Deep dive #4 · 3 November 2026

Prostate cancer

For health professionals Movember

One Australian man in five by age 85, and at autopsy over half of men past 79. A cancer that kills about 420,000 men a year and also, in a large share of cases, would never have been noticed. Grow a moustache by all means; the PSA argument needs better evidence. The numbers, the new Australian guideline and the frontier.

At a glance

Burden: about 1.55 million new cases and 420,000 deaths a year worldwide (GLOBOCAN 2024). The Lancet Commission projects 2.9 million cases and almost 700,000 deaths by 2040.
Australia: the most diagnosed cancer. Estimated 28,868 cases and 3,975 deaths in 2025; 5-year relative survival 96%; 1 in 5 men diagnosed by 85, 1 in 55 die of it.
Screening works, modestly: ERSPC at 23 years, rate ratio 0.87. Invite 456 and diagnose 12 extra men to prevent one death. A single PSA invitation (CAP): 0.69% vs 0.78% at 15 years.
New guideline: NHMRC approved it in May 2026. Two-yearly PSA after shared decision-making at 50–69 (repeat if ≥3.0 µg/L), from 45 if higher risk, and no routine DRE.
Low-risk disease rarely kills: 15-year prostate cancer death was 2.7% across all arms of ProtecT. An estimated 41% of Australian diagnoses are overdiagnosis.
Metastatic disease has moved fast: intensified first-line therapy, PARP inhibitors, Lu-PSMA and now AKT inhibition, but overall survival data for some new approvals are immature.

Biggest take-home: the question is no longer only "PSA or not". It is who to test, how to avoid the biopsy, and who not to treat.

Definition

What it is: a malignancy of the prostate, usually acinar adenocarcinoma. Neuroendocrine disease is a rare, aggressive variant, seen mostly after hormone therapy.
Grading (ISUP grade groups, 2014 consensus): GG1 Gleason 3+3; GG2 3+4; GG3 4+3; GG4 4+4; GG5 4+5, 5+4 or 5+5.
Clinically significant: GG2 or higher in the 2026 Australian guideline.
Androgen states: castration sensitive or resistant. FDA's 2026 label calls the first "androgen pathway modulation-naïve or -sensitive".
A raised PSA is not a diagnosis: around 8 in 10 men with a raised PSA don't have prostate cancer (UK National Screening Committee material via Cancer Research UK).

History

1853, J. Adams (The London Hospital): reported a case in The Lancet as "a very rare disease".
1904, Hugh Hampton Young (Johns Hopkins): first radical perineal prostatectomy.
1941, Charles Huggins and Clarence Hodges: castration or oestrogen shrank metastatic disease. Four of the first 21 men lived more than 12 years. Huggins, born in Nova Scotia, shared the 1966 Nobel Prize.
1966, Donald Gleason (Minneapolis VA): five glandular patterns, a score of 2–10, still the basis of grading today.
1966–1979, PSA, and a credit fight: Japanese forensic scientist Mitsuwo Hara described "gamma-seminoprotein" (1966), Richard Ablin reported prostate antigens (1970), and Ming Wang and T. Ming Chu at Roswell Park purified PSA (1979). Catalona's account disputes the earlier claims.
1971–1977, Andrew Schally: worked out the structure of LHRH; agonists then cut testosterone by 75% in advanced disease.
1974, Julianne Imperato-McGinley: studied children in Salinas, Dominican Republic, with 5α-reductase deficiency. Affected men keep a small prostate. Merck's Roy Vagelos read her work, and the line led to finasteride and dutasteride.
1982, Patrick Walsh and Pieter Donker: mapped the cavernous nerves. Walsh did the first nerve-sparing prostatectomy on 26 April 1982.
2020, proPSMA: a trial across ten Australian hospitals put PSMA PET ahead of CT and bone scan for staging.

