Histopathology and proteomics are synergistic for High-Grade Serous Ovarian Cancer platinum response prediction.

Patients with High-Grade Serous Ovarian Cancer (HGSOC) exhibit varied responses to treatment, with 20-30% showing de novo resistance to platinum-based chemotherapy. While hematoxylin-eosin (H&E) pathological slides are used for routine diagnosis of cancer type, they may also contain diagnostically useful information about treatment response. Our study demonstrates that combining H&E-stained Whole Slide Images (WSIs) with proteomic signatures using a multimodal deep learning framework significantly improves the prediction of platinum response in both discovery and validation cohorts. This method outperforms the Homologous Recombination Deficiency (HRD) score in predicting platinum response and overall patient survival. The study sets new performance benchmarks and explores the intersection of histology and proteomics, highlighting phenotypes related to treatment response pathways, including homologous recombination, DNA damage response, nucleotide synthesis, apoptosis, and ER stress. This integrative approach has the potential to improve personalized treatment and provide insights into the therapeutic vulnerabilities of HGSOC.

A multimodal deep learning architecture for smoking detection with a small data approach.

Covert tobacco advertisements often raise regulatory measures. This paper presents that artificial intelligence, particularly deep learning, has great potential for detecting hidden advertising and allows unbiased, reproducible, and fair quantification of tobacco-related media content. We propose an integrated text and image processing model based on deep learning, generative methods, and human reinforcement, which can detect smoking cases in both textual and visual formats, even with little available training data. Our model can achieve 74% accuracy for images and 98% for text. Furthermore, our system integrates the possibility of expert intervention in the form of human reinforcement. Using the pre-trained multimodal, image, and text processing models available through deep learning makes it possible to detect smoking in different media even with few training data.

[Neurosurgical treatment of tumors of the pineal region - literature review and overview of cases at OMIII]. / A pineális régió tumorainak idegsebészeti ellátása ­ irodalmi áttekintés és az OMIII beteganyagának vizsgálata.

Pineal region tumors account for less than 1% of adult supratentorial tumors. Their treatment requires a multimodality approach. Previously, the treatment of choice was direct surgery, which is associated with high surgical risk. Advances in minimally invasive techniques and onco-radiotherapy offer a safe and multimodal personalized therapy. The aim of our study was to describe the practice of our Institute based on combined endoscopic and radiotherapy techniques. We performed a retrospective clinical study. We processed data from 23 adult patients who underwent endoscopic third ventricle fenestration and pineal tumor biopsy between 2014 and 2023. Descriptive statistics, t-test, Fisher's exact test and Kaplan-Meier analysis were performed. Clinical improvement with endoscopic intervention was achieved in 78.3% of cases. Significant increase in preoperative performance status was observed in the postoperative period (p=2.755e-5), and radiotherapy resulted in regression or stable disease. Our results suggest a safe treatment with good clinical outcome and an excellent alternative to direct surgery.

Are Artificial Intelligence-Assisted Three-Dimensional Histological Reconstructions Reliable for the Assessment of Trabecular Microarchitecture?

Objectives: This study aimed to create a three-dimensional histological reconstruction through the AI-assisted classification of tissues and the alignment of serial sections. The secondary aim was to evaluate if the novel technique for histological reconstruction accurately replicated the trabecular microarchitecture of bone. This was performed by conducting micromorphometric measurements on the reconstruction and comparing the results obtained with those of microCT reconstructions. Methods: A bone biopsy sample was harvested upon re-entry following sinus floor augmentation. Following microCT scanning and histological processing, a modified version of the U-Net architecture was trained to categorize tissues on the sections. Detector-free local feature matching with transformers was used to create the histological reconstruction. The micromorphometric parameters were calculated using Bruker's CTAn software (version 1.18.8.0, Bruker, Kontich, Belgium) for both histological and microCT datasets. Results: Correlation coefficients calculated between the micromorphometric parameters measured on the microCT and histological reconstruction suggest a strong linear relationship between the two with p-values of 0.777, 0.717, 0.705, 0.666, and 0.687 for BV/TV, BS/TV, Tb.Pf Tb.Th, and Tb.Sp, respectively. Bland-Altman and mountain plots suggest good agreement between BV/TV measurements on the two reconstruction methods. Conclusions: This novel method for three-dimensional histological reconstruction provides researchers with a tool that enables the assessment of accurate trabecular microarchitecture and histological information simultaneously.

