Investigating multi-radiomic models for enhancing prediction power of cervical cancer treatment outcomes
| Autor: | Young-Chul Kim, Kujtim Latifi, B.A. Altazi, Puja Venkat, Samuel H. Hawkins, Eduardo G. Moros, D.C. Fernandez, Geoffrey Zhang, Matthew C. Biagioli, Syeda Mahrukh Hussnain Naqvi, Dylan Hunt |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Oncology
Adult medicine.medical_specialty Biophysics General Physics and Astronomy Uterine Cervical Neoplasms Standardized uptake value Feature selection Logistic regression Cross-validation Article 030218 nuclear medicine & medical imaging 03 medical and health sciences 0302 clinical medicine Internal medicine Positron Emission Tomography Computed Tomography medicine Humans Radiology Nuclear Medicine and imaging Aged Cervical cancer Aged 80 and over Models Statistical Receiver operating characteristic medicine.diagnostic_test business.industry General Medicine Middle Aged medicine.disease Tumor Burden Logistic Models Treatment Outcome Positron emission tomography Feature (computer vision) 030220 oncology & carcinogenesis Female business |
| Zdroj: | Phys Med |
| ISSN: | 1724-191X |
| Popis: | Quantitative image features, also known as radiomic features, have shown potential for predicting treatment outcomes in several body sites. We quantitatively analyzed (18)Fluorine–fluorodeoxyglucose ((18)F-FDG) Positron Emission Tomography (PET) uptake heterogeneity in the Metabolic Tumor Volume (MTV) of eighty cervical cancer patients to investigate the predictive performance of radiomic features for two treatment outcomes: the development of distant metastases (DM) and loco-regional recurrent disease (LRR). We aimed to fit the highest predictive features in multiple logistic regression models (MLRs). To generate such models, we applied backward feature selection method as part of Leave-One-Out Cross Validation (LOOCV) within a training set consisting of 70% of the original patient cohort. The trained MLRs were tested on an independent set consisted of 30% of the original cohort. We evaluated the performance of the final models using the Area under the Receiver Operator Characteristic Curve (AUC). Accordingly, six models demonstrated superior predictive performance for both outcomes (four for DM and two for LRR) when compared to both univariate-radiomic feature models and Standard Uptake Value (SUV) measurements. This demonstrated approach suggests that the ability of the preradiochemotherapy PET radiomics to stratify patient risk for DM and LRR could potentially guide management decisions such as adjuvant systemic therapy or radiation dose escalation. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect