An exploratory assessment of early and delta PET radiomic features for outcome prediction in locally advanced cervical cancer

Anita Florit ^1,2, Wyanne A Noortman ³, Nicolò Bizzarri ⁴

¹Nuclear Medicine Unit, Department of Diagnostic Imaging and Oncological Radiotherapy, Fondazione Policlinico Universitario A. Gemelli IRCCS, Largo A. Gemelli, 8, 00168, Rome, Italy.
²Biomedical Photonic Imaging Group, University of Twente, Enschede, The Netherlands.
³Department of Medical Oncology, University Medical Center Groningen, Groningen, The Netherlands.
⁴Gynaecologic Oncology Unit, Department of Women, Children and Public Health Sciences, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy.
⁵Research Core Facility Data Collection G-STeP, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy.
⁶Section of Hygiene, University Department of Life Sciences and Public Health, Università Cattolica del Sacro Cuore, Rome, Italy.
⁷Section of Nuclear Medicine, University Department of Radiological Sciences and Haematology, Università Cattolica del Sacro Cuore, Rome, Italy.
⁸Section of Nuclear Medicine, Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands.
⁹Department of Radiation Science and Technology, Delft University of Technology, Delft, The Netherlands.
¹⁰Institute of Neuroscience and Medicine, INM-4, Forschungszentrum Jülich GmbH, Jülich, Germany.
¹¹Department of Radiology and Nuclear Medicine, Amsterdam UMC, Location VU University Medical Center, Amsterdam, The Netherlands.
¹²Radiation Oncology Unit, Department of Diagnostic Imaging and Oncological Radiotherapy, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy.
¹³Section of Radiology, University Department of Radiological Sciences and Haematology, Università Cattolica del Sacro Cuore, Rome, Italy.
¹⁴Gynaecopathology Unit, Department of Women, Children and Public Health Sciences, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy.
¹⁵Section of Pathology, University Department of Life Sciences and Public Health, Università Cattolica del Sacro Cuore, Rome, Italy.
¹⁶Section of Obstetrics and Gynaecology, University Department of Life Sciences and Public Health, Università Cattolica del Sacro Cuore, Rome, Italy.
¹⁷Advanced Radiodiagnostics Centre, Department of Diagnostic Imaging and Oncological Radiotherapy, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy.
¹⁸Nuclear Medicine Unit, Department of Diagnostic Imaging and Oncological Radiotherapy, Fondazione Policlinico Universitario A. Gemelli IRCCS, Largo A. Gemelli, 8, 00168, Rome, Italy. vittoria.rufini@unicatt.it.
¹⁹Section of Nuclear Medicine, University Department of Radiological Sciences and Haematology, Università Cattolica del Sacro Cuore, Rome, Italy. vittoria.rufini@unicatt.it.

⁰Nuclear Medicine Unit, Department of Diagnostic Imaging and Oncological Radiotherapy, Fondazione Policlinico Universitario A. Gemelli IRCCS, Largo A. Gemelli, 8, 00168, Rome, Italy.

European Journal of Nuclear Medicine and Molecular Imaging +

|

June 18, 2025

View abstract on PubMed

Summary

Radiomic features from [18F]FDG-PET scans did not predict disease-free survival in locally advanced cervical cancer patients. While early and delta radiomic features showed slightly improved prediction for overall survival, their clinical utility remains limited.

Area Of Science

Oncology
Radiology
Medical Imaging

Background

Locally advanced cervical cancer (LACC) requires effective prognostic markers.
Neoadjuvant chemoradiotherapy (CRT) followed by surgery is a standard treatment for LACC.
Predicting treatment response and patient outcomes is crucial for optimizing LACC management.

Purpose Of The Study

To evaluate if radiomic features from baseline and early [18F]FDG-PET scans, and their changes, can predict prognosis in LACC patients.
To assess the prognostic value of radiomic features for disease-free survival (DFS) and overall survival (OS).

Main Methods

Retrospective analysis of 95 LACC patients treated with neoadjuvant CRT and surgery.
[18F]FDG-PET/CT scans were acquired before (baseline) and two weeks after (early) neoadjuvant CRT.
Radiomic features were extracted from PET images; delta radiomics quantified changes. Models (radiomic, clinical, combined) were built and validated using 5-fold cross-validation.

Main Results

None of the models could predict DFS (C-indices ≤ 0.72).
Models predicting OS showed slightly better performance.
Mean C-indices for OS were 0.75 (early radiomic/combined), 0.79 (delta radiomic), 0.78 (delta combined), and 0.76 (clinical models).

Conclusions

Early and delta radiomic features from [18F]FDG-PET scans did not predict DFS in LACC patients.
While radiomic models showed marginal improvement for OS prediction over clinical models, their added value for clinical practice is limited.

Keywords:

Delta radiomics Locally advanced cervical cancer Prognosis [18F]FDG-PET/CT