Thirty-six thousand American men will die of prostate cancer this year. Not because the disease is untreatable. Not because it strikes without warning, or progresses too fast to catch, or hides from every diagnostic tool medicine has developed. They will die because, for more than a decade, the medical establishment told their doctors not to look.
That is the blunt summary of where prostate cancer screening stands in 2026. The longer version involves a flawed trial, a federal task force that built policy on contaminated data, and a cascade of consequences now visible in the cancer registries of two countries. It also involves a solution — several of them, in fact, each more accessible than the last. The tragedy is not biological. It is administrative.
The Second Killer
Prostate cancer is the second leading cause of cancer death in American men, behind only lung cancer. Roughly one in eight men will be diagnosed with it during their lifetime — a 12.8% lifetime incidence that makes it among the most common malignancies in the developed world (Kratzer et al., CA Cancer J Clin, 2025). Most of those men will survive it. The fifteen-year survival rate from first diagnosis approaches 97%, a figure that reflects both the availability of effective treatments and, historically, the benefit of catching the disease before it has spread.
The operative word is “before.” Prostate cancer diagnosed at stages I, II, or III carries a five-year survival rate above 99%. Diagnosed at stage IV — once the cancer has metastasized, spreading beyond the prostate into lymph nodes, bone, or distant organs — that figure collapses to 38%. Median survival at stage IV is thirty months (Schoen et al., JAMA Netw Open, 2024). The disease does not become a different disease at that point. It becomes an incurable one.
This distinction between early and late diagnosis is not incidental to the prostate cancer story. It is the entire story. Prostate cancer, unlike pancreatic cancer or many forms of breast cancer, tends to progress in a relatively predictable, stepwise fashion — from localized, non-aggressive disease through increasingly malignant stages before achieving metastatic potential. That biological characteristic makes it, in principle, one of the most screenable cancers in existence. A cancer that announces itself gradually, through a measurable biomarker, years before it becomes lethal, is a cancer that medicine should be catching early in virtually every case.
It is not.
What PSA Is, and What It Isn’t
Prostate-specific antigen is a protein the prostate gland produces to maintain the fluidity of seminal fluid. A small amount leaks through the prostate’s vasculature into the bloodstream, where it can be measured with a standard blood draw. As the prostate enlarges — which it does in most men as they age — PSA levels in the blood rise proportionally. When the prostate’s architecture is disrupted by malignant growth, or when cancerous tissue creates abnormally permeable blood vessels, PSA leaks into the bloodstream at an accelerated rate. The test that measures it costs a few dollars and is covered by most insurance plans without a copay.
The original clinical logic was direct: elevated PSA prompted an ultrasound; an enlarged or suspicious prostate on ultrasound prompted a biopsy; a biopsy showing cancer prompted treatment. Trials conducted over fourteen and sixteen years found that routine PSA screening reduced prostate cancer mortality by 44% and 64%, respectively, preventing as many as 3.1 cases of metastatic disease for every 1,000 men screened (Hugosson et al., Lancet Oncol, 2010; Alpert, Urology, 2018).
The problem was not the test. The problem was what happened after a positive result.
The most common biopsy method threaded a needle through the rectal wall to sample prostate tissue — a transrectal approach that punctures through a bacteria-rich environment into what is otherwise a sterile organ. Prophylactic antibiotics reduced the risk, but between 5% and 7% of these biopsies still resulted in infection, sometimes serious enough to require hospitalization (Liss et al., J Urol, 2017). Beyond infection, the pipeline from elevated PSA to biopsy to treatment moved too quickly and too automatically. Many of the cancers detected in the early screening era were indolent — slow-growing, low-grade tumors that posed no immediate threat and, in retrospect, required no treatment at all. Men were being subjected to surgery, radiation, and hormone therapy, with their attendant risks of erectile dysfunction, urinary incontinence, and bowel damage, for cancers that would never have harmed them.