Pathophysiology

Androgen dependence: normal prostate cells shrink without androgen; hormone-dependent cancer cells die (Huggins's 1966 distinction). That logic still drives first-line therapy.
Inheritance: heritability 58% (Nordic twins, 48,734 pairs). Germline DNA-repair mutations in 11.8% of metastatic disease vs 4.6% of localised disease (Pritchard, NEJM 2016; nearly all white participants).
Genes: BRCA2 2–6-fold risk; Lynch syndrome 2–5-fold. HOXB13 G84E gives 33–60% risk by age 80, vs 12% in the general population.
Fusions: TMPRSS2-ERG is commoner in men ≤50 than ≥60 (Chalmers 2021, 10,189 patients).
Castration resistance: tumours reactivate androgen signalling or leave it behind. Treatment targets now exploit homologous recombination repair (HRR) defects, PTEN loss and EZH2-driven plasticity.
Germline mutations are rare in men on surveillance: ATM, BRCA1 or BRCA2 carriers were just 1.8% of 437 men on active surveillance.

Epidemiology

Global

GLOBOCAN 2024: 1,546,112 new cases (4th of all cancers), 419,849 deaths (8th). Age-standardised incidence 29.3 and mortality 7.5 per 100,000.
Skewed deaths: Africa has 7.2% of cases but 14.7% of deaths; Asia 27.6% and 31.6%.
Highest incidence (2022): Guadeloupe 157.5, Lithuania 135.0, Martinique 134.3 per 100,000. High mortality: Zimbabwe 71.4 (2024).
GBD 2021: age-standardised mortality 12.63 and DALYs 217.83 per 100,000. Prevalent cases rose from 3.6 million to 10.4 million from 1990 to 2021, driven mostly by population growth and ageing.
Lancet Commission: 1.4 million cases (2020) to 2.9 million (2040); deaths 375,000 to almost 700,000 (+85%), mostly in low- and middle-income countries.
At autopsy: 5% under 30 and 59% over 79 harbour incidental cancer (Bell 2015). Odds rise 1.7-fold per decade.

Australia

Latest actual figures: 26,469 cases (2021) and 3,726 deaths (2023). Second-commonest cancer death in men.
Five-year relative survival 61.1% (1988–1992) to 96% (2017–2021). PSA detection and lead time account for some of that rise, not only better treatment.
First Nations men: lower recorded incidence and mortality, but 5-year survival 86% vs 92% (2007–2014). In NSW, adjusted HR for prostate cancer death 1.49, unexplained by stage or comorbidity; less surgery too (aOR 0.60).
Rural and regional: mortality 56.9 vs 45.8 per 100,000 in cities (2007 data, the latest figure we could verify).
Overdiagnosis: 41% of 2012 diagnoses, or a lifetime risk of about 1 in 13 (Pathirana, BMJ Open 2019).

Causes and risk factors

Age: the strongest. See the autopsy data.
African ancestry: in the US, incidence 60–70% and cancer-specific mortality 100–120% above white men. Parity appears in equal-access settings after adjustment, with caveats. Nine further risk variants were reported in men of African ancestry, seven found mainly or only in men of African ancestry.
Higher risk in the Australian guideline: brother with prostate cancer; father diagnosed before 65; two or more second-degree relatives who died of it (strong). Father at any age (conditional). Two or more second-degree relatives diagnosed (consensus). Sub-Saharan African ancestry and BRCA2.
Body fat: strong evidence for advanced disease (WCRF review, 104 studies, 9.8 million men).
Smoking: mortality RR 1.24 in current smokers, yet incidence RR 0.90 (Islami 2014). Lower detection may explain some of that.
Diet: calcium RR 1.39, dairy RR 1.11, highest vs lowest intake. Weak, confounded, unproven.
Testosterone therapy: TRAVERSE (5,204 men) found high-grade cancer in 0.19% vs 0.12% (P=0.51). Short follow-up, and not a licence to treat.
Late-life environment: migrants reach local rates whatever their age at migration (Shimizu 1991). Detection may explain part of it.

Preventable factors

Smoking: 3.0% of prostate cancer deaths and 3.5% of DALYs (GBD 2021; UI 1.4–4.9% for deaths).
Weight: about 10% of advanced cases in the UK (a WCRF estimate, not a measured figure).
Vigorous exercise: ≥3 hours a week was estimated to cut lethal disease by 34% in US men over 60 (Kenfield, JNCI 2015; modelled from two cohorts).
Drugs: finasteride cut incidence (18.4% vs 24.4%, PCPT). Dutasteride similar (19.9% vs 25.1%, REDUCE). Neither is licensed for prevention here, and high-grade cancers looked commoner at first (6.4% vs 5.1%, PCPT).
Don't supplement: vitamin E raised incidence 17% (SELECT; 76 vs 65 per 1,000). Selenium showed no benefit. Lycopene hasn't held up in PCPT.
Overall incidence: we found no verified population attributable fraction for all prostate cancer.

