Thyroid cancer overdiagnosis statistics: why women's diagnoses rose while death rates stayed flat

She was 42, no neck mass she could feel, no voice change, no trouble swallowing. A CT ordered for a cough picked up a 7 mm thyroid nodule. Ultrasound, fine-needle aspiration, and a week later she sat in my office with a report that said papillary thyroid carcinoma. She had already googled survival numbers and was terrified. She also had a three-year-old and a job that did not pause for a cancer label.

I told her what the population data keep saying: most small papillary cancers in women never kill anyone. Finding them does not automatically mean we prevented a death. Sometimes it means we handed a healthy person a diagnosis, a surgery, and lifelong thyroid hormone replacement for a tumor that would have stayed quiet forever.

That pattern is overdiagnosis. Thyroid cancer is the textbook case in women's health. Incidence climbed for decades while mortality barely moved, and women absorb most of the extra diagnoses. This page gathers the numbers reporters and clinicians keep needing: incidence versus death rates, the female-to-male gap, estimated overdiagnosis fractions, screening intensity, and guideline shifts meant to slow the cascade.

Year	Incidence per 100,000
1975	5.00
1985	5.15
1995	6.28
2005	11.10
2009	14.49
2014	14.94
2019	14.30
2023	14.86

Source: NCI SEER Cancer Stat Facts: Thyroid Cancer. SEER 8 age-adjusted rates of new cases, all races, both sexes. Values rounded as published in the Stat Facts trend table.

Measure	Earlier period	Later period	Source
U.S. incidence (both sexes)	4.9 per 100,000 (1975)	14.3 per 100,000 (2009)	Davies & Welch, 2014
U.S. incidence, women	6.5 per 100,000 (1975)	21.4 per 100,000 (2009)	Davies & Welch, 2014
U.S. incidence, men	3.1 per 100,000 (1975)	6.9 per 100,000 (2009)	Davies & Welch, 2014
Mortality (both sexes)	~0.5 per 100,000 (1975)	~0.5 per 100,000 (2009 / 2019)	Davies & Welch, 2014; Chen et al., 2025
Women's mortality	0.6 per 100,000 (1975)	0.5 per 100,000 (2019)	Chen et al., 2025
Metastasis rate at diagnosis	0.4 per 100,000 (1975)	0.4 per 100,000 (2019)	Chen et al., 2025
Tumors ≤1 cm (share of diagnoses)	25% (1988-1989)	39% (2008-2009)	Davies & Welch, 2014

Sources: Davies L, Welch HG. JAMA Otolaryngol Head Neck Surg. 2014; Chen MM, et al. Lancet Diabetes Endocrinol. 2025. Rates are age-adjusted unless noted by the original authors.

Population / period	Estimated overdiagnosis	Source
U.S. women, PTC, 1991-2019	75%-95% of cases	Francis et al., JAMA Netw Open, 2026
U.S. men, PTC, 1991-2019	63%-90% of cases	Francis et al., 2026
U.S. women, all thyroid cancer, ~2003-2007	70%-80%	Vaccarella et al., NEJM, 2016 (IARC analysis)
Republic of Korea women, ~2003-2007	~90%	Vaccarella et al., 2016
26 countries, 2008-2012	>1 million people (830,000 women)	Li, Dal Maso, Vaccarella, 2020
63 countries, 2013-2017	75.6% of 2.3 million cases	Li et al., 2024

Sources: Francis et al. 2026; Vaccarella et al. 2016; Li et al. 2020; Li et al. 2024. Methods differ (microsimulation vs excess-incidence models); ranges reflect model assumptions, not clinical certainty for any one patient.

Incidence rose. Mortality did not.

Start with the simplest comparison. According to NCI's SEER Cancer Stat Facts, the U.S. rate of new thyroid cancer cases is 13.7 per 100,000 men and women per year (2019-2023 diagnoses). The death rate is 0.5 per 100,000 (2020-2024 deaths). Five-year relative survival is 98.3% for cases diagnosed in 2016-2022. For 2026, SEER cites American Cancer Society projections of about 45,240 new cases and 2,320 deaths.

The historical arc is steeper. In their 2014 SEER analysis covering 1975-2009, Louise Davies and H. Gilbert Welch reported that incidence nearly tripled, from 4.9 to 14.3 per 100,000 adults, while mortality stayed near 0.5 per 100,000. Nearly all of the increase was papillary thyroid cancer (3.4 to 12.5 per 100,000). Tumors got smaller: the share measuring 1 cm or less rose from 25% in 1988-1989 to 39% in 2008-2009. Their conclusion was blunt: an epidemic of diagnosis, not of lethal disease.

