Image Quality Assessment and Reliability Analysis of Artificial Intelligence-Based Tumor Classification of Stimulated Raman Histology of Tumor Biobank Samples

• DOI: 10.3390/diagnostics14232701
• Publication Date: 2024-12-02T14:00:10Z
• Authors: Anna-Katharina Me...; Tobias Blau; David Reinecke; Gina Fürtjes; Lili Leyer; Nina Müller; Niklas von Spreck...; Thomas Stehle; Abdulkader Al Shugri; Reinhard Büttner; Roland Goldbrunner; Marco Timmer; Volker Neuschmelting
• Keywords: Histology; Quality Assessment

Stimulated Raman histology (SRH) is a label-free optical imaging method for rapid intraoperative analysis of fresh tissue samples. Analysis SRH images using Convolutional Neural Networks (CNN) has shown promising results predicting the main histopathological classes neurooncological tumors. Due to relatively low number rare tumor representations in CNN training datasets, valid prediction rarer entities remains limited. To develop new reliable tools, larger datasets and greater variety are crucial. One way accomplish this through research biobanks storing frozen However, there currently no data available regarding pertinency previously samples analysis. The aim study was assess image quality perform comparative reliability artificial intelligence-based classification