Posts Tagged ‘image-interpretation training’

Science Policy Forum: Should we trust healthcare explanations from AI predictive systems?

Some in industry voice their concerns

Curator: Stephen J. Williams, PhD

Post on AI healthcare and explainable AI

   In a Policy Forum article in ScienceBeware explanations from AI in health care”, Boris Babic, Sara Gerke, Theodoros Evgeniou, and Glenn Cohen discuss the caveats on relying on explainable versus interpretable artificial intelligence (AI) and Machine Learning (ML) algorithms to make complex health decisions.  The FDA has already approved some AI/ML algorithms for analysis of medical images for diagnostic purposes.  These have been discussed in prior posts on this site, as well as issues arising from multi-center trials.  The authors of this perspective article argue that choice of type of algorithm (explainable versus interpretable) algorithms may have far reaching consequences in health care.


Artificial intelligence and machine learning (AI/ML) algorithms are increasingly developed in health care for diagnosis and treatment of a variety of medical conditions (1). However, despite the technical prowess of such systems, their adoption has been challenging, and whether and how much they will actually improve health care remains to be seen. A central reason for this is that the effectiveness of AI/ML-based medical devices depends largely on the behavioral characteristics of its users, who, for example, are often vulnerable to well-documented biases or algorithmic aversion (2). Many stakeholders increasingly identify the so-called black-box nature of predictive algorithms as the core source of users’ skepticism, lack of trust, and slow uptake (3, 4). As a result, lawmakers have been moving in the direction of requiring the availability of explanations for black-box algorithmic decisions (5). Indeed, a near-consensus is emerging in favor of explainable AI/ML among academics, governments, and civil society groups. Many are drawn to this approach to harness the accuracy benefits of noninterpretable AI/ML such as deep learning or neural nets while also supporting transparency, trust, and adoption. We argue that this consensus, at least as applied to health care, both overstates the benefits and undercounts the drawbacks of requiring black-box algorithms to be explainable.

Source: https://science.sciencemag.org/content/373/6552/284?_ga=2.166262518.995809660.1627762475-1953442883.1627762475

Types of AI/ML Algorithms: Explainable and Interpretable algorithms

  1.  Interpretable AI: A typical AI/ML task requires constructing algorithms from vector inputs and generating an output related to an outcome (like diagnosing a cardiac event from an image).  Generally the algorithm has to be trained on past data with known parameters.  When an algorithm is called interpretable, this means that the algorithm uses a transparent or “white box” function which is easily understandable. Such example might be a linear function to determine relationships where parameters are simple and not complex.  Although they may not be as accurate as the more complex explainable AI/ML algorithms, they are open, transparent, and easily understood by the operators.
  2. Explainable AI/ML:  This type of algorithm depends upon multiple complex parameters and takes a first round of predictions from a “black box” model then uses a second algorithm from an interpretable function to better approximate outputs of the first model.  The first algorithm is trained not with original data but based on predictions resembling multiple iterations of computing.  Therefore this method is more accurate or deemed more reliable in prediction however is very complex and is not easily understandable.  Many medical devices that use an AI/ML algorithm use this type.  An example is deep learning and neural networks.

The purpose of both these methodologies is to deal with problems of opacity, or that AI predictions based from a black box undermines trust in the AI.

For a deeper understanding of these two types of algorithms see here:


or https://www.bmc.com/blogs/machine-learning-interpretability-vs-explainability/

(a longer read but great explanation)

From the above blog post of Jonathan Johnson

  • How interpretability is different from explainability
  • Why a model might need to be interpretable and/or explainable
  • Who is working to solve the black box problem—and how

What is interpretability?

Does Chipotle make your stomach hurt? Does loud noise accelerate hearing loss? Are women less aggressive than men? If a machine learning model can create a definition around these relationships, it is interpretable.

All models must start with a hypothesis. Human curiosity propels a being to intuit that one thing relates to another. “Hmm…multiple black people shot by policemen…seemingly out of proportion to other races…something might be systemic?” Explore.

People create internal models to interpret their surroundings. In the field of machine learning, these models can be tested and verified as either accurate or inaccurate representations of the world.

Interpretability means that the cause and effect can be determined.

What is explainability?

ML models are often called black-box models because they allow a pre-set number of empty parameters, or nodes, to be assigned values by the machine learning algorithm. Specifically, the back-propagation step is responsible for updating the weights based on its error function.