Annotated dataset for training deep learning models to detect astrocytes in human brain tissue.

Astrocytes, a type of glial cell, significantly influence neuronal function, with variations in morphology and density linked to neurological disorders. Traditional methods for their accurate detection and density measurement are laborious and unsuited for large-scale operations. We introduce a dataset from human brain tissues stained with aldehyde dehydrogenase 1 family member L1 (ALDH1L1) and glial fibrillary acidic protein (GFAP). The digital whole slide images of these tissues were partitioned into 8730 patches of 500 × 500 pixels, comprising 2323 ALDH1L1 and 4714 GFAP patches at a pixel size of 0.5019/pixel, furthermore 1382 ADHD1L1 and 311 GFAP patches at 0.3557/pixel. Sourced from 16 slides and 8 patients our dataset promotes the development of tools for glial cell detection and quantification, offering insights into their density distribution in various brain areas, thereby broadening neuropathological study horizons. These samples hold value for automating detection methods, including deep learning. Derived from human samples, our dataset provides a platform for exploring astrocyte functionality, potentially guiding new diagnostic and treatment strategies for neurological disorders.

Automated prediction of COVID-19 severity upon admission by chest X-ray images and clinical metadata aiming at accuracy and explainability.

In the past few years COVID-19 posed a huge threat to healthcare systems around the world. One of the first waves of the pandemic hit Northern Italy severely resulting in high casualties and in the near breakdown of primary care. Due to these facts, the Covid CXR Hackathon-Artificial Intelligence for Covid-19 prognosis: aiming at accuracy and explainability challenge had been launched at the beginning of February 2022, releasing a new imaging dataset with additional clinical metadata for each accompanying chest X-ray (CXR). In this article we summarize our techniques at correctly diagnosing chest X-ray images collected upon admission for severity of COVID-19 outcome. In addition to X-ray imagery, clinical metadata was provided and the challenge also aimed at creating an explainable model. We created a best-performing, as well as, an explainable model that makes an effort to map clinical metadata to image features whilst predicting the prognosis. We also did many ablation studies in order to identify crucial parts of the models and the predictive power of each feature in the datasets. We conclude that CXRs at admission do not help the predicting power of the metadata significantly by itself and contain mostly information that is also mutually present in the blood samples and other clinical factors collected at admission.

Physical imaging parameter variation drives domain shift.

Statistical learning algorithms strongly rely on an oversimplified assumption for optimal performance, that is, source (training) and target (testing) data are independent and identically distributed. Variation in human tissue, physician labeling and physical imaging parameters (PIPs) in the generative process, yield medical image datasets with statistics that render this central assumption false. When deploying models, new examples are often out of distribution with respect to training data, thus, training robust dependable and predictive models is still a challenge in medical imaging with significant accuracy drops common for deployed models. This statistical variation between training and testing data is referred to as domain shift (DS).To the best of our knowledge we provide the first empirical evidence that variation in PIPs between test and train medical image datasets is a significant driver of DS and model generalization error is correlated with this variance. We show significant covariate shift occurs due to a selection bias in sampling from a small area of PIP space for both inter and intra-hospital regimes. In order to show this, we control for population shift, prevalence shift, data selection biases and annotation biases to investigate the sole effect of the physical generation process on model generalization for a proxy task of age group estimation on a combined 44 k image mammogram dataset collected from five hospitals.We hypothesize that training data should be sampled evenly from PIP space to produce the most robust models and hope this study provides motivation to retain medical image generation metadata that is almost always discarded or redacted in open source datasets. This metadata measured with standard international units can provide a universal regularizing anchor between distributions generated across the world for all current and future imaging modalities.

