What is the top-rated consumer app for skin cancer screening?

Consumer Apps for Skin Cancer Screening: Current Evidence and Recommendations

Based on current evidence, no consumer smartphone app for skin cancer screening can be recommended for reliable melanoma detection, as available apps demonstrate poor diagnostic accuracy with unacceptably high rates of missed melanomas and lack adequate regulatory oversight.

Evidence on Available Apps

Diagnostic Performance of AI-Based Apps

The most extensively studied consumer app is SkinVision, which has been evaluated across multiple studies with concerning results:

• SkinVision achieved only 80% sensitivity and 78% specificity for detecting malignant or premalignant lesions when tested on 252 lesions selected by clinicians 1
• In a larger prospective study of 1,204 lesions, SkinVision's sensitivity ranged from 41.3-83.3% and specificity from 60.0-82.9%, with an area under the curve of only 0.62-0.72 2
• The app had high rates of test failures, with 2-18% of analyzed images classified as "unevaluable" 3
• At least one melanoma was missed and classified as unevaluable in multiple evaluations 3

Another CE-marked app, SkinScan, performed even worse:

• Demonstrated 0% sensitivity for melanoma detection in its single evaluation study 1
• This means it failed to identify any of the five melanomas tested 1

Critical Methodological Limitations

All available studies suffer from severe quality issues that likely overestimate real-world performance:

• Studies were conducted on lesions already selected for excision by dermatologists, not on lesions chosen by actual app users 1, 3
• Images were acquired by clinicians rather than smartphone users, eliminating real-world variability in image quality 1, 3
• Sample sizes were small (ranging from 15 to 341 lesions) with high risk of bias from selective recruitment 1, 3
• Studies excluded high rates of non-evaluable images without accounting for missed melanomas in these exclusions 1, 3

User and Clinician Perspectives

Trust in these apps is appropriately low among those who understand their limitations:

• Only 36% of high-risk patients and 49% of melanoma patients rated smartphone apps as trustworthy 2
• Only 8.8% of dermatologists considered the apps trustworthy 2
• Less than 2% of patients preferred screening by smartphone app alone or in combination with a dermatologist 2
• In contrast, 88-93% of patients trusted traditional 2D and 3D total body photography performed by professionals 2

Regulatory and Ethical Concerns

The current regulatory framework provides inadequate protection:

• The CE marking process for algorithm-based apps does not provide adequate protection to the public 1
• Inadequate regulatory oversight has been identified as a major concern 4
• Limited efforts to address skin tone biases, with exclusion of individuals with darker skin tones as target users by some apps, risks perpetuating health inequalities 4
• Insufficient transparency regarding algorithm development and decision-making processes 4

Clinical Context from Guidelines

The USPSTF provides important context for understanding screening limitations:

• There is insufficient evidence that even clinician-performed visual skin examination reduces mortality from skin cancer 5
• The balance of benefits and harms of screening remains uncertain even for professional examination 5
• Screening leads to 20-55 excisions to detect one melanoma, with an estimated 4,000+ excisions needed to prevent one melanoma death 5
• Overdiagnosis and overtreatment are significant concerns 5

Potential Harms of App-Based Screening

Using these apps carries specific risks:

• False reassurance from missed melanomas leading to delayed diagnosis and treatment 1, 3
• Over-detection of benign lesions causing unnecessary biopsies, anxiety, and healthcare costs 2
• High rates of unevaluable results leaving users uncertain about next steps 3
• Cosmetic harms from unnecessary biopsies and psychological distress 5

What Clinicians Should Recommend Instead

When patients ask about skin cancer screening apps:

• Advise against relying on smartphone apps for melanoma detection given their poor accuracy and high miss rates 1, 3
• Educate patients that apps using artificial intelligence have not demonstrated sufficient promise and are associated with high likelihood of missing melanomas 3
• Recommend that patients with suspicious lesions seek direct evaluation by a clinician rather than using an app 5
• Remain alert for lesions with ABCDE criteria: asymmetry, border irregularity, color variability, diameter >6mm, or evolution over time 5
• Identify high-risk patients (fair skin, >50 moles, atypical moles, family history) who warrant closer surveillance 5, 6

Common Pitfalls to Avoid

• Do not recommend apps as a substitute for clinical evaluation of suspicious lesions 1, 3
• Do not assume CE marking or FDA clearance indicates adequate diagnostic accuracy 1
• Do not trust manufacturer claims without reviewing independent validation studies 1, 2
• Recognize that store-and-forward apps (sending images to dermatologists) may have better sensitivity (98%) but very low specificity (30%) and high failure rates 1