These were real costs. The United States Preventive Services Task Force, the federal body responsible for issuing evidence-based screening recommendations, took them seriously. In 2008, the USPSTF ended its recommendation for PSA testing in men over seventy. In 2012, it recommended against screening for men of all ages (USPSTF, 2012).
The reasoning was utilitarian and, on its face, defensible: if screening saves roughly one life per thousand men tested while causing significant harm in a meaningful proportion of others, the math doesn’t favor screening. The USPSTF’s primary evidence for that calculus came from a large American trial called the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial — the PLCO.
The Trial That Wasn’t
The PLCO randomized participants into two groups: one assigned to regular PSA screening, one assigned to no screening. The study found virtually no difference in prostate cancer mortality between the two groups, which the USPSTF interpreted as evidence that PSA screening offered minimal survival benefit.
What the PLCO authors did not adequately account for — and what the USPSTF did not adequately scrutinize — was that both groups underwent PSA screening.
Not metaphorically. Literally. Annual surveys of participants in the “no screening” control arm found that between 39% and 64% of them reported having received a PSA test within the past year, every single year of the study. Between 60% and 84% reported a test within the past three years, again annually, throughout the trial’s duration. By the study’s final year of survey responses, 90% of men in the supposedly unscreened control group reported having received at least one PSA test in their lifetime. In some years, the surveys showed more men in the control arm reporting recent PSA testing than men in the intervention arm (Shoag et al., N Engl J Med, 2016).
A randomized controlled trial comparing screened men to unscreened men, in which the unscreened men were being screened, cannot produce a meaningful comparison. The finding of “no difference” between groups was not evidence that screening doesn’t work. It was evidence that the study had no functional control group.
This was not a subtle methodological footnote. The contamination data were published in a letter to the New England Journal of Medicine in 2016 — four years after the USPSTF’s 2012 recommendation and eight years after its 2008 recommendation. When researchers subsequently reanalyzed the PLCO data using statistical methods that accounted for the actual screening rates in both arms, the same dataset that had been used to argue against PSA testing showed a 27–32% reduction in prostate cancer mortality attributable to regular screening (Tsodikov et al., Ann Intern Med, 2017).
The USPSTF issued its most recent recommendation in 2018, still declining to directly recommend PSA testing for men aged 55–69 and maintaining its recommendation against testing for men over 70. The 2017 reanalysis appears to have moved the needle modestly — the 2018 guidance softened from a flat recommendation against screening to a “C” grade suggesting the decision be individualized — but the underlying posture remained cautious in ways the evidence no longer supported.
What the Registries Show
The downstream consequences of a decade-plus of screening skepticism are now documented in the cancer registries of both the United States and Canada.
In the United States, since the USPSTF’s recommendations shifted in 2012, there has been a 0.1% annual decrease in stage II prostate cancer diagnoses. That decline has been accompanied by a 3.3% annual increase in stage III diagnoses and a 6.0% annual increase in stage IV — increases that substantially outpace the 0.8% annual growth in total prostate cancer cases detected (Weiner et al., Prostate Cancer Prostatic Dis, 2016). The pattern is not one of rising cancer incidence. It is one of cancers being found later.
Canada tells the same story. A 2025 analysis found a 3.7% year-over-year increase in stage IV diagnoses in men under 75 from 2010 to 2017, even as total cancer detections in that age range declined (Wilkinson et al., Curr Oncol, 2025). Fewer cancers found, but more of them fatal. The authors noted the grim arithmetic directly: new treatments have extended survival for stage IV disease, but overall survival has decreased from the screening era, because the proportion of cases that are incurable at diagnosis has grown.
The USPSTF’s original projection was that its recommendations would affect no more than 0.1% of prostate cancer outcomes. The actual figures — 3.3% and 6.0% annual increases in late-stage diagnoses — represent a failure of prediction by an order of magnitude.
A More Precise Tool
The USPSTF built its guidelines on the assumption that PSA testing meant what it meant in the 1990s: a single value, taken periodically, triggering an automatic cascade toward biopsy. That assumption is now obsolete.
