Diffusion-weighted magnetic resonance imaging in breast cancer

DISSERTATIONS | OTSO ARPONEN | DIFFUSION-WEIGHTED MAGNETIC RESONANCE IMAGING IN… | No 547

uef.fi

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND Dissertations in Health Sciences

ISBN 978-952-61-3282-2

Dissertations in Health Sciences

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

OTSO ARPONEN

DIFFUSION-WEIGHTED MAGNETIC RESONANCE IMAGING IN BREAST CANCER

Breast cancer is the most common cancer in women worldwide and a major cause of cancer-related mortality. Diffusion-weighted magnetic resonance imaging (DWI-MRI) may

make it possible to reclassify as benign some of those lesions that have been determined suspicious according to conventional MRI thus decreasing the number of unnecessary biopsies. This study helped in standardizing DWI-MRI and concluded that DWI-MRI can improve the characterization of incidental

MRI-detected lesions.

OTSO ARPONEN

DIFFUSION-

WEIGHTED MAGNETIC RESONANCE IMAGING IN BREAST CANCER

Otso Arponen

DIFFUSION-

WEIGHTED MAGNETIC RESONANCE IMAGING IN BREAST CANCER

To be presented by permission of the Faculty of Health Sciences, University of Eastern Finland for public examination in

Kuopio University Hospital Auditorium 1, on January 17th, 2019, at 12 o’clock noon Publications of the University of Eastern Finland

Dissertations in Health Sciences No 547

University of Eastern Finland Kuopio

2020

Series Editors

Professor Tomi Laitinen, M.D., Ph.D.

Institute of Clinical Medicine, Clinical Physiology and Nuclear Medicine Faculty of Health Sciences

Associate Professor (Tenure Track) Tarja Kvist, Ph.D.

Department of Nursing Science Faculty of Health Sciences Professor Kai Kaarniranta, M.D., Ph.D.

Institute of Clinical Medicine, Ophthalmology Faculty of Health Sciences

Associate Professor (Tenure Track) Tarja Malm, Ph.D.

A.I. Virtanen Institute for Molecular Sciences Faculty of Health Sciences

Lecturer Veli-Pekka Ranta, Ph.D.

School of Pharmacy Faculty of Health Sciences

Distributor:

University of Eastern Finland Kuopio Campus Library

P.O.Box 1627 FI-70211 Kuopio, Finland

www.uef.fi/kirjasto

Grano Oy Jyväskylä, 2020

ISBN: 978-952-61-3282-2 (print) ISBN: 978-952-61-3283-9 (PDF)

ISSNL: 1798-5706 ISSN: 1798-5706 ISSN: 1798-5714 (PDF)

Author’s address: Department of Clinical Radiology Kuopio University Hospital University of Eastern Finland KUOPIO

FINLAND

Doctoral programme: Doctoral Programme of Clinical Research Supervisors: Professor Ritva Vanninen, M.D., Ph.D.

Department of Clinical Radiology Kuopio University Hospital University of Eastern Finland KUOPIO

FINLAND

Adjunct Professor Mazen Sudah, M.D., Ph.D.

Department of Clinical Radiology Kuopio University Hospital University of Eastern Finland KUOPIO

FINLAND

Adjunct Professor Juhana Hakumäki, M.D., Ph.D.

Medical Imaging and Physiology Karolinska Institutet

SOLNA SWEDEN

Reviewers: Professor Riitta Parkkola, M.D., Ph.D.

Department of Diagnostic Radiology Turku University Hospital

University of Turku TURKU

FINLAND

Adjunct Professor Peeter Karihtala, M.D., Ph.D.

Department of Oncology and Radiotherapy Oulu University Hospital

University of Oulu OULU

FINLAND

Opponent: Adjunct Professor Tuomo Meretoja, M.D., Ph.D.

Breast Surgery Unit

Helsinki University Hospital Comprehensive Cancer Center University of Helsinki

HELSINKI FINLAND

Arponen, Otso

Diffusion-weighted magnetic resonance imaging in breast cancer Kuopio: University of Eastern Finland

Publications of the University of Eastern Finland Dissertations in Health Sciences 547. 2020, 142 p.

