H&NSCC SCOPE

Prediction of metastasis in head and neck squamous cell carcinoma

Issue

Head and Neck Squamous Cell Carcinoma (HNSCC) is the 6th most common cancer globally, creating a heavy burden on healthcare systems. A major issue is that the cancer can spread to other parts of the body (metastasis), which often leads to high death rates. Many patients only receive care to manage symptoms because treatment options are limited once the cancer has spread. Detecting the risk of metastasis early is critical to providing timely treatment and improving survival rates.

Approach

A new diagnostic tool has been developed that uses advanced artificial intelligence (AI) to predict the likelihood of the cancer spreading in HNSCC patients. This AI analyzes tissue samples from patients using deep learning technology to spot important signs that indicate a high risk of metastasis. (The technical stuff: By analyzing histopathological Hematoxylin and Eosin (H&E) stained Whole Slide Images (WSIs), the framework leverages a Self-Attention mechanism combined with a Residual Neural Network (ResNet) to identify key features indicative of metastasis risk.)

Result

This AI-powered tool helps doctors detect the risk of metastasis early and provides treatment suggestions. It has shown impressive results, with a 92.14% accuracy rate (an AUC ‘Area under the Curve’ of 90.61 and an F1 score of 0.92) . By identifying patients at higher risk early on, this tool can help healthcare providers take action sooner, leading to better treatment outcomes for patients. Efforts are underway to expand this AI technology by including more patient data (retrospective and prospective) to further improve early detection and intervention, ultimately helping save more lives.