PSA levels are noisy. A single reading can vary by as much as 15% from day to day in the same individual, and can jump by as much as 40% acutely following ejaculation (Ornstein et al., J Urol, 1997; Tchetgen et al., Urology, 1996). There is also substantial variation between healthy men: the median PSA for a man in his sixties without prostate cancer is between 1.0 and 1.2 ng/mL, but the 95th percentile for the same age group reaches 4.9 ng/mL — still in the absence of cancer (Kalish & McKinlay, Urology, 1999). A single elevated reading, interpreted against population norms rather than an individual’s own history, will generate false alarms. That was the legitimate problem with early PSA screening.
The solution is not to abandon the test. It is to use it longitudinally.
PSA velocity — the rate of change in PSA levels over time — filters out the noise that makes single readings unreliable. The prostate does enlarge gradually throughout life, and PSA levels rise with it. But rapid increases, particularly those that exceed 0.35 ng/mL per year over an eighteen-month period in men with baseline PSA below 4.0 ng/mL, flag a pathological process that gradual aging does not explain. For men with baseline PSA above 4.0, the threshold rises to 0.75 ng/mL per year. The metric is calibrated to the individual rather than the population — a patient whose PSA has been stable at 1.3 ng/mL for years is not the same clinical picture as a patient whose PSA has risen from 1.3 to 1.7 to 1.9 over eighteen months, even though both values fall within population norms.
This is not a sophisticated or expensive refinement. It requires only that the test be run regularly and that the results be tracked over time. The test itself remains a standard blood draw.
The MRI Intermediary
When PSA velocity or density raises a flag, the next step is no longer automatically a biopsy. MRI has become an increasingly standard intermediary — a way to assess the prostate’s architecture and identify suspicious lesions before any tissue is sampled.
PSA density, calculated by dividing the PSA concentration in the blood by the prostate’s volume as measured on MRI, adds another layer of precision. PSA levels rise as the prostate grows, but they should rise proportionally to that growth. When PSA is increasing faster than prostate volume, it suggests that PSA is leaking into the bloodstream at an elevated rate — a sign of disrupted prostate architecture or abnormally permeable vasculature, both of which are hallmarks of malignant growth. PSA density can therefore indicate cancer risk before a biopsy is ever considered.
The practical barrier to MRI as a routine pre-biopsy tool has historically been cost and complexity. The gold-standard prostate MRI — multiparametric imaging using T2-weighted, diffusion-weighted, and dynamic contrast-enhanced sequences — requires gadolinium contrast injection, physician monitoring during the scan, and thirty to forty minutes of scanner time. These requirements limit throughput and add cost.
A 2025 trial published in JAMA addressed this directly. The PRIME trial compared multiparametric MRI to a simpler, contrast-free biparametric MRI — using only T2-weighted and diffusion-weighted sequences — in 490 patients flagged by PSA screening. The biparametric method detected clinically significant prostate cancer in 143 patients; the gold-standard multiparametric method detected it in 145. The difference of 0.4 percentage points fell well within the trial’s noninferiority margin (95% CI: −1.2 to 0.4) (Ng et al., JAMA, 2025). The contrast-free scan takes fifteen to twenty minutes rather than thirty to forty, requires no injection, and eliminates the need for physician monitoring throughout the procedure.
The clinical implication is straightforward: a meaningful barrier to pre-biopsy MRI has been removed. More patients can be assessed with less resource expenditure and no reduction in diagnostic accuracy.
Cleaning Up the Biopsy
If MRI identifies a lesion warranting tissue sampling, the biopsy itself has also changed substantially.
The transrectal approach — needle through the rectal wall — remains the most commonly performed technique in the United States, used by 99.4% of urologists in a recent census (Rezaee et al., Urology, 2026). Its limitations are structural: it introduces bacteria from the digestive tract into the prostate, carries a 5–7% infection rate even with prophylactic antibiotics, and biases sampling toward the posterior prostate, potentially missing tumors at the gland’s base and anterior surface.
