Lausanne University Hospital

italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Goal: The gold standard for detecting the presence of apneic events is a time and effort-consuming manual evaluation of type I polysomnographic recordings by experts, often not error-free. Such acquisition protocol requires dedicated facilities resulting in high costs and long waiting lists. The usage of artificial intelligence models assists the clinician's evaluation overcoming the aforementioned limitations and increasing healthcare quality. Methods: The present work proposes a machine learning-based approach for automatically recognizing apneic events in subjects affected by sleep apnea-hypopnea syndrome. It embraces a vast and diverse pool of subjects, the Wisconsin Sleep Cohort (WSC) database. Results: An overall accuracy of 87.2 $\pm$ 1.8% is reached for the event detection task, significantly higher than other works in literature performed over the same dataset. The distinction between different types of apnea was also studied, obtaining an overall accuracy of 62.9 $\pm$ 4.1%. Conclusions: The proposed approach for sleep apnea events recognition, validated over a wide pool of subjects, enlarges the landscape of possibilities for sleep apnea events recognition, identifying a subset of signals that improves State-of-the-art performance and guarantees simple interpretation.

Purpose To implement a flexible framework, named HydrOptiFrame, for the design and optimization of time‐efficient water‐excitation (WE) RF pulses using B‐spline interpolation, and to characterize their lipid suppression performance. Methods An evolutionary optimization algorithm was used to design WE RF pulses. The algorithm minimizes a composite loss function that quantifies the fat–water contrast using Bloch equation simulations. In a first study, B‐spline interpolated optimized (BSIO) pulses designed with HydrOptiFrame with durations of 1 and 0.76 ms were generated for 3 T and characterized in healthy volunteers' knees. The femoral bone marrow SNR was compared to that obtained with to 1–1 WE and lipid insensitive binomial off resonant excitation (LIBRE) pulses. In a second study, in the heart at 1.5 T, the water–fat contrast ratio and coronary artery vessel length obtained with a 2.56 ms BSIO pulse was compared to 1–1 WE and LIBRE pulses in free‐running cardiovascular MR. Results The 1 ms BSIO pulse resulted in higher fat suppression and lower contrast ratio (CR) in the bone marrow than the state‐of‐the‐art pulses (4.1 ± 0.2 vs. 4.7 ± 0.4 and 4.4 ± 0.3 for the BSIO, the 1–1 WE and LIBRE respectively, p < 0.05 vs. both) at 3 T. At 1.5 T, the BSIO pulse resulted in a higher blood‐epicardial fat CR (3.8 ± 1.3 vs. 1.6 ± 0.6 and 2.4 ± 1.1 for the BSIO, 1–1 WE and LIBRE, respectively, p < 0.05 vs. both) and longer traceable left coronary artery vessel length (8.7 ± 1.4 cm vs. 7.0 ± 1.0 cm [p = 0.04] and 7.5 ± 1.2 cm [p = 0.09]). Conclusion The HydrOptiFrame framework offers a new opportunity to design WE RF pulses that are robust to B0 inhomogeneity at multiple magnetic field strengths and for variable RF pulse durations.

Aims Mutations affecting exon 3 of the β-catenin ( CTNNB1 ) gene result in constitutive activation of WNT signalling and are a diagnostic hallmark of several tumour entities including desmoid-type fibromatosis. They also define clinically relevant tumour subtypes within certain entities, such as endometrioid carcinoma. In diagnostics, β-catenin immunohistochemistry is widely used as a surrogate for CTNNB1 mutations. Yet, it is often difficult to assess in practice, given that the characteristic nuclear translocation may be focal or hard to distinguish from the spillover of the normal membranous staining. Methods We therefore examined lymphoid enhancer-binding factor 1 (LEF1) immunostaining, a nuclear marker of WNT activation that serves as a potential surrogate for CTNNB1 mutations. Results In a cohort of endometrial carcinomas with known mutation status (n=130) LEF1 was 85% accurate in predicting CTNNB1 mutation status (64% sensitivity, 90% specificity) while β-catenin was 76% accurate (72% sensitivity; 77% specificity). Across a variety of entities characterised by CTNNB1 mutations as putative drivers, we found diffuse and strong expression of LEF1 in 77% of cases. LEF1 immunostaining proved easier to interpret than β-catenin immunostaining in 54% of cases, more difficult in 1% of cases and comparable in the remaining cases. Conclusion We conclude that LEF1 immunostaining is a useful surrogate marker for CTNNB1 mutations. It favourably complements β-catenin immunohistochemistry and outperforms the latter as a single marker.

