) (I ² = 96.1%, p < 0.001) (Fig. 3A). Additionally, 46.3% (CI 32.3–60.3) of the incidentalomas were 1 cm or larger (I ² = 97.5%, p < 0.001) (Fig. 3B) and 28.6% (CI 19.9–37.3) of the incidentalomas were at least 1.5 cm (I ² = 85.7%, p < 0.001) (Fig. 3C). Subgroup analysis did not reveal the effect of reconstruction thickness on the reporting of incidentalomas of 1 cm or larger (Supplementary Fig. S3).

FIG. 3.

Pooled percentage of having multiple thyroid incidentalomas on CT (A), size of the incidentaloma of 1 cm or larger (B), size of the incidentaloma of 1.5 cm or larger (C), follow-up thyroid ultrasound (D), follow-up biopsy (E), and follow-up surgery (F). Square represents the weight of each study, and diamond represents the effect size of each study. DL, DerSimonian and Laird estimator.

Follow-up of thyroid incidentalomas on CT

Sixteen studies^{11
–13,18,20,22,34,36,37,42,45,48,49,52,54,55} reported follow-up data for 2683 patients with thyroid incidentalomas detected on CT. The percentage of patients who had a follow-up thyroid ultrasound was 34.9% (CI 26.1–43.7) (I ² = 89.1%, p < 0.001) (Fig. 3D). The percentage who underwent a follow-up biopsy was 28.4% (CI 19.9–36.9) (I ² = 95.9%, p < 0.001) (Fig. 3E). The percentage who underwent thyroid surgery was 8.2% (CI 2.1–14.4) (I ² = 87.1%, p < 0.001) (Fig. 3F).

ROM of thyroid incidentalomas on CT

Twenty-five studies^{11

–14,16
–18,20,22,25,28,30
–32,34,35,40,42,45,47,48,52

–55} reported the malignancy rates of thyroid incidentalomas detected on CT in 6272 patients. The pooled malignancy rate was 3.9% (CI 3.0–4.9) (I ² = 92.3%, p < 0.001) (Fig. 4A). Both Egger’s test (p = 0.95) and Begg’s test (p = 0.051) demonstrated no publication bias.

FIG. 4.

Pooled malignancy rate of thyroid incidentalomas found on CT: overall (A), subgroup analysis by means of identifying incidentaloma (B), body area that CT covered (C), use of contrast (D), mean age of the study population (E), risk of bias (F), largest diameter at least 1 cm (G), and largest diameter at least 1.5 cm (H). Square represents the weight of each study, and diamond represents the effect size of each study.

Subgroup analysis revealed that the malignancy rate in incidentalomas identified by radiologist reevaluation (PP 4.2%, CI 2.0–6.5) was similar to that in studies where incidentalomas were identified from reports (PP 6.0%, CI 4.3–7.8) (Fig. 4B). The p-value for meta-regression was 0.432. Additionally, the malignancy rate of incidentalomas was comparable in neck CTs (PP 6.5%, CI 3.2–9.9) and chest CTs (PP 3.6%, CI 2.1–5.2) (Fig. 4C), with a p-value for meta-regression of 0.562. Further subgroup analysis showed that the malignancy rate in noncontrast CT (PP 5.5%, CI 1.7–9.3) was similar to that in contrast-enhanced CT (PP 3.6%, CI 2.0–5.2) (Fig. 4D), with a p-value for meta-regression of 0.692. The analysis also revealed that thyroid incidentalomas have a higher malignancy rate in studies where the average age of the population was under 60 years old, with a PP of 12.5% (CI 6.3–18.7), compared with studies with an average age of 60 years or older, which have a PP of 3.0% (CI 1.7–4.3) (Fig. 4E), with a p-value for meta-regression of 0.38. Studies with a low risk of bias show a higher pooled malignancy risk of 9.1% (CI 6.2–12.1) compared with studies with a high risk of bias, which had a pooled malignancy risk of 1.8% (CI 0.5–3.1) (Fig. 4F). The p-value for meta-regression was 0.31.

Five studies^{13,14,18,31,45} provided stratified malignancy risk of thyroid incidentalomas on CT by size. Incidentalomas of 1 cm or larger had a malignancy rate of 11.7% (CI 3.9–19.4), compared with 0.1% (CI 0–0.8) for those smaller than 1 cm. Similarly, incidentalomas of at least 1.5 cm had a malignancy rate of 24.9% (CI 0–52.7), compared with 0% (CI 0–0.2) for those less than 1.5 cm.

Additionally, a leave-one-out sensitivity analysis was conducted for the pooled malignancy rate (Supplementary Table S5). The funnel plot is shown in Supplementary Figure S4. Excluding two studies^52,54 resulted in a pooled malignancy rate outside the original CI. After excluding these two studies, the pooled malignancy rate became 6.9% (CI 5.2–8.6) (I ² = 88.8%, p < 0.001).

Discussion

Our systemic review and meta-analysis provide a comprehensive characterization of the prevalence, follow-up, and malignancy rate of thyroid incidentalomas detected on CT. CT was selected as the imaging modality of interest in this study because it has been shown to detect the highest number of thyroid incidentalomas compared with other imaging modalities.^14,17 We found that the prevalence of thyroid incidentalomas on CT scans is 8.3%, which falls between that of ultrasound (50%)⁵⁶ and positron emission tomography scan (2%).⁵⁷

Overall CT detection of thyroid incidentalomas can be influenced by patient characteristics (age and scan indication), imaging properties (contrast enhancement), and interpretation methods (radiologist reevaluation). We demonstrated that only 34.9% of thyroid incidentalomas identified on CT were followed up with ultrasound. Further 3.9% were found to be malignant, with a higher malignancy rate observed in incidentalomas of 1 cm or greater and patient age under 60. These data help to better inform physicians about how often a thyroid incidentaloma may be expected to be found on CT imaging and the concomitant expected risk of malignancy.

We found that nodule size and patient age are significant discriminators of malignancy risk.^58,59 The ROM in thyroid incidentalomas measuring at least 1 cm or 1.5 cm increased twofold and sixfold, respectively, compared with the baseline risk of malignancy. In contrast, the ROM of thyroid incidentalomas below both thresholds approached zero, with low-to-moderate heterogeneity. Our findings seem to favor a recommendation for a 1.5 cm threshold for additional evaluation as it may avoid unnecessary workups without risking missing thyroid cancers. This needs to be further validated with future prospective studies. Furthermore, we demonstrated an increased ROM in younger populations, consistent with other reports showing a negative correlation between age and thyroid cancer incidence.⁶⁰ Our finding supports the inclusion of age in a potential CT thyroid incidentaloma follow-up guideline criteria, aligning with the ACR guidelines.

One common reason why thyroid incidentalomas smaller than 1 cm underwent workup in many studies included in our review was the presence of suspicious features on CT, such as irregular margins, calcification, and increased Hounsfield units.¹³ While some studies have suggested an association between these CT features and malignancy,^61,62 others have found that CT findings only match sonographic findings 53% of the time.⁴⁷ This has led the ACR to conclude that CT lacks adequate spatial resolution to display certain helpful sonographic findings, like cystic changes, microcalcifications, and lobulated or irregular margins.⁶ Recently, a study utilizing radiomics, which extracts numerous features from medical images, developed a prediction model that achieved a sensitivity of 84% and specificity of 94.1% in predicting the malignancy of thyroid nodules based solely on multiphase CT features.⁶³ Although this technology is not yet applicable to all thyroid incidentalomas, advancements in CT radiomics will likely provide more information about incidentalomas in the future.