Diagnosis essentials

PSA in general practice (Australia 2026)

Ages 50–69, usual risk: two-yearly after shared decision-making. Repeat within 1–3 months if ≥3.0 µg/L; refer if confirmed.
Higher risk: two-yearly from 45. Thresholds ≥1.0 µg/L at 45–49 and ≥2.0 µg/L at 50–69.
Ages 45–49, usual risk: one baseline can be offered if interested. Below 1.0 µg/L, wait until 50.
Age 70+: offered after clinical assessment; threshold 5.5 µg/L. Not recommended with life expectancy of 7 years or less.
DRE: conditionally recommended against as a routine add-on in primary care.
First Nations men: same recommendations, built into annual health assessments, with point-of-care PSA and telehealth to support access.
Short shelf life: the PSA chapter is approved only to May 2028.

After a raised PSA

MRI first: PRECISION (n=500): significant cancer 38% vs 26%, insignificant 9% vs 22%, and 28% avoided biopsy.
Australian pathway: biopsy PI-RADS 4–5, and PI-RADS 3 with PSA density ≥0.15. PI-RADS 3 below that may not need biopsy.
Screening with MRI: Göteborg-2 (men 50–60): MRI-targeted biopsy cut GG1 detection (RR 0.43) without clear loss of significant cancer (RR 0.84, 0.66–1.07).
Route: transperineal preferred in Australia. TRANSLATE (n=1,087): 60% vs 54% significant cancer with transrectal. Infection admissions were about 1% with each, with few events.
Staging: proPSMA (n=302): PSMA PET-CT accuracy 92% vs 65% for CT and bone scan, management change 28% vs 15%, and less radiation (8.4 vs 19.2 mSv).
Germline testing: advised in metastatic disease by the Pritchard authors, given 11.8% carriage and PARP inhibitor options.

Current treatment

Localised

Low risk: active surveillance is the recommended management in Australia. PCOR-ANZ: 69% of low-risk men chose it in 2015–16.
ProtecT (n=1,643; 15 years): prostate cancer death 3.1% monitoring, 2.2% surgery, 2.9% radiotherapy. Metastases 9.4%, 4.7%, 5.0%. All-cause death about 21–23%. Only 24.4% of the monitoring arm were alive and never treated.
An older era: SPCG-4 (n=695; 29 years): prostate cancer death RR 0.55, 19.6% vs 31.3% at 23 years, mean 2.9 years gained, NNT 8.4 for all-cause death. A different disease mix from screen-detected cancer today.
Radiotherapy schedule: PACE-B (n=874): five-fraction SBRT vs conventional, 5-year freedom from failure 95.8% vs 94.6%.
Price (ProtecT at 12 years): pads 18–24% after surgery, 3–8% radiotherapy, 9–11% monitoring. Nocturia 34%, 48%, 47%. Faecal leakage 6%, 12%, 6%. Erections fell to low levels in all arms.

Biochemical recurrence

EMBARK (high-risk recurrence): enzalutamide plus leuprolide 8-year OS 78.9% vs 69.5%, HR 0.60. Monotherapy 73.1%, HR 0.83, not significant.

Metastatic, hormone sensitive

Intensify beyond ADT: ARASENS (n=1,306): darolutamide with ADT and docetaxel, 4-year OS 62.7% vs 50.4%, HR 0.68.
PEACE-1 long term (median 7.7 years): adding abiraterone, HR 0.75. Median OS 4.6 years (standard), 5.7 (abiraterone), 6.4 (abiraterone plus radiotherapy).
Treat the primary if low burden: STAMPEDE (n=2,061): 3-year survival 81% vs 73% in low-burden disease; no benefit in high burden.
Adjuncts: metformin added weight control (2.00 vs 4.40 kg gain at 104 weeks) but no OS benefit (HR 0.91, 0.80–1.03).

