2 N E W S & M A NAG E M E N T Tools: tracing, tracking, relocating – no problem A truly smart children’s hospital In a busy hospital, thousands of devic- es and equipment are on the move every hour of every day. Beds, blood pressure monitors, wheelchairs, or infusion pumps can be taken to dif- ferent locations and, at times, even ‘lost’. Keeping track of vital tools is a challenge, particularly given the mas- sive throughput of patients and staff shift changes. However, innovative tracking sys- tems are evolving that not only iden- tify where items are, but they also prompt for redelivery to where they are next required. Fujitsu beacons and tags The Dutch Children’s Hospital in the Netherlands gave Fujitsu and partners the opportunity to provide a unique solution to track and trace hospital assets. This involves mobile tags, sensors and anchor nodes that automatically connect with one another to keep track of assets. The initiative is part of a broader project to improve efficiency and quality of care throughout the hos- pital. Dennis van Doorn, Marketing Manager Wireless Solutions at Fujitsu, explained how Fujitsu is working on these projects. From the first successful proof-of- concept pilot for tracing and track- ing, a follow-up pilot is underway, with equipment connected through battery-powered asset tags that send a signal to a mesh network. As the mesh network can also dovetail with smart lighting systems, further opportunities were suggest- ed. This progressed when it emerged that the Dutch Hospital was prepar- ing a project to upgrade existing lighting to sustainable LED smart lighting – where, as part of the Fujitsu connectivity solution, that sys- tem can be interconnected with the other solutions offered. The team began to imagine a smart light system throughout the hospital, connected wirelessly with mesh tech- nology, sensors and asset tags. ‘By rolling out the whole infrastructure of lights, we can do asset track- ing throughout the whole building very easily,’ van Doorn said. ‘From two different groups in the Dutch Hospital – one asking for asset tracking, the other for sustainable lighting – we are now suddenly offering much more.’ renova- tion work, an empty ward presented an ideal opportunity – test bed Due to location in which to install the tech- nology and demonstrate its capabili- ties without impacting on care and other work. The total solution also includes decision-making software and dash- boards for task management within the hospital; IoT Solutions can also benefit the patient and visitor experi- ence – an indoor navigation solution has been added to guide and inform patients and visitors. The solution has been expanded to include energy- harvesting wireless switches that can be addressed to any action, such as a request for cleaning, a call nurse, etc. ‘We are feeding out asset tracking information and location information and visualising medical equipment on the floor plan for nurses and other staff,’ van Doorn said. ‘If stock levels in a certain area gets too low there will be a notification on the dash- board and staff mobile devices.’ With the proof-of-concept under way, four further pilot projects are being established at the hospital. One involves tracking the whereabouts of wheelchairs, others are on the geriatric and gynaecology wards to keep track of beds, infusion pumps and blood pressure measurement devices, while a further project is in a ‘smart operation theatre’. Dashboard and inventory control software ‘There, we are going to connect occupancy sensors to the network, to see if we can save money on occupancy of operation rooms,’ van Doorn explained. ‘If nobody is detected, the lights and ventilation should not be on. We’re going to use sensor data combined with the operation schedule, to see if we can make it more efficient.’ A further proposed study will look at staff movements in intensive care, to remove unnecessary distractions Learn like a human, deduce like a machine AI in diagnostics Artificial Intelligence (AI) is like a huge blanket that can cover anything from innocuous chess computers to robots which, depending on your viewpoint, could save, oppress or obliterate humanity. However, not every jar labelled AI contains AI. So what is intelligence and can it be cre- ated artificially, synthesised like a nature-identical flavouring substance? At the Siemens Diagnostik Campus, Dr Martin Hirsch, co-founder and CSO of Ada Health GmbH, explored why the term artificial intelligence is frequently misunderstood and why, nevertheless, this technology has the potential to be a major driver in healthcare, Daniela Zimmermann reports. For some people, machine learning is the computer equivalent of intel- ligence, but Dr Martin Hirsch begs to differ. ‘For me, first and foremost, intelligence is the ability to generate sensible behaviour in an entirely new situation. An algorithm that iterates a procedure thousands of times and then deduces certain rules from these runs is not necessarily intelligent.’ In other words: intelligence is the ability to generate hypotheses from existing knowledge and, in a next step, to Co-founder and Chief Scientific Officer Martin Hirsch PhD qualified in Neuroscience and has a Diploma in Physiology. A medical researcher turned serial entrepreneur; Dr Hirsch shifted from theory to innovation after publishing his work on nerve modeling in the scientific journal Nature. His first venture was a nerve modeling program that saved thousands of animals from lab testing. He developed the first version of Ada for doctors and continues to shape the way Ada learns today. Hirsch is a grandson of the celebrated Nobel Laureate Werner Heisenberg. recognise possible courses of action. While a computer requires innu- merable data sets to master a certain procedure reliably, humans in many cases need but a single illustrative example to recognise and under- stand complex relationships. ‘That’s due to our ability to abstract,’ Hirsch explains. While building knowledge in humans may take much longer than feeding an algorithm with data sets, the brain of a physician, for example, will acquire a deep under- standing that can offer a sound solu- tion even in unfamiliar cases. Pattern recognition – all the rage in medical applications – might be suf- ficient to provide a working hypoth- esis, but never a reliable diagnosis. A pattern recognised by the computer may indeed be due to patient physiol- ogy, but it may also be due to poor data, such as image artefacts. ‘When building a decision support system, therefore, we shouldn’t start from square one, but build upon validated scientific knowledge,’ the neurosci- entist underlines. Ideally, he says, we use existing knowledge and incorpo- rate insights from pattern recognition. This approach is rather similar to the “human model” where junior physi- cians first acquire knowledge during their medical school studies and later expand and adapt this knowledge with each day of practical experience. ‘How did you figure this out?’ Ada, an AI-based platform developed at Hirsch’s eponymous company, works pretty much according to this pattern. It analyses certain parameters, weighs them and generates a hypoth- esis from these data, accompanied by a probability score for the validity of the hypothesis. ‘This includes collect- ing positive and negative evidence, very much like during a differential diagnosis. Ada tries to compare and contrast similar pathologies to differ- entiate and exclude.’ This procedure is meant to make the ‘thought process’ of the algorithm understandable for the human brain. ‘Transparency is cru- cial,’ according to Hirsch, because ‘AI systems will only be able to prevail in a hospital and healthcare system if the machine can explain why it arrived at a certain conclusion.’ Currently, the system covers 18,000 disease codes from ICD-10 – a number far beyond the capacities of a human physician. ‘Physicians can’t know the entire range of lab diagnostics, since the human brain is unable to process this degree of complexity,’ Hirsch said. Even if the appropriate test is selected, the physician might draw the wrong conclusions from results. This is where AI can help. A recent study published in the Orphanet Journal of Rare Diseases found that, in 54% of cases investigated, AI could indicate a The AI-based platform ADA analyses certain parameters, e.g. if the tonsils have been taken out. The system covers 18,000 disease codes from ICD-10, like viral sinusitis diagnosis faster than a human physi- cian. Chatbot anamnesis While many physicians are rather wary of chatbots, patients appear to have fewer reservations, as the expe- rience with Ada shows. The patients describe their symptoms to a chatbot, just as they do during a conventional anamnesis, and AI algorithms draw a working hypothesis – a preliminary diagnosis – from these data. ‘To date, we have delivered over 12 million such assessments to more than seven million users,’ says Hirsch, who is convinced that his system meets a real need. The chatbot offers patients something they often miss in their doctor’s office: time. The AI assistant is not under pres- sure and asks as many questions as necessary to be able to arrive at a con- clusion. Moreover, many people find it easier to talk about delicate person- al health issues when the counterpart is a machine, rather than a human being. ‘The future healthcare path- way will not start in the physician’s waiting room, but in the patient’s palm. People really do want to find out more about their conditions and possible therapies and this, in my opinion, is the point where AI meets the patients,’ Hirsch says. However, he EUROPEAN HOSPITAL Vol 28 Issue 4/19

