Korea diagnoses thyroid cancer too early and prostate cancer too late. Half of prostate cases are already high-risk when found

South Korea is the world's reference case for finding cancer too early. Its thyroid cancer incidence rose tenfold on the back of ultrasound scans that people were not sent for, while the number of deaths did not move.

Prostate cancer in the same country, in the same decade, looks like the opposite problem.

What the national picture looks like#

Korea's insurance claims database records that a cancer was diagnosed but not how advanced it was, so risk distribution has not been available nationally. A team at Ewha Womans University assembled it directly, collecting the clinicopathological variables — PSA, clinical stage, Gleason grade, biopsy core involvement — for every biopsy-confirmed case at 51 general hospitals in six sample years between 2010 and 2020. The 27,075 cases represent 37.9 percent of the 71,403 registered nationally in those years.

The direction of travel is the same. Low-risk disease, including potentially low-risk cases, fell from 11.4 percent of diagnoses in 2010 to 7.6 percent in 2020. Korea is finding proportionally fewer of the prostate cancers that are easiest to treat, not more.

Where a man lives changes what he is diagnosed with#

High-risk disease was the most prevalent category in nine of Korea's seventeen administrative regions. It accounted for 55.4 percent of cases among rural residents and 47.7 percent among urban residents, with low-risk disease correspondingly more common in the capital metropolitan area.

The authors read that gap as a disparity in access to early detection and diagnosis rather than as a difference in the disease. It is the same shape this publication found in thyroid cancer, where higher income predicted more screening, more diagnoses and — for a cancer that mostly did not need finding — more harm. Here the gradient runs the other way: the people with less access are the ones presenting with the disease that needed finding early.

How many men could safely be watched#

The same cohort was used to estimate an upper bound on active surveillance — monitoring a confirmed cancer rather than treating it immediately. Defining eligibility as generously as the evidence allows, by including favourable intermediate-risk disease alongside low and potentially low risk, 6,586 of the 27,075 men were eligible: 24.3 percent.

That share fell over the decade, from 27.6 percent in 2010 to 22.0 percent in 2020, which follows arithmetically from the shrinking proportion of low-risk diagnoses. Eligibility also varied significantly by administrative district and by whether a man lived in a city.

The number is an upper bound and the authors describe it as one. It says how many men could be offered surveillance, not how many should be, and nothing in the study records how many actually were.

What this dataset cannot tell us#

1. These are the 37.9 percent of national cases treated at 51 general hospitals with 200 or more beds. Men diagnosed elsewhere, or not diagnosed at all, are not represented, and the direction of that bias is not established.
2. Six sample years between 2010 and 2020 describe a decade in snapshots. A change that occurred between sampled years is invisible here.
3. A rising share of high-risk disease is a proportion, not a count. It is consistent with more high-risk cancer, with less low-risk cancer being found, or with both, and this data cannot separate them.
4. Two papers from the same team on the same 27,075 cases report the high-risk share as 53.94 percent and as approximately 50 percent, under classifications that differ in their details. Both are cited here and neither figure should be quoted as the single national number.
5. Risk category at diagnosis is not an outcome. Nothing here follows these men to find out what happened to them.