Physical examination of the thyroid: accuracy in detecting thyroid nodules and frequency of additional findings

Abstract

Objective

Thyroid palpation is a common clinical practice to detect thyroid abnormalities. However, its accuracy and potential for additional findings remain unclear. This study aimed to assess the diagnostic accuracy of physical exams in detecting thyroid nodules.

Methods

A retrospective observational study was conducted on a random sample of adult patients who underwent their first-time thyroid ultrasound between January 2015 and September 2017, following a documented thyroid physical exam. The study assessed the performance of thyroid palpation in detecting one or multiple thyroid nodules, as well as the proportion of additional findings on ultrasounds due to false positive thyroid palpation.

Results

We included 327 patients, mostly female (65.1%), white (84.1%) and treated in a primary care setting (54.4%), with a mean age of 50.8 years (SD 16.9). For solitary thyroid nodules, the physical exam had a sensitivity of 20.3%, specificity of 79.1%, an accuracy of 68.5%, negative predictive value (NPV) of 81.8%, and positive predictive value (PPV) of 17.6%. For detecting a multinodular goiter, physical exams demonstrated a sensitivity of 10.8%, specificity of 96.5%, accuracy of 55.4%, NPV of 53.9, and PPV of 73.9%. Among 154 cases with palpable nodules, 60% had additional nodules found in subsequent thyroid ultrasound.

Conclusion

Thyroid physical exam has limited diagnostic performance and leads to additional findings when followed by a thyroid ultrasound. Future efforts should be directed at improving the accuracy of thyroid physical exams or re-evaluating its routine use.

Introduction

Increased diagnosis of thyroid nodules and thyroid cancer has been attributed to the use of medical technology—in particular, the use of thyroid ultrasounds, which have high accuracy in detecting small thyroid lesions that are unlikely to cause harm or metastasize in the vast majority of patients.^(1-3) Several studies have shown an association between higher use of thyroid ultrasound and increased detection of small thyroid nodules, which are the type of thyroid lesions driving the increased incidence of thyroid cancer diagnoses.^(4-7) Research on the drivers of thyroid ultrasound use suggests that these tests are often ordered as part of a workup for thyroid abnormalities that are initially detected during a physical exam of the neck and thyroid.^(8,9) A population-based study found that in four out of 10 patients, the reason for undergoing thyroid ultrasound and subsequent fine needle aspiration biopsy was an abnormality palpated in their thyroid. Of particular interest from this study is the finding that clinical observations from these routine physical exams rarely matched the results from the thyroid ultrasound in terms of the position, size, and type of findings. In other words, the ultrasound was requested for a physical exam finding that ultimately did not correlate with the findings on the ultrasound. This suggests that one root cause behind increased detection of thyroid nodules and diagnosis of thyroid cancer may be the physical exam that precedes the use of thyroid ultrasound. However, despite the potentially significant role of physical exams in the diagnostic cascade that leads to detection of thyroid nodules and thyroid cancer, the research literature to date has provided little insight into the diagnostic properties of thyroid physical examinations and how their use may be driving incidental ultrasound findings.

The limited evidence on the diagnostic accuracy of the physical exam was summarized in a recent review of 8 studies, where it was noted that thyroid palpation had very low sensitivity in detecting thyroid nodules (sensitivity of 6.3-16.7% for nodules < 1 cm and 48.1-57.6% for larger thyroid nodules).⁽¹⁰⁾ However, the evidence supporting this review did not represent the current practice, as the most recent included study was published in 1998. Furthermore, included studies did not report the accuracy of the physical exam in different subgroups (e.g., patients with and without symptoms of nodular disease); the physical exam in most cases was mostly conducted as part of a controlled, experimental intervention (i.e., not necessarily a reflection of real-world clinical practice); and study authors did not report the rate of incidental findings triggered by false positive thyroid palpation results. The latter is a key diagnostic measure that could better clarify the harms and benefits of thyroid palpation as well as the potential impact of interventions to curtail the excessive diagnosis of thyroid nodules and cancer.

