PAGE 1

The pediatric radiologist has an important role as the child's advocate since the pediatric patient has unique physiological, anatomical, and maturation factors that must be considered before selecting an imaging test. For example, the neonate's lack of renal tubules reduces its ability to concentrate urine, the small size of an infant's airway allows inflammation and edema to more rapidly compromise respiratory efforts, and many young infants and some older children need to be immobilized and sedated.

Since the fetus, neonate, and young child have rapid growth with numerous mitotic cells, these individuals are most sensitive to many forms of injury. There is abundant evidence that irradiation of the fetus and the young child can have profound effects including genetic mutations, radiation-induced carcinogenesis, and early cell death. These are not theories. The evidence is found in the radiation therapy literature when such misguided efforts as to decrease the size of the thymus in neonates with radiation resulted in an excess relative incidence of thyroid cancer ¹. Other examples of a radiation treatment of benign disease causing cancer are those of children with tinea capitus and hemangiomas of the skin ².

We also find examples of diagnostic imaging causing excess relative risk of cancer. The most compelling non-fetal study is the excess mortality from breast cancer in a cohort of 5,466 female scoliosis patients who received multiple diagnostic x-rays during childhood and adolescent ³. Therefore, we must understand dose descriptors, conversion of absorbed and dose equivalent radiation units, and know what dose of radiation our patients are getting. The largest component of man-made radiation is from medical procedures. This is the most controllable factor and the primary source of this radiation (approximately 70%) comes from CT, though it is only 10% of the diagnostic radiation-producing tests⁴.

The dose descriptors allow us to understand the values and used correctly will allow us all to speak the same language. Nuclear medicine and positron emission tomography examinations are described by radioactivity injected. Exposure to x-rays is measured in roentgens (R) but this term is now rarely used. The two methods of expression of absorbed dose are the gray (Gy=new), and the rad (older term). The sievert (Sv-new) and rem (old) are units used in radiation measurement which take into consideration the type of radiation in producing biological damage. These equivalent dose measurements are based on radiation weighting factor. Fortunately for x-rays, gamma rays, beta particles, and electrons, the radiology weighting factor is 1 and therefore a rem equals a rad.

Equivalent doses are those used in measurement of devices or estimates of dose. Absorbed dose relates to measurement on patients. Each time one uses an absorbed unit (not equivalent unit) one must have a modifier such as skin dose (absorbed by the skin), whole body dose (absorbed by the whole body), or specific organ dose. In pediatrics a very important concept is the effective dose as it considers specific tissues and their radiosensitivity.

"To compute (the effective dose) look at the patient organ dose, take into account relative radiosensitivity using published (organ) weighting factors and you get a number that you can specify in terms of rem or mSv. It is not a risk factor"⁵.

PAGE 2

Units	Radioactivity	Absorbed Dose	Dose Equivalent	Exposure
Old units	curie (Ci)	rad	rem	Roentgen (R)
System international units	becquerel (Bq)	gray (Gy)	sievert (Sv)	coulombs/kg

Conversion Equivalents

1 millicurie (mCi) = 37 megabecquerels (MBq)*

100 rad = 1 Gy

1 rad = 1 cGy

100 rem = 1 Sv

1 rem = 10 mSv

Conversion Rules

To convert mBq to mSv, use conversion table 1 mrad = millirad=1/1000 of a rad Rem = rad x quality factor (for gamma and x-ray quality factor is 1, therefore rem = rad x 1)

Background radiation dose is approximately 1 millirad/day (300 mrad/year)

PAGE 3

Exam	mrad or mrem	Site Measured
Chest - 2 views	10-20	entrance (skin)
Abdominal - 2 views	50-100	entrance (skin)
Fluoroscopy (Non-Pulsed)	300-500/min	entrance (skin)
Fluoroscopy (Pulsed)	100-150/min	entrance (skin)
Computed tomography1 (Head)	6000 (2000-3000)	mid-diameter of phantom of 16 cm
Computed tomography1 (Abdomen)	3000 (1000)	mid-diameter of phantom of 32 cm
Nuclear medicine2 (99mTcMAG3-renal)	120	effective dose
Positron emission tomography2 (Brain FDG)	185	effective dose whole body

¹Scan explained as CT dose index (CTDI). First dose is with adult factors; second in parenthesis is examination adjusted for children.

