Publications

Background
Mutational signatures are increasingly used to understand the mechanisms causing cancer. However, their important applications in predicting prognosis and stratifying patients for therapy are hampered by inaccurate inference of the various featureless, dense trinucleotide mutational spectra, which are often confounded with one another. One of them is the homologous recombination deficiency (HRd)-associated signature SBS3, relevant because of its association with prognosis in ovarian and breast cancer and because of its potential as a biomarker for synthetic lethality therapies.

Methods
Here, we highlight strong benefits of a multimodal approach for mutational signature extraction, applied on top of standard bioinformatic pipelines. By jointly operating on single-base substitution (SBS) and indel (ID) spectra, this method enables accurate identification of various DNA repair deficiency signatures and patient survival prediction.

Results
Across four different cohorts of whole-genome sequenced high-grade serous ovarian cancers (HGSOC), the multimodal SBS + ID approach correctly distinguished the commonly confused signatures SBS3, SBS5, SBS8, SBS39, and SBS40. Importantly, we robustly identified two different multimodal SBS3 signatures, m-SBS3a and m-SBS3b, with distinct patterns in the indel spectrum. Multimodal SBS3b signature was strongly predictive of longer survival in ovarian cancer patients, replicating across four cohorts, with effect sizes greatly exceeding other genetic markers. Our m-SBS3 also predicted survival in platinum-treated patients with various cancer types, and moreover, the SBS + ID joint inference was successfully applied to mismatch repair-deficient colorectal cancer and immunotherapy response, supporting a general utility of the multimodal mutational signatures approach.

Conclusions
Overall, combining SBS and ID mutations improves detection of HR deficiency-associated signatures and reveals distinct SBS3 subtypes with prognostic value. This multimodal approach outperforms existing markers and is readily applicable to therapy stratification.

Precision and personalized medicine for disease management necessitates real-time, continuous monitoring of biomarkers and therapeutic drugs to adjust treatment regimens based on individual patient responses. This study introduces a wearable Microneedle-based Continuous Biomarker/Drug Monitoring (MCBM) system, designed for the simultaneous, in vivo pharmacokinetic and pharmacodynamic evaluation for diabetes. Utilizing a dual-sensor microneedle and a layer-by-layer nanoenzyme immobilization strategy, the MCBM system achieves high sensitivity and specificity in measuring glucose and metformin concentrations in skin interstitial fluid (ISF). Seamless integration with a smartphone application enables real-time data analysis and feedback, fostering a pharmacologically informed approach to diabetes management. The MCBM system’s validation and in vivo trials demonstrate its precise monitoring of glucose and metformin, offering a tool for personalized treatment adjustments. Its proven biocompatibility and safety suit long-term usage. This system advances personalized diabetes care, highlighting the move towards wearables that adjust drug dosages in real-time, enhancing precision and personalized medicine.

As part of the LUCIA project, funded under Horizon Europe, Yaghma facilitated Social Lab workshops to identify and address challenges in implementing novel lung cancer screening technologies including an AI-driven risk prediction algorithm. Engaging healthcare professionals, policymakers, and patient representatives, these participatory sessions helped uncover key barriers of implementing LUCIA technologies to lung cancer screening.

Through structured discussions, 80 actionable barriers have been identified across 10 key categories, including trust, feasibility, ethical considerations, and regulation. These insights were mapped in an Impact-Influence map through a prioritization survey, ensuring that the project focuses on the most critical and actionable challenges.

Yaghma’s role in this process underscores the importance of stakeholder engagement in responsible AI adoption, ensuring that AI technologies in cancer screening are practical, inclusive, and aligned with real-world healthcare and patient needs.

Cancer is among the most common causes of mortality worldwide. Screening for cancer involves surveillance of populations towards identification of cancer that was unknown to the patient. Therapeutic treatments for cancer generally are more effective when applied earlier on in the condition’s development, as such, screening for cancer has the potential to improve the standard for patient care and improving mortality and morbidity associated with the disease. Cancer screening has increasingly become dependent on advanced technologies to assist in the identification and characterization of tumours. This article Collection showcases current research towards the development of new methods for cancer screening, including the development of new advanced technologies in this domain such as novel methods incorporating liquid biopsies, tailed primer isothermal amplification assays, and infrared spectroscopy. As is common in modern research, approaches make use of computational techniques as a critical component in cancer screening, and research is highlighted on the use of artificial intelligence, which is now a common technological innovation contributing to the overall cancer screening process in research studies. Future directions for cancer screening are discussed.