Peripheral gene interactions define interpretable clusters of core ASD genes in a network-based investigation of the omnigenic theory.

According to the recently proposed omnigenic theory, all expressed genes in a relevant tissue are contributing directly or indirectly to the manifestation of complex disorders such as autism. Thus, holistic approaches can be complementary in studying genetics of these complex disorders to focusing on a limited number of candidate genes. Gene interaction networks can be used for holistic studies of the omnigenic nature of autism. We used Louvain clustering on tissue-specific gene interaction networks and their subgraphs exclusively containing autism-related genes to study the effects of peripheral gene interactions. We observed that the autism gene clusters are significantly weaker connected to each other and the peripheral genes in non-neuronal tissues than in brain-related tissues. The biological functions of the brain clusters correlated well with previous findings on autism, such as synaptic signaling, regulation of DNA methylation, or regulation of lymphocyte activation, however, on the other tissues they did not enrich as significantly. Furthermore, ASD subjects with disruptive mutations in specific gene clusters show phenotypical differences compared to other disruptive variants carrying ASD individuals. Our results strengthen the omnigenic theory and can advance our understanding of the genetic background of autism.

HunCRC: annotated pathological slides to enhance deep learning applications in colorectal cancer screening.

Histopathology is the gold standard method for staging and grading human tumors and provides critical information for the oncoteam's decision making. Highly-trained pathologists are needed for careful microscopic analysis of the slides produced from tissue taken from biopsy. This is a time-consuming process. A reliable decision support system would assist healthcare systems that often suffer from a shortage of pathologists. Recent advances in digital pathology allow for high-resolution digitalization of pathological slides. Digital slide scanners combined with modern computer vision models, such as convolutional neural networks, can help pathologists in their everyday work, resulting in shortened diagnosis times. In this study, 200 digital whole-slide images are published which were collected via hematoxylin-eosin stained colorectal biopsy. Alongside the whole-slide images, detailed region level annotations are also provided for ten relevant pathological classes. The 200 digital slides, after pre-processing, resulted in 101,389 patches. A single patch is a 512 × 512 pixel image, covering 248 × 248 µm2 tissue area. Versions at higher resolution are available as well. Hopefully, HunCRC, this widely accessible dataset will aid future colorectal cancer computer-aided diagnosis and research.

[New, innovative prognosis calculator for patients with metastatic spinal tumors]. / Új, innovatív prognóziskalkulátor áttétes gerinctumoros betegek számára.

Background and purpose: The aim of our research was to create a scoring system that predicts prognosis and recommends therapeutic options for patients with metastatic spine tumor. Increasing oncological treatment opportunities and prolonged survival have led to a growing need to address clinical symptoms caused by meta-stases of the primary tumor. Spinal metastases can cause a significant reduction in quality of life due to the caused neurological deficits. A scoring system that predicts prognosis with sufficient accuracy could help us to achieve personalised treatment options. Methods: Methods - We performed a retrospective clinical research of data from patients over 18 years of age who underwent surgery due to symptomatic spinal metastasis at the National Institute of Mental Disorders, Neurology and Neurosurgery between 2008 and 2018. Data from 454 patients were analysed. Survival analysis (Kaplan-Meier, log-rank, Cox model) was performed, network science-based correlation analysis was used to select the proper prognostic factors of our scoring system, such that its C value (predictive ability index) was maximized. Results: Multivariate Cox analysis resulted in the identification of 5 independent prognostic factors (primary tumour type, age, ambulatory status, internal organ metastases, serum protein level). Our system predicted with an average accuracy of 70.6% over the 10-year study period. Conclusion: Our large case series of surgical dataset of patients with symptomatic spinal metastasis was used to create a risk calculator system that can help in the choice of therapy. Our risk calculator is also available online at https://emk.semmelweis.hu/gerincmet.