²This is expressed as effective dose. These are rough guidelines for dose given to a 5-year-old with normal renal function. From International Commission on Radiation Protection: Radiation dose to patients from radiopharmaceuticals³⁶

^99mTcMAG3 = 99mtechnetium mercaptoacetyltriglycine

FDG=(F-18) fluoro-2-deoxyglucose

PAGE 4

Tissue or Organ	Tissue Weighting Factor37
Gonads	0.20
Bone Marrow (red)	0.12
Colon	0.12
Lung	0.12
Stomach	0.12
Bladder	0.05
Breast	0.05
Liver	0.05
Esophagus	0.05
Thyroid	0.05
Skin	0.01
Bone Surface	0.01
Remainder	0.05

PAGE 5

There are two types of biological radiation effects: deterministic and stochastic. Deterministic effects have a threshold below which the specific biological effect will not occur. The effect is dose dependent. That is, the higher the dose (above the threshold), the more severe the effect. An example of a deterministic effect is development of cataracts. Stochastic effect is not dose dependent. It is more random injury of DNA and cytoplasm. It is a probability: that is, the probability of the event occurring increases with increased radiation dose. The severity of a stochastic event is not dose dependent as are deterministic effects.

Radiation induced cancer is a stochastic event, resulting from random injury of DNA and cytoplasm. Stochastic effects can theoretically occur at any dose but chances of occurring increase with increasing dose. This lack of proven threshold and the probability basis (e.g., small risk with small dose) of stochastic effects has led to the linear no threshold theory (LNT) which states that since any dose of radiation can cause DNA injury, we can assume that no level of radiation is safe⁶.

There is another term, hormesis, which has been discussed with respect to radiation biological effects. This concept states that low dose radiation is beneficial and aids adaptive responses, i.e., increased immune response⁷. This concept is not generally accepted for pediatric radiation exposure and is irrelevant to the long-term effects of low dose radiation.

PAGE 6

The three biological effects of radiation are:

• Genetic mutations
• Radiation carcinogenesis
• Cell death (which helps remove damaged potentially mutable cells)

Genetic mutations do occur but the incidence and long-term effects are not well defined. The issue of major concern is rate of excess fatal cancer-radiation-induced carcinogenesis⁹.

Biological effects of radiation result from damage to DNA (the chromosome). The breaking of both strands of DNA (especially when the break in each strand is in close proximity) may result in faulty repair-deletions, reciprocal translocation or aneuploidy⁸. The biochemical and physiological changes follow over the next hours but the induction of cancer takes many years.

The greatest risk occurs in those tissues, organs, or organisms undergoing the largest number of mitotic events such as in the growing fetus, neonate, and child. Since radiation-induced solid tumors (leukemia is usually manifest in a shorter timeframe of 10 years) take 30 or more years to be manifest-it is apparent that a child's long life span after exposure is conducive to developing these solid neoplasms. The effects of each radiation episode are also cumulative-life-long-another factor that puts the child at risk.

Radiation-induced carcinogenesis is a multi-step process---morphological changes initiated by radiation, cellular immortality, and tumorgenicity. "Most human cells are clonal in origin. That is, the cells with the tumor are descendants of a single cell that has undergone the process of neoplastic transformation"⁸. Radiation exposure induces cellular genomic instability which is transmitted to the progeny of the affected cells. This instability is transmitted to their progeny and leads to a "persistent enhancement in the rate of which genetic changes arise in descendants of the irradiated cells after many generations of replication...(this process) has been termed non-targeted effects of radiation as genomic damage occurs in the cells that in themselves had received no direct radiation exposure"⁸.

Most childhood tumors occur sporadically but 10-15% have a familial association in which there are chromosomal deletions. For example, in retinoblastoma the patient is born with one affected allele and it only takes one hit of the wild allele to cause cancer. This is the two-hit theory of Knudsen¹⁰. Those children with certain diseases are very sensitive to radiation-induced carcinogenesis and we should perform examinations with radiation only when a non-radiation producing test cannot be used.

