Search

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. Machine-Learning based advanced-stage Hodgkin lymphoma (aHL) prognosis surpasses International Prognostic Score, competes with aHL international prognostic index in overall survival and progression-free survival prediction. [ABSTRACT FROM AUTHOR]

Purpose Our aim was to capture the biology of classical Hodgkin lymphoma (cHL) at the time of relapse and discover novel and robust biomarkers that predict outcomes after autologous stem-cell transplantation (ASCT). Materials and Methods We performed digital gene expression profiling on a cohort of 245 formalin-fixed, paraffin-embedded tumor specimens from 174 patients with cHL, including 71 with biopsies taken at both primary diagnosis and relapse, to investigate temporal gene expression differences and associations with post-ASCT outcomes. Relapse biopsies from a training cohort of 65 patients were used to build a gene expression-based prognostic model of post-ASCT outcomes (RHL30), and two independent cohorts were used for validation. Results Gene expression profiling revealed that 24% of patients exhibited poorly correlated expression patterns between their biopsies taken at initial diagnosis and relapse, indicating biologic divergence. Comparative analysis of the prognostic power of gene expression measurements in primary versus relapse specimens demonstrated that the biology captured at the time of relapse contained superior properties for post-ASCT outcome prediction. We developed RHL30, using relapse specimens, which identified a subset of high-risk patients with inferior post-ASCT outcomes in two independent external validation cohorts. The prognostic power of RHL30 was independent of reported clinical prognostic markers (both at initial diagnosis and at relapse) and microenvironmental components as assessed by immunohistochemistry. Conclusion We have developed and validated a novel clinically applicable prognostic assay that at the time of first relapse identifies patients with unfavorable post-ASCT outcomes. Moving forward, it will be critical to evaluate the clinical use of RHL30 in the context of positron emission tomography-guided response assessment and the evolving cHL treatment landscape.