Deep learning neural networks architectures such Multi Layer Perceptrons (MLP) and Convolutional blocks still play a crucial role in nowadays research advancements. From a topological point of view, these architecture may be represented as graphs in which we learn the functions related to the nodes while fixed edges convey the information from the input to the output. A recent work introduced a new architecture called Kolmogorov Arnold Networks (KAN) that reports how putting learnable activation functions on the edges of the neural network leads to better performances in multiple scenarios. Multiple studies are focusing on optimizing the KAN architecture by adding important features such as dropout regularization, Autoencoders (AE), model benchmarking and last, but not least, the KAN Convolutional Network (KCN) that introduced matrix convolution with KANs learning. This study aims to benchmark multiple versions of vanilla AEs (such as Linear, Convolutional and Variational) against their Kolmogorov-Arnold counterparts that have same or less number of parameters. Using cardiological signals as model input, a total of five different classic AE tasks were studied: reconstruction, generation, denoising, inpainting and anomaly detection. The proposed experiments uses a medical dataset \textit{AbnormalHeartbeat} that contains audio signals obtained from the stethoscope.
Arkadeep Acharya, Akash Ghosh, Pradeepika Verma
et al.
With the increasing use of RetrievalAugmented Generation (RAG), strong retrieval models have become more important than ever. In healthcare, multimodal retrieval models that combine information from both text and images offer major advantages for many downstream tasks such as question answering, cross-modal retrieval, and multimodal summarization, since medical data often includes both formats. However, there is currently no standard benchmark to evaluate how well these models perform in medical settings. To address this gap, we introduce M3Retrieve, a Multimodal Medical Retrieval Benchmark. M3Retrieve, spans 5 domains,16 medical fields, and 4 distinct tasks, with over 1.2 Million text documents and 164K multimodal queries, all collected under approved licenses. We evaluate leading multimodal retrieval models on this benchmark to explore the challenges specific to different medical specialities and to understand their impact on retrieval performance. By releasing M3Retrieve, we aim to enable systematic evaluation, foster model innovation, and accelerate research toward building more capable and reliable multimodal retrieval systems for medical applications. The dataset and the baselines code are available in this github page https://github.com/AkashGhosh/M3Retrieve.
Precision medicine is a promising approach for accessible disease diagnosis and personalized intervention planning in high-mortality diseases such as coronary artery disease (CAD), drug-resistant epilepsy (DRE), and chronic illnesses like Type 1 diabetes (T1D). By leveraging artificial intelligence (AI), precision medicine tailors diagnosis and treatment solutions to individual patients by explicitly modeling variance in pathophysiology. However, the adoption of AI in medical applications faces significant challenges, including poor generalizability across centers, demographics, and comorbidities, limited explainability in clinical terms, and a lack of trust in ethical decision-making. This paper proposes a framework to develop and ethically evaluate expert-guided multi-modal AI, addressing these challenges in AI integration within precision medicine. We illustrate this framework with case study on insulin management for T1D. To ensure ethical considerations and clinician engagement, we adopt a co-design approach where AI serves an assistive role, with final diagnoses or treatment plans emerging from collaboration between clinicians and AI.
Estimating the quality of published research is important for evaluations of departments, researchers, and job candidates. Citation-based indicators sometimes support these tasks, but do not work for new articles and have low or moderate accuracy. Previous research has shown that ChatGPT can estimate the quality of research articles, with its scores correlating positively with an expert scores proxy in all fields, and often more strongly than citation-based indicators, except for clinical medicine. ChatGPT scores may therefore replace citation-based indicators for some applications. This article investigates the clinical medicine anomaly with the largest dataset yet and a more detailed analysis. The results showed that ChatGPT 4o-mini scores for articles submitted to the UK's Research Excellence Framework (REF) 2021 Unit of Assessment (UoA) 1 Clinical Medicine correlated positively (r=0.134, n=9872) with departmental mean REF scores, against a theoretical maximum correlation of r=0.226. ChatGPT 4o and 3.5 turbo also gave positive correlations. At the departmental level, mean ChatGPT scores correlated more strongly with departmental mean REF scores (r=0.395, n=31). For the 100 journals with the most articles in UoA 1, their mean ChatGPT score correlated strongly with their REF score (r=0.495) but negatively with their citation rate (r=-0.148). Journal and departmental anomalies in these results point to ChatGPT being ineffective at assessing the quality of research in prestigious medical journals or research directly affecting human health, or both. Nevertheless, the results give evidence of ChatGPT's ability to assess research quality overall for Clinical Medicine, where it might replace citation-based indicators for new research.