Background Several studies have confirmed the safety and efficacy of robotic assisted coronary interventions for simple and complex coronary lesions. However, in the real‐world clinical setting the currently available evidence is still inconclusive with regard to clinical outcomes. In terms of financial sustainability, the emerging use of pre‐PCI Coronary Computed Tomography Angiography (CCTA) may potentially impact the overall patient journey and related costs. However, the role of CCTA guidance in elective robotic‐assisted revascularizations and its potential impact on hospital economics has never been investigated. Therefore, we aimed to assess the clinical impact of R‐PCI according to procedural complexity and its potential financial sustainability when integrated with CCTA guidance. Methods Major adverse cardiovascular events (MACE) at the latest clinical follow‐up available have been prospectively evaluated in a single center cohort of 111 patients undergoing elective R‐PCI (CorPath, Corindus GRX Robotic Drive, Boston, MA). The study population was subsequently divided into two groups according to either the median Syntax Score (SS = 14) as a surrogate of procedural complexity or the adoption of CCTA‐guidance. Additionally, both periprocedural and in‐hospital outcomes have been investigated in the overall cohort and in the subgroups. Results Overall, MACE occurred in 5.4% of the patients at a median follow‐up of 309 days. Procedural complexity was associated with significantly longer procedural time and radiation exposure (SS ≥ 14: 94.5 ± 32.1 vs. 78.7 ± 25.5 min, p = 0.011, and 32.7 ± 30.9 vs. 22.5 ± 19.8 mSv, p = 0.010, respectively), but not with increased in‐hospital and long‐term clinical outcomes. Additionally, patients undergoing CCTA‐guided R‐PCI had a significantly higher rate of same‐day‐discharge (SDD—64.6% vs. 44.2% respectively, p = 0.034) than those without CCTA support. Conclusion R‐PCI appears safe and effective regardless the procedural complexity. Moreover, the integration of robotics and CCTA‐guidance appears to have beneficial impact on hospital economics by optimizing resource utilization and improving patient selection for percutaneous revascularization.

This chapter explores the convergence of various digital technologies and their potential impacts on problematic online behaviors. Incorporation of gaming elements into many online activities (e.g., gambling, on-demand video streaming, social media, and pornography) has created concerns about problematic use, particularly among youth). We review the current state of knowledge, which suggests that problematic engagement in different activities represents distinct conditions. Nonetheless, despite their apparent distinctiveness, online addictive activities have shared risk factors, including personality factors, comorbid psychopathological symptoms, cognitive processes, and neurobiology. A common approach to prevention is therefore sensible, even if treatment of problematic cases should follow individualized approaches. Approaches actively involving parents are encouraged. We call for expanded research on the impact of specific design features of online activities on problematic behavior and risk, particularly relating to gamified activities. Longitudinal research is needed to better understand the use and impacts of online activities across developmental stages. Given the limited understanding of long-term impacts and the dynamic nature of related technologies, greater evidence is needed to inform policy and regulation relating to exposure of children and adolescents to content and activities that might in some cases promote problematic behaviors.

This chapter critically considers the research literature on the experiences and potential effects of gaming among young people. The effects of gaming on young people’s well-being appear complex and multidirectional, and depend on patterns, content, and context of gaming activities. Research into media effects appears to have become specialized in discrete areas (e.g., cognition, mental health, social effects), with few studies synthesizing data across subfields. Priority research areas include studies of short- and long-term effects of gaming; studying artificial intelligence-driven designs, monetized, and gambling-like designs in games; evaluating assessment and interventions for problematic gaming; and testing regulatory measures and safeguards for young users and their families. Methodological improvements include the use of holistic frameworks and designs to examine positive and negative effects; considering gaming activity as a unit of analysis in psychological and social terms as distinct from “screen time;” the use of objective measures to complement self-report measures; and the use of industry data. As gaming products evolve and become increasingly embedded into young people’s lives, understanding the psychosocial effects on young users remains a global priority.

