Advancing Microwave-Based Imaging Techniques for Medical Applications in the Wake of the 5G Revolution - Emerald
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Advancing Microwave-Based Imaging Techniques for Medical Applications in the Wake of the 5G Revolution Jean-Charles Bolomey University Paris Sud, France jch.bolomey@gmail.com Special Session CS33 Horizon 2020 research and innovation session (EMERALD): ElectroMagnetic imaging for a novel genERation of medicAL Devices (http://www.msca-emerald.eu/)
OUTLINE • FEW QUESTIONS ON MICROWAVE-BASED IMAGING FOR MEDICAL APPLICATIONS • PART I: WHERE ARE WE ARRIVED ? • PART II: WHAT COULD BE TRIED/DONE ? • CONCLUSIONS Special Session CS33 Horizon 2020 research and innovation session (EMERALD): ElectroMagnetic imaging for a novel genERation of medicAL Devices J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 2
MICROWAVE-BASED MEDICAL IMAGING FACTS AND QUESTIONS … • SINCE EARLY 80’s… “MICROWAVES ARE CLAIMED TO OFFER PROMISES AS IMAGING MODALITY” L.E. Larsen and J.H. Jacobi (Diagnostic Imaging in Clinical Medicine, 11, 44-47,1982) • WHAT EXACTLY MICROWAVES ARE PROMISING OF ? AND BY WHEN ? • INDEED, ALMOST 40 YEARS LATER, MICROWAVES ARE STILL CLAIMING TO OFFER PROMISES… BUT ONLY GAINED A MODEST CLINICAL ACCEPTANCE • IN FACT, THE JOURNEY WAS MORE COMPLICATED THAN EXPECTED BUT THE RESULTS COLLECTED FROM RECENT CLINICAL TRIALS SEEMS SUGGESTING A POSSIBLE EXIT OF THE TUNNEL… J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 3
SIMPLIFIED CLASSIFICATION CHART OF MICROWAVE-BASED APPLICATIONS ACTIVE/PASSIVE Industrial SHORT RANGE LONG RANGE Scientific DIELECTOMETRY RADAR Medical I S M COMMUNICATIONS TRANSFER/HARVESTING MICROWAVE ISM APPLICATIONS AT THE CROSS-ROAD OF SIGNALING/SENSING/POWERING TECHNOLOGIES J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 4
VARIOUS ASPECTS OF MICROWAVES FOR MEDICAL APPLICATIONS « Low » « High » Power Medical Power [1] [2] DIAGNOSIS THERAPY BIO- HYPER- SENSING THERMIA IMAGING ABLATION [1] M.F. Iskander and C.H. Durney, “Electromagnetic Techniques [2] A.W. Guy, “History of Biological Effects and Medical Applications for Medical Diagnosis: a Review”, Proc. IEEE, 68, 226-132, Jan. 1980 Of Microwave Energy”, IEEE Trans. MTT-32, 226-132,Sept. 1984 MICROWAVE-BASED MEDICAL IMAGING IS THE “LAST COMER” ISM APPLICATION J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 5
TIME LINE OF MAJOR IMAGING MODALITIES MICROWAVES AS “LAST COMER” 1940 1950 1960 1970 1980 1990 2000 2010 DIGITAL BREAST MAMMO- CT X-rays RADIOGRAPHY GRAPHY SCANNER RADIOGRAPHY TOMO- SYNTHESIS [1] SONAR NDT US US SCANNER [2] MAGNETIC MRI MRI RESONANCE SCANNER [3] RADIO- NUCLEAR PET PET-CT PET NUCLEIDES MEDECINE SCANNER [4] 1ST DIAG. EARLY IMAGING MW APPL. INVESTIGATIONS [5] RADAR NDT ORGAN MVTS. BIOLOGICAL TARGETS PATHOLOGY ORIENTED LUNG DISEASE ISOLATED ANIMAL ORGANS (BREAST, BRAIN, LUNG, BONES…) … HUMAN IN-VIVO (FOREARM) HYPERTHERMIA CONTROL [1] R. Ciernak, “X-Ray Computed Tomography in Biomedical Engineering”, Springer-Verlag London Limited, 2011 [2] J. Woo, “A