CT radiomics nomogram for prediction of the Ki-67 index in head and neck squamous cell carcinoma

Objectives To construct and validate a contrast-enhanced computed tomography (CECT)–based radiomics nomogram to predict Ki-67 expression level in head and neck squamous cell carcinoma (HNSCC). Methods A total of 217 patients with HNSCC who underwent CECT scans and immunohistochemical examination of...

Full description

Saved in:
Bibliographic Details
Published in:European radiology 2023-03, Vol.33 (3), p.2160-2170
Main Authors: Zheng, Ying-mei, Chen, Jing, Zhang, Min, Wu, Zeng-jie, Tang, Guo-Zhang, Zhang, Yue, Dong, Cheng
Format: Article
Language:English
Subjects:
Computed tomography
Decision analysis
Decision making
Diagnostic Radiology
Effectiveness
Head & neck cancer
Head and Neck
Head and neck carcinoma
Head and Neck Neoplasms - diagnostic imaging
Humans
Imaging
Internal Medicine
Interventional Radiology
Ki-67 Antigen
Mathematical analysis
Medicine
Medicine & Public Health
Neuroradiology
Nomograms
Patients
Quality
Radiology
Radiomics
Squamous cell carcinoma
Squamous Cell Carcinoma of Head and Neck
Tomography, X-Ray Computed - methods
Training
Ultrasound
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Objectives To construct and validate a contrast-enhanced computed tomography (CECT)–based radiomics nomogram to predict Ki-67 expression level in head and neck squamous cell carcinoma (HNSCC). Methods A total of 217 patients with HNSCC who underwent CECT scans and immunohistochemical examination of their Ki-67 index were enrolled in this study. The patients were divided into a training set ( n = 140; Ki-67: ≥ 50% [ n = 72] and < 50% [ n = 68]) and an external test set ( n = 77; Ki-67: ≥ 50% [ n = 38] and < 50% [ n = 39]). The least absolute shrinkage and selection operator method was used to select key features for a CECT-image-based radiomics signature and a radiomics score (Rad-score) was calculated. A clinical model was established using clinical data and CT findings. The independent clinical factors and Rad-score were then combined to construct a radiomics nomogram. The performance characteristics of the Rad-score, clinical model, and nomogram were assessed using ROCs and decision curve analysis. Results Twenty features were finally selected to construct the Rad-score. The radiomics nomogram incorporating the Rad-score, low histological grade, and lymphatic spread showed higher predictive value for the Ki-67 index (≥ 50% vs. < 50%) than the clinical model on both the training (AUC, 0.919 vs. 0.648, p < 0.001) and test (AUC, 0.832 vs. 0.685, p = 0.030) sets. Decision curve analysis demonstrated that the radiomics nomogram was more clinically useful than the clinical model. Conclusions A CECT-based radiomics nomogram was constructed to predict the expression of Ki-67 in HNSCC. This model showed favorable predictive efficacy and might be useful for prognostic evaluation and clinical decision-making in patients with HNSCC. Key Points • Accurate pre-treatment prediction of Ki-67 index in HNSCC is crucial. • A CECT-based radiomics nomogram showed favorable predictive efficacy in estimation of Ki-67 expression status in HNSCC patients.
ISSN:1432-1084
0938-7994
1432-1084
DOI:10.1007/s00330-022-09168-6