Purpose Breast cancer (BC) is the most prevalent cancer in women and radiotherapy (RT) is an integral part of its treatment. High-level evidence guides clinical decisions, but given the abundance of guidelines, a need to navigate within the evidence has been identified by the board of the Scientific Association of Swiss Radiation Oncology (SASRO). A pilot project was initiated aiming to create an overview of recent clinically relevant evidence for BC RT, to make it easily available to (radiation) oncologists and radiation oncologists in training. Methods A panel of 10 radiation oncology experts for BC RT, one expert in BC surgery, and one expert in BC medical oncology critically reviewed the relevant literature. The panel comprehensively represented different geographical regions of Switzerland as well as university, cantonal, and private institutions. We sought to create a consensual overview of the most relevant questions in BC RT today, accompanied by the most recent and relevant available evidence. Results From January 2023 to January 2024, the panel met four times to review and work on an initial draft. The final draft was reviewed and accepted by all panelists. We hereby publish this work to make it available to international audiences. After publication, the work will be made available to all SASRO members on the SASRO website. This work is to be updated every 2 years. Conclusion The identified need was addressed with a successful pilot project and will be further expanded upon in other tumor pathologies.

Purpose. This work proposes a new method to assess the performance of radiographic anti-scatter grids (ASGs) without the use of a narrow primary beam, which is difficult to achieve. Method. Three general purpose ASGs were evaluated, two marketed ASGs and a low frequency and high ratio prototype ASG with molybdenum lamellae. A range of high scatter x-ray beams were used in a standardized geometry, with energies ranging from 60 kV to 121 kV, for five beam sizes between 10 × 10 and 30 × 30 cm². The scatter fraction (SF) of each beam was measured in the image plane with and without ASG using the lead beam stop method with an extrapolation function derived from the scatter point spread function (PSF). Results. The primary, scatter and total transmissions of the three ASGs measured for the different x-ray beams allowed the calculation of the grid factor, contrast improvement factor and detective quantum efficiency (DQE) as functions of the input SF. The results obtained for the three ASGs are consistent with those obtained with the standard narrow-beam method and data published in the literature, confirmed the prime importance of the ASG primary transmission and revealed important variations in ASG performance, especially as a function of the input SF and beam size. The break-even input SFs at which the imaging system efficiency was improved by the ASG ranged between 0.18 and 0.52 for the different ASGs and beam characteristics. Significance. The method is proposed as an alternative to current ASG characterization techniques.

Background Early worsening of plasma lipid levels (EWL; ≥5% change after 1 month) induced by at-risk psychotropic treatments predicts considerable exacerbation of plasma lipid levels and/or dyslipidaemia development in the longer term. Aims We aimed to determine which clinical and genetic risk factors could predict EWL. Method Predictive values of baseline clinical characteristics and dyslipidaemia-associated single nucleotide polymorphisms (SNPs) on EWL were evaluated in a discovery sample ( n = 177) and replicated in two samples from the same cohort (PsyMetab; n 1 = 176; n 2 = 86). Results Low baseline levels of total cholesterol, low-density lipoprotein cholesterol (LDL-C) and triglycerides, and high baseline levels of high-density lipoprotein cholesterol (HDL-C), were risk factors for early increase in total cholesterol ( P = 0.002), LDL-C ( P = 0.02) and triglycerides ( P = 0.0006), and early decrease in HDL-C ( P = 0.04). Adding genetic parameters ( n = 17, 18, 19 and 16 SNPs for total cholesterol, LDL-C, HDL-C and triglycerides, respectively) improved areas under the curve for early worsening of total cholesterol (from 0.66 to 0.91), LDL-C (from 0.62 to 0.87), triglycerides (from 0.73 to 0.92) and HDL-C (from 0.69 to 0.89) ( P ≤ 0.00003 in discovery sample). The additive value of genetics to predict early worsening of LDL-C levels was confirmed in two replication samples ( P ≤ 0.004). In the combined sample ( n ≥ 203), adding genetics improved the prediction of new-onset dyslipidaemia for total cholesterol, LDL-C and HDL-C ( P ≤ 0.04). Conclusions Clinical and genetic factors contributed to the prediction of EWL and new-onset dyslipidaemia in three samples of patients who started at-risk psychotropic treatments. Future larger studies should be conducted to refine SNP estimates to be integrated into clinically applicable predictive models.

Graves’ disease is the most common form of hyperthyroidism, especially in younger people. Current European guidelines recommend antithyroid drugs as initial treatment for a period limited to 12–18 months. Definitive treatment such as surgery or radioactive iodine is proposed in the case of contraindication to antithyroid drugs or in the case of recurrence after medical treatment. However, more recent studies show that long-term antithyroid treatment is associated with reduced risk of recurrence without an increase in adverse effects. Such data support the option of long-term treatment of Graves’ disease with antithyroid drugs and suggest the necessity for a change to long-standing practices in the field. Herein, after reviewing some general knowledge on Graves’ disease treatment, we discuss the evidence regarding long-term treatment of Graves’ disease with antithyroid drugs for endocrinologists, internists, and other specialists involved in the management of these patients. We consider the main studies in the field, outline their respective strengths and limitations, and, finally, present our opinion on when, in the light of this new evidence, endocrinologists should consider long-term treatment with antithyroid drugs.

