OCBE newsletter, June 2025

In our publications this month, you will find meta-analysis results of COVID-19 clinical trials, a new machine learning method to prioritize drug combinations in leukemia patients, a new statistical method for high-dimensional survival analysis, and much more.

Publications

  • Alain Amstutz, Stefan Schandelmaier, Hannah Ewald, Benjamin Speich, Johannes M Schwenke, Christof M Schönenberger, Stephan Schobinger, Thomas Agoritsas, Kay M Tomashek, Seema Nayak, Mat Makowski, Alejandro Morales-Ortega, David Bernal-Bello, Giovanni Pomponio, Alessia Ferrarini, Monireh Ghazaeian, Frances Hall, Simon Bond, María Teresa García-Morales, María Jiménez-González, José R Arribas, Patricia O Guimarães, Caio A M Tavares, Otavio Berwanger, Yazdan Yazdanpanah, Victoria C Simensen, Karine Lacombe, Maya Hites, Florence Ader, Evelina Tacconelli, France Mentré, Drifa Belhadi, Clément R Massonnaud, Cédric Laouénan, Alpha Diallo, Aliou Baldé, Lambert Assoumou, Dominique Costagliola, Erica Ponzi, Corina S Rueegg, Inge C Olsen, Marius Trøseid, Matthias Briel. Effects of Janus kinase inhibitors in adults admitted to hospital due to COVID-19: a systematic review and individual participant data meta-analysis of randomised clinical trials. The Lancet Respiratory Medicine. 2025 May 13.

This individual participant data meta-analysis of RCTs in adults admitted to hospital due to COVID-19 found that JAK inhibitors reduced mortality across all levels of respiratory support, independent of dexamethasone or tocilizumab, and probably decreased serious and severe adverse events compared with no JAK inhibitors.

 

This is a study design for an RCT with the aim of improving coverage in influenza vaccination among older adults in Norway, especially in immigrant groups

 

In this paper of the Physical Activity in Childhood Cancer Survivors (PACCS) study we compared physical fitness, physical function and physical activity (PA) in adolescent childhood cancer survivors (CCSs) to age- and sex-matched controls and across different cancer diagnoses. We found that adolescent CCSs had 4.2% to 11% lower physical fitness and function compared to controls, despite comparable PA levels. Survivors of brain tumors performed poorest.

 

Using wearable activity trackers shows promise in measuring physical activity in patients with axial spondyloarthritis (axSpA). This study aimed to explore the feasibility of recording physical activity using a wearable activity tracker and describe wear-time patterns among patients with axSpA.

 

In this study, we have introduced a novel method that adapts the mixture cure frailty model (MCFM) to account for high-dimensional covariates in both the incidence and the latency components of the model. The method extends the classical MCFM to higher dimensions using a multi-step adaptive elastic net penalty term inside the EM algorithm. Extensive simulation studies are conducted across various scenarios to demonstrate the performance of the model. We apply the novel approach to analyze RNAseq gene expression data from bulk breast cancer patients included in The Cancer Genome Atlas database.

 

We propose a Gaussian mixture model for detecting commuting patterns from automat road traffic counters. We apply the method to detect variations in commuting between the Nordic countries during the disruptions caused by the COVID-19 pandemic. The results from the study align with the literature when it comes to direction of commuting and the effect of holidays on the commuting pattern. At the beginning of the pandemic, the total traffic had about a 90% decrease relative to the pre-pandemic baseline, while the estimated commuting had a smaller decrease of about 42% and 71% for Norway–Sweden and Finland–Sweden borders, respectively. Post-pandemic, the traffic between Finland and Sweden is back to a pre-pandemic level. For Norway and Sweden, on the other hand, the traffic post-pandemic has only recovered to about 73–76% of the pre-pandemic level.

 

Using Scandinavian data from Norway, Sweden and Denmark we looked at the optimal anticoagulant treatment regimes to prevent pregnancy-related venous thrombosis (VT) in women with antithrombin (AT) deficiency. In high-risk AT deficiency pregnancies with previous VT, our results support prophylaxis with high prophylactic doses of LMWH from confirmed pregnancy. Hence, AT concentrate should be given peri-/postpartum

 

We have previously shown a marked reduction and change in the use of chemotherapy by incorporating a gene expression profiling test (the Prosigna assay) to assist adjuvant treatment decisions among HR+ HER2- early breast cancer patients in the EMIT-1 trial. The current study aims to evaluate the economic consequences of using the Prosigna-assay compared to standard immunohistochemical (IHC) markers for adjuvant treatment decisions. For the entire study population of HR+ HER2- lymph node negative patients, the Prosigna-assay was cost-effective from the societal perspective (but not from a healthcare perspective). However, restricting the analysis to the large majority of the patients considered either as uncertain or as clear chemotherapy candidates by IHC-based classification, the Prosigna-assay was cost-effective and cost-saving from both perspectives.

 

Future occurrence of a disease can be highly influenced by some specific risk factors. This work presents a comprehensive approach to quantify the event probability as a function of each separate risk factor by means of a parametric model.

 

Cervical intraepithelial neoplasia (CIN), caused by HPV, can progress to cancer, though some cases regress. HPV16 carries the highest risk for progression. In this study, 21% of CIN2-3 cases regressed. Lesions caused by HPV16, 31, 52, and 58 had significantly lower regression rates than those caused by HPV18, 33, 35, 39, and 45. We found that a combination of HPV genotype, expression of tumor suppressors (pRb and p53), immune cell proportions, and lesion length can predict the regression of high-grade CIN.

 

​​​​​​​This study was conducted at the NORMENT centre at OUS and based on data on bipolar patients with self-registration in the “MinDag” app, developed in collaboration with USIT (UIO). Time-series of daily mood registrations and weekly substance use for up to 6 months, were analysed with a vector autoregressive model (VAR) to assess bidirectional associations and Granger-causality between mood and alcohol/nicotine use in 32 patients. A large average proportion of the variance in alcohol use explained by mood (29%) and vice versa (28%) were found. No adjustment for time-varying confounding was possible and limits the interpretation.

 

​​​​​​​​​​​​​​We developed a systematic combinatorial design strategy that uses machine learning to prioritize the most promising drug combinations for patients with relapsed/refractory (R/R) acute myeloid leukemia (AML). The predictive approach leveraged single-cell transcriptomics and single-agent response profiles measured in primary patient samples to identify targeted combinations that co-inhibit treatment resistant cancer cells individually in each AML patient sample. 

Published Aug. 4, 2025 12:46 PM - Last modified June 12, 2026 2:43 PM
E-mail this page