Deep Learning-based Classification and Prediction of Radiation Dermatitis in Head and Neck Patients

Recruiting

• Conditions: Head and Neck Cancer; Radiation Dermatitis

• Registration Number: NCT05607225
• Lead Sponsor: Cancer Institute and Hospital, Chinese Academy of Medical Sciences

Brief Summary

to develop a deep learning-based model to grade the severity of radiation dermatitis (RD) and predict the severity of radiation dermatitis in patients with head and neck cancer undergoing radiotherapy, so as to provide support for doctors' diagnosis and prediction.

Detailed Description

1. Image acquisition The images of the neck area were collected from the enrolled patients one week before and every week during radiotherapy. The photographs were taken from three angles (front, left and right oblique) of the neck area.

2. Grading evaluation Each image was individually graded by three experienced radiotherapy experts according to the RD criteria of RTOG

3. Data analysis Construct a dermatitis grading model basing on deep learning. Evaluate the performance of model using accuracy, precision, recall, F1-measure, dice value.

Study Timeline

• Study Start: 2022-07-01
• Status Verified: 2022-10
• Study First Submitted: 2022-10-24
• Study First Submitted QC: 2022-10-31
• Last Update Submitted: 2022-10-31
• Study First Posted: 2022-11-07
• Last Update Posted: 2022-11-07
• Primary Completion: 2025-06-30
• Study Completion: 2025-06-30

Recruitment & Eligibility

• Status: RECRUITING
• Sex: All
• Target Recruitment: 300

Inclusion Criteria

• Age ≥ 18 years old.
• Histologically or cytologically confirmed head and neck carcinoma confirmed by pathology.
• Receive radical radiotherapy including neck area
• Informed consent.

Exclusion Criteria

• unable to cooperate with image acquisition

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
Precision	July 1, 2022 to June 30, 2025	The proportion of positive samples in the positive prediction result
ROC curve	July 1, 2022 to June 30, 2025
Accuracy	July 1, 2022 to June 30, 2025	Evaluate the rate of deep learning based rating model in accordance with experts' assessment.
F1-measure	July 1, 2022 to June 30, 2025	The harmonic average of precision and recall
Recall	July 1, 2022 to June 30, 2025	The proportion of positive samples that were predicted to be positive

Secondary Outcome Measures

Name	Time	Method
dice value	July 1, 2022 to June 30, 2025	Ratio of overlap and distance between artificial and automatic neck segmentation regions

Trial Locations

Locations (2)

Shenzhen Cancer Hospital, Chinese Academy of Medical Sciences; 🇨🇳 Shenzhen, Guangdong, China

Cancer Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College; 🇨🇳 Beijing, China