In our study, the percentage of patients who received follow-up ultrasounds was lower than the percentage of incidental thyroid nodules larger than 1 cm. Even with a 1.5 cm cutoff, which results in a lower percentage of incidentalomas requiring further investigation, a significant portion of the studies did not meet this threshold, potentially leaving patients at risk of developing more advanced malignancies. Studies have reported that variability in reporting formats, guideline adherence, and reporting thresholds among radiologists can impact the management decisions of thyroid specialists.^22,33,64 Therefore, a standardized management protocol for thyroid incidentalomas, followed by both radiologists and thyroid specialists, may help promote better patient care. A recent study demonstrated that utilizing a natural language processing algorithm to identify adrenal gland incidentalomas directly from imaging reports supplying detailed guidance for ordering hormonal evaluation and follow-up imaging, improved follow-up and adequate evaluation for adrenal masses exceeding 1 cm in size.⁶⁵ This comprehensive approach could inform future strategies for improving the detection and management of thyroid incidentalomas on CT.

This study has several limitations. First, the results showed substantial heterogeneity. The heterogeneity in prevalence can be partially explained by the body area covered by CT as a significant discriminator. We believe the heterogeneity in the ROM might be caused by different definitions of thyroid incidentalomas across various studies, which we could not fully capture. Additionally, the inclusion criteria for thyroid incidentalomas, and thus the patient populations at risk for thyroid cancer, vary. Some studies began their analysis with all patients receiving CT scans. While these studies are less likely to be subject to selection bias, they tend to suffer from incomplete follow-up. Other studies focused on patients who were found to have thyroid incidentalomas. Although these studies have adequate follow-up, the populations are prone to preselection, where the cohort is inherently at a higher risk of malignancy, leading to an overestimation of the risk of malignancy. Second, only one-fifth of the included studies were suitable for analysis of malignancy risk by size, which is particularly evident from the high variance in malignancy risk of thyroid nodules of 1.5 cm or larger. This is due to most studies not providing data to stratify the malignancy rate. Third, the subgroup analysis on age was based on the mean age of the study population rather than a detailed breakdown of malignancy risk by specific age groups. Lastly, our study does not report on prognostic data. Since thyroid cancer is mostly an indolent cancer, we do not have data showing how the management of thyroid incidentalomas is directly related to morbidity and mortality.

Conclusions

Thyroid incidentalomas are not uncommon findings on CT and generally have a relatively low risk of malignancy. However, follow-up is often inadequate. Implementing a protocol that involves close collaboration between radiologists and thyroid specialists may improve the care for patients with thyroid incidentalomas.

Footnotes

Authors’ Contributions

Z.S.: Conceptualization (lead), writing—original draft (lead), writing—review and editing (equal), formal analysis (lead), and methodology (lead). C.W.: Methodology (equal) and writing—review and editing (equal). J.K.: Formal analysis (equal) and writing—review and editing (equal). A.G.: Formal analysis (equal) and writing—review and editing (equal). J.F.: Formal analysis (equal) and writing—review and editing (equal). B.L.: Writing—review and editing (equal). H.C.: Conceptualization (lead) and writing—review and editing (equal).

Author Disclosure Statement

The authors have no financial interests to disclose. All authors have reviewed the final draft of this article and are fully accountable for the content and integrity of this work.

Funding Information

No funding was received for this study.

Supplementary Material

Supplementary Data S1

Supplementary Data S2

Supplementary Data S3

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

Supplementary Table S5

References

Smith-Bindman

, Kwan

, Marlow

, et al. Trends in use of medical imaging in US health care systems and in Ontario, Canada, 2000-2016. JAMA, 2019; 322(9):843–856; doi: 10.1001/jama.2019.11456

Booij

, Budde

RPJ

, Dijkshoorn

, et al. Technological developments of X-ray computed tomography over half a century: User’s influence on protocol optimization. Eur J Radiol, 2020; 131:109261; doi: 10.1016/j.ejrad.2020.109261

Niedermeier

, Wania

, Lampart

, et al. Incidental CT findings in the elderly with low-energy falls: Prevalence and implications. Diagnostics (Basel), 2022; 12(2); doi: 10.3390/diagnostics12020354

Rositzka

, Hoffmann

, Meyer

, et al. Incidental findings on CT angiography of the head and neck in stroke patients. Evaluation of frequency and detection rate. Clin Neurol Neurosurg, 2020; 193:105783; doi: 10.1016/j.clineuro.2020.105783

