Nasopharyngeal Carcinoma
Conditions
Interventions
Training and Validation:Observational study with no interventions. For model development and training.
Standalone Testing:Observational study with no interventions. Also applicable to the multi-reader, multi-case study.
Clinical validation group:Conduct a multi-reader, multi-case (MRMC) clinical validation study, based on an independent test set of 200 cases, by inviting multiple radiologists to participate in a full
Sponsors
Eye and ENT Hospital of Fudan University
Eligibility
Sex/Gender
All
Age
18 Years to 100 Years
Inclusion criteria
Inclusion criteria: 1. Age >= 18 years; 2. Patients with suspected or newly diagnosed, untreated nasopharyngeal carcinoma (NPC) at the time of consultation, or control individuals diagnosed as non-NPC (including adenoidal hypertrophy, chronic inflammation/hyperplasia, etc.); 3. Availability of a complete head and neck MRI with contrast enhancement (T1WI, T2WI, T1c); 4. Availability of a reference standard: NPC-positive cases confirmed by histopathology, and NPC-negative cases confirmed by negative follow-up (>=6 months) or histopathology showing non-neoplastic lesions.
Exclusion criteria
Exclusion criteria: 1. History of head and neck malignancy, nasopharyngeal surgery, or radiotherapy/chemotherapy; 2. Severe anatomical variations or extensive postoperative changes that interfere with nasopharyngeal assessment; 3. Contraindications to MRI or incomplete key clinical information (e.g., missing age, sex, or diagnostic conclusion); 4. Incomplete MRI sequences, insufficient field of view, or poor image quality that impairs image interpretation; 5. Any other condition deemed unsuitable for inclusion by the investigators.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The model achieved the target sensitivity and specificity of 80% in internal testing;In clinical validation, the sensitivity and specificity of T1+T2 combined with AI assistance achieved non-inferiority (non-inferiority margin: –10%);In clinical validation, the AUC of T1+T2 combined with AI assistance achieved non-inferiority (non-inferiority margin: –5%); | — |
Secondary
| Measure | Time frame |
|---|---|
| Derived metrics in internal testing of the model, including AUC, positive predictive value (PPV), and negative predictive value (NPV);Sensitivity, specificity, and AUC of the model under different comparative tests;Sensitivity, specificity, and AUC in clinical validation across different reader types and T-stage case subgroup analyses; | — |
Countries
China
Contacts
Public ContactYu Hongmeng
Eye and ENT Hospital of Fudan University
Outcome results
None listed