Gautam Rajendrakumar Gare, Tom Fox, Beam Chansangavej
et al.
Accurate and interpretable diagnostic models are crucial in the safety-critical field of medicine. We investigate the interpretability of our proposed biomarker-based lung ultrasound diagnostic pipeline to enhance clinicians' diagnostic capabilities. The objective of this study is to assess whether explanations from a decision tree classifier, utilizing biomarkers, can improve users' ability to identify inaccurate model predictions compared to conventional saliency maps. Our findings demonstrate that decision tree explanations, based on clinically established biomarkers, can assist clinicians in detecting false positives, thus improving the reliability of diagnostic models in medicine.
Background: The efficacy of the 14-day esomeprazole–amoxicillin (EA) dual therapy in eradicating Helicobacter pylori ( H. pylori ) has been widely discussed previously. Vonoprazan, a novel potassium-competitive acid blocker, presents rapid, potent, and long-lasting acid inhibitory effects compared to esomeprazole. However, there is currently a scarcity of direct comparisons between the 10-day vonoprazan–amoxicillin (VA) and the 14-day EA dual therapy for H. pylori eradication. Objectives: This study aimed to compare the efficacy and safety of the 10-day VA and the 14-day EA dual therapy for H. pylori first-line eradication. Design: This study was a prospective, multicenter, open-label, randomized controlled trial. Methods: The study was conducted at 10 hospitals in China. In total, 570 newly diagnosed H. pylori -infected patients were recruited from April 2023 to February 2024. These patients were randomly assigned to either the 10-day VA group (vonoprazan 20 mg twice daily + amoxicillin 1000 mg three times daily) or the 14-day EA group (esomeprazole 20 mg four times daily + amoxicillin 750 mg four times daily). The primary outcome was the eradication rate, with secondary outcomes including adverse events and compliance. Results: The 10-day VA regimen outperformed the 14-day EA regimen in terms of eradication rates in intention-to-treat (ITT) analysis (85.4% vs 76.7%, p = 0.008), modified ITT analysis (90.7% vs 84.8%, p = 0.036), and per-protocol (PP) analysis (91.1% versus 85.5%, p = 0.047). The non-inferiority p -values in all three analyses were less than 0.001. No statistically significant difference was observed in the incidence of adverse events between the two groups (9.1% vs 11.7%, p = 0.308). The 10-day VA regimen demonstrated higher compliance compared to the 14-day EA regimen ( p = 0.006). Conclusion: The 10-day VA dual therapy showed a satisfactory eradication rate of 91.1% (PP analysis), demonstrating good safety and better compliance compared to the 14-day EA dual therapy as the first-line eradication. Trial registration: This trial was registered in the Chinese Clinical Trial Registry (registration number: ChiCTR2300070475) on April 12, 2023.
Diseases of the digestive system. Gastroenterology
Ramona Schweitzer, Stephan Schlögl, Marco Schweitzer
Non-communicable diseases are the leading cause of global deaths. The risk of their development and progression is increased by modifiable behavioral risk factors. Yet, despite the known benefits of primary and secondary prevention, people often do not follow recommendations for a healthier lifestyle. To this end, mobile health (mHealth) applications offer features for behavioral interventions. Yet, reported user engagement is often low. The objective of the work presented in this article is thus to evaluate the suitability of Design Thinking (DT) as a means to inform the development of an mHealth application that helps increase long-term engagement, and consequently supports individuals in sustainably changing their lifestyle. Applying the DT approach, key user needs and challenges were investigated and used to design a first low-fidelity mHealth application prototype. Think-Aloud analysis, task completion, and post-test interviews were then used to evaluate the prototype and generate early-stage insights. Subsequently, a structured, retrospective analysis of this process, evaluating the insight-generation potential of each step in the DT process cycle, was used to reflect on its suitability to inform mHealth application development. The respective results highlight (1) the distinct value of the DT method, particularly in the early stages of a development project; (2) the strong need for interdisciplinary collaboration in such projects, so as to capture realistic end-user requirements and improve the overall effectiveness of the application design; and (3) the significance of integrating behavioral change theories into the design of mHealth applications, in order to promote long-term engagement.
Massimo Fioranelli, Maria Grazia Roccia, Bianca Przybylek
et al.
