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.
BackgroundThe independent prognostic impact of diabetes mellitus (DM) and prediabetes mellitus (pre‐DM) on survival outcomes in patients with chronic heart failure has been investigated in observational registries and randomized, clinical trials, but the results have been often inconclusive or conflicting. We examined the independent prognostic impact of DM and pre‐DM on survival outcomes in the GISSI‐HF (Gruppo Italiano per lo Studio della Sopravvivenza nella Insufficienza Cardiaca‐Heart Failure) trial.Methods and ResultsWe assessed the risk of all‐cause death and the composite of all‐cause death or cardiovascular hospitalization over a median follow‐up period of 3.9 years among the 6935 chronic heart failure participants of the GISSI‐HF trial, who were stratified by presence of DM (n=2852), pre‐DM (n=2013), and non‐DM (n=2070) at baseline. Compared with non‐DM patients, those with DM had remarkably higher incidence rates of all‐cause death (34.5% versus 24.6%) and the composite end point (63.6% versus 54.7%). Conversely, both event rates were similar between non‐DM patients and those with pre‐DM. Cox regression analysis showed that DM, but not pre‐DM, was associated with an increased risk of all‐cause death (adjusted hazard ratio, 1.43; 95% CI, 1.28–1.60) and of the composite end point (adjusted hazard ratio, 1.23; 95% CI, 1.13–1.32), independently of established risk factors. In the DM subgroup, higher hemoglobin A1c was also independently associated with increased risk of both study outcomes (all‐cause death: adjusted hazard ratio, 1.21; 95% CI, 1.02–1.43; and composite end point: adjusted hazard ratio, 1.14; 95% CI, 1.01–1.29, respectively).ConclusionsPresence of DM was independently associated with poor long‐term survival outcomes in patients with chronic heart failure.Clinical Trial Registration
URL: http://www.clinicaltrials.gov. Unique identifier: NCT00336336.
Background: Hypofractionated stereotactic radiotherapy (HFSRT) is indicated for large brain metastases (BM) or proximity to critical organs (brainstem, chiasm, optic nerves, hippocampus). The primary aim of this study was to assess factors influencing BM local control after HFSRT. Then the effect of surgery plus HFSRT was compared with exclusive HFSRT on oncologic outcomes, including overall survival. Materials and methods: Retrospective study conducted in Léon Bérard Cancer Center, included patients over 18 years-old with BM, secondary to a tumor proven by histology and treated by HFSRT alone or after surgery. Three different dose-fractionation schedules were compared: 27 Gy (3 × 9 Gy), 30 Gy (5 × 6 Gy) and 35 Gy (5 × 7 Gy), prescribed on isodose 80%. Primary endpoint were local control (LC). Secondary endpoints were overall survival (OS) and radionecrosis (RN) rate. Results: A total of 389 patients and 400 BM with regular MRI follow-up were analyzed. There was no statistical difference between the different dose-fractionations. On multivariate analysis, surgery (p = 0.049) and size (< 2.5 cm) (p = 0.01) were independent factors improving LC. The 12 months LC was 87.02% in the group Surgery plus HFSRT group vs 73.53% at 12 months in the group HFSRT. OS was 61.43% at 12 months in the group Surgery plus HFSRT group vs 50.13% at 12 months in the group HFSRT (p < 0.0085). Prior surgery (OR = 1.86; p = 0.0028) and sex (OR = 1.4; p = 0.0139) control of primary tumor (OR = 0.671, p = 0.0069) and KPS < 70 (OR = 0.769, p = 0.0094) were independently predictive of OS. The RN rate was 5% and all patients concerned were symptomatic. Conclusions: This study suggests that HFSRT is an efficient and well-tolerated treatment. The optimal dosefractionation remains difficult to determine. Smaller size and surgery are correlated to LC. These results evidence the importance of surgery for larger BM (> 2.5 cm) with a poorer prognosis. Multidisciplinary committees and prospective studies are necessary to validate these observations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.