To predict when a person might die—the fun gamble one might play when calculating a life insurance premium, and the strange bet a person makes against their own life when purchasing a life insurance package—a model will take in its inputs, and output a percent chance the given person has at living to age 80.

Below is an image of a neural network. The inputs are the yellow; the outputs are the orange. Like a rubric to an overall grade, explainability shows how significant each of the parameters, all the blue nodes, contribute to the final decision.

In this neural network, the hidden layers (the two columns of blue dots) would be the black box.

For example, we have these data inputs:

  • Age
  • BMI score
  • Number of years spent smoking
  • Career category

If this model had high explainability, we’d be able to say, for instance:

  • The career category is about 40% important
  • The number of years spent smoking weighs in at 35% important
  • The age is 15% important
  • The BMI score is 10% important

Explainability: important, not always necessary

Explainability becomes significant in the field of machine learning because, often, it is not apparent. Explainability is often unnecessary. A machine learning engineer can build a model without ever having considered the model’s explainability. It is an extra step in the building process—like wearing a seat belt while driving a car. It is unnecessary for the car to perform, but offers insurance when things crash.

The benefit a deep neural net offers to engineers is it creates a black box of parameters, like fake additional data points, that allow a model to base its decisions against. These fake data points go unknown to the engineer. The black box, or hidden layers, allow a model to make associations among the given data points to predict better results. For example, if we are deciding how long someone might have to live, and we use career data as an input, it is possible the model sorts the careers into high- and low-risk career options all on its own.

Perhaps we inspect a node and see it relates oil rig workers, underwater welders, and boat cooks to each other. It is possible the neural net makes connections between the lifespan of these individuals and puts a placeholder in the deep net to associate these. If we were to examine the individual nodes in the black box, we could note this clustering interprets water careers to be a high-risk job.

In the previous chart, each one of the lines connecting from the yellow dot to the blue dot can represent a signal, weighing the importance of that node in determining the overall score of the output.

  • If that signal is high, that node is significant to the model’s overall performance.
  • If that signal is low, the node is insignificant.

With this understanding, we can define explainability as:

Knowledge of what one node represents and how important it is to the model’s performance.

So how does choice of these two different algorithms make a difference with respect to health care and medical decision making?

The authors argue: 

“Regulators like the FDA should focus on those aspects of the AI/ML system that directly bear on its safety and effectiveness – in particular, how does it perform in the hands of its intended users?”

A suggestion for

  • Enhanced more involved clinical trials
  • Provide individuals added flexibility when interacting with a model, for example inputting their own test data
  • More interaction between user and model generators
  • Determining in which situations call for interpretable AI versus explainable (for instance predicting which patients will require dialysis after kidney damage)

Other articles on AI/ML in medicine and healthcare on this Open Access Journal include

Applying AI to Improve Interpretation of Medical Imaging

Real Time Coverage @BIOConvention #BIO2019: Machine Learning and Artificial Intelligence #AI: Realizing Precision Medicine One Patient at a Time

LIVE Day Three – World Medical Innovation Forum ARTIFICIAL INTELLIGENCE, Boston, MA USA, Monday, April 10, 2019

Cardiac MRI Imaging Breakthrough: The First AI-assisted Cardiac MRI Scan Solution, HeartVista Receives FDA 510(k) Clearance for One Click™ Cardiac MRI Package


Read Full Post »

Ultrasound in Radiology – Results of a European Survey

Reporter and Curator; Dror Nir, PhD

Ultrasound is by far, the most frequently used imaging modality in patient’s pathway being used by office-based clinicians and in most of hospitals’ departments. This is also true for cancer patients. As the contribution of imaging to the clinical assessment of patients becomes more substantial, the argument around “who is qualified” to perform such assessment is becoming louder and definitely more relevant!

Both the European and the North America Radiology societies are pushing towards establishment of centralized ultrasound services within the hospitals radiology department, still most ultrasound machines are spread between the different departments and being used by all practitioners. ESR’s working group on ultrasound published a report on the status of ultrasound-practice in European hospitals. Quite a shame; only 13% of the hospital addressed for participation in the survey reacted positively. I would like to highlight the most relevant conclusion from this survey, which is valid no matter which hand is holding the probe: Technique-oriented teaching, time and examinations are necessary to learn how to use Ultrasound properly within the framework of organ-oriented and disease training. Personally, I would support the idea that when it comes to management of cancer patients, this will become a “quality requirement” by law, similar to rules applicable to using radio-active substances.

