Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Towards Fairness in AI for Melanoma Detection: Systemic Review and Recommendations

Laura N Montoya, Jennafer Shae Roberts, Belen Sanchez Hidalgo

TL;DR

It is indicated that while AI can enhance melanoma detection, there is a significant bias towards lighter skin tones and the need for diverse datasets and robust evaluation metrics to develop AI models that are equitable and effective for all patients.

Abstract

Early and accurate melanoma detection is crucial for improving patient outcomes. Recent advancements in artificial intelligence AI have shown promise in this area, but the technologys effectiveness across diverse skin tones remains a critical challenge. This study conducts a systematic review and preliminary analysis of AI based melanoma detection research published between 2013 and 2024, focusing on deep learning methodologies, datasets, and skin tone representation. Our findings indicate that while AI can enhance melanoma detection, there is a significant bias towards lighter skin tones. To address this, we propose including skin hue in addition to skin tone as represented by the LOreal Color Chart Map for a more comprehensive skin tone assessment technique. This research highlights the need for diverse datasets and robust evaluation metrics to develop AI models that are equitable and effective for all patients. By adopting best practices outlined in a PRISMA Equity framework tailored for healthcare and melanoma detection, we can work towards reducing disparities in melanoma outcomes.

Towards Fairness in AI for Melanoma Detection: Systemic Review and Recommendations

TL;DR

It is indicated that while AI can enhance melanoma detection, there is a significant bias towards lighter skin tones and the need for diverse datasets and robust evaluation metrics to develop AI models that are equitable and effective for all patients.

Abstract

Early and accurate melanoma detection is crucial for improving patient outcomes. Recent advancements in artificial intelligence AI have shown promise in this area, but the technologys effectiveness across diverse skin tones remains a critical challenge. This study conducts a systematic review and preliminary analysis of AI based melanoma detection research published between 2013 and 2024, focusing on deep learning methodologies, datasets, and skin tone representation. Our findings indicate that while AI can enhance melanoma detection, there is a significant bias towards lighter skin tones. To address this, we propose including skin hue in addition to skin tone as represented by the LOreal Color Chart Map for a more comprehensive skin tone assessment technique. This research highlights the need for diverse datasets and robust evaluation metrics to develop AI models that are equitable and effective for all patients. By adopting best practices outlined in a PRISMA Equity framework tailored for healthcare and melanoma detection, we can work towards reducing disparities in melanoma outcomes.

Paper Structure

This paper contains 20 sections, 4 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Works Supporting Rationale
  3. Assessing the Generalizability of Deep Neural Networks-Based Models for Black Skin Lesions Barros2023AssessingLesions
  4. Fairness of Classifiers Across Skin Tones in Dermatology KinyanjuiFairnessDermatology
  5. Revisiting Skin Tone Fairness in Dermatological Lesion Classification Kalb2023RevisitingClassification
  6. Skin Deep: Investigating Subjectivity in Skin Tone Annotations for Computer Vision Benchmark Datasets Barrett2023SkinDatasets
  7. Out-of-Distribution Detection in Dermatology using Input Perturbation and Subset Scanning Kim2021Out-of-DistributionScanning
  8. Lack of Transparency and Potential Bias in Artificial Intelligence Data Sets and Algorithms A Scoping Review Daneshjou2021LackAlgorithms
  9. Materials and Methodology
  10. Objectives
  11. Eligibility Criteria
  12. Information Sources and Search Strategy
  13. Data Collection and Analysis
  14. Limitations
  15. Results
  16. ...and 5 more sections

Figures (4)

  • Figure 1: Depicts how hue undertone can affect skin color dimension. Image altered to protect model privacy and the text was translated from Portuguese to English. [Public domain], via Google Search for beauty palette demonstration. (https://i.pinimg.com/564x/1a/5c/18/1a5c18940a453ac538708c5562834ff6.jpg). Skin tone palette range with Values (light, medium, dark), and Hues (cold, neutral, warm, olive) depicted as a range of dimensionality.
  • Figure 2: Number of publications related to AI and melanoma published during the last 20 years. The x-axis represents the years, and the y-axis represents the number of publications per year. The bins of the histogram correspond to individual years.
  • Figure 3: Number of citations related to publications about AI and melanoma over time from 2004 to 2023. The x-axis denotes the years, while the y-axis indicates the total number of citations for that year.
  • Figure 4: Comparison of skin tone scales that can be used for skin cancer detection utilizing AI. Recreation of Fitzpatrick Skin Type Scale, Monk Skin Tone Scale, and Sampling of L'Oreal Color Chart Map for reference.