1 Annual Lower limit for Background known effects Radiation of - - PDF document

1

Radiation Risks in low-Dose Imaging

Michael K O’Connor, Ph.D.

• Dept. of Radiology, Mayo Clinic

Radiation Units

Grey (Gy) = amount of radiation absorbed in any material Sievert (Sv) = estimates biological effect from the absorbed radiation Various weighting factors used to convert absorbed dose (in Grey) and dose to one organ (in Sievert) to “effective dose” to the whole body (in Sievert) e.g. Mammogram delivers absorbed dose of ~4 mGy to breast tissue. This gives an effective dose to the whole body

• f ~0.5 mSv

Converting radiation dose (mGy) to an estimate of its biological effect (mSv)

Head CT = 30 mGy to brain Digital mammogram = 4 mGy to breast tissue

+

Digital mammogram = 0.5 mSv Head CT = 2 mSv

Estimated biological effect of a head CT is 4 times that of a digital mammogram

Examples of Effective Doses in the µSv range

Eating 1 banana (0.1 µSv) Dental x-ray (5 µSv) 2 hours in Denver (1.2 µSv) X-ray of the wrist (1 µSv) 1 day in Minnesota (10 µSv) Airport x-ray screen (0.25 µSv) Bag of chips (0.3 µSv) Flight from New York to LA (40 µSv)

Examples of Effective Doses in the mSv range

Living in a brick or stone building (70 µSv) Cumulative 2-week dose at Fukushima Town Hall after accident (100 µSv) Annual dose from potassium in your body (400 µSv) Mammogram ~ 500 uSv Maximum external dose from 3-mile Island accident (1 mSv)

• r

Annual radiation level inside US Capital Building (1 mSv) Dose from a Head CT scan (2 mSv) Average yearly background dose In Minnesota (~3.5 mSv)

Examples of Effective Doses in the mSv range

Chest CT Scan (7 mSv)

• Max. permitted dose for US Radiation workers (50 mSv)

2

Examples of Effective Doses in the Sv range

Lowest 1-year dose clearly linked to increased cancer risk (100 mSv) Severe radiation poisoning, in some cases fatal (2000 mSv) Fatal dose, even with treatment (8000 mSv)

Radiation Dose (mSv)

Annual Background Radiation Lower limit for known effects

• f radiation

Low dose range

No reliable data on detrimental effects of radiation

Very low dose range

Known carcinogenic effects

?

9

Radiation is a poor carcinogen !

Risk Estimates require “Impracticably Large” sample requirements

• Does the radiation from mammography

(about 1 mSv) cause breast cancer?

 Cohort study: about 100 million (20-year

follow-up)!

 Case-control: about 1 million cases (4:1 ratio)

10 National Research Council (1995) Radiation Dose Reconstruction for Epidemiologic Uses (Natl. Acad. Press, Washington, DC).

Sample size required to detect a significant increase in cancer mortality, assuming lifetime follow-up

Dose (mSv))

Required number of subjects in study Dose range of relevance for radiology)

Arch Intern Med. 2009;169(22):2078-2086 Using LNT & BEIR VII report, estimated radiation-related incident cancers Estimated that 29,000 future cancers could be related to CT scans performed in the U.S. in 2007…..and would translate into about 14,500 cancer deaths.

Why are we concerned with Radiation Risks?

3

Consequences: Heightened fear of radiation = Negative impact on care

• Patients declining needed exams or procedures
• Physicians ordering alternate exams, which may

be less accurate or more expensive

Radiophobia Fukushima

The number of deaths indirectly related to the earthquake in Fukushima Prefecture was >1700 . Deaths were due to the physical / mental stresses related to the evacuation.

Where does the estimate of 29,000 cancers come from ?

Based on BEIR VII

+

risk estimates for 56,900,000 patients

What is the BEIR VII Report

An estimate of cancer risk from low doses of ionizing radiation!

