A multi-institutional and case-matched control study on treatment outcomes of consolidative radiotherapy after a full course of R-CHOP compared with R-CHOP alone in Stage I-II diffuse large B-cell lymphoma (KROG 17-02)

The prognostic value of end-of-treatment FDG-PET/CT in diffuse large B cell lymphoma: comparison of visual Deauville criteria and a lesion-to-liver SUVmax ratio-based evaluation system

PURPOSE

The aim of this study was to determine a better criterion for end-of-treatment PET (EoT-PET) assessment and prognostic evaluation of patients with diffuse large B cell lymphoma (DLBCL).

METHOD

EoT-PET scans were assessed using the visual Deauville 5-point scale (5PS) and LLR, the maximum standard uptake value ratio between the lesion and the liver. The cutoff value of LLR was obtained by receiver operator characteristic curve analysis. Patient outcomes were compared using Kaplan-Meier survival analysis. Prognostic indexes of different criteria were compared. Multivariate Cox regression analysis was performed to evaluate the prognostic factors.

RESULTS

Four hundred forty-nine newly diagnosed DLBCL patients who received rituximab-based immunochemotherapy were included, and the median follow-up duration was 41.4 months. Patients with Deauville score (DS) 4 displayed significantly longer PFS and OS compared with patients with DS 5 (both p < 0.001), and they had significantly shorter PFS (p < 0.01) but similar OS (p = 0.057) compared with patients with DS 1-3. The differences in PFS and OS between groups were all significant whether positive EoT-PET was defined as DS 4-5 or DS 5 (all p < 0.001). The optimal cutoff of LLR was 1.83, and both PFS and OS were significantly different between EoT-PET-positive and EoT-PET-negative patients as defined by the cutoff (both p < 0.001). LLR-based criterion displayed higher specificity, positive predictive value, and accuracy than 5PS-based criterion in the prediction of disease progression and death events. In the multivariate analysis, positive EoT-PET (as defined by LLR) was related to unfavorable PFS and OS (both p < 0.001). Additional treatment was not correlated with outcomes of EoT-PET-negative patients either defined by LLR or 5PS or EoT-PET-positive patients classified by 5PS, but it was the only beneficial factor for OS (p < 0.05) in EoT-PET-positive patients with LLR ≥ 1.83.

CONCLUSION

The optimal cutoff of LLR may be superior to Deauville criteria in identifying low-risk DLBCL patients with negative EoT-PET after the first-line immunochemotherapy and sparing them the cost and toxicity of additional treatment.

Treatment Modalities and Survival Outcomes for Sinonasal Diffuse Large B-Cell Lymphoma

OBJECTIVES/HYPOTHESIS

This study utilizes a large population national database to comprehensively analyze prognosticators and overall survival (OS) outcomes of varying treatment modalities in a large cohort of sinonasal diffuse large B-cell lymphoma (SN-DLBCL) patients.

STUDY DESIGN

Retrospective database study.

METHODS

The National Cancer Database was queried for all SN-DLBCL cases diagnosed from 2004 to 2015. Kaplan-Meier log-rank test determined differences in OS based on clinical covariates. Cox proportional-hazards analysis was used to determine clinical and sociodemographic covariates predictive of mortality.

RESULTS

A total of 2,073 SN-DLBCL patients were included, consisting of 48% female with a mean age of 66.0 ± 16.2 years. Overall, 82% of patients were Caucasian, 74% had early-stage disease, and 49% had primary tumors in the paranasal sinuses. Early-stage patients were more likely to receive multi-agent chemoradiotherapy compared to multi-agent chemotherapy alone (P < .001). Multivariable Cox proportional-hazards analysis revealed chemoradiotherapy to confer significantly greater OS improvements than chemotherapy alone (hazard ratio [HR]: 0.61; P < .001). However, subset analysis of late-stage patients demonstrated no significant differences in OS between these treatment modalities (P = .245). On multivariable analysis of chemotherapy patients treated post-2012, immunotherapy (HR = 0.51; P = .024) demonstrated significant OS benefits. However, subset analysis showed no significant advantage in OS with administering immunotherapy for late-stage patients (P = .326). Lastly, for all patients treated post-2012, those receiving immunotherapy had significantly improved OS compared to those not receiving immunotherapy (P < .001).

CONCLUSIONS

Treatment protocol selection differs between early- and late-stage SN-DLBCL patients. Early-stage patients receiving chemotherapy may benefit from immunotherapy as part of their treatment paradigm.

LEVEL OF EVIDENCE

III Laryngoscope, 2021.

Classifying 2-year recurrence in patients with dlbcl using clinical variables with imbalanced data and machine learning methods

BACKGROUND

Treatments are limited for patients with relapsed/refractory Diffuse large B-cell lymphoma (DLBCL), and their survival rate is low. Prediction of the recurrence hazard for each patient could provide a reference regarding chemotherapy regimens for clinicians to extend patients' period of long-term remission. As current strategies cannot satisfy such need, we have established predictive models to classify patients with DLBCL with complete remission who had recurrences in 2 years from ones who did not.

METHODS

We assessed 518 patients with DLBCL and measured 52 variables of each patient. They were treated between January 2011 and July 2016. 17 variables were first selected by variable selection methods (including Lasso, Adaptive Lasso, and Elastic net). Then, we set classifiers and probability models for imbalanced data by combining the SMOTE sampling, cost-sensitive, and ensemble learning (consisting of AdaBoost, voting strategy, and Stacking) methods with the machine learning methods (Support Vector Machine, BackPropagation Artificial Neural Network, Random Forest), respectively. Last, assessed their performance.

RESULTS

The disease stage and other 5 variables are significant indicators for recurrence. The SVM with AdaBoost ensemble learning method modeling by SMOTE data performs the best (Sensitivity=97.3%, AUC=96%, RMSE=19.6%, G-mean=96%) in all classifiers. The SVM with AdaBoost method(AUC=98.7%, RMSE=17.7%, MXE=12.7%, Cal mean=3.2%, BS0=2.5%, BS1=4%, BSALL=3.1%) and random forest (AUC=99.5%, RMSE=19.8%, MXE=16.2%, Cal mean=9.1%, BS0=4.8%, BS1=2.9%, BSALL=3.9%) both modeling by SMOTE sampling data perform well in probability models.

CONCLUSIONS

This predictive model has high accuracy for almost all DLBCL patients and the six indicators can be recurrence signals.