Comparison of hesitancy between COVID-19 and seasonal influenza vaccinations within the general Hungarian population: a cross-sectional study.

BACKGROUND: The willingness to get COVID-19 or seasonal influenza vaccines has not yet been thoroughly investigated together, thus, this study aims to explore this notion within the general adult population. METHODS: The responses of 840 Hungarian participants were analysed who took part in a nationwide computer-assisted telephone interviewing. During the survey questions concerning various demographic characteristics, perceived financial status, and willingness to get the two types of vaccines were asked. Descriptive statistics, comparative statistics and word co-occurrence network analysis were conducted. RESULTS: 48.2% of participants were willing to get a COVID-19 vaccine, while this ratio for the seasonal influenza was only 25.7%. The difference was significant. Regardless of how the participants were grouped, based on demographic data or perceived financial status, the significant difference always persisted. Being older than 59 years significantly increased the willingness to get both vaccines when compared to the middle-aged groups, but not when compared to the younger ones. Having higher education significantly elevated the acceptance of COVID-19 vaccination in comparison to secondary education. The willingness of getting any type of COVID-19 vaccine correlated with the willingness to get both influenza and COVID-19. Finally, those who were willing to get either vaccine coupled similar words together to describe their thoughts about a COVID-19 vaccination. CONCLUSION: The overall results show a clear preference for a COVID-19 vaccine and there are several similarities between the nature of willingness to get either type of vaccine.

Hierarchy and control of ageing-related methylation networks.

DNA methylation provides one of the most widely studied biomarkers of ageing. Since the methylation of CpG dinucleotides function as switches in cellular mechanisms, it is plausible to assume that by proper adjustment of these switches age may be tuned. Though, adjusting hundreds of CpG methylation levels coherently may never be feasible and changing just a few positions may lead to biologically unstable state. A prominent example of methylation-based age estimators is provided by Horvath's clock, based on 353 CpG dinucleotides, showing a high correlation (not necessarily causation) with chronological age across multiple tissue types. On this small subset of CpG dinucleotides we demonstrate how the adjustment of one methylation level leads to a cascade of changes at other sites. Among the studied subset, we locate the most important CpGs (and related genes) that may have a large influence on the rest of the sub-system. According to our analysis, the structure of this network is way more hierarchical compared to what one would expect based on ensembles of uncorrelated connections. Therefore, only a handful of CpGs is enough to modify the system towards a desired state. When propagation of the change over the network is taken into account, the resulting modification in the predicted age can be significantly larger compared to the effect of isolated CpG perturbations. By adjusting the most influential single CpG site and following the propagation of methylation level changes we can reach up to 5.74 years in virtual age reduction, significantly larger than without taking into account of the network control. Extending our approach to the whole methylation network may identify key nodes that have controller role in the ageing process.

Supplementary valproate therapy for glioma patients: An alternative opportunity to enhance the efficiency of radio-chemotherapy / A valproátterápia túlélésre gyakorolt hatása gliomás betegekben: Alternatív terápiás lehetoség a radiokemoterápia eredményességének javítására