short history of the Real-time ultrasound scanner”, http://www.ob-ultrasound.net/history-realtime.html [3] T. Geva, “Magnetic Resonance Imaging: Historical Perspective”, Journ. Cardiovascular Magnetic Resonance, 8, 573-580, 2006 [4] G. Bernal, “History of PET scanners”, http://large.stanford.edu/courses/2014/ph241/bernal1/ [5] J.Ch. Bolomey, “Crossed Viewpoints on Microwave-based Imaging for Medical Diagnosis: From Genesis to Earliest Clinical outcomes ”, in The World of Applied Electromagnetics, A. Lakhtakia, C.M. Furse (eds), Springer 2018 J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 6
GLOBAL VIEW OF MEDICAL IMAGING MARKET WHAT PLACE FOR MICROWAVES ? • SALIENT MARKET FEATURES: MEDICAL IMAGING – Global Market size worth 29.48 billions and NUCLEAR Advanced Modalities EQUIPMENT MARKET projected to reach 46.18 by 2024, (Transparency Market Research, MRI – Market growth supported by the rise of geriatric 49% transparencymarketresearch.com) population and the increased prevalence of CV diseases, CT – Expected Growth Rate: 6.56%, over the forecast Basic period 2016 - 2024 US Modalities GLOBAL MEDICAL – X-ray first; fastest growth expected for CT/PET IMAGING MARKET (earlier disease detection), 51% (Pictures of the Future, Medical Imaging:Facts and Forecasts, – Larger growth expected for emerging markets. X-RAY siemens.com) – Global market shared by only a few big companies. • KEY GROWTH DRIVERS: A PRIORI – Miniaturization and portability, ✔ MICROWAVES Emerging Markets Developped – Digitization of measured data, ✔ LOOK ALMOST 36% Markets – Hybrid imaging systems, ✔ PERFECT 64% – Use of non-ionizing modalities, ✔ – Non-invasive and easy-to-use equipment, ✔ – Inexpensive, energy saving, ergonomic medical ✔ BUT, A equipment with low maintenance requirements … POSTERIORI… NOT SO SIMPLE ! J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 7
sensitivity for breast cancer detection, especially for cases where the use of mammography and ultrasound has been equivocal. However, MRI is almost always considered as an addition to, rather than a replacement for, digital mammography in Europe today. Constraints of cost and availability for MRI mean it is reserved only for high-risk patients; only in Austria and Italy MICROWAVES AS “CHALLENGER” TECH ? is MRI more commonly offered for regular screening in addition to digital mammography. Given MRI resource constraints, it is hard to see MRI realistically becoming a significant challenger to digital mammography anytime soon. Comparison of 'challenger' breast screening technologies 'Challenger' Measure DBT ABUS MRI tech From: Stephen Holloway, AuntMinnie.com contributing writer March 6, 2019 “DBT is well-positioned as successor in European breast screening” European Congress of Radiology, Vienna, Feb. 27 – March 3, 2019 J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 8
