Hyperspectral Imaging for Resection Margin Assessment during Cancer Surgery

Area of Science:

• Oncology
• Medical Imaging
• Biomedical Engineering

Background:

• Accurate intraoperative margin assessment is crucial for complete tumor removal in cancer surgery.
• Current techniques for margin assessment lack sufficient accuracy, posing challenges in surgical procedures.
• Hyperspectral imaging (HSI) presents a potential solution for real-time margin assessment.

Purpose of the Study:

• To evaluate the efficacy of hyperspectral imaging (HSI) for intraoperative margin assessment in fresh human breast specimens.
• To develop and validate a classification algorithm for HSI data to differentiate between tumor and healthy tissue.
• To assess the clinical applicability of HSI in identifying malignancies on the resection surface during breast-conserving surgery.

Main Methods:

• Hyperspectral data were acquired from tissue slices of 18 patients and correlated with histopathology.
• A support vector machine classification algorithm was developed and tested using over 22,000 spectra.
• HSI was applied to the resection surface of six lumpectomy specimens to detect malignancies.

Main Results:

• The HSI classification algorithm achieved high diagnostic performance: 93% for invasive carcinoma, 84% for ductal carcinoma in situ, 70% for connective tissue, and 99% for adipose tissue.
• Imaging and data analysis of the resection surface were completed within 10 minutes, without requiring an experienced operator.
• HSI detected 19 of 20 malignancies located within 2 mm of the resection surface.

Conclusions:

• Hyperspectral imaging demonstrates significant potential for improving intraoperative margin assessment during breast-conserving surgery.
• The developed HSI technique offers a fast, accurate, and operator-independent method for real-time margin evaluation.
• Successful application of HSI can lead to enhanced surgical precision and improved patient outcomes by ensuring complete tumor removal.