 Here below is the full report:

Organisation and practice of radiological ultrasound in Europe: a survey by the ESR Working Group on Ultrasound

European Society of Radiology (ESR) 

Neutorgasse 9/2, AT-1010 Vienna, Austria

European Society of Radiology (ESR)

Email: communications@myesr.org

URL: http://www.myESR.org

Received: 25 April 2013Accepted: 26 April 2013Published online: 29 May 2013



To gather information from radiological departments in Europe assessing the organisation and practice of radiological ultrasound and the diagnostic practice and training in ultrasound.


A survey containing 38 questions and divided into four groups was developed and made available online. The questionnaire was sent to over 1,000 heads of radiology departments in Europe.


Of the 1,038 radiologists asked to participate in this survey, 123 responded. Excluding the 125 invitations to the survey that could not be delivered, the response rate was 13 %.


Although there was a low response rate, the results of this survey show that ultrasound still plays a major role in radiology departments in Europe: most departments have the technical capabilities to provide patients with up-to-date ultrasound examinations. Although having a centralised ultrasound laboratory seems to be the way forward, most ultrasound machines are spread between different departments. Ninety-one per cent of answers came from teaching hospitals reporting that training is regarded as an art and is needed in order to learn the basics of scanning techniques, after which working in an organ-oriented manner is the best way to learn how to integrate diagnostic US within the clinical context and with all other imaging techniques.

Main Messages

• Hospitals should introduce centralised ultrasound laboratories to allow for different competencies in US under the same roof, share human and technological resources and reduce the amount of equipment needed within the hospital.

• Technique-oriented teaching, time and examinations are necessary to learn how to use US properly within the framework of organ-oriented training.

• A time period of about 6 months dedicated solely to learning US scanning techniques is deemed sufficient in most cases.


The Working Group on ultrasound (US) of the European Society of Radiology was founded in 2009 with the aim of supporting increased quality and visibility of US within radiological departments as well as strengthening the position of US within the radiology community.

Among the many practical goals assigned to the group, one of the most important has been to gather information about the organisation and practice of radiological US in Europe.

This article reports the results of a survey assessing how diagnostic US is practiced and how training in US is organised in radiological departments of European hospitals. Questions were also aimed at evaluating the practice of US within both radiology and other hospital departments in order to understand the relationships among the different users of this technique. A comparison with the results of a previous survey on the US activities within 17 academic radiological departments throughout Europe published in 1999 by Schnyder et al. [1] was also attempted.


A questionnaire was developed to obtain data about the practice of diagnostic US within radiology departments in Europe.

The survey contained 38 questions that were divided into four groups:


Related to the hospital: location; dimensions; presence or absence of teaching duties.


Related to the workload of US: number of US examinations/year, amount of US equipment available; state of available technology; types of most frequent examinations; organisation of the US laboratory; presence of sonographers; methods of reporting and archiving US examinations.


Related to the teaching of US to radiology residents: organisation and duration of training programmes; number of examinations to be performed before completion of the training period; presence of training programmes dedicated to sonographers or other non-radiology residents.


Related to the US examinations performed outside radiology in each hospital; clinical specialists most often involved in performing directly US; availability of special techniques, such as contrast-enhanced ultrasound (CEUS); methods of reporting and archiving US examinations.

The questionnaire was made available online and an invitation to fill it in was sent to all 1,038 heads of radiology departments throughout Europe within the database of the European Society of Radiology. The invitation was repeated three times over a period of 3 months, between June and August 2011.


There were 123 responses to the questionnaire. Considering that 125/1,038 e-mail messages were reported as “undelivered”, the response rate to the invitation was 13 %. Many responders did not answer all the questions presented in the questionnaire, and some answers and comments were somewhat difficult to understand and evaluate.

First group of questions

Answers were gathered from different parts of Europe; 63.4 % were from five nations (Germany, Austria, France, Spain and Italy). The distribution according to countries is presented in Table 1.

