In LACC treated with CRT, radiomics features such as EntropyGLCM and GLNUGLRLM from functional imaging DWI-MRI and PET, respectively, are independent predictors of recurrence and loco-regional control with significantly higher prognostic power than usual clinical parameters. Further research is warranted for their validation, which may justify more aggressive treatment in patients identified with high probability of recurrence.

The aim of this study was to validate previously developed radiomics models relying on just two radiomics features from 18 F-fluorodeoxyglucose positron emission tomography (PET) and magnetic resonance imaging (MRI) images for prediction of disease free survival (DFS) and locoregional control (LRC) in locally advanced cervical cancer (LACC). Methods Patients with LACC receiving chemoradiotherapy were enrolled in two French and one Canadian center. Pretreatment imaging was performed for each patient. Multicentric harmonization of the two radiomics features was performed with the ComBat method. The models for DFS (using the feature from apparent diffusion coefficient (ADC) MRI) and LRC (adding one PET feature to the DFS model) were tuned using one of the French cohorts (n = 112) and applied to the other French (n = 50) and the Canadian (n = 28) external validation cohorts. Results The DFS model reached an accuracy of 90% (95% CI [79-98%]) (sensitivity 92-93%, specificity 87-89%) in both the French and the Canadian cohorts. The LRC model reached an accuracy of 98% (95% CI [90-99%]) (sensitivity 86%, specificity 100%) in the French cohort and 96% (95% CI [80-99%]) (sensitivity 83%, specificity 100%) in the Canadian cohort. Accuracy was significantly lower without ComBat harmonization (82-85% and 71-86% for DFS and LRC, respectively). The best prediction using standard clinical variables was 56-60% only. Conclusions The previously developed PET/MRI radiomics predictive models were successfully validated in two independent external cohorts. A proposed flowchart for improved management of patients based on these models should now be confirmed in future larger prospective studies.

Multicenter studies are needed to demonstrate the clinical potential value of radiomics as a prognostic tool. However, variability in scanner models, acquisition protocols and reconstruction settings are unavoidable and radiomic features are notoriously sensitive to these factors, which hinders pooling them in a statistical analysis. A statistical harmonization method called ComBat was developed to deal with the "batch effect" in gene expression microarray data and was used in radiomics studies to deal with the "center-effect". Our goal was to evaluate modifications in ComBat allowing for more flexibility in choosing a reference and improving robustness of the estimation. Two modified ComBat versions were evaluated: M-ComBat allows to transform all features distributions to a chosen reference, instead of the overall mean, providing more flexibility. B-ComBat adds bootstrap and Monte Carlo for improved robustness in the estimation. BM-ComBat combines both modifications. The four versions were compared regarding their ability to harmonize features in a multicenter context in two different clinical datasets. The first contains 119 locally advanced cervical cancer patients from 3 centers, with magnetic resonance imaging and positron emission tomography imaging. In that case ComBat was applied with 3 labels corresponding to each center. The second one contains 98 locally advanced laryngeal cancer patients from 5 centers with contrast-enhanced computed tomography. In that specific case, because imaging settings were highly heterogeneous even within each of the five centers, unsupervised clustering was used to determine two labels for applying ComBat. The impact of each harmonization was evaluated through three different machine learning pipelines for the modelling step in predicting the clinical outcomes, across two performance metrics (balanced accuracy and Matthews correlation coefficient). Before harmonization, almost all radiomic features had significantly different distributions between labels. These differences were successfully removed with all ComBat versions. The predictive ability of the radiomic models was always improved with harmonization and the improved ComBat provided the best results. This was observed consistently in both datasets, through all machine learning pipelines and performance metrics. The proposed modifications allow for more flexibility and robustness in the estimation. They also slightly but consistently improve the predictive power of resulting radiomic models.