The purpose of our study was to examine the diagnostic outcomes of thyroid palpation in a contemporary cohort of patients with and without thyroid nodule symptoms who had a thyroid ultrasound. In addition, to quantify the accuracy of the physical exam and its impact on subsequent patient clinical trajectories, we calculated the percentage of patients who had additional findings on the ultrasound ordered due to abnormal thyroid palpation.

Methods

Study Design and Data Source

We conducted a retrospective observational study and identified adult patients (>18 years old) who underwent thyroid or neck ultrasounds (TUS) for the first time between January 1, 2015, and September 30, 2017, from our ultrasound database. An initial search identified 5,938 patients with a TUS, from which we randomly selected a cohort of 500 patients who had a TUS ordered for any reasons (e.g., screening, workup of thyroid dysfunction). Of these, 327 patients were found to have both a TUS ordered and documented physical examination of the thyroid, which was performed either as part of the thyroid-related workup (e.g., thyroid function test abnormalities, personal or family history of thyroid cancer) or as a routine neck examination. In all those instances, the TUS was ordered after the physical examination had been conducted.

Variables and Data Collection

We piloted and applied an extraction form to collect patient demographics and the specialty of the clinician who conducted the physical exam. From the physical exam documentation, we extracted the types of findings (solitary nodule, multinodular goiter, enlargement without nodularity, normal thyroid, and an "other category" for patients with indeterminate findings such as thyroid asymmetry). Additionally, we documented the precise locations (right lobe, left lobe, isthmus or bilateral) where these findings were encountered, whenever the data was accessible. In a similar fashion, we carefully extracted the TUS findings (solitary nodule, multinodular goiter, enlargement without nodularity, no finding, and "other category" to group patients who had thyroiditis without the presence of nodules or lymphadenopathy) and noted the corresponding locations of the described findings. When there was a nodular finding in the ultrasound, we extracted the largest diameter from the largest nodule. Given that the accuracy of thyroid physical exam may be affected by previous knowledge of thyroid-related symptoms, we also documented whether there was presence of symptoms associated with thyroid nodules or masses (e.g., neck fullness, dysphagia, odynophagia, choking, etc).

Statistical analysis

We summarized continuous variables using means and standard deviations, and categorical variables using frequencies and percentages. Using the TUS findings as the reference standard, we calculated the accuracy of the findings noted from the thyroid palpation. First, we calculated the sensitivity, specificity, accuracy, and positive and negative predictive value (PPV and NPV) of the thyroid palpation in the identification of any abnormal thyroid findings (any abnormality yes/no in thyroid palpation compared with any abnormality yes/no in thyroid ultrasound). Then, we estimated the accuracy of the physical exam in detecting thyroid nodules (single thyroid nodule [yes/no] detected by thyroid palpation compared to thyroid nodule [yes/no] detected on confirmatory thyroid ultrasound), and multinodular goiters (multiple nodules [yes/no] found by thyroid palpation compared to multiple nodules [yes/no] identified on a thyroid ultrasound). Additionally, we calculated the accuracy and PPV of the thyroid physical exam to determine the location (e.g., right lobe, left lobe, isthmus or bilateral) of the thyroid findings when data was available in the medical records. Finally, we calculated the proportion of additionally findings derived from a positive thyroid palpation as 1) the number of patients who had a solitary nodule palpated and were found to have multiple nodules on the subsequent ultrasound, or 2) patients who had a thyroid enlargement in the physical exam and found to have one or more thyroid nodules on the subsequent ultrasound. When possible, we conducted subgroup analyses by patient sex, patient weight (obese or overweight vs. normal weight based on body mass index (BMI) categories), size of the thyroid nodule (≤2 cm or >2cm), symptoms of nodular disease (yes and no), and by clinician specialty (primary care vs. specialties). We conducted a test interaction to evaluate for statistically significant subgroup differences. All statistical analyses were performed using JMP 16.0.