Table 1

Nationality of responders

Germany (DE)


Austria (AT)


France (FR)


Spain (ES)


Italy (IT)


Hungary (HU)


Switzerland (CH)


The Netherlands (NL)


Turkey (TR)




Czech Rep (CZ)


Poland (PL)


Denmark (DK)


Romania (RO)


Norway (NO)


Croatia (HR)


Portugal (PT)


Belgium (BE)


Greece (GR)


Montenegro (ME)


Lithuania (LT)


Ireland (IE)


Serbia (RS)


Sweden (SE)


There were 25 responses (20.3 %) from hospitals with fewer than 400 beds, 52 (42.3 %) from hospitals with between 400 and 1,000 beds and 46 (37.4 %) from hospitals with more than 1,000 beds. Most answers were from teaching hospitals (91.1 %).

Second group of questions

Most radiology departments (77 %) have fewer than 10 working US units; 22 % have between 10 and 20 US machines; only 0.8 % have more than 20 machines. Small, portable units are available in 64.5 % of departments, 3D/4D capabilities are present in 52 % and elastography in 48.2 %, and 67.3 % have the possibility to perform CEUS examinations.

Up to 57.6 % of radiology departments perform more than 10,000 examinations per year; between 3,000 and 10,000 examinations per year are performed in 33.1 % of cases; only 9.3 % of departments perform fewer than 3,000 examinations.

Abdominal US is the most frequent exam (51.51 %), followed by breast (14.46 %), musculoskeletal (11.59 %), pelvic (10.88 %) and vascular (10.42 %) US examinations. Contrast-enhanced US (CEUS) studies constitute about 4.39 %. US is used by radiologists in emergency in 96.6 % of cases and in paediatrics in 74.6 %. Comments indicate that most of those who answered “no” did not have a paediatric section in their hospital.

Transvaginal US is used in obstetric examinations by 15.8 % of responders and in gynaecological studies by 50.7 %. Endoscopic US is used by radiologists in 13.4 % and intravascular US in 14.6 %; radiologists are called by surgeons for intraoperative US in 64.2 % of cases.

There were 49 responders who indicated the actual number of US examinations performed/year. The characteristics of hospitals in which the radiology department performs more than 20,000 ultrasound examinations/year are presented in Table 2.

Table 2

Characteristics of the hospitals in which the radiology department performs more than 20,000 US examinations/year (nationality, presence/absence of teaching duties, number of inpatients, number of US machines available, ratio between number of US examinations performed by non-radiology specialists vs. radiologists)


Those who reported fewer than 5,000 US examinations/year are reported in Table 3.

Table 3

Characteristics of the hospitals in which the radiology department performs less than 5,000 US examinations/year (nationality, presence/absence of teaching duties, number of inpatients, number of US machines available, ratio between number of US examinations performed by non-radiology specialists vs. radiologists)


Third group of questions

The first question in this group was whether the hospital was organised with a centralised US laboratory where physicians from all specialties work together.

There were 13/110 positive answers (11.8 %) from Germany (5), Spain (3), Austria (2), Hungary (2) and Croatia (1). All other hospitals have US machines scattered throughout the different radiological and non-radiological departments. The centralised US laboratory is organised together by the radiology and the internal medicine departments in three cases; it is truly multidisciplinary, with all specialties concurring, in three others; it is run by radiology in two. The remaining two positive answers did not provide further detail about their organisation.

The second question related to the role of sonographers. Only 15/110 (13.6 %) department heads stated they work with sonographers. They are located in Spain (3), Germany (2), UK (2), The Netherlands (2), Austria (1), Belgium (1), Ireland (1), Lithuania (1) and Montenegro (1). In all others, US examinations are done directly by the radiologists. There were 12 comments describing how the work of sonographers is organised. Sonographers do both the examination and the report, with the radiologist checking difficult cases only in four hospitals; sonographers do the studies and the radiologist takes a final look and writes the reports in six; two departments state they use sonographers for vascular examinations only.

The third question related to the organisation of training programmes in US. Radiology residents are trained in 91.1 % of responders. Some centres organise a theoretical course on basic principles of US before starting practical activity. Then, clinical practice is usually performed according to organ/systems training schemes. Residents work under close supervision of a senior radiologist: they approach the patient, perform a preliminary examination and issue a first report, which is then checked by the expert. The aim is to obtain progressive growth of competences: from scanning capabilities, to reporting capabilities, to complete independence.

The length of the period of training within the US laboratory in the various teaching hospitals and the minimum number of US examinations required before the end of the residency period are summarised in Tables 4 and5.