Background
Electronic cigarettes (e-cigarettes) have gained popularity in recent years. While initially introduced as a safe alternative for tobacco and a bridge for smoking cessation, subsequent studies found that they contain toxic substances. We aimed to assess the acute effect of a single session of e-cigarette smoking on cardiorespiratory parameters, exhaled volatile organic compounds (VOCs) and markers of inflammation.

Methods
A prospective single-centre study was carried out. Participants (healthy volunteers, former e-cigarette users) were assessed before and after a 30-min session of e-cigarette smoking. Evaluations included vital signs, pulmonary functions – spirometry and fractional exhaled nitric oxide (FeNO) – blood and exhaled breath condensate (EBC) cytokines and electronic nose (e-nose) for analysis of exhaled VOCs profile.

Results
30 participants aged 27.9±4.4 years were enrolled in the study. Post-smoking observations revealed a significant increase in heart rate (77.5±10.9 to 85.5±12.1 beats·min−1, p=0.002), respiratory rate (15.4±2.2 to 17.1±1.8 breaths·min−1, p=0.002) and blood pressure (systolic 118±8.1 to 123.5±11.9, p=0.017; diastolic 73.9±8.4 to 78.5±6.3 mmHg, p=0.011). FeNO decreased significantly (median of 11 (7.5–15.5) to 9.7 (7.3–17.3) ppb, p=0.024). Analysis of e-nose found a significant change of exhaled VOC pattern after e-cigarette smoking. No significant changes were found in spirometry and cytokine levels in blood or EBC.

Conclusions
A single session of 30 min of e-cigarette smoking caused significant cardiorespiratory effects, decreased FeNO and altered exhaled VOC pattern, similar to the effect seen with cigarette and water-pipe smoking.‏ The observed acute effects, together with the well-known chronic risks, highlight the importance of effective regulation of e-cigarettes.

Breath analysis offers a non-invasive approach to modern diagnostics by capturing volatile organic compounds (VOCs) in exhaled breath. However, current breath analysis technologies face challenges like humidity sensitivity, high costs, and biodegradable solutions, limiting their scalability and environmental sustainability. This study presents a paper-based, biodegradable, humidity-insensitive electronic nose (e-nose) sensor array integrated into a face mask for real-time breath analysis. The sensors, coated with hydrophobic polymer coating, ensure robust insensitivity to humidity, enabling reliable detection of VOCs even in high-moisture environments. The mask-integrated e-nose facilitates real-time breath monitoring for applications such as alcohol consumption tracking and respiratory health assessment. For the latter, Tuberculosis (TB) detection is selected as a representative use case, achieving 89% accuracy in disease diagnosis and recovery monitoring using a pre-trained deep-learning model. The fully-biodegradable paper-based sensor naturally degrades in soil within months, underscoring its eco-friendly design and suitability for disposable health monitoring. This work introduces a sustainable, user-friendly approach to breath analysis with potential applications in non-invasive disease detection and personalized healthcare monitoring.

Organoids mimic human organ function, offering insights into development and disease. However, non-destructive, real-time monitoring is lacking, as traditional methods are often costly, destructive, and low-throughput. In this article, a non-destructive chemical tomographic strategy is presented for decoding cyto-proteo-genomics of organoid using volatile signaling molecules, hereby, Volatile Organic Compounds (VOCs), to indicate metabolic activity and development of organoids. Combining a hierarchical design of graphene-based sensor arrays with AI-driven analysis, this method maps VOC spatiotemporal distribution and generate detailed digital profiles of organoid morphology and proteo-genomic features. Lens- and label-free, it avoids phototoxicity, distortion, and environmental disruption. Results from testing organoids with the reported chemical tomography approach demonstrate effective differentiation between cyto-proteo-genomic profiles of normal and diseased states, particularly during dynamic transitions such as epithelial-mesenchymal transition (EMT). Additionally, the reported approach identifies key VOC-related biochemical pathways, metabolic markers, and pathways associated with cancerous transformations such as aromatic acid degradation and lipid metabolism. This real-time, non-destructive approach captures subtle genetic and structural variations with high sensitivity and specificity, providing a robust platform for multi-omics integration and advancing cancer biomarker discovery.

Background: Interstitial lung diseases (ILDs) are an heterogeneous group of infiltrating lung pathologies, for which prompt diagnosis and continuous assessment are of paramount importance. While chest CT is an established diagnostic tool for ILDs, there are no formal guidelines on the follow-up regimen, leaving the frequency and modality of follow-up largely at the clinician’s discretion.

