文献速递烧伤创面修复

优化热成像用于烧伤深度临床评估与愈合监测

Optimising thermal imaging for clinical assessment of burn depth and healing monitoring.
PubMed 文献 · Burns · 2026-09-17

本研究在乌克兰哈尔科夫烧伤中心开展前瞻性纵向队列,纳入65例成人115处烧伤创面,利用自研红外相机分析伤后3天内温度参数并建立预后模型。相对温度阈值ΔTcut-off=-2.4°C可区分深二度及全层烧伤与浅度及混合烧伤,模型敏感度83.3%、特异度94.4%、AUC=0.942,与最终临床结局一致性显著(Cohen's ϰ=0.84)。

要点

  • 伤后3天内相对温度阈值ΔTcut-off=-2.4°C有助于区分需≥21天愈合的深二度/全层烧伤。
  • 热成像预后模型敏感度83.3%、特异度94.4%、AUC=0.942,与临床结局一致性显著。
  • 热图像处理可实现烧伤深度的客观评估及创面愈合的无创监测。

核心结论

  • 所提出的热图像处理可客观评估烧伤深度并实现创面愈合的无创监测。
  • 预后模型表现出高 prognostic 性能,与最终临床结局具有显著一致性。

研究设计与样本

  • 前瞻性纵向队列研究,在乌克兰哈尔科夫烧伤中心开展。
  • 纳入65例成人共115处烧伤创面,使用自研红外相机(阵列384×288、灵敏度0.07°C、空间分辨率1 mrad)。
  • 以伤后3天内记录的温度参数建立预后模型,以最终临床结局为参考标准,采用ROC分析评估预后性能。

主要发现

  • 65处创面中,48%在21天内愈合,52%未愈合或在21天内接受皮肤移植。
  • 相对温度阈值ΔTcut-off=-2.4°C(p < 0.00001)可区分深二度及全层烧伤(τ ≥ 21天)与浅度及混合烧伤。
  • 预后模型敏感度83.3%、特异度94.4%、AUC=0.942,预后与观察结局一致性显著(Cohen's ϰ = 0.84)。

关键数值与推荐

  • 相对温度阈值ΔTcut-off=-2.4°C(p < 0.00001)用于区分深二度及全层烧伤(τ ≥ 21天)与浅度及混合烧伤。
  • 预后模型敏感度83.3%、特异度94.4%、AUC=0.942。
  • 预后与观察结局一致性Cohen's ϰ = 0.84。

临床怎么用

  • 伤后3天内可使用红外热成像相对温度参数辅助判断烧伤深度及预测愈合时间。
  • ΔTcut-off=-2.4°C可作为区分深二度/全层与浅度/混合烧伤的参考阈值。
  • 热成像可用于无创监测创面愈合过程中的相对温度变化。

局限性

  • 需在更大规模的前瞻性队列中进一步验证,以支持标准化方案的制定。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: Infrared thermography (IRT) is a promising method for burn depth assessment and monitoring of wound healing. However, its clinical application is limited by the absence of standardised imaging and processing protocols. This study aimed to optimise thermal image analysis for heterogeneous burns and to develop a quantitative model for burn depth assessment and monitoring of wound progression using IRT. METHODS: A prospective longitudinal cohort study was conducted at a burn centre in Kharkiv, Ukraine. Patients were examined using a self-developed infrared camera (array 384 × 288, 0.07 °C sensitivity, 1 mrad spatial resolution). Temperature parameters of burns recorded within three days post-injury were used to develop a prognostic model, with final clinical outcome serving as the reference standard. Prognostic performance was assessed by ROC analysis. Burn healing was monitored by changes in the relative temperature of the wound. RESULTS: Sixty-five adults with 115 burn wounds were enrolled. Of 65 wounds assessed within the first three days, 48% healed within 21 days, while 52% did not heal or underwent skin grafting within 21 days. A relative temperature threshold of ΔTcut-off = -2.4 °C (p < 0.00001) distinguished deep dermal and full-thickness burns (τ ≥ 21 days) from superficial and mixed burns. The prognostic model demonstrated 83.3% sensitivity, 94.4% specificity, and AUC = 0.942. Agreement between prognostic and observed outcomes was substantial (Cohen's ϰ = 0.84). CONCLUSIONS: The proposed thermal image processing enabled objective assessment of burn depth and facilitated non-invasive monitoring of wound healing. The developed model showed high prognostic performance and substantial agreement with the final clinical outcomes. Further validation in larger prospective cohorts is needed to support the development of standardised protocols.

来源:Burns,查看原文。版权归原作者所有。

引用

  1. Fomenko Y, Shustakova G, Gordiyenko E, et al. Optimising thermal imaging for clinical assessment of burn depth and healing monitoring. Burns. 2026. PMID: 42879825. doi:10.1016/j.burns.2026.108230

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。