Table 4

Length of the period of training within the US laboratory in the 84 teaching hospitals that reported it

No. of teaching hospitals

Length of training


<4 months


4–6 months


6–12 months


>1 year

Table 5

Minimum number of US examinations to be performed before the end of the residency period in the 75 teaching hospitals that reported it

No. of teaching hospitals

Minimum no. of US examination









There was a direct correlation between the number of US exams performed in the department and the depth of US involvement during training: training programmes in the two hospitals where the lowest number of US examinations/year is performed indicate a period of 3 months and 250 and 500 examinations. However, a hospital with a workload of 45,000 US studies per year (in which, however, the examinations are performed by sonographers) suggested only 2–3 months of training and 100 exams before the end of the residency period.

Training is also provided for non-radiology residents in 37 hospitals. It is most frequently offered to internal medicine, gastroenterology, surgery, anesthesiology, vascular surgery and paediatrics. Comments indicate that these radiology courses allow only theoretical teaching, since observation, but not direct contact with patient, is provided for non-radiologists.

All 15 departments working with sonographers provide, or are planning to provide, starting in 2012, training courses for these professionals. These include both theory and practice; the theoretical part is done, in some cases, together with radiology residents.

As an important technical point, it must be noted that US images performed by radiologists are recorded into PACS systems in 85.6 % of cases. Comments on this question indicated that not all equipment is linked to PACS and that only selected images or videos are often archived; furthermore, technical problems in archiving videos have been reported.

A final group of questions pertained to the US examinations performed outside the radiology department in each hospital.

One question asked about the proportion of US examinations performed by radiologists vs. those performed by non-radiologists. European radiologists, as a whole, still perform a higher number of examinations (61.27 %) than non-radiologists (38.32 %). Differences in the percentage of studies performed in the different hospitals are presented in Table 6.

Table 6

Proportion of US examinations performed by radiologists vs. non-radiologists. Although radiologists, as a whole, perform more US examinations than non-radiologists, the table shows there are differences among different departments, with slightly more than 50 % performing more than 70 % of the studies

% of hospital US exams performed by radiologists

No. of radiology departments

≥90 %

25 (20.32 %)

70–90 %

37 (30.08 %)

10–70 %

57 (46.35 %)

<10 %

4 (3.25 %)

Comments indicate that most OB/GYN, neurology, vascular, urology, internal medicine, anaesthesiology and gastroenterology departments run their own US units in their wards. CEUS is used in 35.1 % of gastroenterology departments, in 15.1 % of internal medicine, in 10.6 % of transplant units and in 10.4 % of nephrology departments.

The examinations performed out of the radiology department are formally reported in 64.4 % of cases only. Comments indicate that reports are fully stored within the Hospital Information System (HIS) in 31 cases; storage is only partial in 24; no HIS storage is used in 5 cases.

US images obtained outside of the radiology department are recorded into the PACS system of the hospital in 18.3 % of cases only.


Several considerations are raised from the results of this survey.

First, there was a low response rate to the survey itself. There were only 123 answers to the 913 received messages asking for information from radiology department heads (a mere 13 %). It is hoped that this low response rate relates to the many committments on their side and not to low interest in the role of US within radiology [23].

Second, most responders indicated that US is still an important part of the activities of the radiology department. Only 9.3 % report fewer than 3,000 examinations/year. It must be noted that there may be a bias in these figures, since it is conceivable that responders were more interested in US than those who did not answer the questionnaire (even if there were responders who indicated that, in their hospital, US is done mostly outside of the radiology department). Most of the workload is due to abdomino-pelvic exams, followed by breast, musculoskeletal and vascular applications. Furthermore, state-of-the-art equipment is used in about 50 % and CEUS can be performed in 64.2 %. Portable machines are available in 64.5 %, transvaginal US examinations of the pelvis are used in 50.7 %, and radiologists are still involved in intraoperative US examinations in 64.2 % of cases. Most departments still have the technical capabilities to provide up-to-date US answers to the requests they receive.

Another consideration relates to the organisation of US within the hospital. In most cases US machines are scattered throughout the different departments, and only 13 hospitals have organised a centralised US laboratory where all physicians from different specialities come to examine their patients. Although centralisation seems the best way to run a US service, there are several factors that can explain why this is not the case, many of which stem from tradition. US laboratories, in fact, commonly arose separately from one another, following the initiatives of the different specialists who started introducing this technique in their practice. Then, there is a disposition to maintain independence and separate departmental income from the activities as well as the desire to control all aspects of patients’ care.

