Hybrid Imaging research and clinical practice David W. Townsend, PhD A*STAR-NUS
Medical Imaging Modalities using EM radiation
SPECT
Imaging ranks near bottom of Medicare spending growth drivers SPECT/CT PET
X-radiographs Computed Reversing its role as a major contributor to the growth in Medicare Magnetic Resonance (MR) Tomography (CT) spending in the early 2000s, medical imaging spending growth Radiofrequency Infrared placed in the percent of spending growth categories Gamma ray bottom 2 Ultraviolet (RF) in 2011, according to a study published in the December issue of the Microwave American Journal X-ray of Roentgenology . Non-Ionizing Ionizing
Electromagnetic Spectrum
PET/CT
Hybrid imaging modalities
PET/MR
2
Where in the world is…..the lesion?
P. Brueghel
X-ray CT: anatomy
PET: function
CT + PET: anatomy + function
The best of both worlds!
3
First PET/CT device: 1995 - 1998 ECAT ART PET
+
Medical Invention of the Year TIME , December 2000
Somatom AR.SP CT
2010
IEEE Innovations in Healthcare Technology
PET/CT
CT console
PET console
PET/CT imaging, 1998-2001 University of Pittsburgh 4
5
2001: PET/CT in the clinic
+ 6
1984 - 2013
The evolution of PET/CT
Gunma 1984
UPMC
2000
1998 1999
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
PET/CT PET
2 0 0 2
2 0 0 3
2 0 0 4
2 0 0 5
2 0 0 6
2 0 0 7
2 0 0 8
2 0 0 9
2 0 1 0
2 0 1 1
2 0 1 2
NEMA - US Shipments ($M)
7
Vulnerable plaque imaging with 18F-sodium fluoride (NaF) November 2013
Intense 18F-NaF uptake at site of the culprit plaque shown on PET/CT scan (red arrow)
Angiogram of culprit (red, LADA) and non-culprit (white, circumflex) plaque
The PET/CT using 18F-NaF shows uptake in culprit plaque but not non-culprit
The PET/CT using 18FDG shows no uptake in either of the two plaques
Corresponding 18FDG scan of same region shows no uptake in the plaque (red arrow) 8
Imaging ischemia with PET/CT
Courtesy: Gustav von Schulthess, University Hospital, Zurich
9
Therapy response: PERCIST
3.5 3
SUV
2.5
PERCIST: PET Response Criteria in Solid Tumours
2 1.5 1
Responder
Soft 0.5 tissue sarcoma 0 1
2
3
4
5
6
7
SUV
Week 8 7 6 5 4 3 2 1 0
Non-responder 1
2
3
4
Breast cancer
5
6
7
Baseline
Lymphoma + 2 months
A PET/CT scan acquired each week during chemo maps a decreasing SUV in responders compared to patients who do not respond that demonstrate an unchanging or increasing SUV. For these patients with stage IIIb lung cancer, the PET/CT determined that if the patient did not respond to the chemotherapy by week 3 they would never respond to that treatment
Probability of survival
Week
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
Metabolic Responder Metabolic Responders
Stage IIIb Lung Cancer Metabolic Non-responders Metabolic Non-Responder
0
10
20
30 40 50 60 Survival (weeks)
70
80
90
10
High temporal resolution dynamic PET scanning 60 s sinogram data
1 s sinogram data
59 s sinogram data
Josh Schaefferkoetter and Inki Hong
Complementary frame reconstruction (CFR)
60 s image
59 s image
1 s difference image
1 s image
18F-FLT
Arterial
Venous 1 sec per frame
0.1 sec per frame
Dynamic frames (0.1 s)
11
Has PET/CT made a real difference? PET/CT vs PET and CT: average over all cancers: 10-15% accuracy improvement
• Head and neck Accuracy: 95% vs 83% PET; 73% CT
• Thyroid Accuracy: 93% vs 78% CT
• Esophageal cancer Accuracy: 92% vs 86% PET
• Colorectal cancer Accuracy: 89% vs 78% PET
• Lymphoma PubMed on PET/CT: 6500+ publications
• Solitary lung nodules
Accuracy: 96% vs 81% CT
• Lung cancer Accuracy: 98% vs 80% PET (T stage)
• Breast cancer Accuracy: 90% vs 79%
Accuracy: 93% vs 78% CT
• Melanoma Accuracy: 97% vs 93% PET
• Unknown primary No difference; 20-40% detected
Czernin, Allen-Auerbach, Schelbert. J Nucl Med 48 (1, Supplement) 2007: 78S – 88S
12
D. Karantis et al, JNM 2012
19,053 authors of papers in major oncology journals 5.2% response rate
66% felt consultation with imaging expert before a scan to be useful
