John M. Boone, Ph.D., FAAPM, FSBI, FACR

Professor and Vice Chair (Research) of Radiology Professor of Biomedical Engineering

University of California Davis Medical Center Sacramento, California

Patient Dosimetry in Mammography

and Tomosynthesis: What to measure, why and how

2013 ICTP/IAEA Training Course on Radiation Protection of Patients Trieste

1

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

2

U.S. Breast Cancer Statistics (2006)

212,290 new cases 40,970 deaths 1 / 8 women will get breast cancer (12.5%)

Breast Cancer Lung Cancer

Incidence*

31% 12%

Mortality*

15% 26%

* of all cancers Ravdin, et al., NEJM 2004 3

Jemal A, et al., Cancer Statistics 2006

dea

th r

ate

per

1

00

,00

0

1930 2002

U.S. Cancer Mortality (1930-2002)

4

5

CC

CC MLO

MLO

Mammography: Standard of Care

Annual screening dose accumulation

average size breast

6

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

7

Digital mammography unit at UC Davis

8

Breast Dose is highly dependent upon the x-ray spectrum used. Different Anode/Filter combinations are used.

ANODE Mo Mo Rh W W W

FILTER Mo Rh Rh Rh Ag Al

S/F

Tomo

Digital

9

10

In addition to the Anode / Filter combo, the kV and HVL need to be well characterized

compression paddle

10

11

X-ray Tube Voltage Evaluation

non-invasive kV meter

11

12

In addition to the Anode / Filter combo, the kV and HVL need to be well characterized

0.1 mm thick Al filters (99.999%)

ion chamber calibrated to mammo energies

air gap

12

HVL as a function of tube voltage

13

14

ion chamber

Establishing x-ray tube output

mGy air kerma per 100 mAs at 50 cm

14

X-ray output (@50 cm) versus tube voltage

15

kV

HVL’s

glandular fraction

breast thickness

anode / filter combo

16

Validation of Breast Thickness Accuracy

breast phantoms (slabs)

4.1 cm

17

From Bloomquist, et.al. Med Phys 33: 719-736, 2006

Estimating Entrance Air Kerma

18

Estimating Entrance Air Kerma

Recorded in patient’s record (digital)

anode filter kV mA exposure time compressed breast thickness

19

Estimating Entrance Air Kerma

air kerma

dgN values normalized glandular dose coefficients

20

Estimating Entrance Air Kerma

@ Anode/Filter kV:

air kerma per 100 mAs @ 50 cm = k

ESK = k × (mAs/100) × ISL

50 cm 2

50

SID T

ISL

breast thickness usually ~65 cm

SID - T

21

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

22

Mean Glandular Dose

23

Mean Glandular Dose

glandular tissue

skin

adipose tissue

24

Mean Glandular Dose (Monte Carlo Calculations)

glandular tissue?

skin

adipose tissue?

glandular fraction

homogeneous mixture

25

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

26

DgN values are based on MC Studies

assumed 4 mm skin thickness

27

DgN values are based on MC Studies

assumed 4 mm skin thickness

28

Monoenergetic MC eval of E deposition

29

MGD =

men

r

men

r glandular

breast comp

Energy Imparted

Mass

Converting Energy imparted to dose….

f-factor (SI)

30

Mean Glandular Dose (Monte Carlo Calculations)

ESK

MGD

DgN = ESK

MGD

31

Dose Calc with Normalization by ESE

32

To convert the monoenergetic DgN values to realistic polyenergetic values, spectral models are used to weight the

monoenergetic values

33

DgN Tables (poly)

The units used in DgN Tables have varied over the years and by the country of origin. These units are mRad / R, but mGy/mGy

is in more common usage today 34

35

2 cm 4 cm

6 cm 8 cm

DgN versus monoenergetic x-ray energy

36

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

37

t = 4 mm?

38

t

Developed algorithms to segment skin from breast CT images and measure thickness

39

Skin Thickness measurements were relatively precise for each women (18%)

Measurements from the same women

40

Measurements from 51 different women

Skin Thickness measurements ranged from 0.9 mm to 2.3 mm (m = 1.45 mm)

41

Left / Right Comparison provided a consistency check, with good results

42

43

DgN values versus skin thickness

44

Change (in %) in DgN values versus skin thickness, Relative to the assumption of t = 4 mm

45

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

46

Mean Glandular Fraction

100% 0% 50%

high low medium

47

48

49

risk assessment & dosimetry

validation of 2D approaches (M. Yaffe)

%

Breast Density Analysis

segmentation

glandular

adipose

including skin

excluding skin

UC Davis breast CT N = 191

50

Data from UC Davis Breast CT

Den

sity

an

alys

is d

on

e o

n m

amm

ogr

ams

at

Un

iver

sity

of

Toro

nto

51

UCD & UT data combined (N = 2831)

Median ≈ 16 %

52

UCD & UT data combined (N = 2831) Cumulative Distribution

50%

75%

90%

95%

53

Breast density decreases about 5 % per decade

54

Areal Fraction versus Volume Fraction

areal fraction (%)

volu

me

frac

tio

n (

%)

0%

100%

0% 100%

55

DgN for 0% Glandular Breast

56

DgN for 100% Glandular Breast

57

2 cm 4 cm 6 cm 8 cm

5 cm

(265-237)/237 = 12%

50%

16%

increase in “average” DgN coefficient of:

12%

58

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

59

60

Tomosynthesis (limited angle tomography)

61

Conventional versus Tomosynthesis

Conventional Tomosynthesis Conventional Tomosynthesis

Similar display characteristics

62

63

Tomosynthesis (limited angle tomography)

±7.5°

Combo Tomo / Mammo mode output

mammography

tomosynthesis

mR/mAs or mGy per 100 mAs

64

65

Physical measurements of dose in Tomo/Mammo

Polyethylene is an excellent surrogate for adipose

66

Output of real time exposure meter

filter changes & grid is inserted

mammo exposure 15 tomo exposures

67

Point dose measures: Tissue Air Ratios

68

Mammography and Tomosynthesis Dosimetry

Mammography

Why measure breast dose? Basic Concepts of Breast Dosimetry (how) Mean Glandular Dose (MGD) (what) DgN coefficients Skin Thickness Issues Breast Density Issues

Tomosynthesis

Differences between tomo and mammo

Summary

69

Mammography and Tomosynthesis Dosimetry Summary

Complete characterization of the x-ray system is necessary (kV accuracy, HVL, and air kerma / 100 mAs @ anode/filter) Ionization chambers must be mammo beam compatible thin windowed and calibrated Practical Assessment requires table of DgN values specific to the conditions of the actual exam Current DgN tables may slightly under-estimate dose due to % glandular fraction and skin thickness issues Tomosynthesis is similar enough to the geometry of mammo anode/filter combo may differ; different DgN tables needed 70

71