5. Plateau of Productivity Mainstream adoption starts to take off. Criter ia for assessing HYPE CYCLE FOR EMERGING TECHNOLOGIES legitim ate provider s are m or e clearly defined. The technology’s APPLICATION TO MICROWAVE-BASED MEDICAL IMAGING broad m arket applications and relevance are under stood. NUMERICAL & PHANTOM MODELS ? CLINICAL ADOPTION AND DISSEMINATION EXPECTATION CLINICAL ▪ Acceptance TRIALS threshold FUNCTIONAL PROTOTYPES ▪ prototype-to-product PHANTOMS, conversion, ISOLATED ▪ regulatory approval, ORGANS ▪ commercialization EXPERIMENTS ▪ instrument development, …all in the face ▪ preclinical testing, ▪ clinical prototyping of competing RF technologies ! DOSIMETRY TRIGGER Slope of TIME Peak of Through MATURITY of Innovation Inflated Disilliusonment Enlightenment Trigger Expectations Machines are becom ing better at under standing hum ans and the (Adapted from https://www.gartner.com/en/research/methodologies/gartner-hype-cycle) environm ent — for exam ple, recognizing the em otion in a person’s J.Ch. Bolomey voice — and hum ans are becom ing better at 13th EuCAP, Krakow, March 2019 9
DEVELOPMENT MICROWAVES FLOW-CHART FOR BIOMEDICAL APPLICATIONS FOR MICROWAVE-BASED IMAGING SIMPLIFIELD FLOW-CHART “INSIDE” MICROWAVE Encouraging results… COMMUNITY MICROWAVE EXPERTISE Good candidate… EARLY PROOF (HARD/SOFTWARE) OF CONCEPT IDENTIFIED NUMERICAL SIMULATION & CLINICAL ON-PHANTOM EXPERIMENT SUPPOSED SPECIFIC TARGET ADVANTAGES OF MICROWAVES DUE TO LIVING TISSUE DIELECTRIC TARGET PROTOTYPE PROTO PROPERTIES ADJUSTMENT REVISION TEST CASES “Technology” pull PRE-CLINICAL ASSESSMENT CLINICAL NEEDS ? NO « MAJOR » CLINICAL EXISTING PERF/COST TRIALS HEALTH MODALITIES BENEFIT ISSUES YES INTER-COMPARISON MULTI-MODALITY LARGE SCALE VALIDATION “Market” push ACCESS TO CLINICAL “Technology” transfer “OUTSIDE” MICROWAVE MARKET PRACTICE COMMUNITY J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 10
MICROWAVE SCANNERS FOR EARLY IN-VIVO IMAGING SOME EXAMPLES… Carolinas Heart Inst (USA). Keele Univ. (UK) INVERSE F=2.45GHz F=1GHz F=1-2.3GHz SCATTERING Nt=Nr=32 N=32 N=24 Univ. of Calgary (CA) (TOMOGRAPHY) F=50MHz-15GHz T N=1 RADAR MULTI- MU VIEW MO MONO- whole-body functional animal MO breast imaging VIEW T T animal imaging imaging T case study 1990 1995 2000 2005 2010 2015 human forearm NIT on animal breast imaging on volunteers and patients F=0.3-3GHz F=4-8GHz F=2-4GHz N=16 N=60 N=16 F=2.3GHz T T T MU MU N=64 UPC Barcelona (SP) Thayer School of Engng. Dartmouth (USA) Univ. of Bristol (UK) McGill Univ (CA). J.Ch. Bolomey COST MiMed, 20-22 June, 2016 11
MICROWAVE IMAGING SYSTEMS ENGAGED IN (PRE)CLINICAL TRIALS MANY DATA ACQUISITION AND PROCESS ACQUISITION TIME MODALITIES NUMBER OF 25000 CHANNELS MARIA MICRIMARIA (2006) 1.5-8 GHz MARIA 11 MIST (1995) MIST Multi-static, 60 slot antennas 0.5-3 1/2 spherical RADAR DARTMOUTH Coll. GHz Multi-static,16 monopoles SUST Cylindrical, Variable height, 4-8.5 GHz SUST-SHENZEN TOMOGRAPHY Multi-static, , 2 antennas 5 rings, 360 positions/ring, mech./scan, Cylindrical RADAR Mc Gill Univ. Multi-static, 16 antennas Wearable, RADAR TSAR UoCALGARY MEDFIELDS DIAG. (2005) MGU Monostatic, 2 vivaldi antennas 2-4 TSAR Cylindrical/Conformal, Head adjustable, 8 slot antennas 1.5-9 GHz mech./scan RADAR CLASSIFICATION GHz EM-Tensor (2012) Cylindrical 5x32=160 slot antennas TOMOGRAPHY FREQUENCY (GHz) Adapted from: “Microwave Breast Imaging: Clinical Advances and Remaining Challenges”, D. O’Loughlin, M. O’Halloran, B.M. Moloney et al.; IEEE Transactions on Biomedical Engineering ( Volume: 65 , Issue: 11 , Nov. 2018 ) J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 12
