Background: Efficient identification of individuals at high risk of skin cancer is crucial for implementing personalized screening strategies and subsequent care. While Artificial Intelligence holds promising potential for predictive analysis using image data, its application for skin cancer risk prediction utilizing facial images remains unexplored. We present a neural network-based explainable artificial intelligence (XAI) approach for skin cancer risk prediction based on 2D facial images and compare its efficacy to 18 established skin cancer risk factors using data from the Rotterdam Study. Methods: The study employed data from the Rotterdam population-based study in which both skin cancer risk factors and 2D facial images and the occurrence of skin cancer were collected from 2010 to 2018. We conducted a deep-learning survival analysis based on 2D facial images using our developed XAI approach. We subsequently compared these results with survival analysis based on skin cancer risk factors using cox proportional hazard regression. Findings: Among the 2810 participants (mean Age = 68.5 ± 9.3 years, average Follow-up = 5.0 years), 228 participants were diagnosed with skin cancer after photo acquisition. Our XAI approach achieved superior predictive accuracy based on 2D facial images (c-index = 0.72, 95% CI: 0.70-0.74), outperforming that of the known risk factors (c-index = 0.59, 95% CI 0.57-0.61). Interpretation: This proof-of-concept study underscores the high potential of harnessing facial images and a tailored XAI approach as an easily accessible alternative over known risk factors for identifying individuals at high risk of skin cancer. Funding: The Rotterdam Study is funded through unrestricted research grants from Erasmus Medical Center and Erasmus University, Rotterdam, Netherlands Organization for the Health Research and Development (ZonMw), the Research Institute for Diseases in the Elderly (RIDE), the Ministry of Education, Culture and Science, the Ministry for Health, Welfare and Sports, the European Commission (DG XII), and the Municipality of Rotterdam. G.V. Roshchupkin is supported by the ZonMw Veni grant (Veni, 549 1936320).
BACKGROUND: There is lack of nationwide data on the cumulative incidence and timing of subsequent cutaneous squamous cell carcinomas (cSCCs) among patients with a first cSCC. OBJECTIVE: To investigate the cumulative incidence and timing of subsequent cSCCs. METHODS: Patients with a first cSCC in 2007/2008 from the Netherlands Cancer Registry were linked to the Netherlands Pathology Registry for subsequent cSCCs and the Netherlands Organ Transplant Registry. Cumulative incidence function curves were calculated for subsequent cSCCs and stratified for immune status. RESULTS: Among the 12,345 patients, second to sixth cSCC occurred in 4325, 2010, 1138, 739, and 501 patients, with median time intervals of 1.4, 1.2, 0.9, 0.6, and 0.5 years after the previous cSCC, respectively. The cumulative incidence of a subsequent cSCC at 5 years increased from 28% to 67% for the second to sixth cSCC. For solid organ transplant recipients, the cumulative incidences increased from 74% to 92% and from 41% to 64% for patients with hematologic malignancy. LIMITATIONS: Only histopathologically confirmed cSCCs were included. CONCLUSION: The risk of a subsequent cSCC steeply rises with the number of prior cSCCs and immune status, while the time interval decreases. This can support more informed decisions about follow-up management.