ISBN: 978-952-61-3282-2 (print) ISSNL: 1798-5706

ISSN: 1798-5706

ISBN: 978-952-61-3283-9 (PDF) ISSN: 1798-5714 (PDF)

ABSTRACT

Breast cancer is the most frequently diagnosed cancer in women all around the world and a major cause of cancer-related deaths. Tumor size, axillary lymph node status and certain biological prognostic factors have a significant role in the determination of both disease prognosis and appropriate management in local and locally advanced breast cancers. Radical surgical treatment with a curative intent is the cornerstone of the management of local and locally advanced breast cancers. Neoadjuvant or adjuvant therapies, including chemotherapy, radiation therapy, endocrine therapy and targeted therapies are often prescribed to reduce local and / or distant recurrence and to improve survival.

Magnetic resonance imaging (MRI) is increasingly being utilized in breast imaging for various indications, e.g. preoperative local disease staging and an evaluation of the neoadjuvant therapy response. The diagnostics of breast lesions on MRI is based on lesion morphology as well as contrast enhancement kinetics.

Emerging MRI parameters, such as diffusion-weighted imaging (DWI), are intended to reduce false-positive findings and characterize breast lesions. DWI is based on the movement of water molecules in tissues, which inversely correlates to tissue cellularity and membrane integrity. Restricted diffusion, reflected as low apparent diffusion coefficient (ADC) values, is a recognized biomarker for malignancies.

However, since DWI is a relatively new and still experimental biomarker, it will require a more detailed validation before being introduced into clinical practice.

This doctoral dissertation focuses on DWI, ADC values and their implementation into clinical practice. The results of our first study suggest that ADC values should be obtained using small measurement areas (i.e. regions of interests (ROIs)) as these measurements are more accurate and are more frequently associated with prognostic factors in comparison to those acquired using large whole-lesion covering ROIs. The second study demonstrates that the specificity of MRI in MRI- detected suspicious lesions could be improved by incorporating T2 signal intensity and the DWI assessment into the morphology- and enhancement kinetics-based BI- RADS® assessment. Finally, our third study reveals that when the ADC maps are generated using certain b value combinations, the ADC values were lower in post-

10

contrast vs. pre-contrast ADC maps, emphasizing the importance of the stratification of sequence parameters and acquisition order.

These results may be beneficial in the implementation of DWI in clinical practice. DWI has profound implications on diagnostic breast imaging as a problem- solving tool and can reduce the number of false positive findings.

National Library of Medicine Classification: WN 185, WP 870

Medical Subject Headings: Breast/pathology; Breast Neoplasms/diagnostic imaging;

Diffusion Magnetic Resonance Imaging; Magnetic Resonance Imaging; Female; Prognosis;

Sensitivity and Specificity

Arponen, Otso

Rintasyövän diffuusiokuvantaminen Kuopio: Itä-Suomen yliopisto

Publications of the University of Eastern Finland Dissertations in Health Sciences 547. 2020, 142 s.

ISBN: 978-952-61-3282-2 (nid.) ISSNL: 1798-5706

ISSN: 1798-5706

ISBN: 978-952-61-3283-9 (PDF) ISSN: 1798-5714 (PDF)

TIIVISTELMÄ

Rintasyöpä on maailmanlaajuisesti yleisin naisten syöpä ja merkittävä syöpä- kuolleisuuden aiheuttaja. Paikallisessa ja paikallisesti levinneessä rintasyövässä rintasyövän koko, kainalolevinneisyys sekä ennusteelliset tekijät vaikuttavat sekä rintasyövän ennusteeseen että hoitomenetelmien valintaan. Riittävin leikkaus- marginaalein toteutettu rintasyöpäkirurgia on rintasyövän hoidon kulmakivi.

Liitännäishoitoja, joihin kuuluvat solunsalpaajahoidot, sädehoito, endokriiniset hoidot ja kohdennetut hoidot, annetaan monesti paikallisen ja / tai muualla tapahtuvan rintasyövän uusiutumisen vähentämiseksi sekä ennusteen parantamiseksi.