Haugen

, Alexander

, Bible

, et al. 2015 American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: The American thyroid association guidelines task force on thyroid nodules and differentiated thyroid cancer. Thyroid, 2016; 26(1):1–133; doi: 10.1089/thy.2015.0020

Hoang

, Grady

, Nguyen

. What to do with incidental thyroid nodules identified on imaging studies? Review of current evidence and recommendations. Curr Opin Oncol, 2015; 27(1):8–14; doi: 10.1097/CCO.0000000000000147

Nakanishi

, Wang

, Miangul

, et al. Clinical outcomes of outpatient thyroidectomy: A systematic review and single-arm meta-analysis. Am J Surg, 2024; 236:115694; doi: 10.1016/j.amjsurg.2024.02.037

Song

, Wu

, Wang

, et al. The effects of parathyroidectomy vs medical treatments for secondary hyperparathyroidism in patients undergoing dialysis: A meta-analysis. Endocr Pract, 2024; 30(6):569–576; doi: 10.1016/j.eprac.2024.04.002

Ahmed

, Johnson

, Horton

, et al. Prevalence of unsuspected thyroid nodules in adults on contrast enhanced 16- and 64-MDCT of the chest. World J Radiol, 2012; 4(7):311–317; doi: 10.4329/wjr.v4.i7.311

10.

Andersen

, Ebbesen

, Thygesen

, et al. Impact of spectral body imaging in patients suspected for occult cancer: A prospective study of 503 patients. Eur Radiol, 2020; 30(10):5539–5550; doi: 10.1007/s00330-020-06878-7

11.

Atalay

, Walle

, Egglin

. Prevalence and nature of excluded findings at reduced scan length CT angiography for pulmonary embolism. J Cardiovasc Comput Tomogr, 2011; 5(5):325–332; doi: 10.1016/j.jcct.2011.08.001

12.

Auladell

, Boronat

, Barber

, et al. Thyroid nodules on chest CT of patients with tuberous sclerosis complex. Am J Med Genet A, 2015; 167A(12):2992–2997; doi: 10.1002/ajmg.a.37339

13.

Baez

, Zurakowski

, Vargas

, et al. Incidental thyroid nodules detected on thoracic contrast-enhanced CT in the pediatric population: Prevalence and outcomes. AJR Am J Roentgenol, 2015; 205(3):W360–W365; doi: 10.2214/AJR.14.13895

14.

Bahl

. Incidental thyroid nodules in the national lung screening trial: Estimation of prevalence, malignancy rate, and strategy for workup. Acad Radiol, 2018; 25(9):1152–1155; doi: 10.1016/j.acra.2018.02.016

15.

Baugh

, Weireter

, Collins

. The trauma pan scan: What else do you find? Am Surg, 2014; 80(9):855–859.

16.

Cansu

, Yılmaz

, Toru

, et al. Evaluation of incidental thyroid nodules in cancer patients. J Natl Med Assoc, 2017; 109(4):299–306; doi: 10.1016/j.jnma.2017.02.011

17.

Chaikhoutdinov

, Mitzner

, Goldenberg

. Incidental thyroid nodules: Incidence, evaluation, and outcome. Otolaryngol Head Neck Surg, 2014; 150(6):939–942; doi: 10.1177/0194599814524705

18.

Chan

, Chen

, Lin

, et al. Impact of thyroid incidentaloma on liver transplant: A study of 1010 Recipients at a Single Center. Ann Transplant, 2022; 27:e934988; doi: 10.1265/9/AOT.934988

19.

Chen

, Xue

, Wei

, et al. The undiagnosed potential clinically significant incidental findings of neck CTA: A large retrospective single-center study. Medicine (Baltimore), 2020; 99(43):e22440; doi: 10.1097/MD.0000000000022440

20.

Crockett

, Murphy

, Smith

, et al. Prevalence and clinical significance of extravascular incidental findings in patients undergoing CT cervico-cerebral angiography. Eur J Radiol, 2015; 84(8):1569–1573; doi: 10.1016/j.ejrad.2015.05.014

21.

