Machine Learning-Based Survival Prediction Models for Progression-Free and Overall Survival in Advanced-Stage Hodgkin Lymphoma.

PURPOSE: Patients diagnosed with advanced-stage Hodgkin lymphoma (aHL) have historically been risk-stratified using the International Prognostic Score (IPS). This study investigated if a machine learning (ML) approach could outperform existing models when it comes to predicting overall survival (OS) and progression-free survival (PFS). PATIENTS AND METHODS: This study used patient data from the Danish National Lymphoma Register for model development (development cohort). The ML model was developed using stacking, which combines several predictive survival models (Cox proportional hazard, flexible parametric model, IPS, principal component, penalized regression) into a single model, and was compared with two versions of IPS (IPS-3 and IPS-7) and the newly developed aHL international prognostic index (A-HIPI). Internal model validation was performed using nested cross-validation, and external validation was performed using patient data from the Swedish Lymphoma Register and Cancer Registry of Norway (validation cohort). RESULTS: In total, 707 and 760 patients with aHL were included in the development and validation cohorts, respectively. Examining model performance for OS in the development cohort, the concordance index (C-index) for the ML model, IPS-7, IPS-3, and A-HIPI was found to be 0.789, 0.608, 0.650, and 0.768, respectively. The corresponding estimates in the validation cohort were 0.749, 0.700, 0.663, and 0.741. For PFS, the ML model achieved the highest C-index in both cohorts (0.665 in the development cohort and 0.691 in the validation cohort). The time-varying AUCs for both the ML model and the A-HIPI were consistently higher in both cohorts compared with the IPS models within the first 5 years after diagnosis. CONCLUSION: The new prognostic model for aHL on the basis of ML techniques demonstrated a substantial improvement compared with the IPS models, but yielded a limited improvement in predictive performance compared with the A-HIPI.

Patient trajectories after diagnosis of diffuse large B-cell lymphoma-a multistate modelling approach to estimate the chance of lasting remission.

BACKGROUND: Achieving lasting remission for at least 2 years is a good indicator for favourable prognosis long term after Diffuse large B-cell lymphoma (DLBCL). The aim of this study was to provide real-world probabilities, useful in risk communication and clinical decision-making, of the chance for lasting remissions by clinical characteristics. METHODS: DLBCL patients in remission after primary treatment recorded in the Swedish Lymphoma register 2007-2014 (n = 2941) were followed for relapse and death using multistate models to study patient trajectories. Flexible parametric models were used to estimate transition rates. RESULTS: At 2 years, 80.7% (95% CI: 79.0-82.2) of the patients were predicted to remain in remission and 13.2% (95% CI: 11.9-14.6) to have relapsed. The relapse risk peaked at 7 months, and the annual decline of patients in remission stabilised after 2 years. The majority of patients in the second remission transitioned into a new relapse. The probability of a lasting remission was reduced by 20.4% units for patients with IPI 4-5 compared to patients with IPI 0-1, and time in remission was shortened by 3.5 months. CONCLUSION: The long-term prognosis was overall favourable with 80% achieving durable first remissions. However, prognosis varied by clinical subgroups and relapsing patients seldom achieved durable second remissions.

Trends in the prevalence, incidence and survival of non-Hodgkin lymphoma subtypes during the 21st century - a Swedish lymphoma register study.

Non-Hodgkin lymphoma (NHL) prognosis has improved in recent years, yet the number of patients living with the diagnosis, i.e. the prevalence, has seldom been reported. The prevalence provides a measure of the burden of disease, useful for healthcare planning and to optimise resource allocation. We provide a systematic presentation of temporal trends in absolute numbers of prevalent patients by NHL subtypes, linking them to trends in incidence, survival and mortality. Patients diagnosed 2000-2016 were identified in the national Swedish lymphoma register. Incidence and mortality rates, relative survival and prevalence were estimated for NHL overall and for major clinical and morphological subtypes. Poisson regression was used to test for temporal trends. Increasing incidence and improved survival have led to a 47% increase in the five-year prevalence of NHL overall in 2016 compared to 2004. An increasing prevalence was observed for all investigated subtypes during the study period, but most notably for diffuse large B cell lymphomas among aggressive subtypes (66%), and marginal zone lymphomas among indolent subtypes (135%). This dramatic increase in NHL prevalence underscores the need to develop and evaluate alternative follow-up schemes to use resources efficiently and still ensure optimal care of lymphoma survivors.