Magneettikuvantamista (MK) käytetään enenevästi rintojen kuvantamisessa useissa indikaatioissa, kuten rintasyövän levinneisyyden arvioimisessa ja ennen leikkausta annettujen liitännäishoitojen vasteen seurannassa. Rintamuutosten MK- diagnostiikka perustuu muutosten muodon (morfologia) ja varjoainetehostumis- käyttäytymisen arvioimiseen. Uusilla MK-parametreillä, kuten diffuusio- kuvantamisella, pyritään väärien positiivisten löydösten vähentämiseen ja rinta- muutosten parempaan arvioimiseen. Diffuusiokuvantaminen perustuu kudosten vesimolekyylien liikkeeseen, joka on kääntäen verrannollinen kudosten solukkuuteen ja solukalvojen eheyteen. Rajoittunut diffuusio, joka voidaan havaita matalina ADC-arvoina (apparent diffusion coeffient -arvot), on pahanlaatuisuutta kuvaava biomarkkeri. Diffuusiokuvantaminen on vakiintumaton, uusi magneettikuvausmenetelmä rintojen kuvantamisessa. Sen ottaminen kliiniseen käyttöön vaatii lisää tutkimusta sen käyttökelpoisuudesta.

Tämä väitöskirja käsittelee diffuusiokuvantamista ja ADC-arvoja sekä niiden vakiinnuttamista kliiniseen käyttöön. Ensimmäisen osatyömme tulokset viittaavat siihen, että ADC-arvojen mittaaminen käyttäen pieniä mittausalueita (kiinnostuksen alueet, ROI:t), on tarkempaa ja osoittaa tilastollisen yhteyden ennusteellisiin tekijöihin useammin kuin suurilla, koko muutoksen kattavilla ROI:lla tehdyt mittaukset. Toinen osatyömme osoittaa, että arvioitaessa MK:lla havaittuja pahanlaatuisuuden suhteen epäilyttäviä muutoksia MK:n spesifisyys paranee, kun morfologia- ja varjoainetehostumiskäyttäytymispohjaisen BI-RADS®-arvioinnin

12

lisäksi arvioidaan myös T2-signaali-intensiteettiä ja ADC-arvoja. Osoitamme kolmannessa osatyössämme, että tiettyjä b-arvoja käytettäessä varjoaineen annon jälkeiset ADC-arvot ovat matalampia kuin arvot, jotka on mitattu ennen varjoaineen antoa laadituista ADC-kartoista: kuvantamisparametrien ja sekvenssijärjestyksen vakiinnuttaminen on tärkeää.

Tuloksemme voivat auttaa diffuusiokuvantamisen kliinisessä käyttöönottamisessa. Diffuusiokuvantamisella on merkittävä rooli rinta- diagnostiikassa ongelmanratkaisutyökaluna ja se voi vähentää väärien positiivisten löydösten määrää.

Yleinen suomalainen ontologia: diffuusiokuvaus, magneettikuvaus, naiset, rintasyöpä

ACKNOWLEDGEMENTS

I want to express my greatest gratitude to my supervisors Professor Ritva Vanninen and Adjunct Professor Mazen Sudah for their guidance and encouragement. Their passion for research has been an inspiration to me. I thank Adjunct Professor Juhana Hakumäki for his support. I am grateful to Adjunct Professor Anna Sutela for her time and help. I wish to thank Mikko Taina M.D., Ph.D. for the insights he provided and the push he gave me to reach my research goals. Amro Masarwah M.D., Ph.D.

has been a fellow-researcher and a friend to look up to all of the time – thank you. I highly appreciate Suvi Rautiainen M.D., Ph.D., Mervi Könönen Ph.D. and Adjunct Professor Timo Liimatainen and want to thank them for their help. In addition, I am deeply thankful to my other co-authors and the reviewers of this thesis for their efforts.

I am grateful to my parents, siblings and other relatives for their love and encouragement – I could not love you more. Furthermore, I want to thank, from the bottom of my heart, all of my friends from the elementary and high schools, medical school, our research group and work places. You mean the world to me.