Only 15 departments reported they are working with sonographers. Although it is known that in Europe most radiologists perform US examinations directly, it is believed that this figure underestimates the real contribution of these professionals. A possible explanation is that only three hospitals from the UK answered the questionnaire; in the UK sonographers play a major role in dealing with the US workload.

Most answers to the questionnaire came from teaching hospitals (91.1 %). Comments on how training is organised state that US scanning is commonly regarded as an art, taught from maestro to pupil, with progressive growth in scanning and reporting capabilities. In addition, most report that US is taught within an organ-/system-oriented training system. The “art” of US is highly dependent on the operator’s dedication and technical ability, and this has to be properly taught. Additionally, a period of training within a dedicated US laboratory is probably needed to learn the basics of scanning techniques. After learning the technique, working in an organ-oriented manner is surely the best way to learn how to integrate diagnostic US within the clinical context and with all other imaging techniques.

There were 13 teaching hospitals in which fewer than 4 months is deemed sufficient, and in 20 cases having fewer than 500 examinations before the end of the residency is regarded as complete training.

The low number of US examinations performed in some training centres can jeopardise teaching. The recruitment of patients for adequate training can be impossibile to obtain in low-volume practices, leading to a further decrease of radiological US for future generations of radiologists. Furthermore, the use of sonographers can make teaching the practical skills of US scanning difficult. In a hospital with high-volume US practice (45,000 cases/year) in which the examinations are performed by sonographers, residents are asked to remain in the US laboratory only for 2–3 months and to perform only 100 examinations before the end of training. When in clinical practice in a hospital without sonographers, these radiologists would not be able to carry out even routine diagnostic US examinations. On the contrary, the role of expert sonographers as a resource to provide practical training to radiology residents has not been considered and can be explored.

The results of this survey show a large heterogeneity in the use of US within radiology throughout Europe. There are hospitals in which the majority of US examinations are still performed by radiologists, and others in which radiologists are left with only a small proportions of studies.

Similar findings were observed by Schnyder et al. in 1999 [1]. From their survey in 17 academic radiology departments throughout Europe, these authors reported that in some nations radiologists had full control of US, while this was not the case in Germany, Austria and Switzerland. The situation seems somewhat worse today, since there are 22 hospitals (18.2 %) in different nations (Austria, Poland, Germany, France, UK, Norway, Switzerland and Italy) in which radiologists perform less than 70 % of all US examinations and 5 (4 %) who answered they do less than 10 % of the studies. Since the answers to the questionnaire were provided by radiology departments, the figures for radiological activity can be considered as precise. On the contrary, it is possible that those answers on the US activities out of radiology can be regarded as an estimate. However, to the best of our knowledge, the data in the survey of Schnyder et al. were also obtained in a similar way, and a comparison can thus be made.

The percent decrease in the number of US examinations done in radiology vs. those performed outside radiology is probably related to a marked increase of the use of US by non-radiology clinicians rather than to a decreased attention to this technique by radiologists. In fact, new specialists, such as emergency physicians and anesthesiologists, are now using this technique as a complement to their visit or as a guide to therapeutic manoeuvres, and the so-called “point-of-care US” philosophy, in which US equipment accompanies the physician at the patient’s bedside to guide his/her therapeutic decision making, is gaining popularity.

An additional point to be considered relates to the recording of US reports and images into the hospital informations system and PACS. US examinations performed by radiologists are archived within the PACS system in 85.6 %, while those performed by non-radiologists are stored in only 18.3 % of cases. Furthermore, radiologists provide a formal report in virtually all cases, while examinations performed out of radiology are formally reported in 64.4 %. Costs and technical difficulties in connecting all equipment to PACS and RIS are described as reasons for not recording US images, and this is especially the case for recording of video clips. The use of “point-of-care US” is a further difficulty for connecting equipment to PACS, and, within this framework, the US exam is not regarded as a separate study but as part of the physician visit. However, to have all US images and reports of the patient recorded and available for consultation could greatly help during subsequent studies, and efforts have to be made to develop consensus with clinical colleagues to increase connectivity and to report all US studies, at least as a description within the patients’ charts. Within the framework of the relationships established by the ESR WG in US with the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB), it has been agreed to prepare and publish a recommendation about the necessity, for all US examinations, of a formal report and proper archiving of both report and images.


