Luc Rotenberg, Tomosynthese et densité mammaire 2015

!

Luc Rotenberg

ISHH – RPO Clinique Hartmann – Ambroise Paré Neuilly Sur Seine - France

#drrotenberg [email protected]

Tomosynthesis DBT and Breast density

La valeur de la Tomosynthèse dépend elle de la densité mammaire ?

Tomosynthèse Mammaire – Montpellier 10 & 11 septembre 2015

!

Evaluation de la densité mammaire: Est-ce utile ?

Seins Denses

Facteur de Risque

Limites de l’imagerie RX

!

Historique

S  1976, Dr. Wolff published major paper on breast patterns and risk

“Groups can be isolated with as much as a 37 times greater incidence of the disease”

!

Breast Density

S  The literature links breast density with the risk of breast cancer: S  1976: Wolfe1 discussed a 37× increased risk of

developing breast cancer

S  2004: Harvey2 summarized recent estimates at more than 4× increased risk

1 AJR 126:1130-1139, 1976 2 Radiology 2004; 230:29–41

!

Classification Incidence studies

RR [95% IC] Prévalence studies

RR [95% IC]

Wolfe

N1 1 1

P1 1.8 [1.4-2.2] 1.3 [1-1.5]

P2 3.1 [2.5-3.7] 2 [1.3-3]

DY 4 [2.5-6.3] 2.4 [2-3]

% of Density

<5% 1 1

[5-24%] 1.8 [1.5-2.2] 1.4 [1.1-1.8]

[25-49%] 2.1 [1.7-2.6] 2.2 [1.8-2.8]

[50-74%] 2.9 [2.5-3.4] 2.9 [2.3-3.8]

> 75% 4.6 [3.6-5.9] 3.7 [2.7-5]

BI-RADS

1 1 1

2 2.2 [1.6-3] 1.6 [0.9-2.8]

3 3 [2.2-4.1] 2.3 [1.3-4.3]

4 4 [2.8-5.7] 4.5 [1.9-10.6]

McCormack VA & al : Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2006, 15:1159-1169.

Combined relative risks for breast cancer associated with different classifications of mammographic density, study designs, and study populations from meta-analysis

!

Breast density & diseases

S  Greater risk of benign breast disease

S  RR 12.2 for usual hyperplasia

S  RR 9.7 for ADH or DCIS

S  RR 4.3 for Cancer S  Higher grade

S  ER negative

S  Larger size = rapid growth in dense tissue ? S  Masking ?

S  Growth factors ?

!

Breast Cancer Risk Factors

Risk Factor Min Max x LCIS on biopsy 1.00 8.70 8.70 No. 1st degree relatives with breast cancer 1.00 6.80 6.80 Mammographic density 0.41 1.76 4.29 No. of biopsies 1.00 2.88 2.88 Tamoxifen 0.45 1.00 2.22 Biopsy with atypical hyperplasia 0.93 1.82 1.96 Oral contraceptives / Menop Horm Therapy 1.00 1.49 1.49 Alcohol use / Obesity 0.99 1.41 1.42 Early menarche / late menopause 1.00 1.21 1.21

From http://www.halls.md/breast/gailmods.htm

!

Composants de la densité mammaire ?

S  Canaux galactophoriques

S  Lobules

S  Stroma (facteurs de croissance) S  Collagène

S  Fibroblaste

S  Matrice intercellulaire

S  Vaisseaux sanguins

S  tissus de soutien (FAK-ERK link)* …. Profil juvenile Post-ménopausique Asiatique…..

*Provenzano PP, Inman DR, Eliceiri KW, Keely Oncogene 2009; 28: 4326-43.

!

Aspect mammographique

S  Clarté (noir) = graisse

S  Opacité (blanc) = tissus denses

S  tissus fibro-glandulaire S  éléments fonctionnels S  parenchyme S  Éléments de soutien S  stroma

Qualitative Quantitative

Wolfe BI-RADS Visual estimation

Description

Visual classification of the mammographic image into four categories based on extent and distribution of the parenchyma, including ducts, nodular, homogeneous densities, and fat.