The transperineal approach enters through the perineum — the skin between the anus and scrotum — circumventing the rectum entirely. A multicenter phase 3 trial with over 1,700 patients, the PREVENT trial, compared the two methods directly. Transperineal biopsy increased detection of high-grade cancers while decreasing detection of low-grade ones, suggesting greater sensitivity for the tumors that actually matter clinically. More strikingly, not a single patient in the transperineal arm of the trial developed a post-biopsy infection — a zero-infection rate achieved without prophylactic antibiotics (Hu et al., JAMA Oncol, 2024).
Only 37.2% of American urologists currently perform transperineal biopsies. The gap between what the evidence supports and what is practiced reflects the usual lag between trial publication and clinical adoption, not any remaining scientific uncertainty about which approach is superior.
Finding Cancer Without Treating It
The USPSTF’s concern about overtreatment was not invented. It was real, and it applied to the era in which the guidelines were written. In the 1990s and early 2000s, a cancer detected was a cancer treated — surgery, radiation, or hormone therapy, regardless of the tumor’s grade or the patient’s age.
That automatic pipeline no longer exists.
Tissue from a biopsy is now graded on a five-point scale called the Grade Group system, which classifies tumors by how abnormal their cells appear relative to healthy prostate tissue. Grade Groups 3, 4, and 5 indicate increasing proportions of increasingly malignant cells and typically prompt immediate treatment. Grade Groups 1 and 2 — low-grade, slow-growing tumors — are managed through active surveillance: enhanced PSA testing frequency, additional biomarker assays such as the PHI or 4Kscore, symptom monitoring, and follow-up imaging to track growth. Treatment is initiated only if and when the surveillance data indicate that the tumor is progressing toward higher-grade disease.
Active surveillance is not watchful waiting in the passive sense. It is a structured monitoring protocol that preserves the option of treatment while deferring its costs until they are genuinely necessary. For many men with low-grade disease, that deferral becomes indefinite. The burden of overtreatment that concerned the USPSTF has been substantially addressed — not by screening less, but by treating more selectively.
The Finasteride Blind Spot
There is one complication in the PSA story that has become significantly more urgent in recent years, and it is not adequately understood by most of the physicians ordering the test.
Finasteride — sold under the brand names Propecia and Proscar, and now widely available through online prescribing platforms — is a 5-alpha reductase inhibitor (5-ARI) used to treat male pattern hair loss and benign prostatic hyperplasia. Its mechanism involves blocking the conversion of testosterone to dihydrotestosterone (DHT), the more potent androgen that drives both hair follicle miniaturization and prostate growth. In the prostate, DHT suppression also dramatically reduces PSA production.
The suppression is not subtle. A man who has been taking finasteride for six months can expect his PSA levels to be reduced by roughly one-third. After twelve months, PSA is typically cut in half. That suppressed value is maintained for as long as the medication is continued. A PSA reading taken from a man who has been on finasteride for a year or more does not mean what the same number would mean in a man not taking the drug. It needs to be corrected upward — at minimum doubled — before it can be interpreted clinically.
The correction multipliers are duration-dependent. Data from the Prostate Cancer Prevention Trial and a 2005 follow-up analysis by Ruth Etzioni and colleagues suggest multiplying raw PSA by 2.0 for the first two years of finasteride use, by 2.3 for years two through seven, and by 2.5 for seven years and beyond (Thompson et al., N Engl J Med, 2003; Etzioni et al., J Urol, 2005). The FDA label recommends a flat 2x correction, which is conservative for longer-term users.
PSA velocity under finasteride requires a different interpretive framework entirely. After twelve months on the drug, PSA values should not be rising — the cumulative suppression of DHT should nearly negate the effect of normal prostate growth on PSA output. In a study of 19,000 older men, those taking finasteride who had a biopsy negative for cancer at the end of a seven-year study period showed an average 2% annual decrease in PSA levels. Any sustained increase in PSA velocity in a man on long-term finasteride is therefore a more significant red flag than the same velocity increase would be in an untreated man.