Abstract Background The inflammatory response is fundamental to the maintenance of an organism’s physiological homeostasis. Inflammation is controlled by a series of biological events driven by specific inflammatory molecules. When inflammation is within the homeostatic range, it is considered physiological; however, it becomes pathological when it exceeds the immune system’s homeostatic control. Main text Nowadays, the treatment of chronic pathological inflammation is a challenge for pharmacology, as current anti-inflammatory drugs are intended to control acute inflammation. The aim of this narrative review was to provide an overview of the role of molecular pharmacognosy and to demonstrate how current transcriptomics techniques can make an important contribution to the study of the biological functions of natural products in the context of multicomponent/multitarget medication. From our findings, although very few studies have been identified, encouraging results for low-grade chronic inflammations (LGCIs) of various causes emerged in recent transcriptomic studies on multicomponent medicinal products composed of plant and organ extracts at the level of the skin and the musculoskeletal system (Traumeel: Tr14), the liver (Lycopodium compositum: HC-24), and the joints (Zeel-T: Ze-14). Conclusion For adequate control of LGCI, molecular pharmacognosy may be an effective approach to exploring potentially useful herbal agents that are consistent with both physiotherapeutic tradition and modern pharmacology.
The consumption of seafood is crucial for food security, but poor hygiene along the food production chain can result in low microbiological quality, posing significant risks for public health and seafood quality. Thus, this study aimed to assess the microbiological quality and antimicrobial sensitivity of <i>E. coli</i> from 69 samples of illegally marketed shrimp and mussels in the Vitória Region, Brazil. These foods exhibited poor microbiological quality due to high counts of mesophilic, psychrotrophic, and enterobacteria microorganisms. While this issue is widespread in this area, shrimp samples displayed higher microbial counts compared to mussels, and fresh mussels had elevated counts of enterobacteria compared to frozen ones. Among the 10 <i>E. coli</i> isolates, none carried the genes <i>blaCTX-M-1</i>, <i>blaCTX-M-2</i>, <i>blaCTX-M-3</i>, <i>blaCTX-M-15</i>, <i>mcr-1</i>, <i>mcr-2</i>, <i>mcr-3</i>, <i>mcr-4</i>, and <i>tet</i>, associated with antibiotic resistance. Phenotypical resistance to tetracycline and fosfomycin was not observed in any isolate, while only 20% demonstrated resistance to ciprofloxacin. Regarding ampicillin and amoxicillin with clavulanic acid, 60% of isolates were resistant, 10% showed intermediate susceptibility, and 30% were sensitive. One isolate was considered simultaneously resistant to β-lactams and quinolones, and none were conserved as ESBL producers. These findings highlight the inherent risks to local public health that arise from consuming improperly prepared seafood in this area.
Background: Expectations may modify outcomes. However, studies often fail to measure expectations. This raises the need for a brief valid and reliable expectancy measure. Objectives: To study treatment expectations in individuals entering acupuncture or rest, validity and test re-test reliability of a single-item expectancy measure graded on a category scale, a Numeric Rating Scale (NRS) and a Visual Analog Scale (VAS), and to identify psychometric differences between the scales. Method: In this methodology study, treatment expectations were measured in 363 participants before they received acupuncture (genuine traditional penetrating or non-penetrating telescopic sham acupuncture, n = 239, 98%, responded) or a control treatment involving just rest (n = 120, 100%, responded), aimed to improve level of relaxation. A treatment expectancy measure, graded on a five-grade category scale, an eight-grade NRS and a 100 mm VAS, was tested for test re-test reliability. Level of expectation and relaxation was measured at baseline, pre- and post-therapy (n = 729 expectancy measurements). Results: The participants scheduled for acupuncture or rest believed moderately (Inter Quartile Range, IQR, moderately-much) and much (IQR moderately-much) the treatment to be effective. The Intra-Class Correlation coefficient versus Kappa coefficient between test and re-test was .868/.868 for the category scale, .820/.820 for the NRS, and .856/.854 for the VAS. The middle step “Believe moderately the treatment to be effective” was equivalent with median 4 (IQR, 3-4) on NRS and median 52 mm (IQR 42-52) on VAS. The response rates were 708 (97%) on the category scale, 707 (97%) on the NRS, and 703 (96%) on the VAS. All three scales discriminated that pre-therapy expectations were more positive in the individuals who reported an improvement in relaxation level ( P < .001-.003). The VAS presented higher responsiveness to detect expectancy changes over time (71% increased expectation), compared to the NRS (52% increased) and the category scale (12% increased), P < .001. Conclusions: Individuals entering acupuncture, or a control intervention, presented positive treatment expectations, and the expectancy measure presented satisfactory reliability, validity, high response rates, sensitiveness, and responsiveness. Integrative cancer therapy researchers who want to control for expectancy-related bias in clinical trials should consider measuring expectation using the single-item expectancy measure.