Castration resistant

Lu-PSMA-617 (VISION): OS 15.3 vs 11.3 months, rPFS 8.7 vs 3.4; grade ≥3 AEs 52.7% vs 38.0%.
TheraP (ANZUP, n=200): PSA50 66% vs 37% for cabazitaxel; grade 3–4 AEs 33% vs 53%.
But CCTG PR21 (n=199; vs docetaxel): reported OS 14.3 vs 18.2 months, HR 1.64 (90% CI 1.14–2.35). Lu-PSMA doesn't replace docetaxel for everyone.
TALAPRO-2 (n=805): talazoparib plus enzalutamide, OS 45.8 vs 37.0 months, HR 0.80. HRR-deficient HR 0.55; non-deficient or unknown HR 0.88, not significant. Grade ≥3 anaemia 49%.
Australian access: MBS items for Lu-PSMA began in July 2025 with a prior taxane requirement and partial reimbursement. Novartis has asked MSAC to fund pre-taxane use (status as of mid-2026).

Emerging treatments and trials

PSMAddition (phase 3, n=1,144): Lu-PSMA plus ARPI in PSMA-positive hormone-sensitive disease: rPFS HR 0.72; OS immature. FDA approved 31 July 2026.
CAPItello-281 (phase 3, n=1,012): capivasertib in PTEN-deficient de novo metastatic disease. rPFS HR 0.81. OS at ESMO 2026: 57.4 vs 46.5 months, HR 0.82 (descriptive, post hoc). FDA approved June 2026.
DASL-HiCaP (ANZUP-led, phase 3): darolutamide with ADT and radiotherapy in very high-risk localised disease. Interim MFS HR 0.42 at ESMO 2026.
PROTEUS (phase 3): perioperative apalutamide plus ADT with prostatectomy: event-free survival HR 0.38 and MFS HR 0.49 in a substudy (n=408). Parent-trial MFS HR 0.80.
TALAPRO-3 (phase 3): talazoparib plus enzalutamide in HRR-altered sensitive disease, rPFS HR 0.48; interim OS HR 0.77, not significant.
Mevrometostat (EZH2 inhibitor): phase 1 randomised expansion (n=81): rPFS 14.3 vs 6.2 months, HR 0.51. Phase 3 MEVPRO-1 results not yet reported in anything we could find.
Xaluritamig (STEAP1 T-cell engager): phase 1 (n=97): PSA50 49%, ORR 24%; now in phase 3.
Actinium-225-PSMA: WARMTH (retrospective, n=448): median OS 15.5 months, xerostomia 68%. PSMAction (phase 2/3) is randomising after Lu-PSMA.
Neuroendocrine: KEYNOTE-365 cohort I: ORR 33% vs 6% adding pembrolizumab to chemotherapy. Median OS 11.4 vs 7.8 months.

Theoretical frontiers

Hypotheses, not recommendations.

Causes: theories

Hypothesis: a skin commensal in the gland. Cutibacterium acnes grew from 60% of cancer vs 26% of control prostates (OR 4.46; Davidsson 2016). Against: small, surgical sampling, and contamination is hard to exclude. Test: prediagnostic prostate fluid in cohorts, with typing for virulence plasmids.
Hypothesis: sexually transmitted triggers. Trichomonas seemed plausible, but high-titre seropositivity gave OR 0.97 in PCPT (Sutcliffe 2009). The evidence runs against it. Test: a multi-pathogen panel in ancestrally diverse biobanks.
Hypothesis: nerves start it. Chemical or surgical sympathectomy blocked early tumour development in mice, and nerve density tracked with poor outcomes in 43 men (Magnon, Science 2013). Against: one human series, retrospective. Test: nerve density in serial surveillance biopsies, then whether it predicts grade progression.
Hypothesis: youthful androgens write the fusion. TMPRSS2-ERG clusters in men under 50 (Chalmers 2021), consistent with androgen-driven DNA breaks. Against: association only. Test: lifetime hormone proxies vs fusion status.
Hypothesis: clearing the gland helps. ≥21 ejaculations a month vs 4–7 had 19–22% lower risk (Rider, n=31,925). Against: confounding by health, partners and screening. Test: prostatic-fluid biomarker studies, since randomisation is impractical.
Hypothesis: migration shows a modifiable late-life exposure. Migrants reach local rates whatever their age at arrival (Shimizu 1991). Against: PSA and surgical detection may explain much of it. Test: MRI-screened comparison of migrant and homeland cohorts, counting only GG2+.