4 N E W S & M A NAG E M E N T Everyone’s DNA is recorded for disease risks Genetic analysis for all Estonians about any such particularities with their patients. One genome, to take away please! Participants of a pilot group can even have their genomic data hand- ed over to them so that, if they wish, they can pass it on for analysis to another health service provider. However, within the pilot project, the genome data for all partici- pants will be screened to find risk variants, and if increased risk is detected the participants will be invited for further investigations. In this case, referral pathways and data exchange between the Genome Project, hospitals and family doctors will be elaborated and tested. ‘We want to find out how this data can be handled in the healthcare sys- tem.’ One tricky area is how to han- dle hereditary diseases, says Pikani: ‘Biobank participants have agreed to the testing – but what about their relatives? Are we allowed to point towards existing risks, or must we, even? We must think very carefully about the ethical consequences.’ Dr Jaanus Pikani is chairman of the Tartu Biotechnology Park and the Life Science-, Bio- and Health-Tech-Meta- Cluster Organisation ScanBalt. As an entrepreneur and health advisor he also works for the World Bank and the WHO. Dr Pikani is one of the initiators of the Estonian Genome Project. technological infrastructure to han- dle the data in a lawful manner is also planned. Funding from the European Structural Funds is to be made available for this. Pikani hopes to capture the entire Estonian population in the near future: ‘If we gain 100,000 citizens a year, we will have achieved this objective in only ten years.’ As the project is publicly financed, the population must be convinced. ‘The procedure is becoming increasingly more economic, particularly in rela- tion to the possible treatment costs, which can run easily to several thousand euros in some cases.’ Many nations will be watching the genome project with great interest, but it is no coincidence that Estonia is a trailblazer. ‘We are very open- minded towards new technologies; we have been able to vote elec- tronically for the last 15 years, for instance,’ Pikani points out. The fact that the entire population is only 1.3 million also favours the imple- mentation of a national project. If it succeeds in Estonia, Pikani foresees implementation in other countries, and on a larger scale, may well become an option. a p a n d e n / k c o t s r e t t u h S : e c r u o S Report: Daniela Zimmermann When it comes to genetics, Estonia has long been considered a trailblazer, as the ambitious Estonian Genome Project (Eesti Geenivaramu) shows. Its objective is to test the genome of each and every citizen for the risk of diseases. Where risks are detected, testing is followed up with informa- tion, prevention and, where appro- priate, treatment. Dr Jaanus Pikani, one of the initiators, talks about the initial difficulties which the genome project encountered and about its potential for Estonian – and possibly worldwide – healthcare. When the project was launched in 2000, various hurdles had to be over- come. Pikani recalls: ‘Most people were worried about data protection – and this continues to be the case.’ Therefore, the security standards to protect the genome data from misuse were high from the very beginning. Not everything worked well financially, either. Whilst the project (a private-non-profit part- nership between the Estonian gov- ernment with a company called EGeen) initially benefited from the hype caused by reports of success in the International Human Genome Project (HGP), both investors’ inter- est and support markedly waned over time. The project, then handed to the public domain, is now run by the University of Tartu. 150,000 data set with an upward tendency Meanwhile the project now com- prises the data sets from more than 150,000 people; information such as gender, age, height and weight is stored as well as data on medi- cally relevant factors, such as dietary habits, exercise and smoking as well as individual and family medical history. At the core of the examina- tion is a blood test used to extract DNA. Participants’ leucocytes are also cryo-conserved for later use. There is a good reason for this foresight, explains Pikani: ‘When we started collecting data for the biobank there was no affordable technology we could have used to carry out genotyping or sequenc- ing.’ However, advances in genetic research now facilitate the evalua- tion of the data collated. Genetic early warning sys- tem These datasets not only drive sci- ence but also benefit the donors themselves. ‘If a higher risk of developing cancer is found in the genome, for instance, the case is passed to an oncologist and tar- geted sequencing is carried out.’ This way, in 5-10 years, preventive measures such as mammography screening, could become obsolete, Pikani believes. ‘This eliminates the exposure to radiation during an examination, the number of false- positive results will fall, and it also allows us to target younger women who would not usually participate in screening.’ A further aspect concerns phar- macogenomics: ‘A small percent- age – around 2% of the population – metabolise drugs either much faster or slower than the rest,’ Pikani underlines. ‘This is often evident from the genes. If such a marker is found, prescriptions must be adapt- ed to avoid under- or overdosing.’ In Estonia, where prescriptions are administered electronically, doctors could also automatically be warned Pikani sees a large potential in the genome project: ‘Biomarkers such as BRCA2 can predict the occur- rence of certain diseases with high probability – in this case, breast cancer. There are also similar mark- ers for cardiovascular diseases.’ The early detection of such risks can be utilised for effective prevention. ‘Obviously there are some who won’t take the results seriously. But, for many, the knowledge that they are at double or even three times higher risk of suffering a heart attack is a strong enough incentive to do something about obesity or tobacco consumption.’ Doctors must learn to read Until a comprehensive introduc- tion is achieved, family doctors will need to receive thorough training in how to handle genomic data to ensure that they can interpret it correctly. The development of a Endoprosthetic patients face high risk of infection Clinical practice needs realignment Periprosthetic infections are widely underestimated, despite the fact that they can have a mortal outcome as cancer, for example. Particular attention must be paid here to patients with a high risk of infection – which might be caused through pre-existing conditions or sources of infection already present in the body. How can this risk group be better addressed in clinical practice? What methods are available to pre- vent periprosthetic infections or, at least, to detect and treat them more quickly? Biofilm control is the top priority Infections break out sooner or later in 5-10% of all implants placed. Millions of patients are affected, as just in 2017 more than 117 million implants were inserted worldwide, according to Professor Andrej Trampuz, Charité University Hospital Berlin and founder of the Pro-Implant Foundation. The number of implants inserted, as well as the number of infections diagnosed, is rising steadily. In the USA, the number of periprosthetic infections of hip and knee implants increased almost sixfold between 2001 and 2020; the associated costs for the healthcare system are set to rise from just under US$400 million (2001) to an estimated US$1.6 bil- lion in 2020. Trampuz sees the key to solving the problem in the prevention or elimination of the so-called biofilms formed by bacteria on the surface of implants, serving as protection for pathogens and making them hardly reachable by the immune system and drugs. The interdisci- plinary cooperation of surgeons and infectiologists, as well as the closely meshed coordination of the administration of antibiotics, surgi- cal measures and diagnostics, is indispensable. Periprosthetic infections: a complex problem world- wide The experience of the Pro-Implant Foundation and its consultation por- tal for the treatment of implant- associated infections confirm that periprosthetic infections are a sen- sitive issue worldwide. Since 2018, doctors internationally have taken advantage of the portal in over 3,000 cases of infection in implants. Based on the cases submitted, important conclusions can be drawn about the failure of periprosthetic infection treatment. On the one hand, periprosthetic infections often go undiagnosed or are diagnosed too late. This is partly due to the limited number and performance of diagnostic tools available in this area. However, it is often neglected to establish a sound understanding of the patient’s pathogens prior to surgery. On the other hand, it is often not distinguished whether the infection is acute or chronic. However, this differentiation has a considerable influence on the required treatment regimen, according to Trampuz. Infections that are difficult to treat due to resistant biofilm-forming pathogens (e.g. candida, rifampicin- resistant staphylococci) are particu- larly problematic. The aim must be to determine precisely the treatment regimen that not only considers the type of infection, but also requires the least invasive intervention while providing the best functional results. As an additional problem area, Trampuz mentions overlooking low- threshold infections. There are dif- ferent definitions (MSIS, IDSA), each with different criteria, but both can lead to infections or pathogens not being detected. Trampuz sees the identification of identical microor- ganisms from at least two differ- ent cultures as the most important common factor. A new, improved definition (EBJIS) is currently being discussed. A further challenge lies in prov- ing the systemic origin of an infec- tion. There are still 20% of all periprosthetic infections caused by the transmission of pathogens from a distant organ (e.g. oral cavity) via the patient’s bloodstream. Furthermore, often systemic anti- biotics are not used specifically and for long enough. That is why Trampuz advises against the pre- viously propagated suspension of antibiotics during two-stage pros- thesis replacement and favours longer, continuous administration. Antibiotics that work against bio- films, such as ciprofloxacin or ampi- cillin plus gentamicin, are especially helpful here. Local antibiotics in spacers and bone cement, which – ideally com- bined with systemic antibiotics – are used for both primary interventions and revisions, play a decisive role in fighting periprosthetic infections, according to Trampuz. The advan- tage of locally applied antibiotics lies in the fact that they are specifi- cally released on site in significantly higher concentrations (10-100 times higher) than with systemic adminis- tration. At the same time, the entire organism is spared, as the local use of antibiotics only has a slight sys- temic effect. Misjudgements – the lurking danger All of the above areas must be covered in clinical practice. Even a single misjudgement, e.g. over- looking a low-threshold infection, or a too short administration of antibiotics, can lead to revisions and complications that are actually avoidable. There is still potential in clinical practice to expand col- laboration within multidisciplinary teams in order to address peripros- thetic infections more systematically and comprehensively in the future. *Source: Heraeus Medical satellite symposium ‘Arthroplasty patients with high risk for infection – do we need to change our clinical practice?’. EFORT 2019 EUROPEAN HOSPITAL Vol 28 Issue 4/19