Doyle

, George

, Holloway

, et al. Incidental findings in radiographic imaging for inpatients with acute ischemic stroke. J Stroke Cerebrovasc Dis, 2018; 27(11):3131–3136; doi: 10.1016/j.jstrokecerebrovasdis.2018.07.003

22.

Drake

, Gravely

, Ensrud

, et al. Reporting of incidental thyroid nodules on chest computed tomography and the impact on nodule evaluation: A Retrospective Cohort Study. Thyroid, 2022; 32(12):1529–1534; doi: 10.1089/thy.2022.0349

23.

Drake

, Gravely

, Westanmo

, et al. Prevalence of thyroid incidentalomas from 1995 to 2016: A single-center, retrospective cohort study. J Endocr Soc, 2020; 4(1):bvz027; doi: 10.1210/jendso/bvz027

24.

Ergun

, Lakadamyali

. The prevalence and clinical importance of incidental soft-tissue findings in cervical CT scans of trauma population. Dentomaxillofac Radiol, 2013; 42(10):20130216; doi: 10.1259/dmfr.20130216

25.

Farrá

, Picado

, Liu

, et al. Clinically significant cancer rates in incidentally discovered thyroid nodules by routine imaging. J Surg Res, 2017; 219:341–346; doi: 10.1016/j.jss.2017.06.050

26.

Helvacı

, Ozdemir

, Turan

, et al. Incidental thyroid nodules on COVID-19-related thoracic tomography scans: A giant cohort. Hormones (Athens), 2023; 23(2):227–233; doi: 10.1007/s42000-023-00516-9

27.

Jantarapootirat

, Traiwanatham

, Hirunpat

, et al. Thyroid incidentalomas in hospitalized patients with COVID-19: A single-center retrospective analysis. J Endocr Soc, 2023; 7(6):bvad060; doi: 10.1210/jendso/bvad060

28.

Jin

, Wilhelm

, McHenry

. Incidental thyroid nodule: Patterns of diagnosis and rate of malignancy. Am J Surg, 2009; 197(3):320–324; doi: 10.1016/j.amjsurg.2008.10.006

29.

Khasawneh

, Takeshita

, Hisatomi

, et al. Incidental findings in the thyroid gland on computed tomography images of the oral and maxillofacial region. Oncol Lett, 2020; 19(3):2005–2010; doi: 10.3892/ol.2020.11298

30.

Kroeker

, Le Nobel

, Merdad

, et al. Outcomes of incidentally discovered thyroid nodules referred to a high-volume head and neck surgeon. Head Neck, 2014; 36(1):126–129; doi: 10.1002/hed.23273

31.

Lee

, Chalmers

, Treister

, et al. Thyroid lesions visualized on CT: Sonographic and pathologic correlation. Acad Radiol, 2015; 22(2):203–209; doi: 10.1016/j.acra.2014.08.007

32.

Lee

, Jeong

, Kim

. Clinical significance of incidental thyroid nodules identified on low-dose CT for lung cancer screening. Multidiscip Respir Med, 2013; 8(1):56; doi: 10.1186/2049-6958-8-56

33.

Lehnert

, Sandstrom

, Gross

, et al. Variability in management recommendations for incidental thyroid nodules detected on CT of the cervical spine in the emergency department. J Am Coll Radiol, 2014; 11(7):681–685; doi: 10.1016/j.jacr.2013.12.007

34.

Lin

, Arevalo

, Lin

. Incidental thyroid nodules found during acute stroke angiography: Prevalence, outcomes, and suggested management guidelines. J Diagn Med Sonogr, 2021; 37(5):451–456; doi: 10.1177/87564793211018459

35.

Loomans-Kropp

, Dunn

, Kramer

, et al. Thyroid incidentalomas in association with low-dose computed tomography in the national lung screening trial. Am J Epidemiol, 2020; 189(1):27–33; doi: 10.1093/aje/kwz219

36.