Carcinoma, Squamous Cell , Organ Transplantation , Skin Neoplasms , Humans , Carcinoma, Squamous Cell/epidemiology , Carcinoma, Squamous Cell/pathology , Cohort Studies , Incidence , Skin Neoplasms/epidemiology , Skin Neoplasms/pathology , Organ Transplantation/adverse effects
Background: Cutaneous squamous cell carcinoma (cSCC) is a common skin cancer, affecting more than 2 million people worldwide yearly and metastasising in 2-5% of patients. However, current clinical staging systems do not provide estimates of absolute metastatic risk, hence missing the opportunity for more personalised treatment advice. We aimed to develop a clinico-pathological model that predicts the probability of metastasis in patients with cSCC. Methods: Nationwide cohorts from (1) all patients with a first primary cSCC in The Netherlands in 2007-2008 and (2) all patients with a cSCC in 2013-2015 in England were used to derive nested case-control cohorts. Pathology records of primary cSCCs that originated a loco-regional or distant metastasis were identified, and these cSCCs were matched to primary cSCCs of controls without metastasis (1:1 ratio). The model was developed on the Dutch cohort (n = 390) using a weighted Cox regression model with backward selection and validated on the English cohort (n = 696). Model performance was assessed using weighted versions of the C-index, calibration metrics, and decision curve analysis; and compared to the Brigham and Women's Hospital (BWH) and the American Joint Committee on Cancer (AJCC) staging systems. Members of the multidisciplinary Skin Cancer Outcomes (SCOUT) consortium were surveyed to interpret metastatic risk cutoffs in a clinical context. Findings: Eight out of eleven clinico-pathological variables were selected. The model showed good discriminative ability, with an optimism-corrected C-index of 0.80 (95% Confidence interval (CI) 0.75-0.85) in the development cohort and a C-index of 0.84 (95% CI 0.81-0.87) in the validation cohort. Model predictions were well-calibrated: the calibration slope was 0.96 (95% CI 0.76-1.16) in the validation cohort. Decision curve analysis showed improved net benefit compared to current staging systems, particularly for thresholds relevant for decisions on follow-up and adjuvant treatment. The model is available as an online web-based calculator (https://emc-dermatology.shinyapps.io/cscc-abs-met-risk/). Interpretation: This validated model assigns personalised metastatic risk predictions to patients with cSCC, using routinely reported histological and patient-specific risk factors. The model can empower clinicians and healthcare systems in identifying patients with high-risk cSCC and offering personalised care/treatment and follow-up. Use of the model for clinical decision-making in different patient populations must be further investigated. Funding: PPP Allowance made available by Health-Holland, Top Sector Life Sciences & Health, to stimulate public-private partnerships.
PURPOSE: Early stage lip squamous cell carcinoma (lip SCC) can be treated with conventional excision, Mohs micrographic surgery (MMS), or brachytherapy. The aim of this retrospective study was to describe the medical outcomes, patient-reported outcomes, and costs of these treatments. METHODS: A retrospective cohort study of T1-T2 lip SSCs treated between 1996 and 2019. Medical outcomes, recurrences, and survival were retrieved from medical records. Facial appearance, facial function, and Quality of Life (QoL) were measured with the Face-Q H&N and EQ-5D-5L questionnaires. Costs were also calculated. RESULTS: Of the 336 lip SCCs, 122 were treated with excision, 139 with MMS, and 75 with brachytherapy. Locally, the recurrence rate was 2.7% and regionally 4.8%. There were 2% disease-related deaths. T2-stage and poor tumor differentiation were associated with recurrences. Posttreatment QoL, facial function, and appearance were rated as good. Brachytherapy was the most expensive treatment modality. CONCLUSION: Early-stage lip SCC has a good prognosis, with a disease-specific survival of 98.2% after a median follow-up of 36 months, there was a high QoL and satisfaction at long-term follow-up. Based on the costs and the risk of locoregional recurrences, we believe that, for most noncomplex lip SCCs, MMS would be the most logical treatment option.
Background: Artificial intelligence (AI)-based mobile phone apps (mHealth) have the potential to streamline care for suspicious skin lesions in primary care. This study aims to investigate the conditions and feasibility of a study that incorporates an AI-based app in primary care and evaluates its potential impact. Methods: We conducted a pilot feasibility study from November 22nd, 2021 to June 9th, 2022 with a mixed-methods design on implementation of an AI-based mHealth app for skin cancer detection in three primary care practices in the Netherlands (Rotterdam, Leiden and Katwijk). The primary outcome was the inclusion and successful participation rate of patients and general practitioners (GPs). Secondary outcomes were the reasons, facilitators and barriers for successful participation and the potential impact in both pathways for future sample size calculations. Patients were offered use of an AI-based mHealth app before consulting their GP. GPs assessed the patients blinded and then unblinded to the app. Qualitative data included observations and audio-diaries from patients and GPs and focus-groups and interviews with GPs and GP assistants. Findings: Fifty patients were included with a median age of 52 years (IQR 33.5-60.3), 64% were female, and 90% had a light skin type. The average patient inclusion rate was 4-6 per GP practice per month and 84% (n = 42) successfully participated. Similarly, in 90% (n = 45 patients) the GPs also successfully completed the study. GPs never changed their working diagnosis, but did change their treatment plan (n = 5) based on the app's assessments. Notably, 54% of patients with a benign skin lesion and low risk rating, indicated that they would be reassured and cancel their GP visit with these results (p < 0.001). Interpretation: Our findings suggest that studying implementation of an AI-based mHealth app for detection of skin cancer in the hands of patients or as a diagnostic tool used by GPs in primary care appears feasible. Preliminary results indicate potential to further investigate both intended use settings. Funding: SkinVision B.V.