WHAT HAS ALREADY EMERGED ? OUTCOMES A FEW OPERATIONAL SYSTEMS ENGAGED IN CLINICAL TRIALS : * BREAST, BRAIN, BONES, LUNG * ENCOURAGING” SENSITIVITY * SPECIFICITY TO BE VALIDATED - SIGNIFICANT ACHIEVEMENTS * EXTENSIVE NUMERICAL AND EXPERIMENTAL MODELING, * RECONSTRUCTION ALGORITHMS (TOMOGRAPHY, RADAR, HOLOGRAPHY), * TISSUE CHARACTERIZATION, * PROTOTYPE DEVELOPMENT, * PHANTOM DEVELOPMENT/EXPERIMENT, *ETC… - BUT… SOME PERSISTENTLY “OPEN” ISSUES: * LACK OF SYNTHESIS AND BENCHMARKING * A PRIORI INFORMATION * DIELECTRIC CHARACTERIZATION * PATIENT INTERFACE OPTIMIZATION EFFORTS * IMAGE RELEVANCE (SPECIFICITY) WHAT COULD BE TRIED / DONE ? LOOK AT EMERAD PROGRAM… AND THREE SUGGESTIONS J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 13
SUGGESTION 1 “BETTER” DATA ACQUISITION TECHNIQUES • EXPLOITING ALREADY EXISTING OR SUPPOSEDLY RAPIDLY AVAILABLE MICROWAVE TECHNOLOGIES FOR OBTAINING “BETTER” DATA ACQUISITION TECHNIQUES: – INTRODUCING THE BEST OF WIRELESS TECHNOLOGIES USED FOR EXISTING COMMUNICATION, RADAR, SENSOR GRID SYSTEMS (MODULATION/DEMODULATION SCHEMES, OFDN, MIMO PROCESSING,…) – APPLICATION-DEDICATED AND/OR PATIENT ADAPTIVE OPTIMIZATION OF THE PROBE ARRAY – INCREASING DYNAMIC RANGE TO REDUCE MULTI-PATH CORRUPTION – TAKING PROFIT OF 5G, IoT OR IoE DEVELOPMENTS (MINIATURIZATION, ARRAYS OF INTEGRATED PROBES, …) – REDUCING FABRICATION/EXPLOITATION/MAINTENANCE COSTS BY MEANS OF APPROPRIATE ARCHITECTURES J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 14
MICROWAVES MM-WAVES SPECTRUM MEDICAL IMAGING VS 5G FREQUENCY BANDS VISIBLE (Source: Qualcomm) FREQUENCY (GHz) 0.6 GHz 6 GHz 24 GHz 100 GHz BRAIN TOMOGRAPHY MEDICAL BREAST UWB RADAR IMAGING J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 15
SUGGESTION 2 “BETTER” DATA PROCESSING AND ALGORITHMS • EXPLOITING/ANTICIPATING GROWING COMPUTING POWER • RADAR: TOWARD REAL-TIME, FAST PRE/POST PROCESSING • INVERSE-SCATTERING IMAGING: DECREASING MODEL NOISE, FULL 3D ACCURATE MODELING (INCLUDING INTERACTIONS AND MUTUAL COUPLING…) • INCREASING AMOUNT OF DATA • IMPROVING SPATIAL RESOLUTION • QUANTUM COMPUTERS ? BEYOND MOORE’s LAW ? – USING OTHER IMAGING MODALITIES • A PRIORI INFORMATION (WHEN NEEDED) • MICROWAVE DATA FUSION WITH OTHER MODALITIES FOR AI/ML/DL ALGORITHMS – CONSIDERING “NON IMAGING-BASED” MICROWAVE SENSING PROTOCOLS (SENSOR NETWORKS GRIDS, ARTIFICIAL/AUGMENTED INTELLIGENCE, MACHINE LEARNING, CLASSIFICATION, NEURAL NETWORKS, ETC.) J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 16
MOORE’s LAW EXTRAPOLATIONS OF COMPUTATIONAL POWER J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 17
