Predicting Radiological Extranodal Extension in Oropharyngeal Carcinoma Patients Using AI

Status

Active, not recruiting

Phases

Unknown

Study type

Observational

Source

ClinicalTrials.gov

Registry ID

NCT05565313

Acronym

AI4rENE

Enrollment

900

Registered

2022-10-04

Start date

2022-03-22

Completion date

2026-08-01

Last updated

2025-08-14

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Head and Neck Carcinoma

Brief summary

Development and validation of a model that predicts rENE from radiological imaging using annotated / labeled scans by means of deep learning

Detailed description

Oropharyngeal squamous cell carcinoma (OPSCC) is a rare cancer (incidence \ 700 per year in the Netherlands), originating in the middle part of the throat. In OPSCC, nodal status is an important prognostic factor for survival. In the clinical TNM (tumor node metastases) system, nodal status is mainly defined by the size, number and laterality of nodal metastases. In surgically treated patients the pathological TNM classification includes the presence of pathological extranodal extension (pENE). pENE is a predictor for poor outcome and also an indication for the addition of chemotherapy to postoperative radiation. However, most patients with OPSCC are treated non-surgically by means of radiation or chemoradiation and thus information about pENE is lacking. Recently, extranodal extension on diagnostic imaging has been associated with prognosis in OPSCC patients. It is anticipated that in the near future radiological ENE (rENE) may be included in the cTNM classification system for refinement of outcome prediction in patients with nodal disease. The diagnosis of rENE on radiological imaging is new and not trivial and we hypothesize that Artificial Intelligence (AI) may support the radiologist in detecting rENE. In this study we aim to develop and validate a model that predicts rENE from radiological imaging using annotated / labeled scans by means of deep learning

Interventions

None listed

Study design

Observational model

COHORT

Time perspective

RETROSPECTIVE

Eligibility

Sex/Gender

ALL

Age

18 Years to No maximum

Healthy volunteers

Inclusion criteria

* Non-metastatic (M0) node-positive HPV+ and HPV- oropharyngeal carcinoma * Treated between 2008 to 2019 * Curative intent * Radiation only or concurrent chemoradiation * Modern treatment modality: IMRT / VMAT * diagnostic/staging image scanning protocols available (contrast-enhanced CT with 2-3 mm slice thickness and/or MR with 3 mm slice thickness)

Exclusion criteria

* removal of lymph node (LN) (excisional biopsy or neck dissection \[ND\]) prior to staging CT/MR scan * no available imaging within 2 months prior to radiotherapy (RT)

Design outcomes

Primary

Measure	Time frame	Description
Prediction of rENE as labeled by the radiologist, using the AI model	Baseline	The performance of the model will be evaluated in terms of discrimination through the Harrell's C-index and the area (AUC) under the receiver operator curve (ROC) in predicting rENE.

Secondary

Measure	Time frame	Description
Overall Survival	5 years	Percentage of people who are alive five years after their diagnosis.
Disease Free Survival	5 years	Percentage of people whp who are disease free five years after their diagnosis.

Countries

Canada, Netherlands, United States

Outcome results