Secondary peritonitis with intra-abdominal abscesses (IAA) is difficult to treat because of the supposed low rate of penetration of antimicrobial drugs at the site of infection. However, clinical data about the actual bioavailability of antimicrobial drugs in IAA are scarce. This prospective observational study aimed at assessing the drug penetration in IAA of the antibiotics (piperacillin-tazobactam, carbapenems) and antifungals (fluconazole, echinocandins) that are usually recommended for the treatment of intra-abdominal infections. Patients with IAA who underwent a radiological or surgical drainage procedure were included. Antimicrobial drug concentrations were measured in IAA (C IAA ) and in a simultaneous plasma sample (C plasma ) to assess the C IAA /C plasma ratio. The pharmacodynamic target was defined as a C IAA equal or superior to the clinical breakpoints of susceptibility of the most relevant intra-abdominal pathogens. Clinical outcomes were assessed at hospital discharge. A total of 54 antimicrobial drug measurements were performed in 39 IAA samples originating from 36 patients. Despite important inter-individual variability, piperacillin-tazobactam exhibited the highest C IAA /C plasma ratios (median 2). The rates of target achievement were 75%–80% for piperacillin-tazobactam and meropenem but 0% for imipenem and ertapenem. These results tended to correlate with clinical outcomes (96% success rate versus 73%, respectively, P = 0.07). Among antifungals, fluconazole exhibited higher C IAA /C plasma ratios and rates of target achievement compared to echinocandins. However, no differences in clinical outcomes were observed. These results provide unique information about antimicrobial drug penetration in IAA in real clinical conditions and suggest that piperacillin-tazobactam and meropenem may have better efficacy compared to imipenem or ertapenem.

Background Osteogenesis imperfecta (OI) is a rare hereditary bone disease resulting from a defect in collagen synthesis or processing, leading to bone fragility, frequent fractures and skeletal deformities. OI is associated with increased respiratory morbidity and mortality, but the mechanisms of lung involvement are poorly understood, and there are no data on the natural history of lung function. We studied lung function over time in a cohort of adult OI patients at one center. Methods We used data from OI patients aged 15 and above followed up at the Lausanne university hospital between 2012 and 2023 with available pre-bronchodilator spirometry. Associations between spirometric measurements at first visit and clinical characteristics were studied through linear regression. Changes of spirometric variables over time were analysed through mixed linear regression. Models were adjusted for age, sex, height and OI type (Sillence classification). Results Among 46 subjects, 24% had impaired spirometry at baseline, with similar distribution between restrictive (8.7%), obstructive (8.7%) and mixed (6.5%) ventilatory patterns. At first visit, higher age was associated with lower FEV1 (β = −0.019 l, p = 0.014) and lower FEV1/FVC (β = −0.175%, p = 0.012). A history of asthma was associated with higher FEV1 (β = 0.636 l, p = 0.028) and FVC (β = 0.834 l, p = 0.010). At first visit, FEV1 (β = −0.750 l, p = 0.006) and FVC (β = −0.859 l, p = 0.004) was lower in individuals with OI Sillence types 3, 4 or 5 compared to type 1. Over a mean follow-up of 3.4 years, smokers had a greater decline of FEV1/FVC compared to non-smokers (β = −6.592%, p = 0.007). Individuals with a mutation in the gene COL1A2 had 740 ml lower FVC compared to those with a mutation in COL1A1 (p = 0.037). After adjustment for sex, age, height and OI type, FEV1 increased by 26 ml (95% CI 8; 45) or 1.28%pred (0.51; 2.05) and FVC increased by 25 ml (95% CI 8; 43) or 0.93%pred (0.31; 1.55) per year of follow-up. Conclusions An increase of FEV1 and FVC over time was observed in OI patients after adjustment for other variables, suggesting that the defective collagen synthesis may impact the pulmonary interstitium and lead to increased lung compliance and hyperinflation, in contrast to skeletal deformities, which reduce the thoracic volume. Lung function changes in OI thus result from the interplay of several mechanisms. Supplementary Information The online version contains supplementary material available at 10.1186/s13023-024-03452-y.