Standardized reporting of visual assessment of mammographic findings by the American College of Radiology BI-RADS. Both breasts are used for the BI-RADS

Radiologist or expert reader subjectively assigns a percentage density corresponding to the proportion of breast that is dense.

Categorization

N1 – Completely fatty breast P1 – Mainly fatty breast with prominent ducts, up to 25% density

P2 – Prominent ducts, more than 25% density

DY – No visible ducts, diffuse and extensive nodular density

Category 1 : Almost entirely fatty (<25% dense)

Category 2 : Scattered fibroglandular densities (25%–50% dense)

Category 3 : Heterogeneously dense (51%–75% dense)

Category 4 : Extremely dense (>75% dense)

Visually estimated directly into categories

<10%

10%–<25%

25%–<50%

50%–<75%

>75%)

Classifications mammographiqes de la DM

Un choix : BI-RADS®

Standardized reporting of visual assessment of mammographic findings by the American College of Radiology BI-RADS. Both breasts are used for the BI-RADS

u  Categorie 1 : Almost entirely fatty (<25% dense) u  Categorie 2 : Scattered fibroglandular densities (25%–50%

dense) u  Categorie 3 : Heterogeneously dense (51%–75% dense) u  Categorie 4 : Extremely dense (>75% dense)

Classifications mammographiqes de la DM

From the ACR BI-RADS Atlas

Birads density classification

1 3 2 4

< 25% 25-50 % 50-75 % > 75 %

!

prevalence of increased density in the general population

S  percentage density :

S  26% to 32% had 50% or more

S  parenchymal pattern

S  21% to 55% had the P2 or DY

S  BI-RADS density

S  31% to 43% had a BI-RADS of 3 or 4

Dense breast ≥ 33%

!

What is Breast Density?

S  But how does one classify the density of this breast?

S  By % density?

S  By pattern?

!

Present Practice

S  Radiologists assess breast density: 1.  Visually

2.  From two dimensional images

3.  By estimating the amount of “whiteness”

!

mammographic density percentage visually estimated by two untrained radiologists versus reference-standard density percentage

Harvey, J. A. et al. Radiology 2004;230:29-41

!

Then Why Use Density?

S  Very simple … because until now there was no alternative

S  Then we must : S  standardize methods for measuring density

S  with digital volumic breast density assessment

!

Volumic breast density assessment

S  automatically differentiates dense tissue from fat in the breast

S  measures of total breast fibroglandular tissue volume

S  calculates volumetric breast density

S  Example : Quantra (Hologic)

!

Volumetric vs Area Density

S  Fraction of fibro-glandular tissue (pink) within a fatty breast:

By volume:

By area:

L

L 2L

4L

2L

2L

4L

2L

L

%25%100242

=××

×

LLLL

%5.12%1002242

=×××

××

LLLLLL

!

Risk and Density

S  Any density measure is subject to change if the total volume of the breast changes

S  But if the amount of fibroglandular tissue remains constant, does the risk change?

L

L 2L

4L

2L

2L

L

L 2L 4L

4L

2L

!

The Result

S  A table of measures of volume and density in the breast:

!

N Engl J Med 2005; 353:1773-1783

!

N Engl J Med 2005; 353:1773-1783

conclusions Digital mammography is more accurate in women : •  under the age of 50 years, •  radiographically dense

breasts •  premenopausal or

perimenopausal

•  Clinical trial made in 2004 – 2006 in North America

•  50.000 women enrolled made both exams (FFDM/SFM)

!