Neoplasms. Tumors. Oncology. Including cancer and carcinogens
Brinna Anindita, Department of Internal Medicine, Dr. Soetomo Teaching Hospital, Faculty of Medicine Universitas Airlangga, Surabaya, Indonesia, Titong Sugihartono
et al.
Inflammatory bowel disease (IBD) with a poor prognosis may be due to persistent colitis. According to the latest guidelines, monitoring has become a part of the treatment process for colitis. Adequate monitoring of the patient's condition is necessary to determine the course of the disease to prevent the worsening of the condition and suppress the subclinical inflammatory process. This analytical study with a cross-sectional design was conducted to evaluate the activity of colitis using the results of C-reactive protein (CRP) and fecal calprotectin (FC) assays. FC levels were analyzed by ELISA, while CRP levels were analyzed using Siemens Flex particle-enhanced turbidimetric immunoassay. In 30 subjects with endoscopy and biopsy of colitis, 16 men and 14 women had a median age of 52.5 (18–70) years. The median FC value increased by 67 (7.3–722 g/g) and was positive (≥50 g/g) in 20 subjects (66.7%), and the mean CRP value was 13.64 mg/L, positive (10–15 mg/L) in 13 subjects (43.33%), and negative (<10 mg/L) in 17 subjects (56.67%). This study demonstrated that FC had a significant relationship with CRP (r=0.57; p<0.001) in patients with colitis. Assessing the levels of FC and CRP among patients with colitis can be useful to assess the worsening of symptoms early and reduce mortality and morbidity.
I examine the impacts of extending residency training programs on the supply and quality of physicians practicing primary care. I leverage mandated extended residency lengths for primary care practitioners that were rolled out over 20 years in Canada on a province-by-province basis. I compare these primary care specialties to other specialties that did not change residency length (first difference) before and after the policy implementation (second difference) to assess how physician supply evolved in response. To examine quality outcomes, I use a set of scraped data and repeat this difference-in-differences identification strategy for complaints resulting in censure against physicians in Ontario. I find declines in the number of primary care providers by 5% for up to nine years after the policy change. These changes are particularly pronounced in new grads and younger physicians suggesting that the policy change dissuaded these physicians from entering primary care residencies. I find no impacts on quality of physician as measured by public censure of physicians. This suggests that extending primary care training caused declines in physician supply without any concomitant improvement in the quality of these physicians. This has implications for current plans to extend residency training programs.
The technical landscape of clinical machine learning is shifting in ways that destabilize pervasive assumptions about the nature and causes of algorithmic bias. On one hand, the dominant paradigm in clinical machine learning is narrow in the sense that models are trained on biomedical datasets for particular clinical tasks such as diagnosis and treatment recommendation. On the other hand, the emerging paradigm is generalist in the sense that general-purpose language models such as Google's BERT and PaLM are increasingly being adapted for clinical use cases via prompting or fine-tuning on biomedical datasets. Many of these next-generation models provide substantial performance gains over prior clinical models, but at the same time introduce novel kinds of algorithmic bias and complicate the explanatory relationship between algorithmic biases and biases in training data. This paper articulates how and in what respects biases in generalist models differ from biases in prior clinical models, and draws out practical recommendations for algorithmic bias mitigation.
Raman spectroscopy provides spectral information related to the specific molecular structures of substances and has been well established as a powerful tool for studying biological tissues and diagnosing diseases. This article reviews recent advances in Raman spectroscopy and its applications in diagnosing various critical diseases, including cancers, infections, and neurodegenerative diseases, and in predicting surgical outcomes. These advances are explored through discussion of state-of-the-art forms of Raman spectroscopy, such as surface-enhanced Raman spectroscopy, resonance Raman spectroscopy, and tip-enhanced Raman spectroscopy employed in biomedical sciences. We discuss biomedical applications, including various aspects and methods of ex vivo and in vivo medical diagnosis, sample collection, data processing, and achievements in realizing the correlation between Raman spectra and biochemical information in certain diseases. Finally, we present the limitations of the current study and provide perspectives for future research.
