Ultrasound can see thyroid nodules that fingers cannot feel and that a person would never notice, and that sensitivity is useful when a nodule, symptom, or abnormal lymph node needs evaluation. Used as broad screening, it can start a cascade: repeat imaging, biopsy, indeterminate cytology, molecular testing, surgery, and lifelong thyroid hormone treatment for a tumor that might never have caused harm.
Overdiagnosis is not a mistaken pathology label. It is diagnosis of a real abnormality whose natural course would not have affected the person's health. Because clinicians cannot watch an untreated tumor across an entire counterfactual lifetime, overdiagnosis is estimated at population level rather than certified in one person at diagnosis.
A screening test can find disease without improving health#
Screening tests people without symptoms to identify disease earlier. Earlier detection helps only if treatment during the earlier period improves outcomes enough to outweigh harms.
Thyroid ultrasound finds nodules in a substantial share of adults, with prevalence varying by age, sex, iodine status, equipment, and definitions. Autopsy and imaging studies have also found small papillary thyroid cancers that caused no known illness during life.
This reservoir creates the conditions for overdiagnosis. When a screening program looks harder, incidence can rise rapidly. If the added tumors are mostly indolent, mortality may change little while surgery increases. The USPSTF recommends against screening asymptomatic adults for thyroid cancer. The recommendation does not apply to evaluating a palpable nodule, symptoms, or other clinically concerning findings.
The Korean experience made the pattern visible#
After thyroid ultrasound screening became common in South Korea, thyroid cancer incidence rose dramatically. Ahn and colleagues reported that most of the increase involved papillary cancers, while thyroid-cancer mortality remained stable.
The pattern is difficult to explain by a sudden biological epidemic that creates many new detectable tumors without increasing deaths. Screening intensity and detection of small tumors provide a more coherent explanation.
This natural experiment does not imply that every thyroid cancer found through screening is harmless. Aggressive thyroid cancers exist and require timely care. It shows that population detection can expand the diagnosis far more than the burden of lethal disease.
IARC analyses have estimated substantial overdiagnosis in multiple high-income countries, particularly among women. Estimates depend on expected incidence, calendar trends, age, and assumptions. They describe populations, not a probability assigned automatically to each nodule.
Lead time and survival can create an illusion#
Lead-time bias occurs when screening moves the diagnosis earlier without moving death later. Survival from diagnosis appears longer simply because the clock started sooner.
Length bias occurs because periodic screening is more likely to detect slow-growing tumors that remain in the detectable phase for longer. Rapid aggressive tumors can emerge between screens. Screen-detected cases therefore look more favorable even without a treatment benefit. Five-year survival after diagnosis cannot establish that screening saves lives. Disease-specific mortality in comparable populations, randomized evidence where feasible, and treatment harms are more informative.
TI-RADS is a brake as well as a detector#
ACR TI-RADS standardizes five ultrasound feature groups: composition, echogenicity, shape, margins, and echogenic foci. Points generate categories from TR1 through TR5.
The system then applies size thresholds. A highly suspicious TR5 nodule reaches biopsy at a smaller diameter than a mildly suspicious TR3 nodule. Very small suspicious nodules may receive follow-up rather than immediate sampling. Benign or not-suspicious categories generally avoid routine biopsy.
ACR explicitly notes that its thresholds were informed by recognition that many thyroid cancers are indolent. The goal is not maximal cancer detection. It is detection likely to improve outcomes while reducing procedures for findings unlikely to matter.
The companion TI-RADS and Bethesda guide explains the exact feature and cytology sequence.
A high category is not a command by itself#
TR5 describes a suspicious pattern. Management still depends on size, lymph nodes, symptoms, history, prior imaging, life expectancy, comorbidity, and preferences: a 4 mm TR5 nodule and a 2 cm TR5 nodule do not have the same recommendation.
Likewise, a large TR3 nodule may reach biopsy because size changes the balance even though its feature pattern is only mildly suspicious, because the point system prioritizes consistent action, not a universal cancer probability. Reports should identify the system because Korean, European, American Thyroid Association, and ACR schemes differ. The phrase “TI-RADS 4” without the system can be ambiguous.
Incidental nodules create another entry point#
CT, MRI, vascular ultrasound, and PET imaging can reveal thyroid findings while examining another problem. Automatically sending every incidental nodule to dedicated ultrasound amplifies cascades.
Incidental-finding recommendations use age, size, suspicious imaging findings, and clinical context to decide which nodules need ultrasound. Suspicious lymph nodes or local invasion deserve evaluation regardless of a routine size threshold, while a small nonsuspicious finding in a person with limited life expectancy may offer no benefit from follow-up. And the reason for the original scan should not get lost: chasing a low-risk thyroid nodule can pull attention away from the condition that brought the person to imaging in the first place.
Biopsy creates information and uncertainty#
Fine-needle aspiration is generally safe, but it can cause pain, bruising, bleeding, nondiagnostic samples, and repeated procedures. Cytology may be benign, malignant, or indeterminate.