• Input data:

 Environmental studies  Occupational studies  Medical studies  Atomic bomb studies

• Risk models (based on above input data):

 ERR (excess relative risk)  EAR (excess absolute risk)  LAR (lifetime attributable risk)

• Subjective opinion of committee !
• Linear No Threshold hypothesis
• Latency period / DDREF

Sources of data

• Environmental Radiation Studies
• Occupational Radiation Studies
• Medical Radiation Studies
• Atomic bomb survivor Studies

4

Sources of data used in BEIR VII Environmental Radiation Studies

Populations living near nuclear facilities “..no increased risk…with radiation exposure” Populations exposed to atomic bomb testing “..some studies (4 out of 10) show some effect” Chernobyl High incidence of thyroid cancer “..no evidence of an increase in any solid cancer type to date” Natural background (China / India) “..did not find higher disease rates in geographical areas with high background levels..” Cancer Mortality in High Background Radiation Area of Yangjiang, China, 1979-1995

• Estimated cancer risk

associated with the low level radiation exposure of 6.4 mSv / year

• 20-year study in 125,079

subjects

• Excess Relative Risk

ERR/Sv = -0.10 (-0.67 to 0.69)

• Conclusion: the mortality of

all cancers in Yangjiang was generally lower than that in control group, but not significant statistically.

(Tao et al, Zhonghua Yi Xue Za Zhi, 1999; 79: 487-492)

Radon Levels Lung Cancer

Generated from EPA web site (https://www.epa.gov/radon/find- information-about-local-radon-zones- and-state-contact- information#radonmap) Generated from NCI mortality map (http://ratecalc.cancer.gov/ratecalc/)

Sources of data

• Environmental Radiation Studies
• Occupational Radiation Studies
• Medical Radiation Studies
• Atomic bomb survivor Studies

Occupational Radiation Studies on Workers in the Nuclear Power Industry

“….in most cases, rates for all causes and all cancer mortality in the workers were substantially lower than the reference populations.” Findings explained as “healthy worker effect”

(U.S. Academy of Science, BEIR VII, 2007)

Significant limitation of most occupational studies is absence of an appropriate control group !

Sources of data

• Environmental Radiation Studies
• Occupational Radiation Studies
• Medical Radiation Studies
• Atomic bomb survivor Studies

5

Sources of data used in BEIR VII Focus on therapeutic studies “…most of the information comes from studies

• f populations with medium to high doses”

Lung Cancer – 9 studies, 40,000 subjects average dose ~ 1 Gy Breast cancer – 11 studies, 20,000 subjects average dose ~ 300 mGy

Medical Radiation Studies

Dose (mSv)

Standardized Death Rate / 106 py

Mortality from Breast Cancer after Fluoroscopy in Patients being treated for Tubercolosis

31,710 women treated between 1930 - 1952 40-year follow-up Age range 10-40 years

Miller AB et al, NEJM 1989; 321: 1285-1289.

“Risk was statistically significant for all those who received more than 100 mSv of radiation”

Sources of data used in BEIR VII

• Environmental Radiation Studies
• Occupational Radiation Studies
• Medical Radiation Studies
• Atomic bomb survivor Studies

Atomic bomb survivor Studies

• 120,000 survivors

93,000 present at time of bombings 27,000 from locale, but absent at time

• f the bombing (Not In City group)
• Monitored over 70 years & includes both sexes and

all ages of exposure – mean dose = 200 mSv

• Dose range 37,000

0-5 mSv 32,000 5-100 mSv 17,000 100 mSv – 2000 mSv This is the primary source of data for LNT risk models

Atomic bomb survivor Studies

Preston et al, Rad Res 2007;168: 1-64. (Radiation Effects Research Foundation)

Data from Table 4, Preston et al, 2007 # solid cancers adjusted to per 100,000 people

Atomic bomb survivor Studies

Preston et al, Rad Res 2007;168: 1-64. (Radiation Effects Research Foundation)

Data from Table 4, Preston et al, 2007 # solid cancers adjusted to per 100,000 people

Radiology

6

Radiology

Atomic bomb survivor Studies – 70 yrs follow-up

Ozasa et al, 2013, Doss et al 2012

Risk Models used in BEIR VII

Risk models are generated from high-dose studies ! Excess Relative Risk (ERR) model

The ERR is the rate of disease in an exposed population divided by the rate of disease in an unexposed population, minus 1.0.

Excess Absolute Risk (EAR) model

The EAR is the rate of disease in an exposed population minus the rate of disease in an unexposed population.

Same Data – 2 different Risk Models

Comparison of Lifetime Risk of Cancer using ERR and EAR

100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800 LAR based on EAR Model LAR based on ERR Model

Males Females

Breast Prostate Stomach

Risk Models used in BEIR VII

Excess Relative Risk (ERR) vs. Excess Absolute Risk (EAR) Which model is correct ? For each organ, final risk model = x.ERR + (1-x).EAR where x is determined by committee !

Effect of Low Doses - 3 Hypotheses

Linear No Threshold (LNT) Model Threshold Model Hormesis Model LNT Hormesis

Linear No-Threshold Hypothesis

LNT assumes that

• any amount of radiation exposure, no matter how

small, can increase the chance of cancer.