Two points of action can be suggested.

The first relates to the centralisation of the US laboratory. Although at the moment only a small number of hospitals are working according to this model, radiologists should take the lead in proposing such organisation [4]. This would allow the gathering of all the different competencies in US under the same roof, to share human and technological resources and to reduce the amount of equipment needed within the hospital. In an era of cost containments, a centralised US laboratory can allow each US scanner operate for longer hours and with higher numbers of examinations, resulting in an optimisation of resources. Furthermore, requests to upgrade and/or renovate equipment would possibly be easier if coming from a large laboratory and shared by different hospital departments. Another advantage would be having people with different backgrounds work in the same environment, thus promoting exchange and integration of their knowledge and possibly resulting in better patient care. It would be easier, in this respect, to prepare institutional guidelines and protocols that place US in the correct perspective towards all other imaging modalities and, most importantly, towards patients’ needs. It is not clear from the survey how this way of working is organised on a day-to-day basis, and especially how emergency services are provided (i.e. if all specialists concur in the emergency or if this is left to radiologists only), but an integrated management and organisational infrastructure bears numerous advantages for cost containment, quality standards and efficiency.

The second point of action relates to training in US within radiology residency programmes. In the opinion of the ESR Working Group on US, radiologists need to develop consensus on how many examinations under tutorship residents have to perform and on how much time they have to spend in ultrasound before the end of the training period. The results of the survey vary widely. However, out of 75 training centres that reported on the number of examinations, there were 39 (52 %) providing figures between 1,000 and 2,000 or higher. Therefore, approximately 2,000 seems to be a figure on which consensus can be reached. This figure also complies with what is suggested by the EFSUMB [5]. This federation provides recommendations about the number of examinations for training in the different subspeciality areas of US: the sum of studies for abdomen, breast, musculoskeletal and vascular training is 1,500, while figures for head and neck are not provided. The length of training is more complex to decide. A distinction has to be made here between the time needed to learn the technique of US scanning and the time needed to learn how to use US properly, to integrate it with other imaging techniques and to provide useful reports. In order to perform US, both approaches are needed. Technique-oriented teaching is necessary to learn how to perform the studies and to identify anatomy and pathology. Time and exams are needed to learn how to use US properly within the framework of organ-oriented training. A period of time of about 6 months dedicated solely to learning the US scanning technique can possibly be considered sufficient, as suggested by 76.2 % of responders. The capabilities of residents to perform US examinations have to be assessed during the training period, especially during and at the end of the technique-oriented part. It is known that the learning curve can vary widely among trainees, and longer times and higher numbers of examinations may be needed in some cases [6]. Additional time should be spent, and exams taken, during organ-oriented training. It must be underlined that organ-oriented teaching needs to include the proper role of US in each subspeciality and also take into account technical advances such as CEUS, 3D/4D and elastography and to use them when needed.


This article was kindly prepared by the ESR Working Group on US (M. Bachmann-Nielsen, M. Claudon, L. E. Derchi, S. Elliott, G. Mostbeck, C. Nicolau, S. Yarmenitis, A. Zubarev, Y. Menu–Chair of the ESR Professional Organisation Committee and J.A. Reekers–Chair of the ESR Subspecialty Societies Committee) on behalf of the European Society of Radiology. It was approved by the ESR Executive Council in April 2013.

Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.



Schnyder P, Capasso P, Meuwly I-Y (1999) Turf battles in radiology: how to avoid/how to fight/how to win. Eur Radiol 9:741–748PubMedCrossRef


Lockhart ME (2008) The role of radiology in the future of sonography. AJR 190:841–842PubMedCrossRef


Derchi LE, Claudon M (2009) Ultrasound: a strategic issue for radiology? Eur Radiol 19:1–6PubMedCrossRef


Krestin GP (2009) Maintaining identity in a changing environment: the professional and organizational future of radiology. Radiology 250:612–617PubMedCrossRef


Minimum training recommendations for the practice of medical ultrasond in Europe. http://www.org/guidelines/guidelines01.asp


Hertzberg BS, Kliewer MA, Bowie JD, Carroll BA, DeLong DH, Gray L, Nelson RC (2000) Physician training requirements in sonography: how many cases are needed for competence? AJR 174:1221–1227PubMedCrossRef

Read Full Post »