Artistic interpretation of radiation-induced mutagenesis. Note normal wild-type cells, and mutated cells. B is an example of a cell directly mutated by radiation exposure; the mutation is transmitted to all of its progeny. However, most of the cells in the irradiated population will retain the wild-type phenotype (A). C and D are examples of mutations arising as a result of radiation-induced genomic instability. The irradiated cell and its immediate progeny are wild type, but the frequency with which mutations arise amongst the more distant descendants of the irradiated cell is elevated. Reproduced with permission from Ontario, BC Decker 8.

Artistic interpretation of radiation-induced mutagenesis. Note normal wild-type cells, and mutated cells. B is an example of a cell directly mutated by radiation exposure; the mutation is transmitted to all of its progeny. However, most of the cells in the irradiated population will retain the wild-type phenotype (A). C and D are examples of mutations arising as a result of radiation-induced genomic instability. The irradiated cell and its immediate progeny are wild type, but the frequency with which mutations arise amongst the more distant descendants of the irradiated cell is elevated. Reproduced with permission from Ontario, BC Decker 8.

PAGE 7

Inherited Syndromes with Increased Sensitivity to Radiation include^{9 38}:

• Ataxia-telangiectasia
• Basal cell nevoid syndrome
• Bloom Syndrome
• Cockayne's syndrome
• Down syndrome
• Fanconi's anemia
• Gardner's syndrome
• Nijmegan breakage syndrome
• Hereditary retinoblastoma
• Usher's syndrome

PAGE 8

Diagnostic radiation is referred to as "low dose" (less than or equal to 10 R or 10 mSv). There are many risks associated with our everyday life⁸. Certain risks are acceptable, others not. Our highest risk group is fetuses and a 50-year study of fetuses whose mothers received abdominal radiation in the third trimester is quite revealing. The relative risk of cancer in childhood associated with radiation in utero was found to be 1.38 (the baseline is 1), i.e., a 38% individual increase. This individual increase is a small one when one considers the incidence of cancer is 1 per 1000 (this is an example, not actual incidence)^{11 12 13 14 15}.

The largest (86,611 survivors) longitudinal study of life-long (up to 55 years thus far) excessive cancer secondary to irradiation comes from the Radiation Effects Research Foundation Study of the Survivors of the A-bomb^{16 17 18 19}. The study has 3,184,354 patient years and 45% of the group were still alive in the year 2000. The doses calculated in this study were organ doses, so it is irrelevant whether the patient received whole body (as they did) versus "focal irradiation" by a CT. The low dose group was relatively remote from the epicenter or heavily shielded. The exposure was mostly direct gamma irradiation but there was also some neutron exposure included in the calculation.

The most important factor for sensitivity to radiation-induced cancer is the age at exposure²⁰. Children are up to 10 times more sensitive to radiation-induced carcinogenesis than middle-aged adults. The youngest neonate is more sensitive than the older child. Girls are more at risk than boys because of a higher sensitivity to breast and thyroid cancer.

Data from the A-bomb survivors expressing the lifetime risk of excess cancer per sievert as a function of age at the time of exposure. While the average risk for a population is about 5% per sievert, the risk varies considerably with age: children are much more sensitive than adults. At early ages, girls are more sensitive than boys. Reproduced with permission from Springer Verlag20.

Data from the A-bomb survivors expressing the lifetime risk of excess cancer per sievert as a function of age at the time of exposure. While the average risk for a population is about 5% per sievert, the risk varies considerably with age: children are much more sensitive than adults. At early ages, girls are more sensitive than boys. Reproduced with permission from Springer Verlag20.

Both radiation-induced leukemia (relatively early stochastic effect) and solid tumors (relatively late stochastic effect) were found in excess in this ongoing study. The kinds of solid tumors were similar to those occurring naturally (breast, lung, thyroid, gastrointestinal) but occurred more frequently than expected. The doses causing this excess mortality from low dose of radiation in the atomic bomb study overlaps our current CT doses^{21 22}.