96% felt interaction between referring and interpreting physicians good for patients 39% reported frequent necessity for discussion with imaging expert 61% always read entire PET/CT report
30% reported ambiguity, poor explanation or lack of familiarity with terminology 85% expressed desire for access to PET/CT images 47% raised concerns about high cost of PET/CT scans 41% has concerns about over-interpretation of imaging findings
13
Improving communication
• Implement decision support to increase the quality of the requests • Increase transparency with regard to waiting times • Make reports clearer by including images • Improve the communication of critical results • Implement a portal for easier scheduling • Make radiation dose information visible in both the requests and the report
14
2004: SPECT/CT in the clinic
+ Dale Bailey
“CT is potentially more valuable for SPECT than for PET” Bailey DL. Eur J Nuc Med & Mol Imag 2003; 30(7):1045-1046
15
SPECT/CT scanners
Discovery NM/CT 670
Anyscan SPECT
Hawkeye 4
CORONAL
TRANSVERSE
SAGITTAL
SPECT/CT
Symbia TX
54 y/o patient with adrenal cancer recurrence. 131I-MIBG SPECT shows uptake is in spine and not in adrenal bed 16
Bone imaging
Courtesy: Dale Bailey PhD, Royal North Shore, Sydney
A 65 year-old male received a total knee replacement in 2005. After initial success the knee became progressively painful over several years. A bone scan was performed in March 2010 after injection of 800 MBq of 99mTc-MDP. Both blood pool and SPECT/CT images were acquired. The blood pool images showed increased vascularity on the medial side of the right knee. The SPECT/CT images showed increased uptake in the postero-medial tibial screw in the right knee with no other abnormal uptake. Possible causes include infected screw, screw loosening, or screw osteolysis. In May 2010, the patient underwent replacement of the tibial plate and screw and now experiences 17 significantly reduced pain.
2010: ..and so to PET/MR
18
Current designs for MR/PET (/CT) Vendor
Lateral-docking air-hover tabletop
Discovery PET/CT 710
Discovery MR750w
Installations
Siemens
39
Philips
12
GE
6
TOTAL:
57
3m PET
MRI TOF-PET
Achieva-X (3T MR)
mMR
19
Recurrence of prostate cancer:
PET/CT
18F-choline
MR/PET
MR/PET
20
Malignant and non-malignant diseases Tuberculosis
In India, 1 death every 90 s from MDR TB FDG-PET/MR in patients that are receiving treatment for tuberculosis
PET/MR Ovarian cancer
Therapy response
PET/CT
WB radiation dose for PET/MR: ~ 3mSv
21
PET/MR in metastatic colorectal cancer
PET/CT A) PET B) CT C) PET/CT
PET/MR D) PET E) MR F) PET/MR 66-y/o man; follow-up for colorectal Ca. E) T2-weighted HASTE G) arterial phase CE T1-weighted VIBE H) Corresponding H&E stain
22
From PET/CT to PET/MR? PET/MR will not replace PET/CT in the medium term MR is a complex imaging modality compared to CT
PET/MR protocols need to be defined - as for PET/CT PET/MR instrumentation must be more cost-effective
Reduced throughput compared to PET/CT scanning MR-based attenuation correction still unsolved Patient comfort and acceptance of PET/MR may be issue 23
Summary: PET/CT: A technical evolution and an imaging revolution 2002 – 2004: Units shipped in USA: 360
PET/MR: A technical revolution and an imaging evolution
Johannes Czernin UCLA
2010 – 2012: Units shipped in USA: 14
Is there evidence of real clinical benefit? PET/CT over PET and CT separately
1+1=3
SPECT/CT over SPECT and CT separately 1 + 1 = 4 PET/MR over PET(/CT) and MR - TBD
1+1=?
24
John Roderick Cameron 1922 - 2005
John Cameron (left) c.1958
His Mission - Making a Difference to Society
“I am now almost certain that we need more radiation for better health” John Cameron 25
Most radioactive place in the world: Ramsar, Iran Background radiation: 100-260 mSv / year due to 226Radium No epidemiological evidence of adverse affects Residents demonstrate a marked increase in DNA repair capacity
Proposal: to relocate the inhabitants (~2000) to a lower radioactive area!