Population-based studies available to analyze the prevalence, risk factors, and longitudinal outlook of actinic keratoses (AKs) are limited. These features mentioned earlier were assessed using Rotterdam study participants aged ≥40 years who underwent a full-body skin examination by a dermatology-trained physician. ORs with 95% confidence intervals were calculated for the associations between risk factors and the presence of AK. Among 8,239 eligible participants, the prevalence of one or more AKs was 21.1% (95% confidence interval = 20.2-22.0) and was higher in men. Male sex, age, lighter hair and eye color, baldness, genetic risk score, and digital photoaging measures (digitally assessed pigmented spots, telangiectasias, and global facial wrinkling) had a positive association with AK. Cigarette smokers had reduced odds of having AK, with current smokers having the lowest risk. Among patients with two AK assessments, there was no difference in the presence of AK during follow-up between treated and untreated participants. In conclusion, genetic risk score and digital photoaging measures showed associations with increased lesion count. At the individual level, patients were most likely to decrease in AK severity group over time, possibly regardless of whether or not participants were treated.
Artificial intelligence (AI) based algorithms for classification of suspicious skin lesions have been implemented in mobile phone apps (mHealth), but their effect on healthcare systems is undocumented. In 2019, a large Dutch health insurance company offered 2.2 million adults free access to an mHealth app for skin cancer detection. To study the impact on dermatological healthcare consumption, we conducted a retrospective population-based pragmatic study. We matched 18,960 mHealth-users who completed at least one successful assessment with the app to 56,880 controls who did not use the app and calculated odds ratios (OR) to compare dermatological claims between both groups in the first year after granting free access. A short-term cost-effectiveness analysis was performed to determine the cost per additional detected (pre)malignancy. Here we report that mHealth-users had more claims for (pre)malignant skin lesions than controls (6.0% vs 4.6%, OR 1.3 (95% CI 1.2-1.4)) and also a more than threefold higher risk of claims for benign skin tumors and nevi (5.9% vs 1.7%, OR 3.7 (95% CI 3.4-4.1)). The costs of detecting one additional (pre)malignant skin lesion with the app compared to the current standard of care were 2567. Based on these results, AI in mHealth appears to have a positive impact on detecting more cutaneous (pre)malignancies, but this should be balanced against the for now stronger increase in care consumption for benign skin tumors and nevi.
BACKGROUND: Although cutaneous squamous cell carcinoma (cSCC) is common, lymph node metastases are relatively rare and are usually treated with lymph node dissection (LND). The aim of this study was to describe the clinical course and prognosis after LND for cSCC at all anatomical locations. METHODS: A retrospective search at three centres was performed to identify patients with lymph node metastases of cSCC who were treated with LND. Prognostic factors were identified by uni- and multivariable analysis. RESULTS: A total of 268 patients were identified with a median age of 74. All lymph node metastases were treated with LND, and 65% of the patients received adjuvant radiotherapy. After LND, 35% developed recurrent disease both locoregionally and distantly. Patients with more than one positive lymph node had an increased risk for recurrent disease. 165 (62%) patients died during follow-up of whom 77 (29%) due to cSCC. The 5-year OS- and DSS rate were 36% and 52%, respectively. Disease-specific survival was significantly worse in immunosuppressed patients, patients with primary tumors >2cm and patients with more than one positive lymph node. CONCLUSIONS: This study shows that LND for patients with lymph node metastases of cSCC leads to a 5-year DSS of 52%. After LND, approximately one-third of the patients develop recurrent disease (locoregional and/or distant), which underscores the need for better systemic treatment options for locally advanced cSCC. The size of the primary tumor, more than one positive lymph node, and immunosuppression are independent predictors for risk of recurrence and disease-specific survival after LND for cSCC.
