Artificial Intelligence in Imaging-Based Extranodal Extension Detection: Advancing Prognostication in HPV-Positive Oropharyngeal Cancer

An AI-driven imaging pipeline was developed to automatically segment lymph nodes and classify extranodal extension (iENE) on pretreatment CT scans in HPV-positive oropharyngeal carcinoma (OPC).
The AI model achieved high accuracy (AUC 0.81) in detecting iENE compared to evaluations by expert neuroradiologists, addressing limitations of subjective imaging interpretation.
AI-predicted iENE was independently associated with worse overall survival, recurrence-free survival, and distant control, outperforming radiologist assessments in prognostic discrimination.
This approach offers a reproducible, scalable tool for risk stratification in HPV-positive OPC, highlighting the potential of AI to enhance personalized treatment planning.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.

An AI-driven imaging pipeline was developed to automatically segment lymph nodes and classify extranodal extension (iENE) on pretreatment CT scans in HPV-positive oropharyngeal carcinoma (OPC).
The AI model achieved high accuracy (AUC 0.81) in detecting iENE compared to evaluations by expert neuroradiologists, addressing limitations of subjective imaging interpretation.
AI-predicted iENE was independently associated with worse overall survival, recurrence-free survival, and distant control, outperforming radiologist assessments in prognostic discrimination.
This approach offers a reproducible, scalable tool for risk stratification in HPV-positive OPC, highlighting the potential of AI to enhance personalized treatment planning.

Sign in free to continue reading