iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
AI models may help identify groups of people at higher risk of melanoma over the next five years, but they cannot tell an individual that they will develop cancer. A Swedish study reported in 2026 applied models to registry data for more than six million adults; the University of Gothenburg says further research and policy decisions are needed before the approach could be introduced into healthcare.
What does “predicting melanoma risk years ahead” mean?
These models estimate the likelihood of an outcome in a defined population and time period. They can flag patterns associated with higher risk, but a risk estimate is not a diagnosis, a forecast that a particular person will get melanoma, or proof that acting on the score improves health outcomes.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Dermoscopy: The Essentials | $68.30 | Buy on Amazon |
“Five-year risk” also has a specific meaning: it concerns melanoma diagnosed during a five-year period, not a person’s lifetime risk. Predictions about a first melanoma in the general population are different from predictions about another primary melanoma in someone who has already had one.
Free tools Windows power users keep installed
One-click scans. No signup required.
What did the 2026 Swedish study report?
The University of Gothenburg described a study using routinely collected registry data for Sweden’s adult population. The study population included 6,036,186 people, of whom 38,582 developed melanoma during the five-year study period—0.64% of that cohort. The researchers used factors including age, sex, diagnoses, medications and socioeconomic status to identify small groups with significantly higher risk during that period.
#1 Best Overall
That 0.64% is a result for this particular study population and period; it should not be treated as a general estimate of melanoma risk. The university’s account is a news summary, and it does not establish the full technical methods, external validation, or whether use of the model improves clinical outcomes. The university says more research and policy decisions are needed before the method can be introduced in healthcare.
How strong is the evidence so far?
Different studies answer different questions and use different populations and inputs. Their figures should not be compared as if they measured the same model or established a ready-to-use screening service.
| Study or tool | Population and target | Reported result | What the result does not establish |
|---|---|---|---|
| University of Gothenburg report, 2026 | 6,036,186 adults in Sweden; melanoma diagnoses during a five-year study period; registry variables | 38,582 people developed melanoma (0.64% of this cohort); models identified small groups at significantly higher risk | The news summary does not establish full validation details, clinical benefit, or readiness for routine care. |
| Gillstedt and Polesie, Acta Dermato-Venereologica, 2022 | Proof-of-concept convolutional neural network trained on healthcare registry data; historical records used to predict melanoma over a subsequent five years | On a test set constructed with 1,000 melanoma cases and 5,000 controls, AUC was 0.59 (95% CI 0.57–0.61). At the selected point where sensitivity equaled specificity, both were 56% (sensitivity 95% CI 53–60%; specificity 95% CI 55–58%). | This test-set performance is a research signal, not evidence of a clinically useful screening program. |
| Nature Communications, 2021 | 210,000 research participants who completed surveys about personal and family skin-cancer history, skin susceptibility and UV exposure; a score combined 32 genetic and non-genetic factors | The study reported up to a 13-fold increase in skin-cancer risk for the top-percentile score compared with the middle percentile. | The comparison applies to that score and study; it is not a 13-fold estimate for every person or a result from the Swedish registry model. |
| Population-based subsequent-primary study, 2020 | 2,613 melanomas among 1,266 participants with a prior melanoma; median registry follow-up was 14 years; model used 12 risk factors | C-statistic was 0.73 for predicting a second primary melanoma and 0.65 for third and fourth primaries. | It addresses later primary melanomas among people already diagnosed, not first-melanoma risk in the general population. |
How to read the metrics
The 2022 study’s AUC summarizes how well the model ranked cases above controls in its test set; it does not say that a person with a particular score has a 59% chance of melanoma. Sensitivity and specificity describe classification at one selected threshold, not the accuracy of every possible risk estimate. Because the test set was deliberately constructed with 1,000 cases and 5,000 controls, its case-to-control mix should not be mistaken for the frequency of melanoma in the general population.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchA relative comparison such as “up to 13-fold” also needs its reference group and outcome attached. It describes the top score percentile compared with the middle percentile in the 2021 study, not a person’s absolute probability and not the Swedish model’s performance.
What can people use today?
NCI’s Melanoma Risk Assessment Tool
The National Cancer Institute’s Melanoma Risk Assessment Tool (MRAT) estimates five-year absolute risk and is intended for health professionals. It uses self-reported history and a brief physical examination by a health professional. NCI cautions that no estimate can precisely determine which patient will develop melanoma and advises people who are not health professionals to discuss their risk with a provider.
MRAT’s development data came from a case-control study of 1,663 non-Hispanic White patients at clinics in Philadelphia, Pennsylvania, and San Francisco, California. NCI says the tool has not been validated for all non-Hispanic White individuals and notes that further validation research is underway. Its stated evidence therefore does not support assuming the calculator applies equally to other populations or locations.
Research scores are not the same as a clinical service
A model developed or tested in research is not automatically approved, deployed in clinics, or shown to improve outcomes. The 2026 Swedish report does not establish that its model is available for people to use. A 2024 electronic-health-record model is described in a medRxiv preprint; a preprint should not be treated as peer-reviewed evidence unless its publication status is independently established.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How to judge a claim about a melanoma prediction model
- Outcome and time horizon: Check whether the model predicts a first melanoma within five years, lifetime risk, or a subsequent primary after a prior diagnosis.
- Who was studied: Look for the country, setting, eligibility criteria and groups included or excluded. Performance in one cohort does not automatically transfer to another.
- What information it uses: Registry diagnoses and medications, self-reported history and examination, and genetic or UV-exposure factors are not interchangeable inputs.
- How it was validated: Internal testing, a holdout test set, external validation in a separate population, and prospective evaluation answer different questions.
- Which metric is reported: AUC, sensitivity, specificity, absolute risk and relative risk are different quantities. Ask what population and threshold each figure describes.
- Whether it changes care: Risk stratification is not by itself a diagnosis or proof that model-guided screening prevents illness or improves outcomes.
What to do about melanoma prevention
Do not wait for a risk score to take sun-protective steps. NCI guidance recommends seeking shade, avoiding peak sun hours, wearing a wide-brimmed hat and covering exposed skin, and using sunscreen with SPF 30 or higher. These measures are prevention guidance; sunscreen is not a prediction tool or a diagnostic test. For personal questions about risk or concerning skin changes, speak with a healthcare professional.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
