The impact of PET image acquisition and reconstruction parameters on SUV measurements or radiomic feature values is widely documented. This "scanner" effect is detrimental to the design and validation of predictive or prognostic models and limits the use of large multicenter cohorts. To reduce the impact of this scanner effect, the ComBat method has been proposed and is now used in various contexts. The purpose of this article is to explain and illustrate the use of ComBat based on practical examples. We also give examples in which the ComBat assumptions are not met; thus, ComBat should not be used.

Purpose Accurate prognostic markers are urgently needed to identify diffuse large B-Cell lymphoma (DLBCL) patients at high risk of progression or relapse. Our purpose was to investigate the potential added value of baseline radiomics features to the international prognostic index (IPI) in predicting outcome after first-line treatment. Methods Three hundred seventeen newly diagnosed DLBCL patients were included. Lesions were delineated using a semi-automated segmentation method (standardized uptake value ≥ 4.0), and 490 radiomics features were extracted. We used logistic regression with backward feature selection to predict 2-year time to progression (TTP). The area under the curve (AUC) of the receiver operator characteristic curve was calculated to assess model performance. High-risk groups were defined based on prevalence of events; diagnostic performance was assessed using positive and negative predictive values. Results The IPI model yielded an AUC of 0.68. The optimal radiomics model comprised the natural logarithms of metabolic tumor volume (MTV) and of SUVpeak and the maximal distance between the largest lesion and any other lesion (Dmaxbulk, AUC 0.76). Combining radiomics and clinical features showed that a combination of tumor- (MTV, SUVpeak and Dmaxbulk) and patient-related parameters (WHO performance status and age > 60 years) performed best (AUC 0.79). Adding radiomics features to clinical predictors increased PPV with 15%, with more accurate selection of high-risk patients compared to the IPI model (progression at 2-year TTP, 44% vs 28%, respectively). Conclusion Prediction models using baseline radiomics combined with currently used clinical predictors identify patients at risk of relapse at baseline and significantly improved model performance. Trial registration number and date EudraCT: 2006–005,174-42, 01–08-2008.

Introduction: Metabolic tumor volume (MTV) is a promising biomarker of pretreatment risk in diffuse large B-cell lymphoma (DLBCL). Different segmentation methods can be used which predict prognosis equally well but give different optimal cut-offs for risk stratification. Segmentation can be cumbersome meaning a fast, easy and robust method is needed. Aims were to i) evaluate the best automated MTV workflow in DLBCL ii) determine if uptake time, (non)compliance with standardized recommendations for FDG scanning and subsequent disease progression influenced the success of segmentation iii) assess differences in MTV values and discriminatory power of segmentation methods. Methods: 140 baseline FDG-PET/CT scans were selected from UK and Dutch studies in DLBCL to provide a balance between scans at 60-or 90-minutes uptake, parameters compliant or non-compliant with standardized recommendations for scanning and patients with or without progression. An automated tool was used for segmentation using i) standardized uptake value (SUV) 2.5 ii) SUV 4.0 iii) adaptive thresholding [A50P] iv) 41% of maximum SUV [41%] v) majority vote including voxels detected by ≥2 methods [MV2] and vi) detected by ≥3 methods [MV3]. Two independent observers rated the success of the tool to delineate MTV. Scans that required minimal interaction were rated "success"; scans where > 50% of tumor was missed or required more than 2 editing steps were rated as "failure". Results: 138 scans were evaluable, with significant differences in success and failure ratings between methods. The best performing was SUV4.0, with higher success and lower failure rates than all other methods except MV2 which also performed well. SUV4.0 gave a good approximation of MTV in 105 (76%) scans, with simple editing for a satisfactory result in additionally 20% of cases. MTV was significantly different for all methods between patients with and without progression. SUV41% performed slightly worse with longer uptake times, otherwise scanning conditions and patient outcome did not influence the tool's performance. The discriminative power of methods was similar, but MTV values were significantly greater using SUV4.0 and MV2 than other thresholds except for SUV2.5. Conclusion:SUV4.0 and MV2 are recommended for further evaluation. Automated estimation of MTV is feasible.