6 CA R D I O LO G Y A consensus statement published at EuroPCR 2019 Intracoronary imaging advances Report: Jane MacDougall Interventional cardiologists have been aware of the value of intracoronary (IC) imaging in clinical practice for more than twenty years. However, recent developments and improve- ments in modalities and software have enabled huge strides in its range and scope for both diagnos- tic assessment and in percutane- ous coronary interventions. Imaging options now include intravascular ultrasound (IVUS), optical coher- ence tomography (OCT) and near- infrared spectroscopy. This organic growth has created significant dif- ferences in the use of IC imaging across regions and institutions. Generally speaking, IVUS is the most established IC imaging tech- nique, used in more than 80% of PCI cases in Japan, but OCT the equiva- lence of which has now been con- firmed, is increasing in popularity. In an attempt to homogenise its appli- cation, the European Association of Percutaneous Cardiovascular Interventions (EAPCI) has generated frequent consensus statements on its utility, drawing on clinical evi- dence from imaging studies and best practice experience. The ESC revascularisation guide- lines, published in early May, strengthen the indications for IC imaging by recommendation. In light of these new recommendations there was an official presentation of the consensus statement at EuroPCR and also a training session present- ed by two of the authors, Thomas W Johnson from Bristol, United Kingdom, and Francesco Prati from Rome, Italy. Delineating the extent of coronary artery disease Intracoronary imaging provides an adjunct of information to angiogra- phy and, therefore, is invaluable for delineating the extent of coronary artery disease and characterising the severity of atherosclerotic lesion(s), which may appear ambiguous from angiography alone. Angiography overestimates lumen dimensions and is limited by its two-dimensional view of a three- dimensional structure which, as a consequence, results in poor and/or inaccurate estimations of plaque vol- ume, morphology or lesion severity. All of this information can be more precisely obtained by intravascular imaging, OCT being of particular use in thrombus detection because it can distinguish between red (due to the optical attenuating properties of red blood cells) and white throm- bi, the difference between which dictates the choice of pharmacologi- cal management. At a training session at EuroPCR, the role of IC imaging was consid- ered in three areas: • Defining and optimally treat- ing calcified coronaries: coronary plaques can be classified based on the amount of calcium in a lesion. The presence of segments with partially calcified or calcified plaque has been suggested to have a significant effect on mortality; therefore, the use of intracoronary imaging to assess and guide modi- fication strategies for high-burden calcification is increasing. Calcified plaque is brightly echo-reflective and creates a dense shadow, as well as reverberation markers that appear radially at spaced intervals from the calcified segment. With electronic callipers and computer- ised planimetry, accurate quantita- tive measurements can be made at the narrowest cross-section to determine the minimal luminal area (MLA) and diameter and at another ‘normal’ segment (within 10 mm of the culprit lesion) which is used as a reference for the healthy diameter of the vessel. This allows calcula- tion of the percentage stenosis, plaque area and plaque burden and guides appropriate treatment strategies. • Understanding the potential of intracoronary imaging to define the aetiology of atherosclerotic ACS and to guide management: determining lesion morphology has direct impli- cations for the treatment strategy to be implemented. Historically, treat- ment in ACS has been focused on stabilising thin-cap fibroatheroma associated with thrombus formation. However, the widespread use of IC imaging in research, particularly OCT has shown that in one-third of all ACS and one-quarter of STEACS are caused by plaques with an intact fibrous cap, the majority seen as eroded plaques. Identification of this type of plaque means treatment can be tailored to the individual characteristics of the patient’s dis- ease. Plaque erosion, characterised by endothelial denudation, tends to occur in younger patients includ- ing pre-menopausal women and is associated with active smoking in an absence of traditional fac- tors for heart disease. Generally, its treatment is associated with better outcomes than plaque rupture and if correctly diagnosed can be con- servatively treated. DES stent for practice sessions. Non-atherosclerotic ACS • How to delineate and manage non-atherosclerotic ACS presenta- tions: Historically, IC imaging has concentrated on the use of intra- coronary imaging in pre-PCI assess- ment in patients with acute coro- nary syndrome (ACS). In the stable population angiographic ambiguity occurs under many circumstances, when there is excessive plaque bur- den, calcification, old plaque rup- ture, coronary tortuosity, or aberrant vessel anatomy. Under those conditions, intra- coronary imaging provides clarity, revealing the detail within the ves- sel. Careful acquisition of angio- graphic and IC imaging has created a reliable database that facilitates analysis of IC images, however, correct interpretation still requires a considerable level of expertise, training and experience. Despite being with us for more than 20 years, intracoronary imag- ing is still pushing forward the boundaries of what can be achieved in interventional cardiology. Newer probes, bluetooth and Wi-Fi con- nections mean that imaging is an integral part of the Cath lab, used at all stages of intervention. The rapidly advancing technol- ogy brings higher resolution images, faster acquisition times, combined techniques and a mass of data to mine to gain clinical evidence need- EuroPCR participants enhancing their OCT-guided PCI and physiology assessment techniques in hands-on dummy situations in the Abbott training village. ed to drive the field. given the array of options displayed at EuroPCR, the future of intracoronary imaging- guided PCI is very exciting. * Further info: https://www.youtube.com/ watch?v=7eP1DAiznNo Johnson TW et al. Clinical use of intracoronary imaging. Part 2: an expert consensus document of the European Association of Percutaneous Cardiovascular Interventions European Heart Journal. 2019 ;0:1-19 The quantitative measurement of cardiac troponin I concentration Fast and precise monoclonal antibodies are alkaline phosphatase conjugated instead.’ Mitsubishi Chemical Europe GmbH Willstätter Straße 30, 40549 Düsseldorf, Germany `Phone: +49 211 5 23 92-0 pathfast@mc-e.de www.mitsubishi-chemicals.de PathfastTM hs-cTnI is a chemi-lumines- cent enzyme immunoassay (CLEIA) for the quantitative measurement of cardiac troponin I (cTnI) concentra- tion in whole blood or plasma at the point-of-care (POC). Reagents are single use in all-in-one cartridges and up to six tests in paral- lel can be analysed,’ the manufacturer Mitsubishi reports. ‘Low concentra- tions of cTnI can be analysed by using high sensitivity cardiac troponin (hs-cTnI) assays that meet the criteria for high sensitivity defined by IFCC and ESC.’ Pathfast provides high accuracy and precision of test results similar to central lab analyser, combined with the flexibility of a POCT assay within 17 minutes out of whole blood and plasma by an all-in-one cartridge solution, Mitsubishi confirms. ‘In clini- cal studies Pathfast hs-cTnI has been evaluated for a 99th percentile upper reference limit of 29.0 ng/L at an imprecision of 6.1%, which is less than 10%, and fits for the criteria of hs-cTnI, declared by IFCC. ‘Moreover, gender specific 99th cut off values were identified and 0/1 hour Rule-out and Rule-in algorithms of NSTEMI for Pathfast could be established. In large validation cohorts the NPVs for rule-out of NSTEMI exceeded 98%. Mitsubishi is at ESC Hall 7, Level 2, K200 For Pathfast hs-cTnI, cut off levels have been examined for 669 patients with suspected NSTEMI and NPVs between 98.9% and 100%. The PPVs for rule-in obtained from validation cohorts meet the rule-in criteria of 75-80%. ‘Cut off values with PPVs between 75.2% and 78.6% have been identified and tro- ponin values could be deter- mined above the limit of detection in 66.3% of healthy individuals. Pathfast assays show no biotin inter- ference since utilised EUROPEAN HOSPITAL Vol 28 Issue 4/19

Joseph Schoepf, Professor Uwe for Radiology, Cardiology and Paediatrics and Director of the Department of Cardiovascular Imaging at the Medical University of South Carolina, discusses areas of application for AI-based radiology with Daniela Zimmermann. The car- diothoracic imaging expert and his team were largely involved in the development and early clinical trials of the Siemens AI-Rad Companion Chest CT, a software package for computed tomography (CT) based on artificial intelligence (AI). On CT scans the application can distinguish between thoracic organs, such as the heart, lungs, aorta or ver- tebral bodies, and detect anomalies. Achieving this ability entailed train- ing the software with real clinical cases from Schoepf’s department, where the system was being trialled. Those trials were also used to find answers to further diagnostic ques- tions. ‘Apart from cardiac problems, such as automatic measuring of the aorta, interdisciplinary aspects, such as how to also utilise a chest CT to determine bone density, were also of interest. Furthermore, areas of application for emphysema quan- tification and opportunities for the automatic detection and quantifica- tion of arteriosclerotic plaques, were also examined,’ Schoepf explains. The precision of pulmonary nodule detection achievable with this soft- ware is also currently being tested. Focus on the patient Schoepf believes the value of AI application lies in its versatility: ‘The applications are helpful because they deliver a comprehensive evalu- ation of functional imaging stud- ies, providing us with a complete package of everything that may be important for the patient.’ This is an important step towards patient- centred medicine because the focus shifts from the individual organs to the patient, their needs and the events occurring in different organ sections. This interdisciplinary approach requires efficient communication between the medical specialists involved. Schoepf cites an exam- ple: ‘If a radiologist is looking for plaques and then happens upon pulmonary nodules when the AI-Rad Companion Chest CT is used, they will pass on the result to the doc- examples for this are the quanti- fication of coronary calcium and the quantification of emphysema (which is increasingly important due to advances in emphysema treatment) along with the next, logi- cal step – measuring the aortic root for TAVI. ‘The use of AI applications makes it ever easier to utilise all informa- tion contained in an X-ray or chest CT; this becomes obvious when we look at bone density measure- ments, which would usually require AI-based bone mineral density and vertebra measurements tor in charge of treatment and may recommend further observation, a nuclear-medical examination or a biopsy.’ In the US, unlike in Germany, the patients are much more involved in their treatment. Even smaller hospi- tals offer increasing opportunities for all patients to have direct access to their medical records, so they can inform themselves about results and treatment plans. Schoepf believes this trend is positive: ‘I think it’s good that patients finally can take steps to maintain their health. It empowers them and I believe it’s very important that patients become involved in their treatment plans and healing process.’ AI takes over quantifications One large advantage of the use of AI in radiology is that it frees up valua- ble medical staff from laborious and time-consuming quantifications. The fear that AI will replace the radi- ologist has been sufficiently refuted, Schoepf underlines. ‘This won’t hap- pen. The AI-Rad Companion is a very good example of how AI will help us to carry out work that nobody really likes doing, but which is easy for the computer.’ Some 8 CA R D I O LO G Y AI opens up boundaries between medical disciplines AI-based evaluation of chest CT scans allows comprehensive assessment of different organ sections a separate, additional examination,’ the professor enthuses. All results AI extracts from a chest image are automatically presented as a struc- tured report so that the radiolo- gist and referring doctor can draw conclusions for the treatment plan. This focused view into different organ areas also could be possibly used to comprehensively evaluate the effects of an individual hazard or individual cause of a disease – such as the constellation of calcified plaques, pulmonary nodules and pulmonary emphysema in smokers – in a lung cancer screening popula- tion, for instance. The future – analyses based on deep learning Schoepf believes that, in the future, the possible uses for AI in medicine will go way beyond the current capability of the Rad Companion. With the help of deep learning, computers could take over the inde- pendent evaluation of data sets to find correlations that, as yet, are unknown. ‘This is one of the most exciting applications per se: the computer looks at everything with “fresh eyes” and tries to find cor- relations for which the human brain does not have the neural network capacity,’ he explains. There are already cases where a computer analyses a patient’s entire medical record. If, in future, mil- lions of data sets become available, computers will be able to use their evaluations to make new connec- tions, such as those between fac- tors that cause certain diseases, impact on life expectancy or bear specific risks. Deep learning, how- ever, makes many – and not just Austrian-born Uwe Joseph (Joe) Schoepf is a professor with appointments in Radiology, Cardiovascular Medicine and Paediatrics at the Medical University of South Carolina (MUSC) in Charleston, SC. There he directs the Cardiovascular Imaging Division and is Director of Computed Tomography Research and a Director of the University Designated Centre for Biomedical Imaging. Schoepf grew up in Munich, Germany, where he graduated in medicine at Ludwig Maximilian University (LMU) and received specialist training at its Institute of Clinical Radiology. In 2001, already an accomplished radiologist, he left Munich to pursue his interest in cardiothoracic imaging at Brigham and Women’s Hospital, Harvard Medical School, in Boston, MA. The professor joined MUSC in 2004. lay people – a little uncomfortable because the users can no longer follow the individual steps of the computer’s learning process. The expert is relaxed about this. ‘This black box principle, i.e. the fact that we don’t know why the computer makes connections, can lead to con- fusing results. However, ultimately, this means that we are challenged to scrutinise, examine or eliminate the results.’ He forsees no risk of AI apps generating their own, danger- ous momentum, neither in normal circ-umstances, nor in the distant future Example of coronary stenosis on cCTA (A) and ICA with corresponding CT-FFR (B, C) using computational fluid dynamics and machine learning compared with invasive FFR measurement (D) First BIOMONITOR III case in the Nordic region Diagnosis in just three days It started as a fairly typical case: The 79-year-old patient had suffered unex- plained dizziness for years. To diagnose why, the cardiology team at Sweden’s Kalmar Hospital performed echocar- diograms, Holter ECGs and other tests. However, these tests showed normal sinus rhythm and thus were inconclu- From left: Professor Jörg Carlsson, Dr Hendrik Schreyer and Dr David Olsson of the Department of Cardiology at Kalmar Hospital, Sweden, are the first physicians in the Nordic region to implant the device sive. Dr Hendrik Schreyer, Dr David Olsson and Professor Jörg Carlsson decided to use Biotronik’s recently CE-approved Biomonitor III home monitoring system to finally identify the problem – becoming the first team in the Nordic region to inject the new implant- able cardiac monitor (ICM). Three days later, during the daily data transmission, the monitor detected an arrhythmia and the hospital soon received the relevant clinical and technical information. The physicians then informed their patient that they had found atrial fibrillation and an asystole with an eight-second pause. Based on the diagnosis, they recommended pacemaker implantation. ‘She was very pleased to get a call from us after only three days,’ said Schreyer, who performed the injec- tion. ‘The device detected the problem very reliably. This is important because patients with ICMs often wait around four months until another episode occurs that gives us the information we need to make an accurate diagnosis.’ The Biomonitor III is designed to monitor a patient’s heart rhythm long- term, transmitting data to the physi- cian on a daily basis, to help them detect arrhythmia. Schreyer credits the system’s signal quality for helping to simplify the diagnosis, noting r-wave amplitudes as high as 2 mV being regis- tered on ECG readouts from the device. EUROPEAN HOSPITAL Vol 28 Issue 4/19

1 0 CA R D I O LO G Y New research indicates cardiac regeneration potential Cells combination heals damaged hearts Researchers have discovered a unique combination of cells grown from stem cells that could prove pivotal in helping a heart regenerate after a patient has suffered a myocardial infarction, Mark Nicholls reports. The University of Cambridge research team found that transplanting an area of damaged tissue with a com- bination of heart muscle cells and supportive cells, similar to those that cover the outside of the heart, might help the organ to recover from damage caused by a heart attack. In Cambridge, consultant cardi- ologist Dr Sanjay Sinha and team – collaborating with researchers at the University of Washington, in Seattle – have used supportive epi- cardial cells developed from human embryonic stem cells to help trans- planted heart cells to live longer. To test the combination, they used 3D human heart tissue grown in the lab from human stem cells and found that the supportive epicardial cells helped heart muscle cells to grow and mature and also improve the heart muscle cell’s ability to contract and relax. Transplanted cells restore lost heart muscle In rat models, with hearts dam- aged through myocardial infarction, the combination also allowed the transplanted cells to survive and restore lost heart muscle and blood vessel cells. ‘Basically what we have discovered is what we think is the right combination of cell types that will effectively regenerate a dam- aged heart. Dr Sinha, who is also a consultant cardiologist, said: ‘People who have had a heart attack often do not have much a l l e d r a b m a G e r u a L r D : e g a m I viable muscle left, so the heart does not contract as well. What we would like to do is use stem cells to gener- ate new heart muscle and then actu- ally inject those new heart muscle cells into the heart.’ For the research, the team posed the question of how they could improve the function of cardiomyo- cytes from a stem cell that did not contract particularly well. Realising that epicardial cells are essential for cardiomyocyte to develop normally, already proving significant. ‘We found the grafts that regen- erate muscle were two-and-a-half times as big in these animal models than if we did not have epicardial cells, so they make a huge differ- ence to the amount of new muscle that is basically regenerating that damaged heart. I think that is a big step,’ Sinha said. ‘We found that epicardial cells were as good or better than any other cell type that we tried. The heart cells are more mature, they contract better, whole heart function is better and there are more blood vessels inside new heart muscle. ‘Heart cells do not work best by themselves. They need other cell types to support them. We know these epicardial cells are great sup- porting cells when the heart devel- ops, so we are using that same principle to support the heart cells when we regenerate the heart. ‘The beauty of this approach is that nature has been telling us the answers all along because, in devel- opment, this is what nature does, it gets different cell types to work together. We have just taken that principle from development and applied it now to regeneration.’ Epicardial cells, together with heart muscle cells, used in the study to regenerate the damaged heart. Both cell types are generated from human embryonic stem cells. the team mixed heart cells gener- ated from a stem cell with epicardial cells. With the heart muscle cells and epicardial cells generated from embryonic stem cells, and the two derivatives injected to regenerate a damaged rat heart, the results are More success than other centres Sinha hopes his team’s approach will be more successful than that of other centres that have generally focused only on heart muscle cells. ‘We have shown that a combination Beating tissue regenerates cardiac muscle Dr Sanjay Sinha is a Senior Research Fellow at the University of Cambridge (funded by the British Heart Foundation) and an Honorary Consultant Cardi- ologist at Addenbrooke’s Hospital, Cambridge. His research interests focus on the use of stem cells to understand and treat heart disease. of cell types works much better than heart muscle on its own,’ he con- firmed, although he acknowledged there is some way to go before using the cells in human patients. From trials in rats, the team plan to validate their findings in a larger system, such as a pig model, with a view to first-in-man trials in five years’ time. He is confident the current dis- covery is a major step forward and praised Dr Johannes Bargehr, first author of the study, and the Cambridge team, as well as Professor Chuck Murray in Seattle, USA. ‘We should also acknowledge the patients who have damaged hearts and heart failure,’ Sinha added thoughtfully. ‘I see them in hospital and they are the inspiration for what we do.’ Transplant is the main treatment for heart failure, so finding an alter- native is essential. BHF Medical Director Professor Sir Nilesh Samani said: ‘When it comes to mending broken hearts, stem cells haven’t yet really lived up to their early promise. We hope that this latest research represents the turning of the tide in the use of these remarkable cells.’ Patching up a damaged heart Report: Mark Nicholls Scientists in the UK have developed tiny patches of engineered heart tissue that have the potential to be implanted to help people recover from a heart attack. Measuring approximately 3cm x 2cm, the patches contain up to 50 million human-induced pluripotent stem cell derived cardiomyocytes (hiPSC-CM). Yet, these are pro- grammed to turn into working heart muscle that can beat and gradu- ally be incorporated into damaged muscle to help reverse any loss of pumping ability sustained due to myocardial infarction (MI). Richard Jabbour, Tom Owen and the research team from Imperial College London are now keen to design clinical trials to test the safe- ty of the patches and see whether they are equally effective in humans. The rabbit experiments revealed heart function improvements after MI and after four weeks the left ventricle (which pumps blood out through the aorta) was recovering without developing any abnormal heart rhythms. Most significantly, the lab-grown patches appeared to be nourished by blood vessels growing into them from the recipient heart. The research has been led by Sian Harding, Professor of Cardiac Pharmacology at the National Heart and Lung Institute, Imperial College London (ICL), and Director of the BHF Centre for Regenerative Medicine, who explained that her Stem cell patches rotated team has worked on the beating heart patches since 2013 to create new cardiac muscle and target loss following a heart attack. Their work was inspired after the approach of injecting stem cells directly into the myocardium in research at other centres showed lim- ited success with the heart ejecting the cells and triggering arrhythmias. While injection of large numbers of cells disrupted the delicate con- struction of the heart and produced arrhythmia, Professor Harding said the new beating patches – which have a spontaneous rhythm – have offered more promising results as they are sutured lightly on to the heart surface and then covered by the pericardium and they begin to establish a connection to the heart. ‘We do not know very clearly yet whether they have any electrical coupling, certainly not at an early stage, but there is a possibility that there is a mechano-effect: with the patches on, the heart stretches and relaxes; it could automatically syn- chronise the patch with it.’ The team also noticed that the patches did not produce arrhythmi- as in the rabbit model, unlike what happened when large amounts of cells are injected into the centre of the myocardium. Up to 60 million cells are being produced a week in the ICL lab that are 95-98% pure cardiomyocytes and ‘trained’ to develop stronger contractions in unison with the cells placed in hydrogel, which induces the cells to establish connections. ‘When they are placed together in the hydrogel they start to beat and join up,’ she said. ‘They’re suspend- ed between flexible silicone posts so that when they do beat they pull the posts together and because they are doing that, they are exercising against a load so they develop a much stronger contraction.’ The patches are designed to beat spontaneously after three days and mimic mature heart tissue within a month. They also release natural chemicals that would stimulate the heart cells to repair and recover. ‘We had thought about a patch bigger than 2cm x 3cm, but sur- geons advised us that these were suitable, and that a larger number, perhaps 10-15 of them, could be used and then deployed in areas where they saw most damage.’ The research is carried out at the London BHF Centre of Regenerative Medicine where Imperial College London research teams, along with partner universities in the UL and Hamburg focus on innovative ways to get the heart to regenerate. With a view to first-in-man tri- als in 2-4 years, the next step is in ensuring the cells are GMP (Good Manufacturing Practice) compliant and conducting a further pre-clinical trial in the rabbit model. £11,000 per treatment By using more than one patch, the team is confident the therapy can match the one billion cells lost during a heart attack and will also be cost effective at about £11,000 (12,300 euro) per treatment. ‘We hope the cells will integrate or add to the contractility of the heart,’ Harding added. ‘We have EUROPEAN HOSPITAL Vol 28 Issue 4/19

1 2 U LT R A S OU N D POC ultrasound shines in intensive care Pre-set assets and 4-D needle navigation is Professor Felice Eugenio Agrò the Director and Chairman of Postgraduate School in Anaesthesia and Intensive Care at the University Campus Bio-Medico of Rome. During our recent interview he spoke of the need for safe and accurate methods in ICU procedures. Many strong factors in this are related to ultra- sound use in everyday clinical prac- tice, he stressed, adding that these include portability, ergonomics and user-friendliness. ‘Obviously, a port- able unit with advanced ergonomics becomes very crucial to be effec- tive and precise. That’s why tablet solutions are becoming even more appreciated, since ergonomics is becoming key.’ Equipment chosen for use at his ICU includes, for example, Mindray’s TE7 ultrasound system. ‘In our expe- rience TE7 offers very safe, accurate and convenient solutions in anaes- thesia – not only because we can obtain great imaging in soft tissue and identifying structures, but we also have a real-time feedback and good visualisation of the needles in the imaging plane,’ he explained. ‘The TE7 touch screen system provides very good image qual- ity, which is crucial to make rapid patient care decisions. In the era of smartphones and tablets, intuitive gesture controls provide an impor- tant help in being fast and effec- tive. I believe that efficient focused point-of-care exams would minimise the user learning curve, with no need to navigate a knob keyboard, as in many laptop-like and cart- based ultrasound solutions.’ ‘This solution seems to be very promising and the initial outcome is very safe, fast and accurate, increas- ing efficacy and patient comfort,’ Agrò observed. Locating a needle tip in freehand technique ‘In general, ultrasound is not good for locating a needle tip, either out of plane or in plane, in freehand technique, since it becomes difficult to align the needle with ultrasound plane and plan trajectory pre-punc- ture,’ he explained. On the other hand, using a biopsy kit the needle- guided bracket is not flexible. All these issues make standard ultrasound needle access guid- ance techniques quite difficult to learn. ‘Based on mag- netic field induc- tion technology, real-time position detection and orientation of the needle, dis- played on the ultra- sound enhance ultra- sound guid- ance. This way it’s possible to image, predict the trajectory and best scan- ning approach to reach the target before inserting the needle into the region of interest. With this innova- tive approach we believe that the learning curve of doctors, which is approaching free-hand access, will decrease with the result of very accurate and confident needle accesses.’ ‘It’s important to have systems suitable for a disinfected and sterile environment with the possibility of easy cleaning and sterilisation,’ Agrò added. ‘We are talking about a medical environment where we deal with needles, in which safety and effectiveness should be going in the same direction. ‘We appreciate that we can use the TE7 using gloves, no matter that it’s a touch-screen tablet unit. Also, the screen is non-porous, which can easy be disinfected without turning off the unit, thanks to a locking feature.’ Education and at research The objective of the Postgraduate School of Anaes- thesia & Intensive care the Campus Bio-Medico University of Rome is to pro-mote inte- grated structures of teach-ing, research healthcare, and Agrò said ‘We plan to provide students with the technology to speed up their learning curve and increase their con- fidence for a better patient throughput.’ best Felice Eugenio Agrò is Professor of Anaesthesia and Intensive Care, Chairman of the Postgraduate School of Anaesthesia and Intensive Care and Director of the Department of Anaesthesiology, Intensive Care and Pain Management at the University School of Medicine Campus Bio-Medico of Rome, Italy. He has also been the Medical Director and Director of Health Strategy & Business Development for about a decade. He has authored and edited numerous medical publications and reviews manuscripts in Critical Care Medicine, Anaesthesiology, and for the Journal of Clinical Anaesthesia, European Journal of Anaesthesia, British Journal of Anaesthesia, Minerva Anestesiologica etcetera. His research focus lies on airway management, mechanical ventilation in anaesthesia and intensive care, pain management, body fluid management, the evolution of neuromuscular monitoring, haemodynamic monitoring: invasive and non-invasive clinical application. He has been among the Book of Experts for the Ministry of Education, University and Research since 2002. In 2008, he received the Commander to the Order of Merit of the Italian Republic medal from the President of the Italian Republic. Ultrasound can assess intracranial pressure in emergency A future gold standard tool Report: Mark Nicholls researchers acknowledge Whilst ultrasound, when used as a tool to assess intracranial pressure in an emergency, is not a replacement for current gold standard invasive approaches, they believe it has enor- mous potential as a non-invasive and fast, cost-effective, and patient- friendly way to assess possible brain injury at a patient’s bedside. Consultant anaesthetist Dr Chiara Robba, a specialist in the field, suggests the use of ultrasound for intracranial pressure (ICP) assess- ment can become standard practice in the not too distant future. In recent years, Robba, from the San Martino Hospital in Genoa, has worked with colleagues at the University of Cambridge, United Kingdom, to explore the suitability of ultrasound to conduct brain scans in an emergency setting. One recent study compared the relationship between ultrasound- based non-invasive ICP (nICP) and invasive ICP measurement in neu- rocritical care patients and found that it was a ‘promising and easily available technique for identifying critically ill patients with intracranial hypertension’. She outlined latest developments in the field during her presentation ‘An update on the use of US for the estimation of intracranial pressure in emergency’ at the 21-24 February annual congress of WINFOCUS (World Interactive Network Focused on Critical Ultrasound), held in Dubai. ‘Over the last few years, we’ve realised that the use of ultrasound for the brain is suitable in an emer- gency setting,’ Robba explained. ‘When a patient arrives in the emer- gency department, a doctor per- forms an assessment of the body using ultrasound and that generally includes assessment of the heart, lungs and abdomen, but not the brain. A wide range of pathology ‘But you can use ultrasound to provide a lot of information, even when the patient has just arrived in the emergency department. That includes a wide range of pathology like increases of intracranial pres- sure or reduction of cerebral perfu- Non-invasive ICP can be estimated through waveform analysis of the main cerebral arteries. This is an example of transcranial colour duplex sonography with the insinuation of the middle cerebral artery. sion pressure or you can also have a direct assessment and visualisation of intracranial haematoma or haem- orrhage.’ A key reason why ultrasound has not been used previously in this context is due to difficulties of accessing the brain encased in the skull. ‘Ultrasound waves cannot pass through the bone so sometimes it’s difficult to visualise the brain,’ Robba pointed out. ‘But if you use the “windows” – temporal, transorb- ital, submandibular and suboccipital windows – you can have a proper look at the brain.’ This is also a valid care sandard for patients who undergo decompressive craniecto- my, she added. Use of ultrasound for ICP assess- ment requires training, commitment and study, but Robba believes this should be encouraged because of the benefits. ‘You can visualise cer- ebral pathology; get early identifica- tion of intracerebral haemorrhage or intracranial cerebral complica- tions without having to transfer the patient for a CT scan, which can be painful for a patient who is haemodynamically unstable. Also, CT requires the use of radiation and cannot be performed at a patient’s bedside, so the advantages are many.’ Her study team in Genoa and Cambridge is setting up training programmes and writing a book to complement courses already avail- able at Cambridge University. technique, While there remains a lack of awareness of the potential of this relatively new she believes it can become a care stand- ard for patients in emergency units, intensive care or the operating thea- tre where patients are at risk of neu- rological complication, as well as at the bedside and to be used in the same way ultrasound is deployed to monitor the rest of the body. For a clinician, she explained, performing a brain ultrasound at the bedside can identify issues of concern and prompt further tests – or offer reassurance that all is well. ‘From a patient’s perspective, it is safe monitoring, does not cause any harm but adds information. When it becomes routine clinical practice, it’s going to reduce the number of CT scans, the number of radiation exposures and transfers.’ 20 international experts Over the last two years, expo- nents of the technique – including Robba – have been presenting US potential at conferences, and she is currently working with 20 experts from around the world on a con- sensus paper as well as looking at training requirements, standards and competences for practition- ers to be able to use ultrasound as a technique to assess the brain, including the estimation of ICP in an emergency. Yet, amid the enthusiasm for the EUROPEAN HOSPITAL Vol 28 Issue 4/19

1 4 R A D I O LO G Y UK tests high-speed remote medical diagnosis Ultrasound scanning via a 5G network Report: Mark Nicholls To demonstrate advances in 5G con- nectivity for healthcare, a UK team has linked a paramedic in a simu- lated ambulance to a hospital-based clinician. The paramedic wore a robotic or ‘haptic’ glove, which received sig- nals over the live 5G network. Using a joystick, the clinician remotely directed the paramedic to move the ultrasound sensor to where on the patient the clinician wanted to scan. Really high quality and in real-time From this examination, high-quality ultrasound images were transmitted to the clinician in real-time, over the high-bandwidth 5G connection. In addition, a camera in the ambulance From left: Gerry McQuade, CEO of BT Enterprise, with Fotis Karonis, BT 5G Executive Advisor, and Cameron McVittie, Operations Manager at West Midlands Ambulance Service, with the haptic glove and ultrasound used in the ambulance simulation Dr Omkar Chana is programme director for citizen wellbeing at WM5G, covering health and social care and the emergency services. With a PhD in physics, his career has covered hi-tech data and analytics and business strategy specific to the NHS and healthcare, as well as mergers and acquisition. transmitted images of the paramedic and the patient to a second screen at the clinician’s workstation to offer an overall view. WM5G – which aims to acceler- ate deployment of the infrastructure needed for 5G, and is building health, industry and mobility test- beds in the West Midlands region – pointed out the benefit of enabling ultrasound scans to be performed in the field, and reviewed remote- ly, facilitates quicker diagnosis and onward treatment. The demonstration was hosted by the Medical Devices Testing and the Evaluation Centre (MD-TEC), in the University Hospital Birmingham (UHB) simulation lab, at the Institute of Translational Medicine, along with British Telecom (BT), the West Midlands Ambulance Service, and WM5G. ‘5G will help us to roll out this next generation of healthcare technologies,’ said Tim Jones, UHB Chief Innovation Officer. ‘In the future, our clinicians will be able to deliver holistic specialist advice in real time, potentially forming virtual multi-disciplinary teams to provide the best patient care using intelli- gent IT links.’ Doing what you cannot do with 3G or 4G Dr Omkar Chana, WM5G pro- gramme director, added: ‘The ultra- sound demonstration was a flavour of what we can really do with 5G. You cannot do that with 3G or 4G. And, although it was 5G ultra- sound, what we are demonstrating can lend itself to any aspect of clinical imaging and real time com- Andy Street, Mayor of the West Midlands, tries out the haptic glove alongside Omkar Chana, from WM5G, and Fotis Karonis and Jeremy Spencer from BT, watched by Paramedic Cameron McVittie, Operations Manager at West Midlands Ambulance Service Tim Jones is Executive Chief Innovation Officer at University Hospital Birmingham. He joined UHB in 1995, became Head of Service Improvement in 2002 and led the New Hospital Clinical Redesign Programme, before being appointed to the role of Chief Operating Officer in June 2006. In September 2008, he was appointed Executive Director of Delivery, which incorporates board level responsibility for Research & Innovation, Education and Workforce. munication. Also, it does not have to be in an ambulance, it can be in a care home, or GP surgery, and the image quality is just as good as it would be in a hospital.’ The NHS long-term plan Technological solutions – driven by 5G - are at the forefront of the lat- est NHS Long Term Plan as the NHS endeavours to meet the challenges of increased demand and an ageing population. ‘With low latency and the ability to communicate in real time, coupled with faster speed, massive amounts of data can be put through. With ultrasound, the clini- cian can see everything in real time. If he says “move left”, the paramedic in the ambulance moves the probe left – immediately,’ Chana observed. ‘It means the clinician can decide that the patient may not actually need to go to hospital, or may need to be taken to hospital quickly and go straight into surgery because of what they see on scan.’ Assessing the AI revolution Education and management Professor of radiology and director of the Academic Department of Radiology and Medical Informatics at the University of Geneva, Switzerland, Christoph D Becker also chairs the radiology department at the University Hospital. Formerly a radiologist at the University Hospital of Berne, and accredited as Privat-Docent at Berne University, in 1994 he became an abdominal and interventional radiologist at the university hospital in Geneva, and later was professor and vice-director of the radiology division, and since 2004, its director. He also chairs the Academic Department of Radiology and Medical Informatics. Since 2018, he has participated in the creation of the new Diagnostic Department at the hospital with the long- term objective of combining the information of medical imaging, pathology, and genetics in an integrated report How will artificial intelligence (AI) affect continuing education and management in radiology? This issue was discussed by an expert panel at the ESR AI Premium meeting in Barcelona, Mélisande Rouger reports. Continuing education – It must be clear what radiologists need to learn about AI; one way to go could be to give it more space in the train- ing curriculum, according to Elmar Kotter, deputy director of the radi- ology medical centre at Freiburg University. ‘We need to integrate imaging IT, and especially AI, more into the curriculum than it is today, through dedicated AI and informat- ics modules.’ AI should be treated like a new modality to be properly used in daily routine. ‘This requires continu- ous learning for everyone. We have to define what is the minimum AI learning, so that everybody knows the basics. For those who want deeper knowledge and certifica- tion, the ESR and European Society of Medical Imaging Informatics are preparing a diploma,’ explained vice president of the society Kotter. It should also be clear what AI can and cannot do in a given insti- tution, and how algorithms are vali- dated, a task only radiologists can do. ‘We must understand how AI is integrated into our workflow and how our interaction with AI is working and can be optimised. Knowledge of medical legal aspects and ethical implications is also man- datory,’ he added. Online resources to help are increasing. Micro learning, with con- tent from scientific societies and radiology editors is a trend, along with dedicated apps and social media. Systems that pre-process and make data analysis to access ran- dom information are emerging. For example, so-called cognitive radiol- ogy assistants present the radiolo- gist with the right images, ask the referring clinician relevant questions and allow for efficient answers for radiologists, by providing possible relevant data for image interpreta- tion. ‘Such systems give comparative t o h P i a m a l n o h P / k c o t s r e t t u h S : e c r u o S cases already in your PACS system and could also give access to the relevant part of patient history in this case,’ he explained. Solutions that offer access to both archive and information should allow faster learning curves for radi- ologists, by providing similar cases EUROPEAN HOSPITAL Vol 28 Issue 4/19

1 6 O N C O LO G Y Aiming at lethal paediatric tumours Spanish hospital coordinates €10 million EU project tion to help identify the best treat- ment for each individual patient, instead of collecting sample data, as in prospective trials. Exploiting the information from existing imag- ing biobanks and patient clinical files using advanced computational methods will help significantly to improve decision making in can- cer management, according to GIBI La Fe University and Polytechnic Hospital in Valencia, Spain, is coordinating EU-funded program PRIMAGE, which uses precision information from medical imaging to advance knowledge of the most lethal paediatric tumours, by establishing their prognosis and expected treatment response using radiomics, imaging biomarkers and artificial intelligence (AI), Mélisande Rouger reports. eral internationally recognised insti- tutions, and four leading industrial partners, all of which are working under the aegis of the Imaging Biomedical Research Group (GIBI230) at the La Fe Research Institute in Valencia. The great value brought by PRIMAGE is that the study will use real world data as a founda- About six months ago, the European Commission funded the PRIMAGE* project with over €10M, to help improve treatment and identify a tumour’s main characteristics with- out the need for biopsy, using com- putational processing of medical images on the cloud. The PRIMAGE consortium will create a bank of images obtained through AI, using an open cloud- based platform to support decision- making in the clinical management of Neuroblastoma (NB), the most frequent solid cancer of early child- hood, and Diffuse Intrinsic Pontine Glioma (DIPG), the leading cause of brain tumour-related death in chil- dren. The PRIMAGE platform will implement the latest advancement of in-silico imaging biomarkers and modeling of tumour growth towards a personalised diagnosis, prognosis and therapies follow-up. The ambitious project involves 17 European partners, including sev- The PRIMAGE project team Methods, quality assurance and commercial providers issues Molecular testing In terms of success in revolutionary cancer treatment, molecular genetic examination procedures have developed immensely over recent years. They now range from conventional polymerase chain reactions (PCR) or fluorescence-in-situ hybridisation (FISH) to Next Generation Sequencing (NGS) with analysis of the entire exome or genome (Whole-Exome, WES or Whole-Genome, WGS) and of the transcriptome (RNA- seq) within genetic diagnostics. Here, Professor Wilko Weichert MD, Director of the Institute of General Pathology and Pathological Anatomy at the Technical University of Munich, outlines methods, quality assurance and the role of commercial providers in molecular testing. ‘Fundamentally, conventional in-situ protein analysis is still very impor- tant in predictive diagnostics. In part, it has been carried out for dec- ades and is therefore backed by a wide range of data,’ Professor Wilko Weichert points out. ‘Now, along with protein expression analysis, genomics, i.e. sequencing for the detection of mutations, is playing an increasingly important part, often in the shape of multigene analysis, for example. This is currently the backbone of molecular diagnostics.’ Challenges for molecular testing Weichert emphasises that, despite the increasing differentiation and diversification, molecular diagnos- tics is no end in itself but an integral part of treatment planning. ‘What use would the best molecular profile of a tumour be if it does not result in any clinical consequences? We have to tailor molecular diagnostics in such a way that it generates infor- mation that can be used for thera- peutic purposes in the best possible way.’ This is a technological chal- lenge. ‘Complex molecular assays in high-tech patient care originally came from the world of biological research,’ says the pathologist. ‘Their use in medical diagnostics however calls for different qualitative bench- marks, as we must never be wrong.’ In the case of solid tumours, the integration of molecular diagnostics Verification of molecular changes of importance in malignant tumour development can be utilised for therapy into the entire diagnostic algorithm ensures a choice of the most suit- able molecular verification proce- dure for a problem. According to Weichert, ‘Very different assays can be used for this purpose. Their use also depends on the local environ- ment in which the doctors work and on the expertise available to carry out genetic testing with the respec- tive technologies. ’The financial feasibility of mod- ern molecular diagnostics in health- care is currently another impedi- ment. ‘Fact: As doctors and scientists, we are repeatedly faced with situations where, from a purely theoretical perspective, we could offer some- thing diagnostically, but for reasons of affordability it’s just not pos- sible for certain clinical situations,’ Weichert explains. Based on the understanding of scientific medical societies, the costs of molecular analyses, as long as they are specifically for measures of prevention, screening and assurance of diagnosis and/or treatment, must be covered by the cost bearers. In view of fast knowledge growth not only the unified assessment stand- ard (EBM) for out-patient care must be reviewed regularly, but the new examination and treatment proce- dures (NUB) must also be adapted for in-patient care. ‘Financing is a particular problem in Germany due to the differen- tiation between outpatient care and hospital care with all its in-patient flat rates. If a new treatment proce- dure is beneficial for a patient, but not particularly cheap, these flat rates create funding shortfalls. For example, the diagnostic process for critically ill patients is then trans- ferred from in-patient to out-patient care, where the treatment can be funded without putting extra pres- sure on the flat rate per case system in hospitals. This leads to unneces- sary delays in the treatment process.’ Quality assurance is the be-all and end-all Quality assurance in molecular pathology is extremely important. ‘The German national Accreditation Body ensures excellent monitoring of quality standards in pathology,’ Weichert says. ‘Around 100 patholo- gies have been accredited based on a very strict system; this also includes 40 for molecular pathology. They are basically reviewed on an annual basis. ‘This first cornerstone of qual- ity assurance is complemented by a second one in the shape of its own company: The Quality Initiative Pathology of the German Society for Pathology. ‘This subsidiary of the German Society for Pathology and the Federal Association of German Pathologists carries out numerous ring trials which molecular diagnos- tic institutes, for instance, can and must participate in if they want to become accredited.’ At the end of any diagnostic 2030 Director Luis Martí-Bonmatí, Chairman of Radiology at La Fe Hospital. Predicting disease right at the beginning ‘Using real world data in in-silico models can help establish imaging and molecular data’s capacity to estimate and predict right from the beginning of disease, to ultimately make the best therapeutic deci- sion for each patient. This has, to our knowledge, never been done before for any type of cancer in a multicentric approach and using such disruptive technology,’ Martí- Bonmatí said. Being able to precisely estimate tumour phenotype, aggressiveness, stage and extension, define the best treatment plan and establish the most accurate prognosis will defi- nitely help paediatric oncologists improve their decision-making, he added. PRIMAGE will use the latest avail- able technology in the field of computing and AI, for example machine learning software devel- oped by QUIBIM, a spinoff of La Fe’s Research Institute that has received CE marks for its image procedure comes an interdiscipli- nary discussion about patients by a tumour board, to agree on ther- apy recommendations. ‘We need experts in molecular diagnostics in all regions; pathologists are part of tumour boards and make essen- tial contributions to therapy recom- mendations with their expertise,’ Weichert points out. In recent years molecular tumour boards with par- ticular expertise in the interpretation of molecular-diagnostic results have developed. These also work inter- regionally. ‘We will probably have a multistage system in Germany that’s structured into levels of complexity because not all German pathologists will be able to offer the same level of highly complex molecular diag- nostics. The procedures are already too complicated for this, and, if you only have a few cases to work on, also too expensive, he believes. The problem with commercial providers The use of commercially orientated molecular diagnostics in treatment is viewed critically by the specialist societies. They fear a development as seen in the USA: ‘Commercial providers have discovered large- scale provision for themselves,’ Weichert observes. In the USA, sev- eral private laboratories have been set up by exclusively commercial companies to carry out sequenc- ing. Pathologists send in their tis- sue samples and receive a written molecular diagnosis. This became possible as comprehensive regional care entailing this type of complex analytics was not possible via the healthcare providers. These companies are expanding globally and trying to integrate their business model into the existing German system – a system which, unlike in the USA, is comprehensive- ly structured with high expertise. ‘A cuckoo’s egg, as the technical, EUROPEAN HOSPITAL Vol 28 Issue 4/19