Összefoglaló. Bevezetés: A gliomák, ezen belül a glioblastoma kezelése továbbra is megoldatlan onkológiai problémát jelent. A szekunder szimptómás epilepsziabetegség megjelenése pozitív prognosztikai faktornak tekintheto a korai diagnosztizálás és az antiepileptikumok potenciális tumorellenes hatásának köszönhetoen. A valproát túlélést hosszabbító hatása már több mint 20 éve az alap- és klinikai kutatások tárgyát képezi. Napjainkban ismert citotoxikus, proapoptotikus, antiangiogenetikus és hiszton-deacetiláz-gátló hatásmechanizmusa. Célkituzés: Kutatásunk célja a valproát túlélést hosszabbító hatásának vizsgálata egy hazai gliomás betegcsoportban. Módszer: Egycentrumos, retrospektív klinikai vizsgálatot végeztünk. A vizsgálatba 122 felnott beteget vontunk be, akiknél 2000 januárja és 2018 januárja között supratentorialis glioma miatt mutét történt, és rohamtevékenység miatt antiepileptikumot (valproát, levetiracetám, karbamazepin) szedtek. Egyúttal gyógyszert nem szedo kontrollcsoportot is kialakítottunk. A populációt vizsgálati és kontrollcsoportokra osztottuk 28 : 52 arányban. Leíró statisztikai, Kaplan-Meier- és log-rank analízist végeztünk. Eredmények: A vizsgált szövettani kategóriák túlélési analízise az irodalmi adatokkal megegyezo értékeket mutatott. A progressziómentes (PFS: p = 0,031) és a teljes (OS: p = 0,027) túlélés tekintetében is szignifikáns eltérés mutatkozott a különbözo antiepileptikumot szedo betegcsoportok között, amely még kifejezettebbé vált a valproátot és az egyéb antiepileptikumot szedo betegek túlélési idejének összehasonlítása során (PFS: p = 0,006; OS: p = 0,015). Következtetés: Vizsgálatunkban a valproát betegeink PFS- és OS-idejének meghosszabbodását eredményezte. Az irodalmi adatok és kutatásunk alapján megfontolandónak tartjuk a valproát elso vonalban történo alkalmazását onkoterápiában részesülo, epilepsziás, agyi gliomás betegekben. Orv Hetil. 2021; 162(24): 960-967. INTRODUCTION: Gliomas still prove to be a serious oncological problem. The presence of epilepsy may present a favorable prognosis due to early diagnosis and the potential antitumor effects of antiepileptic drugs. The survival prolongation effect of valproate has been studied for more than 20 years, nowadays its proapoptotic, anti-angiogenetic, cytotoxic and histone deacetylase inhibitory effects are well known. OBJECTIVE: Our goal was to investigate the survival-enhancing effects of valproate in a Hungarian patient cohort of primary brain tumors. METHOD: A single-center based retrospective clinical trial was designed. In our study, we included 122 patients harboring supratentorial glioma who underwent surgery and experienced seizures between 2000 January and 2018 January. The patients were grouped by the antiepileptic therapies and survival analysis was performed. RESULTS: The Kaplan-Meier curves of the histological categories showed the survival values consistent with the data of the literature. The progression-free (PFS: p = 0.031) and the overall (OS: p = 0.027) survival of the antiepileptic drug categories were significantly different. It was performed by comparing the valproate group and the population formed by the other groups which also showed a significant increase in the survival values (PFS: p = 0.006; OS: p = 0.015). CONCLUSION: Our results show that valproate increases the PFS and OS period of glioma patients in comparison to other antiepileptic drugs. Our data suggest that the use of valproic acid should be considered as a first-line antiepileptic agent in certain well-selected epileptic patients with glioma as a supplement to the oncotherapy. Orv Hetil. 2021; 162(24): 960-967.

Countrywide survey on utilization of medical devices by GPs in Hungary: advantages of the cluster-practice model.

BACKGROUND: Expectations towards general practitioners (GPs) are continuously increasing to provide a more systematic preventive- and definitive-based care, a wider range of multidisciplinary team-based services and to integrate state-of-the-art digital solutions into daily practice. Aided by development programmes, Hungarian primary care is facing the challenge to fulfil its role as the provider of comprehensive, high quality, patient-centred, preventive care, answering the challenges caused by non-communicable diseases (NCDs). AIM: The article aims to provide an insight into the utilization of simple, digital, medical devices. We show the relationship between the primary health care (PHC) practice models and the used types of devices. We point at further development directions of GP practices regarding the utilization of evidence-based medical technologies and how such devices support the screening and chronic care of patients with NCDs in everyday practice. METHODS: Data were collected using an online self-assessment questionnaire from 1800 Hungarian GPs registered in Hungary. Descriptive statistics, Wilcoxon's test and χ2 test were applied to analyze the ownership and utilization of 32 types of medical devices, characteristics of the GP practices and to highlight the differences between traditional and cluster-based operating model. FINDINGS: Based on the responses from 27.7% of all Hungarian GPs, the medical device infrastructure was found to be limited especially in single GP-practices. Those involved in development projects of GP's clusters in the last decade reported a wider range and significantly more intensive utilization of evidence-based technologies (average number of devices: 5.42 versus 7.56, P<.001), but even these GPs are not using some of their devices (e.g., various point of care testing devices) due to the lack of financing. In addition, GPs involved in GPs-cluster development model programmes showed significantly greater willingness for sharing relatively expensive, extra workforce-demanding technologies (χ2 = 24.5, P<.001).