Introduction Early adolescence is a time of high psychosocial stress exposure and high stress reactivity, associated with the development of mental disorders. Understanding how the brain reacts to acute and social stressors during this period might help us detect and protect those at risk. Methods We used functional magnetic resonance imaging to investigate acute social stress reactivity in non‐clinical adolescents between ages 13 and 15 years (N = 61) with a range of depression scores (Beck Depression Inventory scores 0–32). Participants underwent a modified Montreal Imaging Stress Task (MIST) with psychosocial stress condition consisting of two parts: acute stress (challenging maths) followed by social feedback (positive or negative), separated by brief recovery periods. The test condition was compared to a non‐stressful control. We examined brain responses to social feedback relative to the acute stressor and feedback valence. Results Psychosocial stress produced differential activation in the paracingulate gyrus, insula, and deactivation in the ventral striatum. Receiving social feedback, compared to acute stress, activated cortical midline regions such as the medial prefrontal cortex and posterior cingulate cortex. Positive feedback increased activity in frontal pole and middle frontal gyrus whereas negative feedback did not show any differential response in the whole group. However, participants with depressive symptoms reacted with higher activation in the posterior cingulate cortex to negative feedback. Conclusion We show that social feedback after an acute stressor activates regions involved in self‐referential processing, with positive feedback eliciting generally higher activation and negative feedback impacting only individuals with vulnerable mood traits during early adolescence.

Background Clinical supervision by psychiatric liaison clinicians is frequently provided in medical settings such as oncology and palliative care, but rarely in endocrinology. Consequently, the specific psychosocial issues faced by endocrinologists in their daily clinical practice and how they deal with them remain largely unknown. We aimed to explore individual supervisions of endocrinologists to gain insight into what kind of clinical situations they present, how they react to them and how this is worked through in supervision. Methods The data set consisted of eight audio-recorded first supervision sessions of endocrinologists conducted by liaison psychiatry clinicians, which were transformed into written core stories accounting for key components of each session. A secondary analysis of these core stories was conducted using an interpretative approach, focusing on (i) the types of clinical situations and (ii) the supervisees’ counter-attitudes towards patients. Additionally, particular attention was given to how the supervisors worked through these elements. Results Endocrinologists presented patients who did not adhere to treatment, behaved inexplicably, or held moral values that differed from their own. Challenged by these situations, supervisees experienced negative emotions (e.g., anxiety, irritation, guilt), associated with behavioral reactions (e.g., avoidance) and/or defensive stances (e.g., denial, rationalization). In half of the supervisions, addressing these difficulties allowed supervisees to link key characteristics of the patient interaction with their own unresolved issues; in the other half, supervisees were less inclined to confront themselves with their own contributions to the patient interaction and the supervisor adopted a more active stance, making specific contributions (e.g. support, advise). Conclusions The findings call for training programs addressing “difficult” patients and advocate for closer collaboration between endocrinologists and liaison psychiatry clinicians.

Background The management of avalanche victims in cardiac arrest (CA) is a challenging situation for rescuers. Despite existing specific management algorithms, previous studies have reported poor compliance with international guidelines and incomplete documentation and transmission of the information required for patient management. The Avalanche Victim Resuscitation Checklist (AVRC) was developed in 2014 in response by the International Commission for Mountain Emergency Medicine. Our aim was to assess the impact of the AVRC on the quality of onsite management of critically buried avalanche victims in CA, i.e. the compliance of management with international guidelines and the completeness of documentation of avalanche specific information. Methods We assessed compliance and documentation in a Swiss helicopter emergency medical service (HEMS) between January 2010 and April 2020. Victims buried for more than 24 h were excluded. Results In the 10-year study period, 87 critically buried avalanche victims in CA were treated by the HEMS, 44 of them after the introduction of the AVRC. Enough information was available to assess management compliance in over 90% of cases ( n = 79). Inadequate management ( n = 25, 32%) and incomplete documentation occurred more often in patients with a long burial duration. After the introduction of the AVRC, the compliance of patient management with the guidelines increased by 36% (from 59 to 95%, p < 0.05) and led to complete documentation of the required information for patient management. Conclusions The use of the AVRC improves the quality of management of critically buried avalanche victims in CA and ensures complete documentation of avalanche specific information. Quality improvement efforts should focus on the management of avalanche victims with a long burial duration. The use of the AVRC enables identification and appropriate treatment of patients with hypothermic cardiac arrest.