Carcinoma, Squamous Cell , Skin Neoplasms , Humans , Carcinoma, Squamous Cell/pathology , Lymphatic Metastasis/pathology , Retrospective Studies , Skin Neoplasms/surgery , Skin Neoplasms/pathology , Lymph Node Excision , Lymph Nodes/surgery , Lymph Nodes/pathology , Neoplasm Staging
BACKGROUND: The increasing group of melanoma survivors reports multiple unmet needs regarding survivorship care (SSC). To optimise melanoma SSC, it is crucial to take into account the perspectives of oncological healthcare providers (HCPs) in addition to those of patients. The aim of this study is to gain an in-depth understanding of HCPs' perspectives on appropriate melanoma SSC. METHODS: Four online focus groups were conducted with mixed samples of oncological HCPs (dermatologists, surgeons, oncologists, oncological nurse practitioners, support counsellors and general practitioners) (total n = 23). A topic guide was used to structure the discussions, focusing on perspectives on both SSC and survivorship care plans (SCPs). All focus groups were recorded, transcribed verbatim, and subjected to an elaborate thematic content analysis. RESULTS: Regarding SSC, HCPs considered the current offer minimal and stressed the need for broader personalised SSC from diagnosis onwards. Although hardly anyone was familiar with SCPs, they perceived various potential benefits of SCPs, such as an increase in the patients' self-management and providing HCPs with an up-to-date overview of the patient's situation. Perceived preconditions for successful implementation included adequate personalisation, integration in the electronic health record and ensuring adequate funding to activate and provide timely updates. CONCLUSIONS: According to HCPs there is considerable room for improvement in terms of melanoma SSC. SCPs can assist in offering personalised and broader i.e., including psychosocial SSC. Aside from personalisation, efforts should be focused on SCPs' integration in clinical practice, and their long-term maintenance.
Melanoma , Survivorship , Humans , Focus Groups , Patient Care Planning , Health Personnel , Melanoma/therapy
BACKGROUND: Merkel cell carcinoma (MCC) is a rare tumour with neuroendocrine differentiation and high associated mortality. Studies that describe the epidemiology of MCC are often limited by small sample size, short duration of follow-up, absence of nationwide data and paucity of data on different risk factors. OBJECTIVES: To determine the incidence, demographics and survival for MCC in England between 2004 and 2018. METHODS: This national retrospective cohort study identified all cases of MCC in England from 2004 to 2018 using national population-based data from the National Disease Registration Service. Crude counts, European age-standardized incidence rates (EASRs) and joinpoint analysis were conducted. Patient demographics and treatments received were described. Multivariable Cox regression analysis was used to study risk factors for MCC-specific mortality, by including a priori defined demographic factors, tumour characteristics and immunosuppression. Treatment data were not included in the Cox regression analysis. RESULTS: A total of 3775 MCC tumours were registered. The median age at diagnosis was 81 years (interquartile range 74-87). Overall, 96·6% of patients identified as White ethnicity, and 8·3% of patients were immunosuppressed. The most common site was the face (27·4%). Patients most often presented with stage one disease (22·8%); however, stage was unknown in 31·0%. In total, 80·7% of patients underwent surgical excision, 43·5% radiotherapy and 9·2% systemic therapy. The EASR increased from 0·43 per 100 000 person-years (PYs) to 0·65 per 100 000 person-years between 2004 and 2018, representing a significant annual percentage change of 3·9%. The EASR was greater in men than in women for all years, with an overall male-to-female ratio of 1·41 : 1. The highest EASR was in South West England. Five-year disease-specific survival was 65·6% [95% confidence interval (CI) 63·8-67·4], with a median follow-up of 767 days. MCC-specific mortality increased with age [hazard ratio (HR) 1·02, 95% CI 1·02-1·03], deprivation (HR 1·43, 95% CI 1·16-1·76), immunosuppression (HR 2·80, 95% CI 2·34-3·34) and stage at diagnosis (HR 8·24, 95% CI 5·84-11·6). CONCLUSIONS: This study presents the largest national MCC dataset in Europe, and the most complete reporting of MCC incidence and survival ever published. With the EASR of MCC increasing and high associated mortality, this study encourages further research into the pathology, diagnosis and therapeutic options for MCC to support management guidelines.