Morgan

, Choi

, Reid

, et al. Frequency of incidental findings and subsequent evaluation in low-dose computed tomographic scans for lung cancer screening. Ann Am Thorac Soc, 2017; 14(9):1450–1456; doi: 10.1513/AnnalsATS.201612-1023OC

37.

Nguyen

, Roy Choudhury

, Eastwood

, et al. Incidental Thyroid nodules on CT: Evaluation of 2 risk-categorization methods for work-up of nodules. AJNR Am J Neuroradiol, 2013; 34(9):1812–1817; doi: 10.3174/ajnr.A3635

38.

O’Leary

, Treacy

, Geoghegan

, et al. Incidental thoracic findings on routine computed tomography in epithelial ovarian cancer. Int J Gynecol Cancer, 2018; 28(6):1073–1076; doi: 10.1097/IGC.0000000000001268

39.

Park

, Choi

, Huh

, et al. Incidental findings on simulation CT images for adjuvant radiotherapy in breast cancer patients. Technol Cancer Res Treat, 2015; 14(5):525–529; doi: 10.7785/tcrt.2012.500415

40.

Park

, Lee

, Cho

, et al. Incidental thyroid nodules on thoracic contrast-enhanced computed tomography in clinical practice during a 10-year period. Medicine, 2017; 96(11); doi: 10.1097/MD.0000000000006388

41.

Rona

, Arifoğlu

, Voyvoda

, et al. Should CT be used for the diagnosis of RT-PCR-negative suspected COVID-19 patients? Clin Respir J, 2021; 15(5):491–498; doi: 10.1111/crj.13332

42.

Rothberger

, Cohen

, Sahay

, et al. Method of detection of thyroid nodules: Correlation with frequency of fine-needle aspiration and malignancy rate. Head Neck, 2020; 42(2):210–216; doi: 10.1002/hed.25984

43.

Sahiner

, Coteli

, Kivrak

, et al. Incidental findings diagnosed during preprocedural evaluation of TAVR. Cardiol Res Pract, 2019; 2019:7478608; doi: 10.1155/2019/7478608

44.

Seah

MKT

, Murphy

, McDonald

, et al. Incidental findings on whole-body trauma computed tomography: Experience at a major trauma centre. Injury, 2016; 47(3):691–694; doi: 10.1016/j.injury.2016.01.012

45.

Seo

, Jin

, Park

, et al. Risk of thyroid cancer in a lung cancer screening population of the National Lung Screening Trial according to the presence of incidental thyroid nodules detected on low-dose chest CT. Ultrasonography, 2023; 42(2):275–285; doi: 10.1436/6/usg.22111

46.

Shahir

, McCrea

, Lozano

, et al. Reduced z-axis technique for CT Pulmonary angiography in pregnancy–validation for practical use and dose reduction. Emerg Radiol, 2015; 22(6):651–656; doi: 10.1007/s10140-015-1340-7

47.

Shetty

, Maher

, Hahn

, et al. Significance of incidental thyroid lesions detected on CT: Correlation among CT, sonography, and pathology. AJR Am J Roentgenol, 2006; 187(5):1349–1356; doi: 10.2214/AJR.05.0468

48.

Short

, Dondlinger

, Wildman-Tobriner

. Management of incidental thyroid nodules on Chest CT: Using natural language processing to assess white paper adherence and track patient outcomes. Acad Radiol, 2022; 29(3):e18–e24; doi: 10.1016/j.acra.2021.02.019

49.

Sich

, Rogers

, Bertozzi

, et al. Filling the void: A low-cost, high-yield approach to addressing incidental findings in trauma patients. Surgery, 2018; 163(4):657–660; doi: 10.1016/j.surg.2017.09.043

50.

Sierink

, Saltzherr

, Russchen

MJAM

, et al. Incidental findings on total-body CT scans in trauma patients. Injury, 2014; 45(5):840–844; doi: 10.1016/j.injury.2013.10.009

51.