SUGGESTION 3 INCREASING INTERACTIONS WITH BIOLOGISTS AND “END-USER” MEDICAL COMMUNITY • RETURNING TO BASIC MW/BIO INVESTIGATIONS ▪ TISSUE DIELECTRIC CHARACTERIZATION (LOCAL PROBE, MRI, SCANNING MICROSCOPY,…) ▪ PHYSICAL OR HEURISTIC MODELS FOR FREQUENCY DEPENDENCE ▪ FOCUSING ON DIELECTRIC SPECIFICITY AND SENSITIVITY ASSESSMENTS ▪ QUANTIFYING DIELECTRIC IMAGES IN TERMS OF BIO-PHYSIOLOGICAL FACTORS • INCREASING INTERACTIONS WITH “END USER” MEDICAL COMMUNITY ▪ IDENTIFYING REAL NEEDS AND LOOKING FOR THE MOST MICROWAVE-FRIENDLY AND CLINICALLY RELEVANT SCENARIOS ▪ CASE BY CASE VALIDATION THANKS TO LARGE SCALE /MULTIMODALITY CLINICAL ASSESSMENT CAMPAIGNS ▪ GUIDING TECHNICAL ADVANCES BY CLINICAL RETURNS ▪ HYBRIDIZING WITH OTHER MODALITIES ▪ FOLLOWING UP IMPROVEMENT OF OTHER IMAGING MODALITIES: MAN-PORTABLE MRI, DIGITAL TOMOSYNTHESIS, LOW-DOSE PHASE CONTRAST X-RAYS, ETC. ▪ NOT UNDERESTIMATING THE CHANGES IN THE MEDICAL PRACTICE INDUCED BY AI / ML DEVELOPMENTS ▪ ATTENDING MEDICAL MEETINGS, LOOKING AT MEDICAL JOURNALS, ETC. J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 18
FUTURE MEDICAL IMAGING FACING REVOLUTIONARY AI AND ML CHALLENGES (EXAMPLES) ▪ “Look Ahead: The Future of Medical Imaging”∑ James H. Thrall, radiology imaging, August 1, 2015 ▪ “How Artificial Intelligence Will Change Medical Imaging” D. Fornel, artificial intelligence, February 2017 ▪ “What is the Future of Medical Imaging Equipment ?” M. Taschetta-millane, radiology imaging, July 2018 ▪ “Technologies to Watch in Breast Imaging” J. Zagoudis, radioimaging, July 2018 ▪ “Artificial Intelligence in Radiology: Hype or Hope ?” M.B. Massat, pp. xx-yy, March, 2018 ▪ “A Promising future for AI in breast cancer screening” M. B. Massat, pp. 22-25, September 2018 ∑ ▪ “Demystifying Artificial Intelligence” Part I, “Simplifying AI and Machine Learning”, E. Siegel, pp. 26-28, May, 2018 Part II, “An Imaging Tool Ready to Explode”, L.N. Tanenbaum,pp. Search26-27, the site... July 2018 Part III, “Going Beyond Escape”, R.B. Shrestha, pp. 8-11, November 2018 ▪ “Machine (Deep) Learning Methods for Image Processing IEEE Trans. Radiation and Plasma Medical Sciences and Radiomics”, M. H. Latim, C. Parmar, J. QiI, I. el Naqa Vol. 3,pp.104-107, March 2019 ∑ SPECIAL ISSUE ON MACHINE LEARNING FOR IMAGE PROCESSING AND RADIOMICS J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 19
CS33 Horizon 2020 research and innovation session (EMERALD): ElectroMagnetic imaging for a novel genERation of medicAL Devices CONCLUSIONS • AFTER A LONG MATURATION PERIOD, MICROWAVE-BASED IMAGING SYSTEMS ARE BEGINNING TO PROVIDE CLINICAL RESULTS • FUTURE DEVELOPMENTS MUST: – TAKING PROFIT OF 5G TECHNOLOGY AND DEEP LEARNING ALGORITHMS – ACCOUNTING FOR THE EVOLUTION OF THE MEDICAL PRACTICE, MORE PARTICULARLY IMAGING MODALITIESUNDER THE IMPACT OF ARTIFICIAL INTELLIGENCE • DUE TO THEIR NECESSARY SPECIFICITY, NEW SYSTEMS MUST BE FOCUSED ON REAL CLINICAL NEEDS FOR: – COMPLETING/SUPPLEMENTING OTHER IMAGING MODALITIES – ADDRESSING SPECIFIC REQUIREMENTS FOR DEVELOPPING COUNTRIES • A J.Ch. Bolomey 13th EuCAP, Krakow, March 2019 20
You can also read