PC 1 - Will emerging technology replace the radiologist?
PC 1 - Will emerging technology replace the radiologist?Wednesday, March 1, 08:30 - 10:00 Room: E1 Session Type: Professional Challenges Session Topics: Physics in Medical Imaging, Hybrid imaging, Computer Applications Moderator: L. Donoso (Barcelona/ES) Add session to my schedule In your schedule (remove)
1. To become familiar with the emerging technologies in the imaging field.
2. To learn about the new concepts behind the computerised image analysis and diagnosis.
3. To understand the potential benefits and threats related to its implementation.
Call it artificial intelligence, deep learning, cognitive computing; whatever its name, it is the same thing, machines recognizing clinical problems in digital images ahead of the radiologists charged with making the diagnosis. Regardless of whether machine- or human-based aids are leveraged, radiology needs such aids. Never has improving performance been so important to its future. The liquid biopsy as a test is done on a sample of blood to look for cancer cells from a tumour that are circulating in the blood or for pieces of DNA from tumour cells that are in the blood. A liquid biopsy may be used to help find cancer at an early stage. It may also be used to help plan treatment or to find out how well treatment is working or if cancer has come back. The clinical impact of these developments together with the ones in molecular imaging, quantification and biomarkers will be discussed in this session.
1. To lean about the different tools related with "computer assisted" diagnosis.
2. To understand the challenges in management and radiologist practice of introducing these technologies.
3. To become familiar with "real-life" implementations.
As computers outperform humans at complex cognitive tasks, disruptive innovation will increasingly remap the familiarity with waves of creative destruction. In healthcare, nowhere is this more apparent or imminent than at the crossroads of radiology and the emerging field of clinical data science. As leaders in our field, we must shepherd the innovations of cognitive computing by defining its role within diagnostic imaging, while first and foremost ensuring the continued safety of our patients. If we are dismissive, defensive or self-motivated, industry, payers and provider entities will innovate around us achieving different forms of disruption, optimized to serve their own needs. To maintain our leadership position, as we enter the era of machine learning, it is essential that we serve our patients by directly managing the use of clinical data science towards the improvement of care. In this session, we will explore the state of clinical data science in medical imaging and its potential to improve the quality and relevance of radiology as well as the lives of our patients. Attendees will learn the basics of clinical data science, understand the potential impact of data science on the field of radiology, understand the transition of radiology from visualization to quantification in preparation for precision healthcare, and understand the value of deep learning in the era of MACRA and MIPS payment reform policies.
1. To understand the concept of liquid biopsy.
2. To learn about the advantages of liquid biopsy in the diagnostic process.
3. To understand the impact that these techniques will have on clinical practice.
Circulating tumour cells and circulating tumour DNA often referred as a ‘liquid biopsy’ are promising tools that have the potential to improve cancer diagnosis, prognosis assessment and real-time monitoring of treatment efficacy. In June 2016, the Food and Drug Administration (FDA) approved a test to screen for EGFR mutations in plasma samples to identify patients with metastatic non-small cell lung cancer that are eligible for treatment with erlotinib. In the future, more liquid biopsy tests are expected to complement other approaches used today for the prediction of treatment efficacy towards precision medicine.
1. To understand the role of hybrid imaging in the current clinical practice.
2. To become familiar with the new hybrid imaging applications in relationship to disease presentations.
3. To learn about quantification in hybrid imaging: its benefits and limitations.
Molecular imaging with hybrid imaging such as PET/CT is integrated in many clinical pathways. The selected tracer for the PET part will determine which biochemical or molecular information the examination with the PET part will return. On the other hand, the section of study protocol for the CT or MR on the case of PET/MR imaging will decide which information that part of the examination will return. The major role for PET/CT or PET/MR is staging of oncologist diseases but applications in cardiac and neuro-imaging are emerging. There are also non-oncologic applications for these modalities. In the field of oncology, we are aware of several hallmarks in cancer that are involved in disease development as well as in treatment strategies and treatment response. A major challenge is to develop imaging so we can visualize the behaviour of these hallmarks and during the talk the possibility of using hybrid imaging to do this will be discussed. The interest and attempts to quantifying biomarkers is higher than ever; however, there are many challenges in this field. Quantification and how it can be used will be briefly discussed.
1. To learn about the specific engineering challenges of developing new quantification methods.
2. To become familiar with the process of adapting the use of biomarkers in the clinical setting.
3. To understand the impact of deep learning on these diagnostic tools.
Quantitative imaging biomarkers are driving the paradigm shift in radiology towards precision medicine. Although the lack of standardization can hinder their appropriate use in clinical practice and drug development trials, alliances like QIBA and EIBALL allow for arriving to a consensus in image acquisition and image processing algorithms, which are the main current sources of uncertainty. The adoption of quantitative imaging solutions in the clinical setting requires, however, from the synthesis of the most relevant information, moving from redundancy to relevancy in the data evaluated by the radiologist and the clinical specialist, avoiding information overload. Therefore, data clustering and data reduction techniques, consisting of machine learning approaches have to be implemented. One of the most promising artificial intelligence techniques in the field of medical imaging is deep learning, which allows for the supervised (based on given features like imaging biomarkers) or non-supervised (learning features from data) by means of convolutional neural networks (CNN). Such networks have been applied for the automated classification of medical images as computer-aided detection (CAD) systems; however, the high number of data (millions of cases) required to train the CNNs and obtain efficacy is influencing the research evolution to new network configurations such as generative adversarial nets (GAN), which are expected to have a highly significant impact in the field of artificial intelligence and medical imaging.