Hyperspectral Imaging in Gynaecological Oncology (HSI-GO)

Aim of the project

Not all tumours have a clearly visible outline. Often, there is a conflicting pull on surgeons to excise tumours broadly to ensure surgical completeness and reduce chances of cancer recurrence, while avoiding the removal of healthy functional tissues, and minimizing complications and postoperative impacts on the patient’s quality of life.

Imaging methods can assist surgeons to objectively delineate tumour volumes. Hyperspectral imaging (HSI) is a novel, wide-field, and label-free technique with promising results in cancer detection. In the HSI-GO study, we aim to perform translational research to effectively implement HSI in gynaecological oncology surgical practice.

Execution of the project

The HSI camera captures a series of images of the underlying tissue in contiguous spectral bands. For each pixel in the image, an intensity spectrum or spectral signature is derived. Image registration methods are used to overlay HS images with histopathological scans, in which tissue type annotations are provided by a pathologist. This way, spectra belonging tothe same tissue type can be clustered, and artificial intelligence (AI) can be used to estimate the most probable cluster (tissue type) to which a newly obtained spectrum belongs. Such an algorithm provides near real-time HSI-based diagnostic feedback on tissue types.

Two indications have been selected for which the clinical gain of HSI is particularly promising:

1. Ovarian cancer is often detected a late stage. As a result, the disease may have spread to other tissues and organs. This can make ovarian cancer surgeries extensive and complex. A wide-field imaging technique to inspect tissues - for instance series of spots on the colon - will improve efficiency, and may profoundly impact the patient’s recovery and quality of life (e.g. the need for a stoma).
2. Vulvar cancer is characterized by high recurrence rates and surgical margins can be difficult to assess, in particular in the presence of lichen sclerosus or pre-malignant changes in the vulvar skin. It will be of value to objectively determine whether midline structures, such as the urethra, clitoris, anal sphincter can be spared, without increasing the risk on recurrence.

Impact on patient care

The camera offers a label-free and non-contact method to evaluate tissue types. At this moment, HSI data will become available only after surgeries, and will not be used for clinical decision-making. The impact on the clinical workflow (during these pre-clinical and translational validations) is therefore very low.

Collaborations within Erasmus MC

• The department of Gynaecological Oncology
• The department of Pathology

• Laurie van de Weerd, PhD position, biomedical engineer
• Nick van de Berg, biomedical engineer
• Heleen van Beekhuizen, gynaecological oncologist
• Lucia Rijstenberg, pathologist
• Gatske Nieuwenhuyzen-de Boer, gynaecological oncologist
• Lena van Doorn, gynaecological oncologist
• Eva Maria Roes, gynaecological oncologist
• Ralf van de Laar, gynaecological oncologist
• Marianne Maliepaard, research nurse