Countrywide population movement monitoring using mobile devices generated (big) data during the COVID-19 crisis.

Mobile phones have been used to monitor mobility changes during the COVID-19 pandemic but surprisingly few studies addressed in detail the implementation of practical applications involving whole populations. We report a method of generating a "mobility-index" and a "stay-at-home/resting-index" based on aggregated anonymous Call Detail Records of almost all subscribers in Hungary, which tracks all phones, examining their strengths and weaknesses, comparing it with Community Mobility Reports from Google, limited to smartphone data. The impact of policy changes, such as school closures, could be identified with sufficient granularity to capture a rush to shops prior to imposition of restrictions. Anecdotal reports of large scale movement of Hungarians to holiday homes were confirmed. At the national level, our results correlated well with Google mobility data, but there were some differences at weekends and national holidays, which can be explained by methodological differences. Mobile phones offer a means to analyse population movement but there are several technical and privacy issues. Overcoming these, our method is a practical and inexpensive way forward, achieving high levels of accuracy and resolution, especially where uptake of smartphones is modest, although it is not an alternative to smartphone-based solutions used for contact tracing and quarantine monitoring.

Generalized entropies, density of states, and non-extensivity.

The concept of entropy connects the number of possible configurations with the number of variables in large stochastic systems. Independent or weakly interacting variables render the number of configurations scale exponentially with the number of variables, making the Boltzmann-Gibbs-Shannon entropy extensive. In systems with strongly interacting variables, or with variables driven by history-dependent dynamics, this is no longer true. Here we show that contrary to the generally held belief, not only strong correlations or history-dependence, but skewed-enough distribution of visiting probabilities, that is, first-order statistics, also play a role in determining the relation between configuration space size and system size, or, equivalently, the extensive form of generalized entropy. We present a macroscopic formalism describing this interplay between first-order statistics, higher-order statistics, and configuration space growth. We demonstrate that knowing any two strongly restricts the possibilities of the third. We believe that this unified macroscopic picture of emergent degrees of freedom constraining mechanisms provides a step towards finding order in the zoo of strongly interacting complex systems.

Research on the predicting power of the revised Tokuhashi system: how much time can surgery give to patients with short life expectancy?

OBJECT: The primary treatment option for symptomatic metastatic spinal tumors is surgery. Prognostic systems are designed to assist in the establishment of the indication and the choice of surgical methodology. The best-known prognostic system is the revised Tokuhashi system, which has a predictive ability of about 60%. In our study, we are attempting to find the reason for its poor predictive ability, despite its proper separation ability. METHODS: We have designed a one-center-based retrospective clinical trial, by which we would like to test the feasibility and the inaccuracy of the revised Tokuhashi system. In our database, there are 329 patients who underwent surgery. Statistical analysis was performed. RESULTS: A significant increase in survival time was observed in the 'conservative' category. Earlier studies reported OS 0.15 at the 180-day control time, in contrast with our 0.38 OS value. The literature suggested supportive care for this category, but in our population, every patient underwent surgery. Our population passes the 0.15 OS value on day 475. We propose an adjustment of the Tokuhashi category scores. We observed significant success in resolving pain. Motor functions were improved or stabilized compared to changes in vegetative dysfunction. CONCLUSION: According to our results, the Tokuhashi scoring system makes very conservative predictions and prefers non-surgical palliative or supportive care. Surgical treatment increases the life expectancy of patients in poor condition. We propose modifying the therapeutic options of the revised Tokuhashi system, taking into consideration modern spine surgery techniques.