Results

We included 327 patients, the majority were female (213/327, 65.1%) and white (275/327, 84.1%) with a mean age (SD) of 50.8 years old (16.9). Most patients were obese or overweight (225/327, 71.2%) and had the thyroid physical exam performed by a primary care clinician (178/327, 54.4%). Additionally, most patients did not present any symptoms of thyroid nodular disease (231/327, 70.6%). Table 1 summarizes the sample demographics. Regarding finding any thyroid abnormality, the thyroid physical exam had a sensitivity of 71.8% (n= 189/263; 95% CI 66%-72.2%), specificity of 35.9% (n= 23/64; 95% CI 24.2%-49.9%), accuracy of 64.8% (212/327; 95% CI 59.4%-70.0%), NPV 23.7% (23/97; 95% 17.5%-31.3%) and PPV 82.1% (189/230; 95% CI 79.1%-84.9%). When the diagnostic accuracy was estimated by an individual thyroid palpation finding (thyroid nodule or MNG), the physical exam had a sensitivity of 20.3% (n=12/59; 95% CI 12.3%-36.2%), specificity of 79.1% (212/268; 95% CI 73.7%-83.2%), accuracy of 68.5% (n=224/327; 95% CI 59.4%-70.0%), NPV 81.8% (n=212/259; 95% 81.5%-85.8%) and PPV 17.6% (n=12/68;95% CI 11.1%-27.1%) in detecting a thyroid nodule; and a sensitivity of 10.8% (n-=17/157; 95% CI 6.4%-16.7%), specificity of 96.5% (n=164/170; 95% CI 92.4%-98.7%), accuracy of 55.4% (n=181/327; 95% CI 49.78%-60.8%), NPV 53.9% (n=164/304; 95% 52.4%-55.4%) and PPV 73.9% (n=17/23; 95% CI 53.4%-87.5%) in finding a multinodular goiter.

Table 1.

Baseline patient characteristics

Age	N=327	(%)
Years, mean (SD)	50.85 (16.9)
Sex
Female	213	(65.10%)
Male	114	(34.86%)
Ethnicity
White	275	(84.10%)
Asian	18	(5.50%)
Unknown/No Documentation	15	(4.59%)
Black or African American	13	(3.98%)
Other	3	(0.92%)
Hispanic or Latino	2	(0.61%)
Native Hawaiian/Other Pacific Islander	1	(0.31%)
BMI
Overweight (25 to <30) or obese (≥30)	225	(71.20%)
Normal (<25)	91	(28.80%)
Clinical setting
Primary care	178	(54.43%)
Endocrinology	75	(22.94%)
Other specialty	74	(22.63%)
Symptoms of nodular disease
Yes	96	(29.36%)
No	231	(70.64%)

Abbreviations: BMI = Body Mass Index.

Out of 327 patients, 60 physical exams included the side of the thyroid nodule. Among these, the overall accuracy, and the positive predictive value (PPV) of the physical examination in identifying the location (e.g., right lobe, left lobe, isthmus, or bilateral) of the thyroid finding were 46.6% and 37.5%, respectively. Physical exams among symptomatic patients demonstrated superior performance than those conducted among asymptomatic patients, with higher specificity (100% vs. 62%; p<0.01), accuracy (83% vs. 62%; p<0.01), and NPV (87% vs. 79%; p<0.01) for detection of solitary nodules, as well as a higher NPV for detection of multinodular disease (61% vs. 51%; p<0.01). In addition, higher PPV for detecting multinodular disease was observed when a specialist conducted the physical exam compared to a primary care provider (100% vs. 60%; p<0.01) and when the nodule detected by physical examination was greater than 2 cm (100% vs. 73% for nodules ≤2 cm; p=0.026). Figure 1 shows diagnostic outcomes in patients with solitary and multinodular goiter in different subgroups.

Figure 1.

Diagnostic outcomes of physical exam by subgroups

We found that out of 68 solitary nodules palpated, 37 (54%) were found to have additional findings such as, multinodular goiter on a subsequent TUS; similarly, out of 88 patients for whom the physical exam identified an enlarged thyroid, 56 (64%) of these were subsequently found to have a solitary nodule (16 cases) or a multinodular goiter (40 cases) on the TUS. Additional details can be found in Supplemental table 1.