26
Radiation dose from radon versus cancer mortality
Radiation levels from radon
Cancer mortality rates
Background Radiation: Differences on Annual Cancer Mortality Rates/100,000 for each U.S. State over a 17-Year Period
dose > 2.7 mSv/yr
States with significantly higher doses, greater than 2.7 mSv/year like Colorado, have lower cancer rates than States with much lower average doses like Georgia, and vice versa. (Frigerio and Stowe, 1976 )
27
Radiation dose ranges • • •
Dose quantities 1 Gy = 1 J/kg = 100 rad 1 Sv = 1 J/kg equivalent = 100 rem
Background radiation (2 – 10 mSv / year worldwide) Mammogram (0.5 – 1.0 mSv) Radiation poisoning and death: 1000 x background Annual Background Radiation
Annual Limit for Radiation Workers
Energy deposited by radiation per unit mass of substance
Lower limit for known effects of radiation
Equivalent dose (Sv) Absorbed dose weighted for the effect of different radiation
No reliable data on effects of radiation
Very low dose range
Absorbed dose (Gy)
Low dose range
Effective dose (Sv) Equivalent dose weighted for susceptibility of different tissues
Known carcinogenic effects
Effective Radiation Dose (mSv)
28
What we really know about the risks of ionising radiation Lifetime Attributable Risk (LAR) is the additional cancer risk above baseline and is based on the sex of patient, magnitude of the single dose and age at time of exposure to radiation. • 100,000 men aged 30 • single dose of 100 mSv • 686 will develop cancer over their lifetime from the radiation exposure No of cases per 100,000 men exposed to a single dose of 100 mSv
29
The Press
30
…and so on…
Could this be over reaction of the Press? Drowning
Industrial Accidents
Road Accidents
Radiation
31
“My main frustration is the fear of cancer from low dose radiation, even by radiologists”
“Too many radiologists still believe there is a risk from a chest x-ray. Few radiologists can explain radiation to the patient in words the patient can understand” John Cameron 32
…and on…
DESPITE great strides in prevention and treatment, cancer 1 mSv rates remain stubbornly highRadiation and may soondose surpass=heart disease as the leading cause of death in the United States. Increasingly, we and many other experts believe that an important culprit may be our own medical practices: We are silently irradiating ourselves to death.
2012: 14 million new cancer cases with 8.2 million deaths 1 in 5 men and 1 in 6 women before the age of 75 Focus areas: stop smoking, reduce obesity, vaccination against HPV (cervical and liver) 33
The BEIR Report: theoretical models
?
The models: 1. Linear No Threshold 2. Threshold 3. Hormesis
2. Threshold 1. LNT
Biological Effects of Ionizing Radiation (BEIR) VII Phase 2 (2007)
Japanese Survivor Data
3. Hormesis
Potential sources of data: A. B. C. D.
Environmental Radiation Studies Occupational Radiation Studies Medical Radiation Studies Atomic Bomb Survivor Studies
No scientific data to distinguish between any of the models 34
Risk models used in LNT Excess Relative Risk (ERR) versus Excess Absolute Risk (EAR)
ERR = (Rex – Run) / Run
EAR = (Rex – Run)
But which model is correct ? Tough decision?
Decision by BEIR VII Committee
For each organ: Risk model = x ∙ ERR + (1-x) ∙ EAR where x is determined by the BEIR committee!
35
The challenge:
Michael O’Connor, Mayo Clinic
We need to reduce radiation dose from imaging procedures: • Not necessarily because it causes cancer • But because people fear it will cause cancer
• And where does this fear come from? - inappropriate use of the BEIR risk models
Consequences: Negative impact on patient care: • Patients declining needed exams or procedures • Physicians ordering alternate exams, which may be less accurate, more expensive, or require anesthesia 36
“In 1970 I realized that there was negligible risk from x-rays but many radiographs had poor image quality so that the risk from a false negative was significant.” John Cameron 37
Image Wisely
38
Radiation doses for clinical imaging procedures Annual Background Radiation
No reliable data on effects of radiation
CT Abdomen / Pelvis PET/CT (FDG) Myocardial Perfusion Scan
Lower limit for known effects of radiation
Chest CT Virtual Colonscopy CT Screening Lung Cancer PET/MR (FDG) MBI (8 mCi)
Radiation poisoning and death: 1000 x background
Mammogram + tomosynthesis Mammogram Chest X-ray
Bone Densitometry Very Low Dose
Low Dose 100
Radiation Dose (mSv)
39
Ali Shabestani Monfared Seyed Mohamad Javad Mortazavi
“It is likely that we need more radiation to improve our longevity” John Cameron 40
Where should we focus our imaging resources? Australia’s Productivity Loss Population Growth
Diabetes
Mental Illness
Ageing Population
Cardiovascular Disease
Obesity
Cancer
Joint Disorders
Predicted Growth in Healthcare Expenditure: 2003 - 2033 Source Goss, J., 2008
41
Thank you for your attention
42 42