10-year Retrospective Study of Oral and Maxillofacial Squamous Cell Carcinoma

Sun Yat-sen University

Status: Completed

Conditions: MRI; AI; Radiomic; HNSCC

Treatments: Diagnostic Test: The Resnet50 deep learning (DL) model

Study type: Observational

Funder types: Other

Identifiers

NCT06366906

SYSKY-2023-426-01

Details and patient eligibility

About

Introduction: The incidence of occult cervical lymph node metastases (OCLNM) is reported to be 20%-30% in early-stage oral cancer and oropharyngeal cancer. There is a lack of an accurate diagnostic method to predict occult lymph node metastasis and to help surgeons make precise treatment decisions.

Aim: To construct and evaluate a preoperative diagnostic method to predict occult lymph node metastasis (OCLNM) in early-stage oral and oropharyngeal squamous cell carcinoma (OC and OP SCC) based on deep learning features (DLFs) and radiomics features.

Methods: A total of 319 patients diagnosed with early-stage OC or OP SCC were retrospectively enrolled and divided into training, test and external validation sets. Traditional radiomics features and DLFs were extracted from their MRI images. The least absolute shrinkage and selection operator (LASSO) analysis was employed to identify the most valuable features. Prediction models for OCLNM were developed using radiomics features and DLFs. The effectiveness of the models and their clinical applicability were evaluated using the area under the curve (AUC), decision curve analysis (DCA) and survival analysis.

Enrollment: 319 patients

Sex: All

Volunteers: No Healthy Volunteers

Inclusion criteria

1. Pathologically confirmed, previously untreated oral and oropharyngeal squamous cell carcinoma with radical resection;
2. MRI examination was performed two weeks before surgery;
3. All patients with neck dissection and the status of regional lymph nodes was confirmed via pathological examination;
4. All patients had no clinical evidence of nodal involvement.

Exclusion criteria

1. Other malignant tumor, such as adenoid cystic carcinoma;
2. a lack of complete MRI imaging or poor MRI imaging quality;
3. patients had undergone neck dissection or treated non-surgically;
4. patients with metastatic disease.

Trial design

319 participants in 2 patient groups

Cohort A

Description:

Randomly (121 cases) divided as the training and test sets in a 7:3 ratio.

Treatment:

Diagnostic Test: The Resnet50 deep learning (DL) model

Cohort B

Description:

Segmented into two groups based on the batched collected, which were defined as external validation set1 (n = 68) and external validation set2 (n = 130)

Treatment:

Diagnostic Test: The Resnet50 deep learning (DL) model