As a matter of fact we know:

•  Breast screening target is EARLY DIAGNOSIS OF BREAST CANCER

•  In most of the cases Screening reaches the goal

•  almost 10 – 15% of the found late cancers is originated in regularly screened women

•  FFDM is “blind” under some particular circumstances :

•  dense breasts •  dense tissues overlapping lesions

FFDM: SUPERIOR TECHNIQUE, BUT NOT PERFECT

RESULTS •  For women ≤ 50 years old and/or dense breast

•  Sensitivity goes from 51% (SFM) to 70 - 78% (FFDM) •  Visualized almost 28% more breast cancers •  More than 1 over 4 cancers were not recognized:

false negatives

!

Pooled BI-RADS–based ROC curves for diagnostic assessment of conventional diagnostic views and tomosynthesis views

Zuley M L et al. Radiology 2013;266:89-95, Pittsburgh

Digital Breast Tomosynthesis - DBT

Detection

!

DBT

ROC curves for average probability of malignancy as assessed by using conventional supplemental diagnostic views and tomosynthesis views.

Zuley M L et al. Radiology 2013;266:89-95

VPP

!

Pooled ROC curves for reader studies 1 and 2 using probability of malignancy scores; curves represent average ROC performance

for 12 readers in study 1 and 15 in study 2.

Rafferty E A et al. Radiology 2013;266:104-113

!

Assessing Radiologist Performance Using Combined Digital Mammography and Breast Tomosynthesis Compared with Digital Mammography Alone: Results of a Multicenter, Multireader Trial

Diagnostic Sensitivity, Specificity, and Positive and Negative Predictive Values

Rafferty E A et al. Radiology 2013;266:104-113, Boston

!

Applications

1.  Detection « screening » 2.  Caracterisation 3.  localisation

!

Applications


!

Detection

!

!

!

44 y, screening

Density ?

Birads 3

!

002. CA 44 y, screening

!


!

Question

Final Birads classification ?

S  1

S  2

S  3

S  4

S  5

S  0

!

Question


S  1

S  2

S  3

S  4

S  5

S  0

Birads 5 • Biospy required

!


IDC G2, HR +, R2-, N-

!

Applications


!

!

!

Breast Cancer Screening Using Tomosynthesis in Combination With Digital Mammography

Sarah M. Friedewald, MD1; Elizabeth A. Rafferty, MD2; Stephen L. Rose, MD3,4; Melissa A. Durand, MD5; Donna M. Plecha, MD6; Julianne S. Greenberg, MD7; Mary K. Hayes, MD8; Debra S. Copit, MD9; Kara L. Carlson, MD10;

Thomas M. Cink, MD11; Lora D. Barke, DO12; Linda N. Greer, MD13; Dave P. Miller, MS14; Emily F. Conant, MD15

JAMA. 2014 Jun 25;311(24):2499-507

S  454 850 examinations S  281 187 digital mammography

S  173 663 digital mammography + tomosynthesis

S  recall rate : S  107 ‰ with digital mammography

S  91 ‰ with digital mammography + tomosynthesis;

S  Difference = – 15% (P < .001)

S  Biopsies

S  18.1‰ with digital mammography

S  19.3‰ with digital mammography + tomosynthesis

S  difference + 6%

Recall Rate – 15%

Biopsy Rate + 6%

!




JAMA. 2014 Jun 25;311(24):2499-507

S  Cancer detection,

S  4.2 ‰ with digital mammography

S  5.4 ‰ with digital mammography + tomosynthesis

S  Difference 1.2 ‰

S  Invasive cancer detection

S  2.9 ‰ with digital mammography

S  4.1 ‰ with digital mammography + tomosynthesis

S  difference +1.2 ‰

S  in situ cancer detection S  1.4 ‰ screens with both methods.

S  Adding tomosynthesis increase in the PPV S  for recall from 4.3% to 6.4% = + 2.1%

S  for biopsy from 24.2% to 29.2% = + 5.0%

CDR Rate + 28,5% 5 , 4 ‰ v s 4 , 2 ‰

CDR InvK + 41,3% 4 , 1 ‰ v s 2 , 9 ‰

P P V 1 + 4 8 , 8 % 6 , 4 % v s 4 , 3 %

PPV3 for biopsy + 20,6% 2 9 , 2 % v s 2 4 , 2 %

!