A 2025 age-period-cohort analysis by Chen and colleagues in The Lancet Diabetes & Endocrinology extended the timeline through 2019. Overall incidence rose from 5.0 per 100,000 in 1975 to 14.6 in 2009, then plateaued (14.1 in 2019). Metastasis at diagnosis held at 0.4 per 100,000 in both 1975 and 2019. Mortality remained about 0.5 per 100,000. SEER's modeled trends now describe incidence as stable from 2014 through 2023 and death rates as stable from 2015 through 2024. That is a plateau at high diagnosis volume, not a return to 1970s rates.

One nuance: Lim and colleagues, writing in JAMA in 2017, found that overall incidence-based mortality rose about 1.1% per year from 1994 to 2013 (about 1.7% per year for papillary thyroid cancer; 2.9% per year for distant-stage papillary) even as the classic death rate stayed near 0.5. The dominant story remains overdiagnosis of small indolent tumors, with a smaller concurrent rise in some advanced papillary disease. Both can be true. Absolute deaths stay low; absolute diagnoses do not.

Thyroid cancer sits next to breast and cervical screening debates for a reason: same analytic question, different evidence. Our reviews of breast cancer screening statistics and cervical cancer and HPV screening cover screening that does reduce mortality. Asymptomatic thyroid screening does not have that evidence base.

Why women bear the diagnosis burden

SEER's sex-specific rates put female incidence at 21.2 per 100,000 and male incidence at 6.1 per 100,000 (2019-2023). That is roughly 3.5 women diagnosed for every man. Median age at diagnosis is 51; median age at death is 74. About 63.5% of cases are localized at diagnosis, with 99.9% five-year relative survival for that stage.

Davies and Welch quantified how uneven the climb was. From 1975 to 2009, women's incidence rose from 6.5 to 21.4 per 100,000 (relative rate 3.3), while men's rose from 3.1 to 6.9 (relative rate 2.2). The absolute increase in women was nearly four times that in men. Chen and colleagues found women's incidence peaking at 22.2 per 100,000 in 2014, and men's at 8.1 in 2012. After the mid-1990s, middle-aged women accounted for a disproportionate share of the rise. That is the age band most likely to receive imaging for other reasons and to have a neck examined in primary or gynecologic care.

Lim's 2017 SEER analysis put women at 75% of diagnosed cases, with papillary histology making up 84% of tumors. Aggressive histologies such as anaplastic and medullary disease do not show the same female predominance. The sex gap is largest for the small, localized papillary cancers that autopsy series find in people who never knew they had them.

Biology may play some role (incidence peaks earlier in women), but it does not explain a tripling of diagnoses without a matching rise in deaths. Access to ultrasound and CT does. Women use ambulatory care more often; more visits mean more incidental findings. That pathway rhymes with patterns in our medical gaslighting and diagnostic delay statistics: women can be over-investigated in some pathways and under-believed in others. Thyroid overdiagnosis is the over-investigation side of that coin.

Women's death rates are quieter. Chen et al. reported female mortality declining slightly from 0.6 per 100,000 in 1975 to 0.5 in 2019, while male mortality edged from 0.4 to 0.5. Diagnosis is gendered; lethal disease is not, to the same degree.

How much is overdiagnosis

Overdiagnosis means detecting a cancer that would not have caused symptoms or death in that person's lifetime. You cannot prove it in one living patient. You estimate it from rising incidence without rising mortality, excess cases relative to historical rates, autopsy reservoirs of latent disease, and models that separate true incidence change from diagnostic expansion.

The 2026 Francis et al. microsimulation in JAMA Network Open is the cleanest recent U.S. estimate for papillary disease. Overall overdiagnosis of U.S. papillary thyroid cancer from 1991 to 2019 was 72%-94%. For women the range was 75%-95%; for men, 63%-90%. Absolute overdiagnosis rates were 13-17 per 100,000 women versus 3-5 per 100,000 men, or roughly 443,000 to 574,000 women and 108,000 to 155,000 men over the study window. Women ages 35-49 had the highest absolute rates (21-24 per 100,000).

