AI Helps Doctors Better Understand Prostate Cancer in Spain

Artificial intelligence (AI) is quietly changing how Spanish hospitals learn from day-to-day clinical practice. A new open-access study in BMC Cancer analysed millions of electronic health record (EHR) notes from eight hospitals across Spain and shows how AI can map real-world care for men with localized and locally advanced prostate cancer — who gets which treatment, how they fare, and where outcomes can be improved.

Why this matters to expats in Spain

Prostate cancer is one of the most common cancers in men. If you’re living in Spain — or planning to — it helps to know how hospitals actually manage this disease in routine practice (not just in clinical trials), what treatments are typically used, and what outcomes look like. This study offers exactly that kind of “real-world” view, using AI to read the unstructured, free-text notes that doctors write every day.

What the researchers did

Researchers examined 65.9 million EHR entries from 2.6 million patients seen at eight Spanish hospitals between 2014 and 2018. From these, they identified 22,166 men with prostate cancer. Focusing on the early stages — localized (confined to the prostate) and locally advanced (spread just beyond the prostate) — they studied 14,434 patients, and then a well-defined subgroup of 5,331 “incident” cases with complete baseline data for outcomes. AI (natural language processing + machine learning) extracted details like PSA levels, Gleason scores, staging terms, treatments, and events (metastasis, progression, death) directly from the doctors’ notes.

Key takeaways at a glance

• Who was treated with what? First treatment after diagnosis was radiotherapy in 40.7% of men, surgery (radical prostatectomy) in 37.1%, active surveillance/watchful waiting in 6.4%, brachytherapy in 4.2%, and androgen-deprivation therapy (ADT) alone in 3.3%.
  
  
• Outcomes track with risk level. At 36 months, real-world overall survival was 98% for low-risk localized disease, 97% for intermediate-risk, 93% for high-risk, and 91% for locally advanced disease. Metastasis-free survival at three years ranged from 96% (low-risk) to 77% (locally advanced).
  
  
• ADT alone performed poorest. Patients started on ADT alone were typically older, had more health issues, and the worst tumour profiles — and they had the weakest outcomes (around 79% overall survival at 36 months), underscoring that ADT monotherapy is not a curative approach for localized disease.
  
  
• Even with curative intent, events happen. Median event-free survival across the cohort was around 34 months, meaning many men experienced a defined “event” (such as PSA failure, progression or other oncologic event) within three years despite standard care.

How AI made this possible

Most of the richest clinical details live in narrative notes — radiology reports, clinic letters, operative summaries — not in neat coded fields. The team used a clinical natural language processing system (EHRead®) to convert those free-text notes into structured data at scale. That allowed them to stratify patients by D’Amico risk categories (low, intermediate, high) or “locally advanced,” and to tie those risk levels to real-world treatments and outcomes.

Who tends to get which treatment?

The pattern reflects everyday clinical decision-making:

• Radiotherapy (RT): More common in older men and in those with more comorbidities or higher PSA at diagnosis.
  
  
• Radical prostatectomy (RP): More common in younger, fitter men with lower PSA and fewer comorbidities.
  
  
• Active surveillance / Watchful waiting: Used for carefully selected low-risk or low-metastatic-potential cases, consistent with European guidelines aimed at avoiding overtreatment.
  
  
• ADT monotherapy: Reserved for men who cannot receive local therapy (surgery, radiotherapy or brachytherapy). In this study, these patients started off older and sicker — and their outcomes were predictably worse.

What did outcomes look like in the real world?

When the team followed the 5,331 incident patients with adequate data for a median of 2.3 years, they saw a clear “risk-gradient” — higher risk meant lower survival and more metastases. Here are the headline numbers (36-month estimates):

• Real-world overall survival: ~98% (low risk), 97% (intermediate risk), 93% (high risk), 91% (locally advanced).
  
  
• Metastasis-free survival: ~96% (low), 95% (intermediate), 87% (high), 77% (locally advanced).
  
  
• Event-free survival: varied by risk and first treatment; median across the whole cohort was ~33.7 months.

By first treatment received, RP and brachytherapy cohorts had the highest overall survival (~98% at three years), while ADT-only patients were lowest (~79%). RP was associated with the best three-year event-free rates (~55% without an event), followed by radiotherapy (~47%) and brachytherapy (~43%). Again, these groups aren’t directly comparable — the RT group was older and sicker at baseline, and the ADT-only group markedly so.

How does this compare with guidelines?

The study’s treatment patterns generally align with the European Association of Urology guidance: active surveillance for select low-risk men; curative local therapy (surgery or radiotherapy) for those with a life expectancy over 10 years and a meaningful recurrence risk; and avoidance of ADT monotherapy unless local therapy isn’t feasible. The real-world numbers support those principles: curative-intent local treatment correlates with better survival and metastasis control, especially in high-risk and locally advanced disease.