Indeterminate results create choices among repeat sampling, molecular testing, surveillance, and diagnostic surgery. Molecular tests can refine probability but add cost and can yield results whose meaning depends on local prevalence and validation. Even a definitive small papillary cancer diagnosis does not answer whether immediate surgery improves the person's life; the diagnosis changes identity, insurance experience, anxiety, follow-up, and decisions even when the tumor is biologically indolent.
Treatment harms are part of screening harm#
Thyroid surgery can injure the recurrent laryngeal nerve, alter voice, cause bleeding, and damage parathyroid glands, producing low calcium. Total thyroidectomy requires lifelong hormone replacement. Complication risk depends on procedure extent and surgeon volume.
Hormone replacement is effective but requires monitoring and dose adjustment. Too much or too little can affect symptoms, heart rhythm, bone, pregnancy, and metabolism. Radioactive iodine has its own indications and harms and is not required for every differentiated cancer. Overtreatment is the treatment side of overdiagnosis. A screening program should count these downstream outcomes, not only cancers found.
Active surveillance changes the choice after diagnosis#
For selected adults with low-risk papillary thyroid microcarcinoma, active surveillance can be an alternative to immediate surgery. Programs use ultrasound, clinical review, and defined triggers such as meaningful growth or lymph-node involvement.
Surveillance is structured care, not abandonment. It requires reliable follow-up, experienced imaging, clear criteria, and a person comfortable living with an untreated diagnosis. Some people choose surgery because uncertainty creates unacceptable distress; others prioritize avoiding an operation.
Delayed surgery after documented progression can remain effective in selected cohorts, but results from specialized programs may not transfer to every setting. Higher-risk tumor features, invasive disease, suspicious nodes, or inability to follow reliably can favor treatment.
Surveillance itself can become a burden#
Repeated imaging can produce measurement noise. A difference of a few millimeters may reflect angle, border selection, or cystic change. Each scan can discover another nodule and reset anxiety.
Follow-up should have a purpose, interval, and stopping rule. What change would lead to biopsy? How many stable years are enough? Does age or health alter the benefit? “Repeat forever” is not a neutral plan. ACR TI-RADS specifies schedules for categories that meet follow-up thresholds and criteria for meaningful growth. Departures may be appropriate, but they should be reasoned rather than driven by fear of any visible nodule.
Overdiagnosis cannot be identified by good prognosis alone#
A tumor with excellent prognosis might still have caused future harm, and effective treatment might be responsible for that prognosis. Overdiagnosis is a counterfactual: the cancer would never have become clinically consequential without detection.
Researchers estimate it by comparing incidence with expected trends, mortality, tumor size, screening uptake, and long follow-up. Assumptions about background incidence and diagnostic practice can change estimates.
The term should not be used to dismiss a person's cancer. Once diagnosed, decisions use current pathology, stage, molecular features, imaging, health, and values. Population evidence guides how aggressively to look in the first place.
A balanced nodule pathway#
Evaluate symptoms and high-risk findings promptly, and use ultrasound when it can answer a clinical question rather than because it is available. Apply a named risk system consistently, and respect its size thresholds unless a clinical feature justifies another route. Connect the cytology back to the imaging and to local risk, and discuss surveillance and treatment in absolute numbers.
The objective is not to miss nothing. No diagnostic system can do that without causing substantial harm. The objective is to find disease whose earlier recognition improves health while allowing low-risk findings to remain low burden.
Communication should make deliberate nonaction visible. “No biopsy now” can mean that a nodule falls below a validated threshold, not that it was overlooked, and the report can name the category, size, recommended interval or absence of follow-up, and the symptoms that would reopen evaluation. This record helps another clinician avoid restarting the cascade from a decontextualized problem-list entry. It also lets you see that observation was chosen because it was expected to do more good than harm, not because your finding was dismissed.
References#
- ACR TI-RADS white paper
- ACR TI-RADS FAQ
- USPSTF thyroid-cancer screening recommendation
- NCI thyroid-cancer screening evidence summary
- IARC thyroid-cancer overdiagnosis analysis
- Korea's thyroid-cancer epidemic
For your own health, talk with your clinician.*
Questions and answers
What is thyroid cancer overdiagnosis?
It is diagnosis of a real thyroid cancer that would not have caused symptoms or death during the person's lifetime if it had remained undiscovered.
Is overdiagnosis the same as a false-positive test?
No. A false positive means the suspected disease is absent. In overdiagnosis, the tumor exists but its detection does not improve the person's health.
Why does TI-RADS use size thresholds?
Size thresholds reduce biopsies of small nodules whose probability of causing harm is often low, while preserving lower thresholds for more suspicious patterns.
Does no TI-RADS follow-up mean a radiologist made a mistake?
No. A small nodule can be accurately described and intentionally receive no follow-up because expected downstream harm exceeds likely benefit.
Who falls outside routine asymptomatic screening advice?
People with a palpable finding, symptoms, abnormal lymph nodes, certain hereditary syndromes, or relevant prior radiation history need individualized diagnostic evaluation rather than population-screening rules.