Smith

, King

, Shealy

, et al. Incidental findings in the trauma population: Interdisciplinary approach and electronic medical record reminder association with pre-discharge reporting and medicolegal risk. J Am Coll Surg, 2021; 232(4):380–385.e1; doi: 10.1016/j.jamcollsurg.2020.11.028

52.

Willems

JIA

, Tummers-de Lind van Wijngaarden

RFA

, Dubelaar

IJM

, et al. Thyroid and adrenal incidentalomas on chest CT: Prevalence, diagnostic work-up and outcomes in a cohort of COVID-19 suspected patients. Eur J Intern Med, 2023; 123:114–119; doi: 10.1016/j.ejim.2023.12.008

53.

Yoon

, Chang

, Choi

, et al. The prevalence and significance of incidental thyroid nodules identified on computed tomography. J Comput Assist Tomogr, 2008; 32(5):810–815; doi: 10.1097/RCT.0b013e318157fd38

54.

Youserm

, Huang

, Loevner

, et al. Clinical and economic impact of incidental thyroid lesions found with CT and MR. Am J Neuroradiol. Sep, 1997; 18(8):1423–1428.

55.

Zhang

, Qiao

, Zhang

. Value of CT features in the diagnosis of papillary thyroid tumors in incidental thyroid nodules. Int J Endocrinol, 2020; 2020:9342317; doi: 10.1155/2020/9342317

56.

Acar

, Ozbek

, Acar

. Incidentally discovered thyroid nodules: Frequency in an adult population during Doppler ultrasonographic evaluation of cervical vessels. Endocrine, 2014; 45(1):73–78; doi: 10.1007/s12020-013-9949-3

57.

de Leijer

, Metman

MJH

, van der Hoorn

, et al. Focal thyroid incidentalomas on 18F-FDG PET/CT: A systematic review and meta-analysis on prevalence, risk of malignancy and inconclusive fine needle aspiration. Front Endocrinol (Lausanne), 2021; 12:723394; doi: 10.3389/fendo.2021.723394

58.

, Holland

, Song

, et al. Very elevated parathyroid hormone levels in patients with primary hyperparathyroidism: Is it cancer? Am J Surg, 2024; 231:140–141; doi: 10.1016/j.amjsurg.2023.12.026

59.

, Zmijewski

, Akhund

, et al. Graves’ disease: Unveiling a novel etiology of secondary hyperparathyroidism. Am J Surg, 2024; 229:182–183; doi: 10.1016/j.amjsurg.2023.10.041

60.

Kwong

, Medici

, Angell

, et al. The influence of patient age on thyroid nodule formation, multinodularity, and thyroid cancer risk. J Clin Endocrinol Metab, 2015; 100(12):4434–4440; doi: 10.1210/jc.2015-3100

61.

Yoon

, Cho

, Lee

, et al. Thyroid incidentalomas detected on 18F-fluorodeoxyglucose-positron emission tomography/computed tomography: Thyroid Imaging Reporting and Data System (TIRADS) in the diagnosis and management of patients. Surgery, 2015; 158(5):1314–1322; doi: 10.1016/j.surg.2015.03.017

62.

Kim

, Jung

, Baek

. Computed tomography features of benign and malignant solid thyroid nodules. Acta Radiol, 2015; 56(10):1196–1202; doi: 10.1177/0284185114552216

63.

Lin

, Gao

, Yang

, et al. CT-based radiomics models for differentiation of benign and malignant thyroid nodules: A multicenter development and validation study. Am J Roentgenol, 2024:1–12; doi: 10.2214/AJR.24.31077

64.

Lawrence

, Prout

, Mankowski-Gettle

. Evaluation of guideline adherence and subsequent follow-up outcomes for incidental thyroid nodules detected in hybrid academic-community practice. J Comput Assist Tomogr, 2022; 46(4):651–656; doi: 10.1097/RCT.0000000000001311

65.

Veazey

, Wu

, Song

, et al. Brave New World: Use of Natural Language Processing to Retrospectively Identify Adrenal Incidentalomas. Unpublished. 2024.