Intraoperative Assessment of Tumor Margins in Tissue Sections with Hyperspectral Imaging and Machine Learning

Abstract

The intraoperative assessment of tumor margins of head and neck cancer is crucial for complete tumor resection and patient outcome. The current standard is to take tumor biopsies during surgery for frozen section analysis by a pathologist after H&E staining. This evaluation is time-consuming, subjective, methodologically limited and underlies a selection bias. Optical methods such as hyperspectral imaging (HSI) are therefore of high interest to overcome these limitations. We aimed to analyze the feasibility and accuracy of an intraoperative HSI assessment on unstained tissue sections taken from seven patients with oral squamous cell carcinoma. Afterwards, the tissue sections were subjected to standard histopathological processing and evaluation. We trained different machine learning models on the HSI data, including a supervised 3D convolutional neural network to perform tumor detection. The results were congruent with the histopathological annotations. Therefore, this approach enables the delineation of tumor margins with artificial HSI-based histopathological information during surgery with high speed and accuracy on par with traditional intraoperative tumor margin assessment (Accuracy: 0.76, Specificity: 0.89, Sensitivity: 0.48). With this, we introduce HSI in combination with ML hyperspectral imaging as a potential new tool for intraoperative tumor margin assessment.

Hyperspectral imaging and evaluation of surgical margins: where do we stand?

“…Different molecular compositions of tumor cells in contrast to nontumor cells result in different spectral characteristics. These characteristics can be used as spectral fingerprints for tumor tissue identification [10 ▪▪ ]. The high amount of data makes manual interpretation of HSI data nearly impossible, which is why automated image analysis and machine learning approaches are already routinely used to interpret HSI data [11].…”

“…In contrast, some of the recent studies worked with tissue sections instead of fresh specimens. Pertzborn et al [10 ▪▪ ] used unstained tissue sections for the HSI measurements before standard staining was performed (Fig. 1).…”

“…Example of hyperspectral imaging (HSI) on tissue sections as performed by Pertzborn et al [10 ▪▪ ]. (a) Unstained head and neck cancer tissue frozen section.…”

Facilitating clinically relevant skin tumor diagnostics with spectroscopy-driven machine learning

“…If an image does not contain the necessary spectral contrast in order to produce a clinically reliable diagnosis, 23 the data quality is automatically insufficient and alternative approaches should be considered. 24,25 In this work we present an alternative ML approach to pre-surgical skin tumor delineation that does not require manual identification of the tumor borders. Rather than analyzing spatial features in standard color images, our model instead builds on recognizing spectral patterns in hyperspectral images [26][27][28] containing information beyond the sensitivity of the naked eye.…”

“…If an image does not contain the necessary spectral contrast in order to produce a clinically reliable diagnosis, 23 the data quality is automatically insufficient and alternative approaches should be considered. 24,25…”