Data from the A-bomb survivors expressing the relevant risk for cancer mortality. Relevant dose range for pediatric CT: 6-100 mSv (0.006=0.1 Sv). "There is direct, statistically significant evidence for risk in the dose range from 0 to 0.1 Sv." Reproduced with permission from Springer Verlag22.

Data from the A-bomb survivors expressing the relevant risk for cancer mortality. Relevant dose range for pediatric CT: 6-100 mSv (0.006=0.1 Sv). "There is direct, statistically significant evidence for risk in the dose range from 0 to 0.1 Sv." Reproduced with permission from Springer Verlag22.

Brenner et al predicted that of the 600,000 abdominal and head CT examinations annually performed in children under the age of 15, a rough estimate is that 500 of these individuals might ultimately die from cancer attributed to CT radiation²¹. Roughly this corresponds to a risk of 1 in a 1000. If estimates are true or even if we curb the amount of radiation per examination so that the estimates are only one-quarter of those predicted (0.25 per 1000), we still end up with 675 cases of excessive cancer per million children irradiated by CT. Currently there are at least 2.7 million children in the world who receive CT irradiation. What began as a small individual risk has become a larger public health problem.

PAGE 9

Activity	Death (per million/year)8
Being a person age 55 years (all causes)	10,000
Smoking a pack of cigarettes daily (all causes)	3,500
Rock climbing for 2 h (accident)	500
Canoeing for 20 h (accident)	200
Motorcycling for 1,000 miles (accident)	200
Traveling 1,500 miles by car (accident)	40
Being a pedestrian (accident)	40
Working 1 week as a firefighter (accident)	15
Working 1 week in agriculture (accident)	10
Fishing (drowning)	10
Eating (choking on aspirated food)	8
Skiing for 10 h (accident)	8
Working 1 month in a typical factory (accident)	5
Traveling 5,000 miles by air (accident)	5
Having a chest radiograph (radiation-induced cancer)	1
Visiting Denver for 2 months (cancer from cosmic rays)	1
Living in the vicinity of a nuclear power plant (radiation-induced cancer)	<0.1

PAGE 10

CT may not be our only problem. To follow the ALARA concept (As Low As Reasonably Achievable) one must know the appropriate radiation dose (or a dose factor such as mA in CT) and balance this against image quality. We will soon be working in entirely digitized imaging departments with computed radiography (CR) or digital radiography (DR). While this digital technology allows for the use of PACS, improved service, and ease of reading images without bulky film, there are some trade-offs:

• CR and DR have less spatial resolution than plain films (plain films have one-half to two-thirds better spatial resolution)²³
• Because of the digital method of exposing the plate, basic radiologic techniques of immobilizing, collimation, and position remain crucial to good image quality and cannot be ignored
• More to the point, there is the uncoupling effect, i.e., if there is adequate radiation to obtain an image, the post-processing of images makes them all look similar and great

If a 2-4 time excess radiation dose is utilized, the image will still be excellent, not a dark film like we used to have. One problem with current imaging is that it is digital with post-processing and the elements (exposure factors) that controlled image quality with film (causing over or under exposed) no longer apply. We have uncoupled the end product from the radiation given. CR and DR both often have higher dose than the old plain films to obtain the same examination. The production of digital "plain films" needs to be understood by the radiologist so we do not overdose our children.

PAGE 11

There are three ways to reduce radiation from diagnostic imaging:

1. Don't do any examination that is not indicated
2. Utilize a non-radiation imaging modality whenever possible to answer the clinical question
3. When performing an examination utilizing radiation, tailor it to the clinical question and the child (i.e., sized-based adjustments in exposure parameters)

Utilizing 1 and 2 reduces the radiation dose to 0! To do this, the pediatric radiologist must spend time discussing the case with the clinician and reach a decision if any examination is indicated and what type of imaging should be done. The American College of Radiology has imaging guidelines which will help us understand what is useful^{20 21 22 23 24 25}. It has been estimated that 30% of CT examinations in children are not necessary or are redundant, or can be done with another non-radiation producing modality²⁶. When a CT examination is indicated, it must be done properly with pediatric parameters^{26 27 28}. Adult CT settings with resultant exposure to children result in a relatively higher dose than that given to adults (higher effective dose to the organ)²⁹. This is explained by the fact that the absorption of x-rays is dependent on a cross-sectional area of the irradiated tissue in addition to tissue characteristics (the latter being similar for adults and children). However, x-rays that would have been absorbed in a near field in an adult passes through the entire child irradiating all the organs.