Time evolution of the hierarchical networks between PubMed MeSH terms.

Hierarchical organisation is a prevalent feature of many complex networks appearing in nature and society. A relating interesting, yet less studied question is how does a hierarchical network evolve over time? Here we take a data driven approach and examine the time evolution of the network between the Medical Subject Headings (MeSH) provided by the National Center for Biotechnology Information (NCBI, part of the U. S. National Library of Medicine). The network between the MeSH terms is organised into 16 different, yearly updated hierarchies such as "Anatomy", "Diseases", "Chemicals and Drugs", etc. The natural representation of these hierarchies is given by directed acyclic graphs, composed of links pointing from nodes higher in the hierarchy towards nodes in lower levels. Due to the yearly updates, the structure of these networks is subject to constant evolution: new MeSH terms can appear, terms becoming obsolete can be deleted or be merged with other terms, and also already existing parts of the network may be rewired. We examine various statistical properties of the time evolution, with a special focus on the attachment and detachment mechanisms of the links, and find a few general features that are characteristic for all MeSH hierarchies. According to the results, the hierarchies investigated display an interesting interplay between non-uniform preference with respect to multiple different topological and hierarchical properties.

Generalised thresholding of hidden variable network models with scale-free property.

The hidden variable formalism (based on the assumption of some intrinsic node parameters) turned out to be a remarkably efficient and powerful approach in describing and analyzing the topology of complex networks. Owing to one of its most advantageous property - namely proven to be able to reproduce a wide range of different degree distribution forms - it has become a standard tool for generating networks having the scale-free property. One of the most intensively studied version of this model is based on a thresholding mechanism of the exponentially distributed hidden variables associated to the nodes (intrinsic vertex weights), which give rise to the emergence of a scale-free network where the degree distribution p(k) ~ k-γ is decaying with an exponent of γ = 2. Here we propose a generalization and modification of this model by extending the set of connection probabilities and hidden variable distributions that lead to the aforementioned degree distribution, and analyze the conditions leading to the above behavior analytically. In addition, we propose a relaxation of the hard threshold in the connection probabilities, which opens up the possibility for obtaining sparse scale free networks with arbitrary scaling exponent.

[Potential applications of deep learning-based technologies in Hungarian mammography]. / A számítógépes mélytanulási technológia várható megjelenése a hazai mammográfiában.

INTRODUCTION AND AIM: The technology, named 'deep learning' is the promising result of the last two decades of development in computer science. It poses an unavoidable challenge for medicine, how to understand, apply and adopt the - today not fully explored - possibilities that have become available by these new methods. METHOD: It is a gift and a mission, since the exponentially growing volume of raw data (from imaging, laboratory, therapy diagnostics or therapy interactions, etc.) did not solve until now our wished and aimed goal to treat patients according to their personal status and setting or specific to their tumor and disease. RESULTS: Currently, as a responsible health care provider and financier, we face the problem of supporting suboptimal procedures and protocols either at individual or at community level. The problem roots in the overwhelming amount of data and, at the same time, the lack of targeted information for treatment. We expect from the deep learning technology an aid which helps to reinforce and extend the human-human cooperations in patient-doctor visits. We expect that computers take over the tedious work allowing to revive the core of healing medicine: the insightful meeting and discussion between patients and medical experts. CONCLUSION: We should learn the revelational possibilities of deep learning techniques that can help to overcome our recognized finite capacities in data processing and integration. If we, doctors and health care providers or decision makers, are able to abandon our fears and prejudices, then we can utilize this new tool not only in imaging diagnostics but also for daily therapies (e.g., immune therapy). The paper aims to make a great mind to do this. Orv Hetil. 2019; 160(4): 138-143.