Prevention: theoretical options

Hypothesis: target finasteride at those who stand to gain. PCPT at 18 years: prostate cancer death HR 0.75 (n=18,880), and the early high-grade worry didn't translate into more deaths. Test: a randomised trial in BRCA2 carriers or men with a top-decile polygenic score.
Hypothesis: gene-first screening finds what PSA misses. BARCODE1 (130 variants; 745 top-decile men screened): cancer in 40% of those biopsied, and 71.8% of significant cancers would have been missed by the UK PSA-MRI route. Against: no mortality data, European ancestry only, biopsy offered regardless of MRI. Test: a mortality-powered randomised trial with GG1 as a harm outcome.
Hypothesis: AI subtracts overdiagnosis. PI-CAI (Lancet Oncol 2024): AUROC 0.91 vs radiologists' 0.86. At equal sensitivity, 50.4% fewer false positives and 20% fewer GG1 cancers. Against: not non-inferior to original clinical reads. Test: randomised AI-first MRI reading in a screening pathway.
Hypothesis: fitness changes risk. Swedish men whose fitness rose ≥3% a year had 35% lower odds of diagnosis than decliners (n=57,652). Against: no mortality signal (46 deaths); fitter men may be tested differently. Test: an exercise trial with significant cancer as the endpoint in high-risk men.
Hypothesis: anti-obesity drugs reduce lethal disease. WCRF links body fat to advanced cancer. Against: metformin's 2.5 kg weight advantage gave no OS win (a different setting and mechanism). We found no randomised GLP-1 data on prostate cancer. Test: prespecified outcomes in long GLP-1 cardiometabolic trials.
Hypothesis: renaming is a public-health intervention. Eggener et al (JCO 2022) argue GG1 should stop being called cancer. Against: GG1 labelled wrong on biopsy could miss worse disease nearby. Test: a randomised vignette study, then registry treatment rates.

Treatment: theoretical options

Hypothesis: treat the gut to delay castration resistance. ADT expands commensals that make androgens. Antibiotics delayed resistance in mice, and stool from resistant patients conferred it in mice (Pernigoni, Science 2021). Against: mice; antibiotics have their own harms. Test: a narrow-spectrum or FMT phase 2 with stool-androgen endpoints.
Hypothesis: denervate the tumour. Sympathectomy blocked tumour initiation, and M1 muscarinic blockade reduced spread, in mice (Magnon 2013). Repurposing cheap β-blockers is tempting. Against: no randomised human data we could verify. Test: a window-of-opportunity trial of propranolol before prostatectomy.
Hypothesis: give the cancer too much testosterone, deliberately. Bipolar androgen therapy swings between extremes. TRANSFORMER (phase 2, n=195): similar PFS and OS to enzalutamide (HR 0.93), but post-crossover PSA50 77.8% vs 23.4% favouring bipolar-first. Against: non-randomised sequence analysis. Test: a randomised sequencing trial.
Hypothesis: T-cell redirection succeeds where checkpoint drugs failed. Engagers like xaluritamig bypass the need for a pre-existing immune response (phase 1 ORR 24%). Against: toxicity and durability unknown; phase 3 pending. Test: earlier-line trials with durable-response endpoints.
Hypothesis: alpha emitters work when lutetium fails. Higher linear energy transfer in smaller volumes; WARMTH suggests activity. Against: retrospective, and 68% xerostomia. Test: PSMAction, then salivary protection trials.
Hypothesis: AI should set ADT length. Pathology AI biomarker in RTOG 9202 (ASCO 2023): positive men, 15-year metastasis 19% on long vs 33% on short ADT; negative men, 11% on each, so 34% might avoid long ADT. A separate model was validated for short-term ADT in five phase 3 trials (NEJM Evidence 2023). Against: retrospective trial data. Test: prospective biomarker-directed de-escalation.
Repurposing lesson: metformin had the cheap-drug rationale, and STAMPEDE found no significant OS benefit (HR 0.91).

Gaps and open questions

Labelled hypotheses, not recommendations.