Limitations you should know about

As with any real-world study, there are caveats:

• Short-to-mid follow-up: Median follow-up was about 2.3 years — useful for early outcomes, but prostate cancer often needs longer follow-up to understand long-term control and side effects.
  
  
• Unbalanced groups: Treatment groups differed at baseline (age, PSA, comorbidities, risk). That means differences in outcomes can’t be read as head-to-head comparisons.
  
  
• Documentation gaps: Not every detail (for example, radiotherapy dose or duration of ADT) is consistently recorded in routine notes, so some nuances are missing.
  
  
• Pre-PSMA-PET era: Care pathways reflect 2014–2018 practice, before widespread PSMA PET/CT staging in Spain, which now detects metastases earlier and may shift decisions.

What this means for patients and families

For many men with localized disease in Spain — including expats — outcomes at three years are excellent. The vast majority are alive and free of metastasis, particularly in low- and intermediate-risk disease. Where vigilance is crucial is in high-risk and locally advanced groups, where the disease is more aggressive and the chance of metastasis is higher, even with curative therapy. This underlines the importance of:

• Accurate risk stratification at diagnosis (PSA, MRI, biopsy with Gleason grading, clinical staging), to guide treatment choice.
  
  
• Curative-intent local therapy (surgery or radiotherapy) for eligible high-risk and locally advanced cases, often alongside hormone therapy.
  
  
• Shared decision-making — discussing trade-offs between treatments (oncologic control, side effects, recovery time) based on your age, health, and preferences.
  
  
• Follow-up planning to monitor PSA and symptoms, because a proportion of men will experience events within a few years and may benefit from timely salvage or systemic therapy.

Practical tips for expats navigating care in Spain

• Primary care first: Start with your family doctor (médico de familia) or private GP for PSA testing and symptom review.
  
  
• Specialist referral: Urologists lead diagnosis and surgical options; radiation oncologists lead radiotherapy. Multidisciplinary tumour boards are common in larger centres.
  
  
• Diagnostics: Expect PSA testing, multiparametric MRI of the prostate, and targeted or systematic biopsy for confirmation and Gleason grading.
  
  
• Treatment pathways: Low-risk men may be offered active surveillance. Intermediate-risk patients may choose surgery or radiotherapy, sometimes with short-term hormone therapy. High-risk/locally advanced disease often involves combined approaches.
  
  
• Know your terms: “Active surveillance” involves scheduled PSA checks, imaging and biopsy, with a plan to treat if the cancer progresses. “Watchful waiting” is different and used when life expectancy or comorbidities make curative treatment unsuitable.

How AI could help you (indirectly)

AI isn’t making your diagnosis or choosing your treatment. But by converting millions of real-world notes into data, it helps health systems and clinical teams see what’s happening across hospitals — where outcomes are strong, where they lag, and which patients may need more support. Over time, this can inform better guidelines, faster adoption of effective strategies, and more personalised care.

Frequently asked questions

Does Spain screen men for prostate cancer?
Spain does not have a universal PSA screening programme. Testing is usually individualised after a discussion about benefits and harms. If you’re 50–70 (or younger with strong family history), ask your doctor about PSA testing.

What is the D’Amico risk system?
It groups localized prostate cancers by PSA, stage, and Gleason score to estimate recurrence risk and guide treatment choices (surveillance vs surgery vs radiotherapy ± hormones).

Is ADT a cure?
No. ADT lowers testosterone to slow cancer growth. For localized disease, ADT alone is generally not curative and is typically reserved for men who cannot receive local therapy.

How quickly do events happen after treatment?
In this study, many men had an oncologic “event” within about three years (median event-free survival ~33.7 months), underscoring the importance of structured follow-up.

Study in context

The BMC Cancer analysis aligns with a growing body of work using AI to unlock insights from routine cancer care. Similar efforts in pathology, imaging, and registry analysis are showing that AI can match expert performance for some diagnostic tasks and help stratify risk — all of which can feed into more informed, patient-centred decisions.

Bottom line

For expats living in Spain, this is encouraging news. Real-world care for localized prostate cancer is robust, and AI is giving doctors a clearer picture of what works best for whom. If you’re facing decisions about screening or treatment, the key steps are: understand your personal risk, discuss options with a specialist team, and commit to follow-up. As AI-driven analyses expand to include newer tools like PSMA-PET and longer follow-up, we should get even sharper insights — and, hopefully, better outcomes over time.

Source
Maroto JP, Puente J, Conde Moreno A, et al. Real-world evidence in localized and locally advanced prostate cancer: applying artificial intelligence to electronic health records. BMC Cancer (2025) 25:1618. Open access full text.