JAMA. 2014 Jun 25;311(24):2499-507. doi: 10.1001/jama.2014.6095.

Conclusions and Relevance

S  Addition of tomosynthesis to digital mammography was associated with

S  a decrease in recall rate

S  an increase in cancer detection rate

S  Further studies are needed to assess the relationship to clinical outcomes.

!

AJR Am J Roentgenol. 2014 Sep;203(3):687-93. Clinical performance metrics of 3D digital breast tomosynthesis compared with 2D digital mammography for breast cancer screening in community practice.

Greenberg JS1, Javitt MC, Katzen J, Michael S, Holland AE.

S  Outcomes from screening mammography between Aug 2011, and Nov 2012

S  using 3D DBT S  n = 23,149 patients

S  versus 2D DM S  n = 54,684 patients

!

AJR Am J Roentgenol. 2014 Sep;203(3):687-93 Clinical performance metrics of 3D digital breast tomosynthesis compared with 2D digital mammography for breast cancer screening in community practice.


RESULTS

S  For patients screened with 3D DBT, the relative change in recall rate was 16.1% lower than for patients screened with 2D DM (p > 0.0001)

S  The overall cancer detection rate (CDR), expressed as number of cancers per 1000 patients screened, was 28.6% greater (p = 0.035) for 3D DBT (6.3/1000) compared with 2D DM (4.9/1000).

S  The CDR for invasive cancers with 3D DBT (4.6/1000) was 43.8% higher (p = 0.0056) than with 2D DM (3.2/1000).

S  The positive predictive value for recalls from screening (PPV1) was 53.3% greater (p = 0.0003) for 3D DBT (4.6%) compared with 2D DM (3.0%).

S  No significant difference in the positive predictive value for biopsy (PPV3) was found for 3D DBT versus 2D DM (22.8% and 23.8%, respectively) (p = 0.696).

Recall Rate - 16%

CDR Rate + 28,6% 6 , 3 ‰ v s 4 , 9 ‰

CDR InvK + 43,8% 4 , 6 ‰ v s 3 , 2 ‰

P P V 1 + 5 3 , 3 % 4 , 6 % v s 3 %

PPV3 for biopsy = 2 2 . 8 % v s 2 3 . 8 %

!

AJR Am J Roentgenol. 2014 Sep;203(3):687-93. Clinical performance metrics of 3D digital breast tomosynthesis compared with 2D digital mammography for breast cancer screening in community practice.


CONCLUSION

S  In community-based radiology practice, mammography screening with 3D DBT yielded compared with 2D DM

S  lower recall rates

S  increased CDR for cancer overall

S  increased CDR for invasive cancer

S  increased PPV1 in the group screened using 3D DBT.

!

S  Mammogram

S  Density S  Birads 3

S  Mass UOQ S  Mixte

S  Fat

S  Regular ?

!

!

!

Question


S  1

S  2

S  3

S  4

S  5

S  0

Breast Lipoma •  Benign •  No biopsy •  Back to screening

!

S  To compare diagnostic performance of S  (2D) mammography

S  2D mammography plus digital breast tomosynthesis (DBT)

S  synthetic 2D mammography plus DBT

S  8869 women

S  age range, 29–85 years ; mean, 56 years

S  from July 2011 to March 2013

Fiona J. Gilbert, FRCR Lorraine Tucker, DCR Maureen G. C. Gillan, PhD Paula Willsher, DCRJulie Cooke, FRCR Karen A. Duncan, FRCR Michael J. Michell, FRCR Hilary M. Dobson, FRCR Yit Yoong Lim, FRCR Tamara Suaris, FRCR Susan M. Astley, PhD Oliver Morrish, MSc Kenneth C. Young, PhDStephen W. Duffy, MSc

Radiology Ahead of Print - 2016

!