Same evidence, opposite policy: Australia now offers two-yearly testing after shared decisions. The UK NSC (May 2026) rejected general screening and recommended only BRCA2 carriers with a relevant family history, 45–61, two-yearly. England plans to start in 2027.
Hypothesis: the mortality case is outdated. ERSPC and CAP pre-date MRI-first pathways and surveillance. We found no screening trial with a mortality result for that modern pathway.
Black men: mortality double in the US, yet the UK NSC found the evidence too weak to recommend screening. The Pritchard germline series was nearly all white.
First Nations men: the HR of 1.49 remains unexplained, and we found no trial of what closes it.
Africa: 7.2% of cases and 14.7% of deaths, yet the trials above recruited mainly in high-income countries. Hypothesis: their thresholds and pathways won't transfer directly.
Surrogate approvals: PSMAddition's OS is immature; CAPItello-281's OS is descriptive. CCTG PR21 argues against assuming Lu-PSMA beats docetaxel everywhere.
Implementation: 31% of Australian low-risk men still chose treatment in 2015–16, and 42% of those under 60. We found no more recent national figure.
Older men: two-yearly testing at 70+ rests on a consensus recommendation.

Prognosis and quality of life

Localised (ProtecT, 15 years): 97.3% overall didn't die of prostate cancer (96.9–97.8% by arm).
Australia: 96% 5-year relative survival, helped by lead time and overdiagnosis.
Metastatic at diagnosis: PEACE-1 medians 4.6–6.4 years, depending on therapy.
After ARPI and taxane: median OS about 15 months with Lu-PSMA in VISION.
Survivorship: ProtecT's erection rates at 7 years were just 18% after surgery, 27% radiotherapy, 30% monitoring. All arms converged on low levels by 12 years.

Weird trivia

The name: Greek prostatēs, "one standing in front", also a word for leader or guardian. It stands in front of the bladder outlet.
Ancient cases: a 2,700-year-old Scythian king had the earliest diagnosed metastatic prostate cancer (reported 2007). A 2,250-year-old Egyptian mummy in Lisbon, catalogued M1, is the oldest known case from ancient Egypt.
Dogs: the only species besides humans where it naturally occurs in significant numbers. In his Nobel lecture, Huggins described them as the only laboratory animal in which prostate tumours occur.
Dogs again: two German Shepherds trained on explosives scored 98.6–100% sensitivity and 97.6–98.7% specificity on 902 urine samples (Taverna, J Urol 2015). Not in any guideline we reviewed.
Nerve sparing began with a stillborn infant: Donker showed Walsh the cavernous nerves in a dissected infant in Leiden on 13 February 1981.
PSA has a day job: it liquefies semen by cutting semenogelins, and zinc reins it in.
Movember: started in a pub in Australia in 2003 with two friends and then 30 mates, not a public-health agency.
Lithuania: Cancer Council Australia's (older) policy page calls it the only country with an organised programme. It also has the world's second-highest incidence.
Nelson Mandela had radiotherapy for prostate cancer at 83 in 2001, announced publicly by his office.

Guidelines

Key trials

SPCG-4 (2018, 29-year): surgery vs watchful waiting, mean 2.9 years gained.
PRECISION (2018): MRI-targeted biopsy finds more that matters, less that doesn't.
STAMPEDE radiotherapy (2018): treat the primary in low-burden metastatic disease.
proPSMA (2020): PSMA PET 92% accurate vs 65%.
TheraP (2021): Lu-PSMA beat cabazitaxel on PSA50.
VISION (2021): Lu-PSMA added 4 months of OS.
ARASENS (2022): triplet beat docetaxel with ADT.
Göteborg-2 (2022): MRI-targeted screening halved GG1 detection.
ProtecT (2023, 15-year): 2.7% prostate cancer death across arms.
CAP (2024): single PSA invite, RR 0.92.
PACE-B (2024): 5-fraction SBRT non-inferior.
TRANSLATE (2025): transperineal found more significant cancer.
BARCODE1 (2025): PRS finds cancers PSA pathways miss.
ERSPC (2025, 23-year): NNI 456, NND 12.
EMBARK (2025): enzalutamide plus leuprolide OS HR 0.60.
TALAPRO-2 (2025): OS HR 0.80; 0.55 if HRR deficient.
STAMPEDE metformin (2025): no OS benefit.
PSMAddition (2026): rPFS HR 0.72.
CAPItello-281 (2026): OS HR 0.82, descriptive.
DASL-HiCaP (2026): interim MFS HR 0.42.

Sources

Information only. Screening, thresholds and drug access differ between countries and change quickly. Check product information, PBS and MBS status, and specialist advice before acting.

Reviewed 3 November 2026. General education for health professionals; not a substitute for local protocols.

maladywise.com/deep-dives/prostate-cancer/