Earlier IARC work reached similar conclusions with different methods. Vaccarella and colleagues, in the New England Journal of Medicine in 2016, estimated that 70%-80% of thyroid cancers in U.S. women around 2003-2007 were overdiagnoses, with roughly 90% in Republic of Korea women after intensive ultrasound screening. That amounted to more than 470,000 women and 90,000 men across 12 high-income countries over two decades. A 2020 IARC analysis by Li, Dal Maso, and Vaccarella put the 2008-2012 tally above 1 million people (830,000 women and 220,000 men) across 26 countries. Their 2024 update across 63 countries attributed 75.6% of 2.3 million cases to overdiagnosis in 2013-2017.

Autopsy series explain the reservoir. Furuya-Kanamori and colleagues' 2016 meta-analysis in the Journal of Clinical Oncology pooled 35 studies and 12,834 autopsies: incidental differentiated thyroid cancer prevalence was 11.2% when whole glands were examined and 4.1% with partial examination. Latent prevalence did not rise in lockstep with clinical incidence after 1970. The reservoir was always there; diagnostic technology found it. Model percentages depend on assumptions, but independent teams keep landing in the same range: majority overdiagnosis in women, higher than in men.

Screening intensity and the Korea lesson

South Korea is the natural experiment every textbook cites. In the early 2000s, thyroid ultrasound was widely offered as an inexpensive add-on during national multi-cancer screening. Ahn and colleagues reported in the New England Journal of Medicine in 2014 that the 2011 diagnosis rate was 15 times the 1993 rate, while thyroid-cancer mortality stayed flat. When opportunistic screening later fell, incidence fell too. That is evidence that detection intensity, not a new carcinogen, drove the curve.

The United States never ran a formal national thyroid screening program. It did something quieter: more neck imaging for other reasons, more ultrasound of nonpalpable nodules, and more fine-needle aspiration of small lesions. Francis and colleagues modeled that pathway. Cutting ultrasound referrals for nonpalpable nodules by 33% reduced projected papillary incidence by 17%; a 67% cut reduced incidence by 41%. Mortality changed by less than 0.1% in both scenarios. Stopping those ultrasounds entirely brought modeled 2019 incidence to about 4.5 per 100,000 with essentially no mortality change.

That is a high-harm, low-benefit cascade. The USPSTF formalized the response in May 2017 with a Grade D recommendation against screening asymptomatic adults by neck palpation or ultrasound, concluding with moderate certainty that harms outweigh benefits. It does not apply to people with symptoms (hoarseness, dysphagia, a felt mass) or high-risk groups such as childhood head-and-neck radiation, certain hereditary syndromes, or a first-degree relative with thyroid cancer. Incidentalomas will keep arriving from CT and MRI; the live question is which nodules get biopsied and which cancers get immediate surgery versus active surveillance.

What guidelines changed

For years the default after a papillary cancer diagnosis was total thyroidectomy, often followed by radioactive iodine. That approach fit clinically apparent disease. It fits poorly when most new cases are subcentimeter, low-risk tumors found on imaging.

The 2015 American Thyroid Association (ATA) guidelines for adult patients with thyroid nodules and differentiated thyroid cancer made that mismatch explicit. A stated goal was to minimize overtreatment in low-risk patients while still treating higher-risk disease. The guidelines accepted lobectomy rather than total thyroidectomy for many low-risk cancers, tightened biopsy criteria for small nodules, and recognized active surveillance for selected papillary microcarcinomas, an approach pioneered in Japanese cohorts with low progression under observation, especially in older adults.

The 2025 ATA guidelines continue that shift: refined risk stratification, less extensive initial surgery for many low-risk cases, and more structured active-surveillance discussion. Clinical culture is still catching up. Many patients hear "cancer" and reasonably want it out. Many surgeons trained when total thyroidectomy was near-automatic. Changing the default is slower than changing a PDF.

Choosing Wisely campaigns and the USPSTF Grade D statement both push against asymptomatic screening. The plateau after about 2009-2014, documented by Chen et al. and later SEER years, is consistent with some restraint. It is not a solved problem. Incidence remains near its peak. Francis et al.'s estimates through 2019 still put most women's papillary diagnoses in the overdiagnosed range.

For women with other thyroid conditions, nodule workups often collide with routine endocrine care. Our overview of thyroid disorders in women covers hypothyroidism and hyperthyroidism (and autoimmune thyroid disease). Those are far more common than thyroid cancer, and a frequent reason ultrasound enters the chart. Separating functional disease from cancer hunting helps not feed the cascade.