• probability of cancer from radiation exposure

increases with cumulative lifetime dose. LNT & Radiation

• 1930’s: developed by Herman Mueller to explain

mutagenesis in fruit flies

• 1950: Mueller persuaded BEIR committee in 1950 to

use his LNT hypothesis to explain carcinogenesis

7

37

For example: Using the LNT model the following are equivalent in terms of their effect 1 person jumps off a 100-foot cliff 100 people jump off a 1-foot cliff 1 person jumps off a 1-foot cliff 100 times

LNT Hypothesis

Cancer incidence from ionizing radiation

 Almost exclusively based on high-dose

data in atomic bomb survivor studies

 Uses risk models (developed from

Japanese population, wartime conditions) then applied to cancer rates for U.S. population

 Uses LNT to extrapolate to low dose

scenarios

The following organizations have clearly stated that the use

• f the LNT Hypothesis to compute the effects of small doses
• n large populations is inappropriate
• International Commission on Radiological Protection
• American Association of Physicists in Medicine
• Health Physics Society
• Academie Nationale de Medecine, France
• National Council on Radiation Protection & Measurement
• United Nations Scientific Committee on Effects of Atomic

Radiation

Radiation Dose (mSv)

Annual Background Radiation

Federal limit for Radiation workers

Virtual Colonoscopy Myocardial Perfusion Scan CT Urogram (3-phase) CT Abdomen / Pelvis CT Coronary Angiogram PET / CT Scan Mammogram (screen / diagnostic) Mammogram + tomosynthesis Molecular Breast Imaging Chest X-ray CT Screening Lung Cancer Bone Densitometry Coronary Calcium Score Breast Stereotactic Biopsy CT Abdomen / Pelvis w & w/o contrast Chest CT Intravenous pyelogram

The 3 most common (radiation-related) reasons patients do NOT want to have an x-ray / CT scan

• 1. “I’ve had too many – I’m worried about the

cumulative effects of so many x-rays”

• 2. “I’m only 25 – the radiation risk is far higher

for me than for a 50-year old person”

• 3. I read a recent newspaper article that said….

8

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 )

Background Radiation vs. Annual Cancer Mortality Rates/100,000 for each U.S. State

Dose delivered instantaneously Dose delivered Over 3 years

(avg. dose / session = 11 mSv)

Lung Cancer Mortality vs. Dose Rate

According to LNT they should be the same !

Excess Relative Risk as a Function of Radiation Dose per day

Gregoire O, Cleland MR. Novel Approach to Analyzing the Carcinogenic Effect of Ionizing Radiations. International Journal of Radiation Biology 82(1), pp.13-19 (2006)

The 3 most common (radiation-related) reasons patients do NOT want to have an x-ray / CT scan

• 1. “I’ve had too many – I’m worried about the

cumulative effects of so many x-rays”

• 2. “I’m only 25 – the radiation risk is far higher

for me than for a 50-year old person”

• 3. I read a recent newspaper article that said….

Are CT scans safe for young women?

Figure 7-3, BEIR VII report

Each dot represents a different / separate study

The 3 most common (radiation-related) reasons patients do NOT want to have an X-ray / CT scan

• 1. “I’ve had too many – I’m worried about the

cumulative effects of so many x-rays”

• 2. “I’m only 25 – the radiation risk is far higher

for me than for a 50-year old person”

• 3. I read a recent newspaper article that said….

9

Incidence risk ratio Lag period Single procedure Multiple procedures 1 year 1.72 9.05 2 years 1.64 5.98 5 years 1.48 2.90

Cohort included total of 12,068,821 individuals. Of these, 1,275,829 individuals (10.6%) were exposed to diagnostic low-dose ionizing radiation

Example of a fatally flawed study

50

Incidence risk ratio Lag period Single procedure Multiple procedures 1 year 1.72 9.05 2 years 1.64 5.98 5 years 1.48 2.90

The lag period for solid cancers is 10 years ! This study is an example of reverse causation

51

Should we reduce radiation does from Medical Imaging Procedures?

Yes (but not too much!) Not necessarily because it causes cancer, but because people fear it will cause cancer Fear has consequences

• missed or delayed diagnoses / treatment
• higher health care costs
• anxiety and emotional distress

Also, because imaging is so beneficial, many people will receive multiple imaging procedures, CT scans, or x-rays

• want them to get the care they need without fear/worry
• want them to get the most effective medical care