The clinical consequences of ignoring this are documented. A retrospective cohort study of Veterans Affairs patients undergoing PSA screening found that men taking finasteride were diagnosed with metastatic prostate cancer more than twice as often as men not taking a 5-ARI — 6.7% versus 2.9% (Sarkar et al., JAMA Intern Med, 2019). The raw PSA values at cancer diagnosis were similar between the two groups: 6.8 ng/mL in the finasteride group versus 6.4 ng/mL in the non-finasteride group. When the authors applied the appropriate correction multipliers to the finasteride group’s values, the adjusted figure was 13.5 ng/mL — more than double the level that should have triggered urgent clinical attention. The physicians ordering those PSA tests appear to have been unaware that the values required correction. The cancers grew to metastatic stage while appearing, on paper, to be within a range that warranted only routine monitoring.
Finasteride prescriptions have increased sharply in recent years, driven largely by its adoption for hair loss in younger men. Many of those prescriptions come through online platforms that operate independently of a patient’s primary care physician. The result is a growing population of men whose PSA values are being systematically misread by the doctors ordering their blood work — not because those doctors are negligent, but because they don’t know their patient is taking the drug.
The Arithmetic of Inaction
The USPSTF’s 2012 recommendation projected that its guidance would affect no more than 0.1% of prostate cancer outcomes. The actual data — a 6.0% annual increase in stage IV diagnoses in the United States since 2012, a 3.7% annual increase in Canada, and an overall survival that has declined from the screening era despite improved treatments for metastatic disease — suggest the projection was wrong by a significant margin.
The tools that existed when those guidelines were written have been substantially refined. PSA velocity and density have replaced the single-value threshold as the primary interpretive framework. Biparametric MRI has removed the cost and logistical barriers that made pre-biopsy imaging impractical at scale. Transperineal biopsy has reduced infection rates to near zero while improving detection of high-grade disease. Active surveillance has replaced the automatic treatment pipeline that made overdiagnosis genuinely harmful.
None of these advances require new infrastructure. PSA testing is a blood draw. MRI scanners exist in every regional medical center. The Grade Group system is standard pathology. The entire pipeline from screening to informed treatment decision can be executed within the existing healthcare system, at costs that are, by any measure, modest relative to the cost of treating metastatic disease.
What it requires is that physicians order the test, track the results longitudinally, understand how to adjust for medications that suppress PSA, and refer patients for imaging before defaulting to biopsy. And it requires that patients — particularly men over forty, men with a family history of prostate cancer, and men currently taking finasteride for any indication — understand that the test exists, that it is cheap, and that the guidelines discouraging it were built on a study that did not have a functional control group.
Thirty-six thousand deaths this year. The biology is not the obstacle.
Sources
- Kratzer TB, Mazzitelli N, Star J, Dahut WL, Jemal A, Siegel RL. Prostate cancer statistics, 2025. CA Cancer J Clin. 2025;75(6):485–497. https://doi.org/10.3322/caac.70028
- Wilkinson AN, Ellison LF, Zhang SX, et al. Canadian prostate cancer trends in the context of PSA screening guideline changes. Curr Oncol. 2025;32(12):669. https://doi.org/10.3390/curroncol32120669
- Hugosson J, Carlsson S, Aus G, et al. Mortality results from the Göteborg randomised population-based prostate-cancer screening trial. Lancet Oncol. 2010;11(8):725–732. https://doi.org/10.1016/S1470-2045(10)70146-7