Treatment harms when disease is indolent

Overdiagnosis becomes overtreatment when a label triggers surgery or radioiodine the person would never have needed. The USPSTF evidence review behind the 2017 recommendation summarized surgical harms that remain relevant. Permanent hypoparathyroidism after total thyroidectomy occurs on the order of 2 to 6 events per 100 operations in better estimates, with wider variation when lymph nodes are dissected. Permanent recurrent laryngeal nerve palsy runs about 1 to 2 per 100 operations. Those are lifelong problems: calcium dependence, voice change, and specialist follow-up.

Radioactive iodine adds salivary damage, dry mouth, and a small excess risk of second primary cancers in some series. Total thyroidectomy commits nearly everyone to daily levothyroxine and lab monitoring through pregnancy and aging. For a woman whose tumor would never have declared itself, that is iatrogenic chronic disease traded for a statistical risk that was already near zero.

Davies and Welch noted that among Americans treated around 2009, more than 90% underwent surgery and about half received radiation therapy, before the full effect of the 2015 ATA shift. Active surveillance uptake has grown but remains uneven. Francis and coauthors' 2026 modeling is useful: fewer ultrasounds of nonpalpable nodules change incidence a lot and mortality almost not at all. Upstream, fewer unnecessary scans. Downstream, lobectomy or observation when guidelines support it.

None of this means thyroid cancer is trivial. Anaplastic disease is catastrophic. Distant-stage differentiated cancer still kills; SEER's five-year relative survival for distant disease is about 48%, not 99%. The error is treating every 6 mm papillary microcarcinoma as if it were that disease.

If you are sorting symptoms that might be hormonal rather than malignant, tools such as our period calculator and cycle length calculator can help track patterns while you work with a clinician. They do not diagnose thyroid cancer and should not replace evaluation of a real neck mass or progressive compressive symptoms.

The 42-year-old with the 7 mm incidentaloma eventually chose active surveillance after a second opinion. Two years later the nodule was unchanged. She still flinches at the word cancer. That reaction is human. The data's job is to keep the system from forcing every woman in her position into the most aggressive path by default.

Frequently asked questions about thyroid cancer overdiagnosis

What is thyroid cancer overdiagnosis?

Overdiagnosis means detecting a cancer that would not have caused symptoms or death during a person's lifetime. In thyroid cancer, most excess cases are small papillary tumors found on imaging. Large incidence increases without matching mortality rises are the epidemiologic signature of overdiagnosis, not a true epidemic of lethal disease.

How much more common is thyroid cancer in women than men?

Per SEER data for 2019-2023, age-adjusted incidence is 21.2 per 100,000 women and 6.1 per 100,000 men, about 3.5 to 1. Women account for roughly three-quarters of U.S. diagnoses in older SEER series. Death rates remain low for both sexes (about 0.5 per 100,000 overall), so the sex gap is mostly diagnosis volume, not mortality.

What share of cases in women may be overdiagnosed?

A 2026 JAMA Network Open microsimulation estimated that 75%-95% of papillary thyroid cancers diagnosed in U.S. women from 1991 to 2019 were overdiagnoses. Earlier IARC analyses put the U.S. female fraction around 70%-80% for the mid-2000s. Exact percentages depend on method; independent teams agree most women's diagnoses in high-resource settings are low-risk detections.

Should asymptomatic women be screened for thyroid cancer?

No, not as a population strategy. The USPSTF gave thyroid cancer screening in asymptomatic adults a Grade D recommendation in 2017, concluding harms outweigh benefits. That advice does not apply to a felt neck mass, progressive voice or swallowing problems, or high-risk histories such as childhood neck radiation or certain genetic syndromes.

Why did South Korea's thyroid cancer rate explode?

Opportunistic ultrasound screening during a national multi-cancer program sharply increased detection of small papillary cancers. Ahn and colleagues reported that the 2011 diagnosis rate was 15 times the 1993 rate while mortality stayed stable. When screening intensity later fell, incidence fell too. That is evidence that detection practices, not a new environmental killer, drove the curve.

Have treatment guidelines changed because of overdiagnosis?

Yes. The 2015 ATA guidelines and the 2025 differentiated thyroid cancer update emphasize less extensive surgery for many low-risk cancers, more selective radioactive iodine, and active surveillance for selected papillary microcarcinomas. The USPSTF Grade D statement is the prevention-side counterpart: do not screen asymptomatic necks without a high-risk indication.