Huda et al has estimated the effective dose to a newborn from a head CT examination at comparable parameters of milliamperage per second (mAs) and kilovoltage (kV) gives four times the dose of an adult³⁰. With abdominal imaging, the dose is increased by two-thirds. In a later publication, the same group determined that value of energy imparted to patients undergoing abdominal CT examinations was a factor of three higher in adults than children but the corresponding patient effective dose was 50% higher in children than adults³¹.

PAGE 12

Designing the appropriate study technique involves many factors. The major focus should be on:

• mAs (both tube current and gantry cycle time)
• kVp
• Thickness of slices (collimator thickness) and table speed (pitch)
• Number of detector rows (single, 4, 8, 16, 32, 64)
• Number of phases - each run (without contrast, with contrast, delayed) can increase the radiation in a multiple of the number of phases performed (eg. a three phase study at the same parameters will be 300% increased over a single phase)
• Length scanned

Radiation dose and mAs in CT are directly proportional. When you lower mAs 50% you reduce the dose 50% ²⁷. A good starting place for mAs and abdominal CT scans in children is found in the following Table²⁷. We also must limit the number of phases to what we absolutely need. Most of us only do a contrast-enhanced study and if we need delayed images, we may reduce the mAs further. Remember, a single phase versus two phases halves the dose.

One must restrict the length of study to our particular area of interest. Significant unnecessary radiation occurs in an abdominal examination for the liver when one includes three quarters of the chest and the entire pelvis. The technologists must understand what you want. The technologist is a vital part of the team. He/she must search for the correct protocol, know what anatomical region is the focus of the procedure, and be familiar with factors that control radiation doses. When possible, vendor specific programs in which radiation is tailored to thickness of tissue irradiated can be utilized.

Thinner slices need more radiation as you need more signal to overcome the noise if you want to maintain the same image quality as with a thicker slice. Basically, it may be best to obtain thinnest collimation configuration and reconstruct desired thickness.

Changing kVp is less familiar to radiologists but can be important to understand. For example, changing kVp is not linearly related to the radiation dose as is mAs. For every change of kVp the radiation dose changes are exponential, and greater in extent (perhaps 1/3 more or less). While maintaining diagnostic image you may have to increase mAs when you decrease kVp. The sum total still can be a substantial reduction in radiation with diagnostic images. The manipulation of kVp is a fertile area for research.

The key factor is obtaining a diagnostic image-not necessarily the most beautiful picture. This is obtained by having enough signal over background (the noise). It has been shown in many instances that cutting the radiation dose in half to adults receiving CT does not reduce the diagnostic efficiency of the study and the radiologist's ability to make the proper diagnosis^{32 33 34 35}.

PAGE 13

1. Make sure indications justify examination
2. Choose the appropriate modality; if possible use one without or with less radiation
3. Shield, when possible, at-risk structures (breast, eyes, gonads)
4. Optimize techniques:

   Conventional (plain film) radiography:

   • Standardize views
   • Follow basic tenants of good technique: immobilization, position, collimation, immobilization
   • Work with manufacturers to obtain diagnostic image at least radiation dose: filter the beam; quality assurance program for digital equipment and PACS

   Fluoroscopy:

   • Use pulsed fluoroscopy
   • Use last frame save
   • Have idealized time/exam
   • Think with your foot off the pedal

   CT:

   • May use variable vendor programs to alter radiation based on body geometry or density of tissue
   • Adjust parameter for child's size: mAs, kVp, number of phases or runs (without and with contrast doubles the radiation), collimator thickness and table speed (pitch), length scanned

PAGE 14

CT and other radiation-producing imaging modalities are invaluable to our armamentarium. Because of the real and potential radiation risks, we must use these methods judiciously. It is the responsibility of the radiologist to know how to do this.