Nobel laureates cluster together. 696 of the 727 winners of the Nobel Prize in physics, chemistry, medicine, and economics belong to one single academic family tree. 668 trace their ancestry to Emmanuel Stupanus, 228 to Lord Rayleigh (physics, 1904). Craig Mello (medicine, 2006) counts 51 Nobelists among his ancestors. Chemistry laureates have the most Nobel ancestors and descendants, economics laureates the fewest. Chemistry is the central discipline. Its Nobelists have trained and are trained by Nobelists in other fields. Nobelists in physics (medicine) have trained (by) others. Economics stands apart. Openness to other disciplines is the same in recent and earlier times. The familial concentration of Nobelists is lower now than it used to be.
In this study we analyzed content and marketing tactics of digital medicine companies to evaluate various types of cross site tracking middleware used to extract health information from users without permission. More specifically we examine how browsing data can be exchanged between digital medicine companies and Facebook for advertising and lead generation purposes. The analysis was focused on a small ecosystem of companies offering services to patients within the cancer community that frequently engage on social media. Some companies in our content analysis may fit the legal definition of a personal health record vendor covered by the Federal Trade Commission, others are HIPAA covered entities. The findings of our analysis raise policy questions about what constitutes a breach under the Federal trade Commission's Health Breach Notification Rule. Several examples demonstrate serious problems with inconsistent privacy practices and reveal how digital medicine dark patterns may elicit unauthorized data from patients and companies serving ads. Further we discuss how these common marketing practices enable surveillance and targeting of medical ads to vulnerable patient populations, which may not be apparent to the companies targeting ads.
Mahdie Kian, Elham Hosseini, Tooba Abdizadeh
et al.
Polycystic ovary syndrome (PCOS) is the most common cause of women’s infertility. Some inflammatory pathways play a pivotal role in the pathogenesis of PCOS. This study aimed to investigate the possible beneficial effects of minocycline on chemokine-like receptor 1 (CMKLR1) and Insulin Receptor (INSR) in a PCOS model. A molecular docking study was implemented using Molecular Operating Environment (MOE) software. The PCOS was induced in NMRI mice (mean body weight 14.47±0.23) by 28 days estradiol valerate injection (2 mg/kg/day). The mice were then divided into six groups (n=8 per group, mean body weight 17.77± 0.26): control (received normal saline), PCOS model, control for minocycline, minocycline treated PCOS (50 mg/kg), letrozole treated PCOS (0.5 mg/kg), and metformin-treated PCOS (300 mg/kg). Serum FSH, LH, estradiol (E2), and testosterone were detected by ELISA. The ovarian tissues were stained by hematoxylin and eosin. The CMKLR1 and INSR expression levels were determined by Real-time-PCR. The molecular docking studies showed scores of -10.92 and -9.30 kcal/mol, respectively, for minocycline with CMKLR1 and INSR. Estradiol valerate treatment led to a significant increase in E2, graffian follicle, and decrease in corpus luteum (CL) numbers (P<0.05), while minocycline treatment improved these PCOS features. The minocycline treatment significantly decreased the CMKLR1 expression and increased the INSR expression (P<0.05) while the CMKLR1 expression was increased in PCOS model. Minocycline may improve ovulation in PCOS model by returning E2 to a normal level and increasing CL number (ovulation signs). These beneficial outcomes may be related to the changes in CMKLR1 and INSR gene expression involved in glucose metabolism and inflammation.
Neoplasms. Tumors. Oncology. Including cancer and carcinogens, Biology (General)
Abstract Due to the excellent radiation hardness and high–temperature endurance, diamond detectors are suitable for intense neutron measurements and promising for neutron diagnostics of scientific fusion devices. In the present work, simultaneous measurement of energy spectrum and fluence of neutrons using a diamond detector was realized for the first time. The absolute response matrix of the diamond detector was simulated based on detailed analysis of the nuclear reactions and the proper selection of nuclear reaction data. Neutron energy spectra as well as neutron fluences for 5.0, 5.5, 8.5, 9.5 and 10.5 MeV neutrons from d–d reaction were measured using the diamond detector based on the absolute response matrix. The measured neutron energy spectra and neutron fluences are reasonable compared with those detected using a EJ-309 liquid scintillator and a 238U fission chamber, respectively, which verifies the reliability of the present work. Furthermore, the energy spectrum and fluence of a 14.2 MeV d–t neutron source were also measured using the diamond detector. The present work demonstrates the ability of simultaneous measurement of energy spectrum and fluence as well as for both d–d and d–t neutrons using a diamond detector, which is of great significance for neutron diagnostics of scientific fusion devices.