Between 72% and 94% of papillary thyroid cancers diagnosed in the United States from 1991 through 2019 were overdiagnosed, meaning the tumors would never have caused symptoms or death if left undetected, according to a modeling study led by David Francis, MD, of the University of Wisconsin School of Medicine and Public Health. The study, published in JAMA Network Open, estimated that the pattern affected between 443,000 and 574,000 women and between 108,000 and 155,000 men over that 28-year span. Papillary thyroid cancer is the most common form of thyroid cancer, and its diagnosis rate in the country has climbed for decades even as death rates from the disease have stayed essentially flat.
The gap between rising diagnoses and flat mortality is the pattern researchers use to identify overdiagnosis in the first place: a cancer is being found more often not because it has become more common or more dangerous, but because better tools are finding versions of it that were always present and were never going to progress. Francis and colleagues built a computer model to estimate how much of the thirty-year rise in papillary thyroid cancer diagnoses fits that description rather than reflecting a genuine increase in disease.
A microsimulation model reconstructs three decades of diagnoses
The team’s tool, called the Papillary Thyroid Carcinoma Microsimulation Model, or PATCAM, is a validated simulation built to reconstruct incidence, stage at diagnosis, and mortality using data from the National Cancer Institute’s SEER registries. Coverage of the findings describes the model running population-level scenarios to estimate how many diagnosed cancers would have stayed silent for a patient’s lifetime absent detection, a figure no single patient’s case can answer on its own.
Applying PATCAM to the 1991-through-2019 window let Francis’s team separate diagnoses driven by an actual rise in disease from those driven by more sensitive detection. The 72%-to-94% range reflects the model’s uncertainty bounds rather than a single point estimate, and the researchers presented it as a range specifically because no simulation can pin the true rate to one number with certainty. Building the model required calibrating it against known U.S. incidence and mortality data first, a step meant to confirm PATCAM reproduces the historical record accurately before using it to estimate a quantity, overdiagnosis, that cannot be observed directly in any registry.
Ultrasound, not new cancer, drove the increase
The overdiagnosed cases identified by the model trace almost entirely to neck ultrasound, a technology that became far more common in general medical practice over the study period and that regularly finds small nodules a physical exam alone would never detect. Many of those nodules are picked up incidentally, during imaging ordered for an unrelated complaint such as a swallowing issue or a routine physical, rather than because a physician suspected thyroid cancer specifically.
Once a nodule appears on an ultrasound, standard practice has generally pushed toward biopsy and, frequently, surgical removal of part or all of the thyroid, even for tumors small and slow-growing enough that they would never have been found, or caused a symptom, without the scan. That treatment pathway is what turns an incidental finding into a formal cancer diagnosis and, often, a thyroidectomy. Some hospital systems have begun offering active surveillance instead of immediate surgery for the smallest, lowest-risk nodules, monitoring them with repeat imaging rather than operating right away, though that approach remains far less common in routine practice than immediate biopsy and surgical referral.
The overdiagnosis burden fell hardest on women
Women accounted for the large majority of the overdiagnosed cases in the model, with an estimated 443,000 to 574,000 affected compared with 108,000 to 155,000 men over the same 28 years. The imbalance mirrors papillary thyroid cancer’s overall diagnosis pattern, in which women are diagnosed roughly three times as often as men, a gap researchers have long attributed partly to women undergoing neck and chest imaging more frequently for unrelated conditions.
Each of those diagnoses carries consequences beyond the number itself. A thyroidectomy typically requires lifelong thyroid hormone replacement, carries a risk of vocal cord or parathyroid injury during surgery, and leaves a permanent scar, costs that fall on hundreds of thousands of patients whose tumors, the model estimates, would never have caused a symptom.
Wisconsin researchers point to fewer, more selective scans
Francis’s team frames the findings as an argument for narrowing when clinicians order thyroid ultrasound and how aggressively they pursue small, nonpalpable nodules found incidentally on scans ordered for other reasons. The paper points to more selective referral criteria, rather than a wholesale retreat from screening, as the more realistic lever available to reduce overdiagnosis without missing the smaller share of cases that do progress.
The paper’s authors also note that thyroid cancer overall carries one of the lowest mortality rates among common cancers, a fact that makes the overdiagnosis question different from a similar debate in lung or breast cancer screening, where missing a genuine case carries a far higher cost. The low mortality baseline is part of why Francis’s team argues that fewer scans, aimed more narrowly, would change how many people carry a cancer diagnosis without changing how many die of the disease.
None of the paper’s findings suggest patients diagnosed with papillary thyroid cancer should ignore a physician’s recommendation or forgo treatment already scheduled. The model estimates a population-level pattern across three decades of diagnoses, not which specific patient’s tumor would have stayed silent, and Francis’s team frames the results as guidance for how clinicians order and interpret scans going forward rather than as a reason for any individual to second-guess a diagnosis already made.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