1 8 L A B O R AT O RY High-throughput screening advances Acoustic mist ionisation mass spectrometry Report: Mark Nicholls Acoustic mist ionisation mass spec- trometry (AMI-MS) has played a pivotal role in the evolution of high- throughput screening (HTS) and steps are being taken to advance this field to other potential appli- cations. AstraZeneca has been an important player in this area, hav- ing already run more than 10 full collection HTS campaigns against a range of enzyme target types. The company is now working on future applications, such as cell screening. The work to date, and future appli- cations, was outlined by the biop- harmaceutical company’s Principal Scientist Dr Jonathan Wingfield dur- ing SLASEurope 2019, in the pres- entation ‘Continued Development of AMI-MS for High-Throughput Mass Spectrometry Applications Beyond Biochemical HTS’. Wingfield told the congress – which revealed recent developments in instrumentation, assays, and diag- nostics – how AstraZeneca has start- ed to explore cell-based applications with AMI-MS. He illustrated this development with data generated from HepG2 cells comparing the mass spec ‘fingerprints’ from cell lysates. So far, AstraZeneca has tested a library of compounds with known liver damage mode of action, and in several cases showed the modula- tion of metabolites by compounds, cal assays are run using substrate concentrations the μM range, there has usually been enough sensitivity to measure both substrate and product. typically in With the system running on a time-of-flight MS all the signal peaks relating to their assay were visible and, in terms of cellular applica- tions, there was the ability to gener- ate ‘fingerprint’ spectra from lysates. ‘This technology enables us to work with small numbers of cells and the ability to sample multi- ple times means we can generate significant technical and biological replicates to improve the confidence we have in the data even when the analyte signal is low,’ he said. From a workflow perspective, the assay is built in the plate and fired directly from the plate into the MS, so there is no requirement for com- plex re-plating or additional sample preparation post assay. AstraZeneca believes the cell lysate work has the potential to open up new areas of application. Low cost and robust AMI-MS is the hyphenation of a Waters Xevo G2XS time-of-flight (ToF) mass detector with an acous- tic sampling interface. This new technology enables direct injection of samples from a standard 384 well plate into the mass spectrometer at very high-throughput making the workflow process relatively simple An AMI-MS system offers potential for future applications such as cell screening which is consistent with the mecha- nism of action of these compounds. For the AMI-MS work a standard Echo 555 transducer, which had been externalised, was used to pro- vide the acoustic tone burst to the 384-well plate, which generates the spray event of multiple 50-100fL droplets. Above the plate, the team has a charging cone which pushes charge into the test well. The charge sets up a polarity gradient in the well and results in pre-charged droplets being formed as the acoustics fires; the charged droplets are pulled into the MS via a heated transfer line. The speed of the acoustics enables samples to be loaded at a rate of three per second and, as it is con- tactless, there is no risk of cross contamination between samples. The lab team note that speed and sample volume are key strengths of the system and, as only a small volume is sampled, there is always the opportunity to re-read a well if necessary. Dr Wingfield said that AstraZeneca has had considerable success with biochemical HTS despite suppres- sion being an issue and, at times, limiting sensitivity. However, he added that since the biochemi- Dr Jonathan Wingfield is Principal Scientist in Discovery Sciences, R&D, AstraZeneca, Discovery Sciences, based at Cambridge in the UK. He joined AstraZeneca in 2000 as part of a team responsible for delivering automation solutions and technology into the disease area of post high-throughput screening. He was later involved in using leading edge technology to deliver high quality data to global projects, including delivery of acoustic droplet ejection technology. as the assay reagents are assem- bled in the one plate which then is passed to the MS to read. As the automation requirement for AMI-MS is minimal and simple, Wingfield indicates that makes it relatively low cost and robust. In addition to building the hard- ware, there has been a significant investment by AstraZeneca in data and data analysis and handling workflows, which enable HT analy- sis alongside a collaboration with GeneData and its Expressionist soft- ware to automate data analysis. This, said Wingfield, has signifi- cantly reduced the analysis time from hours to minutes for a typical HTS batch of data and the ability to analyse data in close to real time. Finding the right solution LC-MS for the clinical laboratory • Platforms that run laboratory developed tests (LDTs) In the latter case, the development and optimisation of LDTs require extensive verification and validation steps. However, for many laborato- ries, the abundance of technologies and system configurations means the challenge often starts well before LDT optimisation begins. Meeting laboratories’ needs The list of medical devices available for LDTs is rather small compared to the LC-MS instruments available LC-MS has established itself as a powerful tool for clinical applications. However, for maximum utility, it is important that laboratories develop a comprehensive understanding of the available platforms and options. to adopt LC-MS have three potential options: • Turnkey solutions that incorpo- rate instrumentation, software and complete assay kits pro- vided by the manufacturer • Platforms that enable commer- cially available and validated LC-MS methods Inserting the reagent rack into the Thermo Scientific Cascadion SM Clinical Analyser*** of LC-MS in clinical use. LC-MS is now routinely used for a wide range of clinical applications, from the quantitation of Vitamin D (and its analogues) and therapeutic drug monitoring assays, to the quantita- tion of immunosuppressants and toxicology/endocrinology assays. Today, clinical laboratories looking Report: Dr Debadeep Bhattacharyya1 Advances in liquid chromatography (LC) and mass spectrometry (MS) instrumentation have seen the pop- ularity of LC-MS grow rapidly. A wide range of component technolo- gies now exist, from single quadru- pole* and triple quadrupole (QqQ) mass spectrometers, to the use of Thermo Scientific Orbitrap sys- tems with high resolution accurate mass (HRAM)** capability. Similar improvements in high-performance LC (HPLC) technologies mean that the analytical power of LC-MS has kept pace with the need for better selectivity and specificity, greater throughput, automation, and most importantly, increased confidence in results. LC-MS as a clinical tool The power of LC-MS is illustrated by its growth as a key tool for clinical analysis. The improved specificity and selectivity of LC-MS compared to immunoassay techniques, its abil- ity to analyse multiple analytes in one run, as well as the lack of dedi- cated immunoassay kits for certain analytes and the variability observed for some available immunoassays, have all contributed to the uptake *** Product is IVD/CE marked but not 510(k)-cleared and not yet available for sale in the U.S. Availability of the product in each country depends on local regulatory marketing authorization status. for clinical research; choosing the optimal combination of platforms is still not an easy task. The desired platform should offer: biological matrices • Quality data across a range of • Reduced instrument downtime • Fast turnaround times to ensure to minimise costs per sample high productivity and rapid reporting of results • Ability to address high through- • Ease of operation regardless of put sample requirements user expertise Whilst most laboratories will strive to deliver against all these criteria, they may prioritise some needs over others. For example, some laborato- ries will desire LC-MS set-ups that offer the flexibility to develop multi- ple methods, whereas others require high-throughput workflows capable of analysing hundreds of samples each day for a specific analyte. Similarly, some laboratories have the expertise to address complex clini- cal assays with LC-MS, while others would appreciate a fully-automated workflow from sample preparation to report generation. 1 Dr Debadeep Bhattacharyya is Senior Marketing Manager for Clinical and Forensic Applications at Thermo Fisher Scientific. *For research use only. **For in vitro diagnostic use. EUROPEAN HOSPITAL Vol 28 Issue 4/19

Leading International Trade Fair DÜSSELDORF, GERMANY 1821 NOVEMBER 2019 www.medica.de Member of BE PART OF THE NO. 1! WORLD FORUM FOR MEDICINE MEDICA 2019 – Always a heartbeat ahead! Messe Düsseldorf GmbH P.O. Box 10 10 06 _ 40001 Düsseldorf _ Germany Tel. +49 211 4560-01 _ Fax +49 211 4560-668 www.messe-duesseldorf.de