Scholarlink

<p>Skin colour is known to confound optical devices, adversely impacting care for patients with darker skin. Photoacoustic imaging (PAI) combines optics and ultrasound for deep tissue imaging, creating a complex relationship between PAI-derived biomarkers and skin melanin concentration, yet no generalisable bias correction has been demonstrated. Drawing on a healthy volunteer cohort of 42 participants spanning Fitzpatrick types I–VI and vitiligo – the most diverse ever assembled in PAI – we characterise optical and acoustic mechanisms driving skin colour-dependent degradation in image quality and biomarker quantification. Wavelength-dependent melanin absorption causes spectral colouring, dominating at low melanin levels, while epidermal ultrasound backscattering dominates at high melanin levels, producing a non-linear relationship between unmixed sO₂ and skin tone. Leveraging this understanding, we propose a practical spectral colouring correction and adapt a plane-wave reconstruction algorithm to resolve backscattered ultrasound artefacts. Our findings underscore the need for advanced reconstruction methods to enable equitable clinical PAI.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The confounding effects of skin colour in photoacoustic imaging

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月17日

查看全文
  • Thomas R. Else,
  • Christine Loreno,
  • Alice Groves,
  • Benjamin T. Cox,
  • Evangelia Vetsiou,
  • Janek Gröhl,
  • Inés Modolell,
  • Sarah E. Bohndiek,
  • Amit Roshan

摘要

Skin colour is known to confound optical devices, adversely impacting care for patients with darker skin. Photoacoustic imaging (PAI) combines optics and ultrasound for deep tissue imaging, creating a complex relationship between PAI-derived biomarkers and skin melanin concentration, yet no generalisable bias correction has been demonstrated. Drawing on a healthy volunteer cohort of 42 participants spanning Fitzpatrick types I–VI and vitiligo – the most diverse ever assembled in PAI – we characterise optical and acoustic mechanisms driving skin colour-dependent degradation in image quality and biomarker quantification. Wavelength-dependent melanin absorption causes spectral colouring, dominating at low melanin levels, while epidermal ultrasound backscattering dominates at high melanin levels, producing a non-linear relationship between unmixed sO₂ and skin tone. Leveraging this understanding, we propose a practical spectral colouring correction and adapt a plane-wave reconstruction algorithm to resolve backscattered ultrasound artefacts. Our findings underscore the need for advanced reconstruction methods to enable equitable clinical PAI.