The Influence of Incidental Detection of Thyroid Nodule on Thyroid Cancer Risk and Prognosis

A Systematic Review

Je Ern Chooi; Abiramie Ravindiran; Saba P. Balasubramanian


Clin Endocrinol. 2022;96(2):246-254. 

In This Article


The incidence of the incidentally detected thyroid nodule is on the rise, while its clinical significance remains controversial.[9] This systematic review aimed to determine the risk of cancer in these nodules and the impact of incidental detection on prognosis.

Apart from two studies published in 1998,[26,27] other studies included in this review were published in the last 20 years. This could be due to the exponential rise in incidental thyroid nodules detected secondary to the relatively recent widespread use of cross-sectional imaging including ultrasound.[10]

Risk of Thyroid Cancer by Mode of Detection

The number of incidentally detected thyroid nodules was lower than symptomatic nodules in all studies included in this analysis. In the case-control studies (n = 6),[20–22,24,26,28] 140 (19%) of 738 malignant tumours and 1324 (53%) of 2508 benign tumours were detected incidentally (odds ratio = 0.64–2.86). In the retrospective cohort studies (n = 3),[23,25,27] 9 (10%) of 91 patients in the incidental group and 56 (14%) of 398 patients in the nonincidental group had malignant nodules. The risk of malignancy ranged from 4% to 24% in the incidental arm and 4% to 29% in the nonincidental arm (relative risk = 0.13–6.27). These results are consistent with the malignancy rate of 7%–29% in other single-arm studies on incidental thyroid nodules.[12,39–41] The wide range is probably due to differing study designs, nature of the population studied and the definitions used for outcomes and exposures.

These studies were limited by the lack of adjustment for potential confounding factors (age, ethnicity, nodule size or patients' body mass index [BMI]) and the absence of follow up for nonoperated or benign nodules in many studies.

Prognosis of Thyroid Cancer by Mode of Detection

In this prognosis part of the review, 4293 patients (2129 in incidental group and 2164 in nonincidental group) and 871 patients (295 in incidental group and 576 in nonincidental group) were included in the 12 retrospective cohort studies[19,20,22,23,27,28,31–36] and two cross-sectional studies,[37,38] respectively. Nonincidental detection was much more common than incidental detection in all studies except for two studies.[19,32] Most retrospective cohort studies (n = 9)[20,22,23,27,28,33–36] were limited by lack of adjustment for potential confounding factors (i.e., patient's age, BMI or ethnicity, size of cancer nodules) except for three studies.[19,31,32]

Overall, incidentally detected thyroid cancer was significantly less likely to have recurrence/residual disease and had higher 5-year recurrence-free and overall survival rates. When comparing prognosis between these groups, it is important to recognize the potential for 'lead time' bias to impact on these results.

Esserman et al.[10] proposed that asymptomatic, nonpalpable and indolent cancers identified incidentally be redefined as indolent lesions of epithelial origin tumours to mitigate the inevitable anxiety surrounding the diagnosis and the potential for overtreatment and the associated morbidity and cost to society. However, accurate identification of patients with tumours that will express an indolent behaviour is challenging and should therefore be the focus of further research.

Limitations of the Review and Future Work

The vast majority of studies were either case-control or retrospective cohort; the limitations of which are widely recognized.[42–44] The number of studies meeting the eligible criteria for this review was limited. Twelve studies were excluded as relevant data on risk or prognosis studies were not available. The quality of the studies included in the review were variable but consistent with the moderate risk of bias demonstrated in two systematic reviews of observational studies in thyroid cancer.[45,46] There was significant variation observed in the definitions of incidental and nonincidental detection (shown in Tables S4 and S7), populations studied and outcomes measured. The review protocol defined thyroid incidentaloma as a lesion identified on imaging for reasons unrelated to the thyroid gland; however, in two manuscripts[23,26] in the risk study, some patients with thyroid incidentaloma detected during imaging for thyroid-related reasons or thyroid cancer screening were unable to be excluded from the incidental detection group during data extraction. Other definitions used by different studies included thyroid nodules found during exploration of the neck or autopsy regardless of any existing benign disease of the thyroid gland.[8,9,47,48] Five studies were excluded at the screening stage of this review because incidental detection was not represented by nodules detected on imaging for nonthyroid indications. There was also variation in eligibility criteria, and in some studies, it was unclear if patients with risk factors were equally represented in both arms. A meta-analysis could not be performed for the prognosis part of the study due to variation in inclusion criteria, assessment of prognosis markers and period of follow up. Reaching a consensus on the definitions of exposures and outcomes will help the standardisation across various studies, improve reporting quality, enable comparisons across populations and reduce heterogeneity in a meta-analysis.