Methods: The study retrospectively evaluated the indication of chest CT in a cohort of 129 ILD patients selected from the ambulatory care polyclinic at University Hospital of Liège. The aim was to determine whether the imagining acquisition had a true impact on clinical course and follow-up. We accepted three different situations for justifying the indication of the CTs: clinical deterioration, a decrease in pulmonary function tests (at least a 10% drop in a parameter), and monitoring for oncological purposes. The other indications, mainly routine follow-up, were classified as “non-justified”. Radiation dose output was evaluated with Computed Tomography Dose Index (CTDI) and Dose Length Product (DLP).

Results: The mean number of CT scans per patient per year was 1.7 ± 0.4, determining irradiation in CTDI (mGy)/year of 34.9 ± 64.9 and DLP in (mGy*cm)/year of 1095 ± 1971. The percentage of justified CT scans was 57 ± 32%, while the scans justified a posteriori were 60 ± 34%. Around 40% of the prescribed monitoring CT scans had no impact on the management of ILD and direct patient care.

Conclusions: Our study identifies a trend of overuse in chest CT scans at follow-up (up to 40%), outside those performed for clinical exacerbation or oncological investigation. In the particular case of ILD exacerbation, CT scan value remains high, underlying the benefit of this strategy.

Background
While different lung cancer risk prediction models have been established as essential tools to identify high-risk participants for lung cancer screening programs, evaluations of their risk discriminatory performances have reported heterogenous findings in different research cohorts. We therefore aimed to summarise results of head-to-head comparisons of the predictive performance of various lung cancer risk models performed within the same study population.

Methods
In this systematic review and meta-analysis, we performed a systematic search of PubMed and Web of Science databases for primary studies published from inception to Oct 16, 2024. Articles comparing the performance of questionnaire-based lung cancer risk models in an independent, external validation cohort of participants with previous or current smoking exposure were included. The main reasons for exclusion of studies were if only one model was assessed in the external population or risk discrimination was not evaluated. Random-effects meta-analyses were conducted to synthesize differences in the area under the curve (AUC) of two models compared in multiple populations. To assess the risk of bias, PROBAST (the Prediction model Risk of Bias Assessment Tool) was used. The study was registered with PROSPERO, CRD42023427911.

Findings
The systematic search yielded 5568 records. In total, 15 eligible studies were included in the meta-analysis, comprising 4,134,648 individuals with previous or current smoking exposure, of whom 45,448 (1.10%) developed LC within 5–7 years. Among the nine models that were compared, AUC differences reached up to 0.050 between two models. The Lung Cancer Risk Assessment Tool (LCRAT), Bach model and PLCOm2012 model consistently had a higher AUC when compared to any other model, with AUC differences ranging between 0.018 (95% CI 0.011, 0.026) and 0.044 (95% CI 0.038, 0.049). The risk of bias and applicability concerns were deemed low in eight, and high in seven of the included studies. Results excluding studies with high risk of bias were mostly consistent. Among eight of the 24 model pairs that were compared, there was notable between-study heterogeneity (I2 ≥50%).

Interpretation
Our systematic review and meta-analyses of head-to-head comparisons disclose major differences in predictive performance of widely used lung cancer risk models. Although our review is limited to the availability of head-to-head comparisons, evidence from current cohort-based model comparisons indicates that the LCRAT, Bach and PLCOm2012 consistently outperformed alternative questionnaire-based risk prediction tools.

The most well-established risk factor for lung cancer (LC) is smoking, responsible for approximately 85% of cases. The Lung Cancer Risk Assessment Tool (LCRAT) is a key advancement in this field, which predicts individual risk based on factors like smoking habits, demographic details, personal and family medical history, and environmental exposures. This paper proposes a model with fewer features that improves state of the art performance, using a simplified stacking ensemble, making it more accessible and easier to implement in routine healthcare practice. The data used in this work were derived from two cohorts in the United States: The National Lung Screening Trial (NLST) and the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. Both our model and LCRAT achieve an AUC of 0.799 and 0.782 on test respectively. In terms of percentage of positives, in the 50% of the population, both detect 0.766 and 0.754 of the cases. The ensemble of different survival models enhances robustness by mitigating the weakness of individual models and directly impacts the efficiency of the model, increasing the efficiency and generalizability.