Carcinoma, Merkel Cell , Skin Neoplasms , Humans , Male , Female , Aged, 80 and over , Carcinoma, Merkel Cell/epidemiology , Carcinoma, Merkel Cell/therapy , Cohort Studies , Skin Neoplasms/epidemiology , Skin Neoplasms/therapy , Skin Neoplasms/diagnosis , Incidence , Retrospective Studies
AIM OF THE STUDY: Cutaneous squamous cell carcinoma (cSCC) incidences are increasing but scarcely available separated. We analysed incidence rates of cSCC over three decades with an extrapolation to 2040. METHODS: Cancer registries from the Netherlands, Scotland and two federal states of Germany (Saarland/Schleswig-Holstein) were sourced for separate cSCC incidence data. Incidence and mortality trends between 1989/90 and 2020 were assessed using Joinpoint regression models. Modified age-period-cohort models were applied to predict incidence rates up to 2044. Rates were age-standardised using the new European standard population (2013). RESULTS: Age-standardised incidence rates (ASIR, per 100,000 persons per year) increased in all populations. The annual percent increase ranged between 2.4% and 5.7%. The highest increase occurred in the age groups ≥60 years, especially in men aged ≥80 years, with a three to 5-fold increase. Extrapolations up to 2044 showed an unrestrained increase in incidence rates in all countries investigated. Age-standardised mortality rates (ASMR) showed slight increases between 1.4 and 3.2% per year in Saarland and Schleswig-Holstein for both sexes and for men in Scotland. For the Netherlands, ASMRs remained stable for women but declined for men. CONCLUSION: There was a continuous increase of cSCC incidence over three decades with no tendency for levelling-off, especially in the older populations as males ≥80 years. Extrapolations point to a further increasing number of cSCC up to 2044, especially among ≥60 years. This will have a significant impact on the current and future burden on dermatologic health care which will be faced with major challenges.
Carcinoma, Squamous Cell , Skin Neoplasms , Male , Humans , Female , Carcinoma, Squamous Cell/epidemiology , Incidence , Skin Neoplasms/epidemiology , Netherlands/epidemiology , Germany/epidemiology , Scotland/epidemiology , Registries
We present two cases of plaque-type trichoblastoma with atypical foci. A rare variant of trichoblastoma is the plaque variant, which is characterized by poor circumscription and locally infiltrative growth pattern. These lesions mostly require multiple stages of Mohs micrographic surgery. Debate still exists whether this variant should be considered as a benign entity or as "low-grade" malignant counterpart of trichoblastoma. In this report we describe two cases of plaque-type trichoblastoma with atypical foci, which harbored somatic mutations in the Hedgehog pathway, thus should be acknowledged as intermediate malignancies. In addition, extensive molecular workup of both the trichoblastic and atypical component in sequential lesions in the same patient was performed.
Hair Diseases , Skin Neoplasms , Humans , Hedgehog Proteins , Skin Neoplasms/pathology , Hair Diseases/pathology , Mohs Surgery , Mutation
Recent studies show promising potential for artificial intelligence (AI) to assist healthcare providers (HCPs) in skin cancer care. The aim of this study is to explore the views of dermatologists and general practitioners (GPs) regarding the successful implementation of AI when assisting HCPs in skin cancer care. We performed a qualitative focus group study, consisting of six focus groups with 16 dermatologists and 17 GPs, varying in prior knowledge and experience with AI, gender, and age. An in-depth inductive thematic content analysis was deployed. Perceived benefits, barriers, and preconditions were identified as main themes. Dermatologists and GPs perceive substantial benefits of AI, particularly an improved health outcome and care pathway between primary and secondary care. Doubts about accuracy, risk of health inequalities, and fear of replacement were among the most stressed barriers. Essential preconditions included adequate algorithm content, sufficient usability, and accessibility of AI. In conclusion, dermatologists and GPs perceive significant benefits from implementing AI in skin cancer care. However, to successfully implement AI, key barriers need to be addressed. Efforts should focus on ensuring algorithm transparency, validation, accessibility for all skin types, and adequate regulation of algorithms. Simultaneously, improving knowledge about AI could reduce the fear of replacement.
General Practitioners , Skin Neoplasms , Humans , Artificial Intelligence , Dermatologists , Skin
To facilitate nationwide epidemiological research on advanced cutaneous squamous cell carcinoma (cSCC), that is, locally advanced, recurrent, or metastatic cSCC, we sought to develop and validate a rule-based algorithm that identifies advanced cSCC from pathology reports. The algorithm was based on both hierarchical histopathological codes and free text from pathology reports recorded in the National Pathology Registry. Medical files from the Erasmus Medical Center of 186 patients with stage III/IV/recurrent cSCC and 184 patients with stage I/II cSCC were selected and served as the gold standard to assess the performance of the algorithm. The rule-based algorithm showed a sensitivity of 91.9% (95% confidence interval = 88.0â95.9), a specificity of 96.7% (95% confidence interval = 94â2-99.3), and a positive predictive value of 78.5% (95% confidence interval = 74.2â82.8) for all advanced cSCC combined. The sensitivity was lower per subgroup: locally advanced (52.3â86.2%), recurrent cSCC (23.3%), and metastatic cSCC (70.0%). The specificity per subgroup was above 97%, and the positive predictive value was above 78%, with the exception of metastatic cSCC, which had a positive predictive value of 62%. This algorithm can be used to identify advanced patients with cSCC from pathology reports and will facilitate large-scale epidemiological studies of advanced cSCC in the Netherlands and internationally after external validation.