Discussion

In this retrospective observational study, we found that, while a thyroid physical exam can detect thyroid abnormalities, these findings were rarely concordant with findings identified on a subsequent thyroid ultrasound. Furthermore, the diagnostic accuracy of specific thyroid physical exam findings was low, particularly for the PPV of solitary thyroid nodules. For instance, out of every 100 patients with a physical exam consistent with a solitary thyroid nodule, approximately 18 had a thyroid nodule, while out of every 100 patients with a physical exam consistent with multinodular goiter, about 74 had the same finding in the ultrasound.

These findings differ in some subgroups of interest. The PPV of solitary nodules was higher for patients with symptoms (42.8%) than for patients without symptoms (14.7%). Similarly, we found that physical exam performance was higher in cases where the nodule was greater than 2 cm (PPV: 100%) compared to nodules ≤2 cm (PPV: 73%). Compelling evidence has demonstrated a direct correlation between the presence of compressive symptoms and size of these abnormalities, potentially affecting the diagnostic yield of the physical exam.⁽¹¹⁾

We aimed to emphasize that patients with thyroid conditions frequently exhibit compressive symptoms, especially those receiving care at tertiary care centers. Notably, our cohort study uncovered a prevalence of 29.36% in patients presenting with symptoms related to compression. While some studies have reported lower prevalence, ranging from 11% to 22%, Bank et al. observed a notably higher prevalence of 52% in patients who had undergone thyroidectomy.^(11,12) Nonetheless, due to the non-specific nature of the compressive symptoms and the potential coexistence of non-thyroid-related findings that exhibit similar characteristics, our study is limited in its capacity to establish a direct relation between these symptoms and the outlined thyroid observations.

Our study revealed that the PPV for multinodular goiter was also higher when specialists conducted the physical exam (100%) when compared when primary care clinicians conducted the physical exam (60%). Moreover, we estimated that the possibility of an additional nodular finding triggered by a physical exam was between 50-60%.

Our study has limitations that affect the generalizability of the results. The diagnostic accuracy estimates may require validation in a cohort receiving care in a different setting. For instance, it is possible that in clinics located in areas with limited availability of or access to ultrasounds, clinicians may have a higher threshold to order thyroid ultrasounds, affecting the diagnostic accuracy of any physical exam and the subsequent diagnostic workup (e.g., biopsy rate).⁽¹³⁾ Furthermore, we did not evaluate the accuracy of thyroid physical exams prospectively, which lowers the quality of outcome ascertainment. In addition, a single nodule in the ultrasound may not be the same thyroid nodule palpated in the physical exam and could technically be considered a discordant rather than an accurate finding.^(14,15) The latter approach would likely decrease even more the positive predictive value of solitary nodule evaluation. Furthermore, using this retrospective cohort for diagnostic test evaluation may lead to biased results since there is already a suspicion of disease that leads to using ultrasound.⁽¹⁶⁾ In these circumstances, however, the bias leads to an overestimation of the diagnostic accuracy, suggesting the diagnostic accuracy of the physical exam could be even lower than shown in this study.

Given our constrained sample size, we were unable to determine the accuracy of thyroid physical examinations in specialized medical fields, including endocrinology. Although the impact of specialty on physical examination precision remains uncertain, a study indicated that surgeons, closely followed by endocrinologists at 40.6% and 37.9% respectively, exhibited the highest involvement in diagnosis and management, suggesting a potential higher level of expertise within this context.⁽¹⁷⁾

Implications for practice and research

The limited diagnostic performance of thyroid physical exam is concerning, as it is often used routinely in care for patients with no risk factors for thyroid cancer or symptoms of thyroid disease.^(18,19) This form of “screening of thyroid nodules or cancer” among asymptomatic or low-risk patients is controversial. The most recent U.S. Preventive Service Task Force guideline for screening of thyroid cancer recommends against routine physical exams in asymptomatic patients.⁽¹⁹⁾ In this recommendation, the Task Force emphasized avoiding the detection and treatment of small thyroid nodules and cancer that are unlikely to lead to mortality in affected patients, but treatment for which may lead to harms in the form of unnecessary health care interventions and costs.

