[18F]FDG PET radiomics to predict disease-free survival in cervical cancer: a multi-scanner/center study with external validation

Abstract

Purpose: To test the performances of native and tumour to liver ratio (TLR) radiomic features extracted from pre-treatment 2-[¹⁸F] fluoro-2-deoxy-D-glucose ([¹⁸F]FDG) PET/CT and combined with machine learning (ML) for predicting cancer recurrence in patients with locally advanced cervical cancer (LACC).

Methods: One hundred fifty-eight patients with LACC from multiple centers were retrospectively included in the study. Tumours were segmented using the Fuzzy Local Adaptive Bayesian (FLAB) algorithm. Radiomic features were extracted from the tumours and from regions drawn over the normal liver. Cox proportional hazard model was used to test statistical significance of clinical and radiomic features. Fivefold cross validation was used to tune the number of features. Seven different feature selection methods and four classifiers were tested. The models with the selected features were trained using bootstrapping and tested in data from each scanner independently. Reproducibility of radiomics features, clinical data added value and effect of ComBat-based harmonisation were evaluated across scanners.

Results: After a median follow-up of 23 months, 29% of the patients recurred. No individual radiomic or clinical features were significantly associated with cancer recurrence. The best model was obtained using 10 TLR features combined with clinical information. The area under the curve (AUC), F₁-score, precision and recall were respectively 0.78 (0.67-0.88), 0.49 (0.25-0.67), 0.42 (0.25-0.60) and 0.63 (0.20-0.80). ComBat did not improve the predictive performance of the best models. Both the TLR and the native models performance varied across scanners used in the test set.

Conclusion: [¹⁸F]FDG PET radiomic features combined with ML add relevant information to the standard clinical parameters in terms of LACC patient's outcome but remain subject to variability across PET/CT devices.

Keywords: Cervical cancer; Disease-free survival; Machine learning; Radiomics; [18F]FDG PET/CT.

Fig. 1

Radiomics pipeline

Fig. 2

Kaplan-Meier curve of the each individual significant feature in univariate and multivariate analysis, after the Cox proportional hazard model. a GLDZM_DZNN_0.5 (OR) (Threshold = 0.59, log-rank test P value 0.079). b GLSZM_HILAE_0.5 (TLR) (Threshold = 0.07, log-rank test P value 0.044), c GLDZM_DZV_0.05 (TLR from interpolated images) (Threshold = 1.28, log-rank test P value 0.47). d Stats_qcod_0.2 (TLR from interpolated images) (Threshold = 1.18, log-rank test P value 0.088). e Histology (Threshold = 0.5, log-rank test P value 0.15)

Fig. 3

Kaplan-Meier curve of the test set for the best OR (a) and TLR (b) model. Red and blue curves represent respectively patients with better and worse prognosis. The log-rank test was used to estimate statistical significance of the difference between survival curves. The P value obtained from the log-rank test is shown in the left down corner of each image. The difference between both the Kaplan-Meier curves is statistical significant (log-rank P value = 0.034 for the OR model and 0.002 for the TLR model)