Carcinoma, Squamous Cell , Skin Neoplasms , Humans , Skin Neoplasms/diagnosis , Skin Neoplasms/epidemiology , Skin Neoplasms/pathology , Carcinoma, Squamous Cell/diagnosis , Carcinoma, Squamous Cell/epidemiology , Carcinoma, Squamous Cell/pathology , Neoplasm Recurrence, Local/epidemiology , Registries , Algorithms
BACKGROUND: Incomplete excision of squamous cell carcinoma (cSCC) is associated with an increased risk of recurrence, metastasis, and mortality. OBJECTIVE: To determine the rate and characteristics of incompletely excised cSCC in a dermatological daily practice setting. METHODS: Prospective study of all patients who gave informed consent, with a cSCC treated with standard excision (SE) at 1 of 6 Departments of Dermatology in the Netherlands between 2015 and 2017. Pathological reports were screened to detect all incompletely excised cSCCs. RESULTS: A total of 592 patients with 679 cSCCs were included, whereby most cases were low risk cSCC (89%). The rate of incompletely excised cSCC was 4% ( n = 26), and the majority were high-risk cSCCs of which 24 invaded the deep excision margin. CONCLUSION: This prospective study showed that in a dermatological setting, the risk of an incompletely excised cSCC is low (4%) for a cohort that was dominated by low-risk cSCCs. Most incompletely excised cSCCs were of high risk, and incompleteness was almost always at the deep margins. These results suggest that for high-risk cSCC, one should pay attention especially to the deep margin when performing SE, and/or microscopic surgery should be considered.
Carcinoma, Squamous Cell , Skin Neoplasms , Humans , Carcinoma, Squamous Cell/pathology , Prospective Studies , Skin Neoplasms/pathology , Margins of Excision , Cohort Studies
Background: The current standard of care for patients without sentinel node (SN) metastasis (i.e., stage I−II melanoma) is watchful waiting, while >40% of patients with stage IB−IIC will eventually present with disease recurrence or die as a result of melanoma. With the prospect of adjuvant therapeutic options for patients with a negative SN, we assessed the performance of a clinicopathologic and gene expression (CP-GEP) model, a model originally developed to predict SN metastasis, to identify patients with stage I−II melanoma at risk of disease relapse. Methods: This study included patients with cutaneous melanoma ≥18 years of age with a negative SN between October 2006 and December 2017 at the Sahlgrenska University Hospital (Sweden) and Erasmus MC Cancer Institute (The Netherlands). According to the CP-GEP model, which can be applied to the primary melanoma tissue, the patients were stratified into high or low risk of recurrence. The primary aim was to assess the 5-year recurrence-free survival (RFS) of low- and high-risk CP-GEP. A secondary aim was to compare the CP-GEP model with the EORTC nomogram, a model based on clinicopathological variables only. Results: In total, 535 patients (stage I−II) were included. CP-GEP stratification among these patients resulted in a 5-year RFS of 92.9% (95% confidence interval (CI): 86.4−96.4) in CP-GEP low-risk patients (n = 122) versus 80.7% (95%CI: 76.3−84.3) in CP-GEP high-risk patients (n = 413; hazard ratio 2.93 (95%CI: 1.41−6.09), p < 0.004). According to the EORTC nomogram, 25% of the patients were classified as having a 'low risk' of recurrence (96.8% 5-year RFS (95%CI 91.6−98.8), n = 130), 49% as 'intermediate risk' (88.4% 5-year RFS (95%CI 83.6−91.8), n = 261), and 26% as 'high risk' (61.1% 5-year RFS (95%CI 51.9−69.1), n = 137). Conclusion: In these two independent European cohorts, the CP-GEP model was able to stratify patients with stage I−II melanoma into two groups differentiated by RFS.