Does this mean that thyroid palpation is an unwarranted practice?

The inaccuracy of physical examination and its likelihood to trigger additional imaging that may detect incidental findings suggests that thyroid palpation is an unwarranted practice in asymptomatic patients if the goal is to screen for thyroid cancer.⁽²⁰⁾ However, thorough physical examination is a cornerstone of the art of medicine, and its harms and benefits extend beyond thyroid cancer detection.^(21,22) For example, one literature review suggested that physical examination might strengthen the doctor-patient relationship and reduce patient anxiety.⁽²³⁾ Furthermore, although rare, a physical examination of the neck might lead to incidental findings of clinically significant disease.⁽²⁴⁾ However, the extent to which the benefits of thyroid palpation outweigh the potential harms of thyroid cancer overdiagnosis in the eyes of patients, clinicians, and other stakeholders is unclear.

A call for research

Given the complex nature of healthcare delivery, it is unlikely we will be able study the harms and benefits of neck palpation as an isolated screening intervention for thyroid cancer.⁽²⁵⁾ However, additional research could uncover the perspective of key stakeholders (e.g., patients, and clinicians) on the potential benefits and harms of thyroid palpation and inform recommendations that maximize benefits and minimize risks. Forthcoming research has the potential to uncover interventions, like physical examination training, designed to elevate the precision of neck palpation, particularly within medical training. Notably, a survey involving 1,387 medical trainees revealed that a considerable two-thirds of participants felt a consultant never supervised them during their physical examination practices. Furthermore, 31% noted that consultants never demonstrated the utilization of physical examination techniques to them.⁽²⁶⁾ Addressing this training gap could involve students engaging in physical examination alongside their consultants in real-life scenarios, particularly with patients known to have thyroid findings.

These and similar interventions have the potential to optimize the detection of thyroid nodules that are likely to require further workup or management (e.g., workshops on thyroid palpation followed by feedback from subsequent thyroid ultrasound confirmation) to establish whether a certain type of clinicians experiences below-average accuracy of thyroid examination, and would benefit from additional physical examination training, or to enhance the understanding of patient subgroups who may experience the greatest benefits or harms from the detection of benign nodules through routine thyroid examination (e.g., older asymptomatic patients with competing comorbidities

Likewise, it is important to better understand and mitigate the potential harms of not performing thyroid palpation. For instance, the COVID-19 pandemic disrupted clinical care by limiting in-person visits, creating a unique opportunity to evaluate the impact of no physical examination on patient experience and outcomes.^(27,28) Anecdotally, we noticed that any associated impacts on patient trust from the absence of the physical exam during virtual visits were ameliorated by better communication with patients and the ability to focus on symptoms. Given the increasing incidence of thyroid cancer and frequency of overdiagnosis, as well as routine performance of thyroid palpation in clinical practice, clarifying the potential benefits and harms of thyroid palpation should be a research priority.

Conclusions

Thyroid physical exam has limited diagnostic performance and leads to additional findings when followed by a thyroid ultrasound. As such, thyroid physical exams in asymptomatic patients may play an important role in the overdiagnosis and overtreatment of thyroid cancer. To determine if any patient subgroups derive some benefit from thyroid physical exams beyond their use as a diagnostic approach, further research is necessary. Additionally, potential interventions should be explored to improve the accuracy of thyroid physical exams.

Highlights.

• Diagnostic accuracy of physical exams for thyroid nodules is limited with low sensitivity.

• Understanding diagnostic accuracy of physical exams reveals their role in overdiagnosing thyroid nodules and cancer.

• Thyroid physical exams have limited accuracy, often resulting in additional ultrasound findings.

• The study investigated diagnostic outcomes of thyroid palpation and its impact on patient clinical paths.

Clinical Relevance section.

The diagnostic yield of thyroid physical examination (TPE) is limited and leads to incidental findings. Furthermore, its performance is linked to specific subgroups, including symptomatic patients and those with larger nodules. Nonetheless, TPE remains a cornerstone in the medical field with implications beyond our study’s scope, necessitating further research.