Figure 2a: Graphs show the ROCreceiver operating characteristics curve analysis curves for (a) all cases in the three arms of the study Fiona J. Gilbert; Lorraine Tucker; Maureen G. C. Gillan; Paula Willsher; & al; Radiology Ahead of Print

§  87% for 2D mammography, §  89% for 2D mammography + DBT, §  88% for synthetic 2D mammography +

DBT

Sensitivity Specificity

!

Figure 2b: Graphs show the ROC receiver operating characteristics curve analysis curves for cases with visually assessed breast density of 50% or more in all three arms of the study

Fiona J. Gilbert; Lorraine Tucker; Maureen G. C. Gillan; Paula Willsher; & al; Radiology Ahead of Print

§  86% for 2D mammography, §  93% for 2D mammography + DBT

Sensitivity Specificity

Density ≥ 50%

!

In women with dense breasts §  DBT increased

§  the sensitivity

§  86% for 2D mammography alone

§  93% for 2DM plus DBT

§  the specificity

§  58% for 2D mammography alone

§  69% for 2DM plus DBT

§  when the dominant radiologic feature was a mass, sensitivity :

§  89% for 2D mammography

§  92% for 2D mammography plus DBT



!

•  reducing the number of false-positive results

•  particular benefit in younger women with dense breasts

•  Synthetic 2D mammography similar to that of 2D mammography when used in conjunction with DBT



!

S  Birads before DBT ?

S  58 y

S  Right Breast cancer 2003

S  UOQ

S  Left VABB OQ 2010 : benign

S  Previous Mammo 2013 : Birads 2v

!

Recurrence IDC G3

!

Christoph I. Lee, MD, MSHS Mucahit Cevik, MS, Oguzhan Alagoz, PhD, Brian L. Sprague, PhD, Anna N. A. Tosteson, ScD, Diana L. Miglioretti, PhD, Karla Kerlikowske, MD, Natasha K. Stout, PhD, Jeffrey G. Jarvik, MD, MPH Scott D. Ramsey, MD, PhD Constance D. Lehman, MD, PhD

Radiology: Volume 274: Number 3—March 2015

To evaluate the effectiveness of combined biennial digital mammography (DM) and tomosynthesis(DBT) screening, compared with biennial digital mammography screening alone, among women with dense breasts.

!

Advances in Knowledge •  Combined biennial DMand DBT, compared with DM alone •  U.S. women aged 50–74 years with dense breast •  would avert :

•  1 additional breast cancer death per 2000 women screened

•  405 false-positive screening examination findings per 1000 women screened.

Comparative effectiveness of combined Digital Mammography and Tomosynthesis screening for

Women with Dense Breasts Christoph I. Lee, MD & all


!

Implications for Patient Care

In women aged 50–74 years with dense breasts Combined biennial DM and DBT screening compared with DM alone is : •  likely to decrease the number of false-positive findings •  increase the number of cancers detected •  likely to improve outcomes at reasonable additional cost :

•  cost-effective if priced around $226 for combined DM+DBT vs $139 for DM alone •  if reported interpretive performance metrics of improved specificity with DBT in routine

Comparative effectiveness of combined Digital Mammography and Tomosynthesis screening for

Women with Dense Breasts Christoph I. Lee, MD & all


!

Conclusion Take Home Messages

S  Best detection

S  Best caracterisation

S  Best localisation

S  Decrease recall rate ++

S  Increase :

S  cancer detection

S  Invasive +++

S  PPV 1

S  Dense Breast +++

S  Risk factor x 4

S  Best detection

S  for mass

S  Best specificity

S  Focal density assymetry

S  Decrease recall rate +++

S  Best location

•  National Screening Program ? •  Pricing ?

!

La valeur de la Tomosynthèse

dépend elle de la densité mammaire ?

Les seins denses bénéficient le plus de l’apport de la tomosynthèse

Conclusion Take Home Messages

Health & Medicine

Luc Rotenberg, Tomosynthese et densité mammaire 2015