- Alpert PF. New evidence for the benefit of prostate-specific antigen screening: data from 400,887 Kaiser Permanente patients. Urology. 2018;118:119–126. https://doi.org/10.1016/j.urology.2018.02.049
- Schröder FH, Hugosson J, Carlsson S, et al. Screening for prostate cancer decreases the risk of developing metastatic disease: findings from the European Randomized Study of Screening for Prostate Cancer (ERSPC). Eur Urol. 2012;62(5):745–752. https://doi.org/10.1016/j.eururo.2012.05.068
- Liss MA, Ehdaie B, Loeb S, et al. An update of the American Urological Association white paper on the prevention and treatment of the more common complications related to prostate biopsy. J Urol. 2017;198(2):329–334. https://doi.org/10.1016/j.juro.2017.01.103
- U.S. Preventive Services Task Force. Prostate Cancer: Screening — 2012 Recommendation. Archived. https://www.uspreventiveservicestaskforce.org/uspstf/document/how-did-the-uspstf-arrive-at-this-recommendation-/prostate-cancer-screening-2012
- Schoen MW, Montgomery RB, Owens L, Khan S, Sanfilippo KM, Etzioni RB. Survival in patients with de novo metastatic prostate cancer. JAMA Netw Open. 2024;7(3):e241970. https://doi.org/10.1001/jamanetworkopen.2024.1970
- Weiner AB, Matulewicz RS, Eggener SE, Schaeffer EM. Increasing incidence of metastatic prostate cancer in the United States (2004–2013). Prostate Cancer Prostatic Dis. 2016;19(4):395–397. https://doi.org/10.1038/pcan.2016.30
- Ornstein DK, Smith DS, Rao GS, Basler JW, Ratliff TL, Catalona WJ. Biological variation of total, free and percent free serum prostate specific antigen levels in screening volunteers. J Urol. 1997;157(6):2179–2182. https://doi.org/10.1016/s0022-5347(01)64708-6
- Tchetgen MB, Song JT, Strawderman M, Jacobsen SJ, Oesterling JE. Ejaculation increases the serum prostate-specific antigen concentration. Urology. 1996;47(4):511–516. https://doi.org/10.1016/s0090-4295(99)80486-5
- Kalish LA, McKinlay JB. Serum prostate-specific antigen levels (PSA) in men without clinical evidence of prostate cancer: age-specific reference ranges for total PSA, free PSA, and percent free PSA. Urology. 1999;54(6):1022–1027. https://doi.org/10.1016/s0090-4295(99)00349-0
- Ng ABCD, Asif A, Agarwal R, et al. Biparametric vs multiparametric MRI for prostate cancer diagnosis: the PRIME diagnostic clinical trial. JAMA. 2025;334(13):1170–1179. https://doi.org/10.1001/jama.2025.13722
- Hu JC, Assel M, Allaf ME, et al. Transperineal vs transrectal prostate biopsy — the PREVENT randomized clinical trial. JAMA Oncol. 2024;10(11):1590. https://doi.org/10.1001/jamaoncol.2024.4000
- Rezaee ME, Meeks W, Galen E, Helsel A, Pavlovich CP. Examining variation in prostate biopsy practices in the United States: an analysis of American Urological Association census data. Urology. 2026;208:9–13. https://doi.org/10.1016/j.urology.2025.10.039
- Shoag JE, Mittal S, Hu JC. Reevaluating PSA testing rates in the PLCO trial. N Engl J Med. 2016;374(18):1795–1796. https://doi.org/10.1056/NEJMc1515131
- Tsodikov A, Gulati R, Heijnsdijk EAM, et al. Reconciling the effects of screening on prostate cancer mortality in the ERSPC and PLCO trials. Ann Intern Med. 2017;167(7):449–455. https://doi.org/10.7326/M16-2586
- Thompson IM, Goodman PJ, Tangen CM, et al. The influence of finasteride on the development of prostate cancer. N Engl J Med. 2003;349(3):215–224. https://doi.org/10.1056/NEJMoa030660
- Etzioni RD, Howlader N, Shaw PA, et al. Long-term effects of finasteride on prostate specific antigen levels: results from the Prostate Cancer Prevention Trial. J Urol. 2005;174(3):877–881. https://doi.org/10.1097/01.ju.0000169255.64518.fb
- Sarkar RR, Parsons JK, Bryant AK, et al. Association of treatment with 5α-reductase inhibitors with time to diagnosis and mortality in prostate cancer. JAMA Intern Med. 2019;179(6):812–819. https://doi.org/10.1001/jamainternmed.2019.0280