I also wish to extend my appreciation to people diagnosed with cancer. Having the privilege of meeting you has profoundly changed my outlook on life.

Tampere, Finland, July 2019

Otso Arponen

14

LIST OF ORIGINAL PUBLICATIONS

This dissertation is based on the following original publications:

I Arponen O, Sudah M, Masarwah A, Taina M, Rautiainen S, Könönen M, Sironen R, Kosma VM, Sutela A, Hakumäki J and Vanninen R. Diffusion-weighted imaging in 3.0 tesla breast MRI: diagnostic performance and tumor characterization using small subregions vs. whole tumor regions of interest. Plos One 10(10): e0138702, 2015.

II Arponen O, Masarwah A, Sutela A, Taina M, Könönen M, Sironen R, Hakumäki J, Vanninen R and Sudah M. Incidentally detected enhancing lesions found in breast MRI: analysis of apparent diffusion coefficient and T2 signal intensity significantly improves specificity. European Radiology 26(12): 4361-4370, 2016.

III Arponen O, Sudah M, Sutela A, Taina M, Masarwah A, Liimatainen T and Vanninen R. Gadoterate meglumine decreases ADC values of breast lesions depending on b value combination. Scientific Reports 8(1): 87, 2018.

Publications I and III were licensed, published and adapted under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA. Publication II was adapted with the permission of Elsevier.

16

CONTENTS

ABSTRACT ... 9

TIIVISTELMÄ ... 11

ACKNOWLEDGEMENTS ... 13

LIST OF ORIGINAL PUBLICATIONS ... 15

ABBREVIATIONS ... 23

1 INTRODUCTION ... 27

2 REVIEW OF THE LITERATURE ... 29

2.1 BREAST DISEASES ... 29

2.1.1 Subtypes and prevalence of benign breast lesions ... 29

2.1.2 Malignant breast diseases ... 29

2.1.2.1 Breast cancer subtypes ... 29

2.1.2.2 Incidence, survival and mortality rates ... 30

2.1.3 Female breast cancer risk factors ... 31

2.1.3.1 Non-modifiable breast cancer risk factors ... 31

2.1.3.2 Lifestyle-related breast cancer risk factors ... 31

2.1.4 Breast cancer prognosis ... 31

2.1.4.1 Anatomical staging of breast cancer ... 32

2.1.4.2 Prognostic biological factors in early breast cancer ... 32

2.2 BREAST CANCER TREATMENTS ... 33

2.2.1 Surgical management of the tumor and the axilla... 34

2.2.2 Radiation therapy of the breast and the lymph node regions ... 34

2.2.3 Systemic therapies ... 35

2.3 BREAST DIAGNOSTICS ... 36

18

2.3.1 Diagnostic breast imaging ... 38

2.3.1.1 Mammography ... 38

2.3.1.2 Tomosynthesis ... 39

2.3.1.3 Ultrasonography ... 39

2.3.1.4 Magnetic resonance imaging ... 40

2.3.1.4.1 Physical basis ... 40

2.3.1.4.2 Clinical imaging technique and sequences ... 40

2.3.1.4.3 Imaging procedure ... 40

2.3.1.4.4 Indications and contraindications for breast MRI ... 41

2.3.1.4.5 Rationale for preoperative MRI staging of the breast ... 41

2.3.1.4.6 The assessment of breast MRI ... 44

2.3.1.4.7 Management of the breast lesions ... 45

2.3.2 Breast cancer screening ... 46

2.3.2.1 Women at average and moderate risk ... 46

2.3.2.2 Women at high risk ... 46

2.3.3 Neoadjuvant chemotherapy response evaluation ... 47

2.4 MOLECULAR AND FUNCTIONAL MR IMAGING ... 47

2.4.1 Emerging MR imaging parameters ... 48

2.4.2 Diffusion-weighted imaging (DWI) ... 51

2.4.2.1 Biophysics of water diffusion ... 51

2.4.2.2 Considerations on b values and apparent diffusion coefficient (ADC) values ... 51

2.4.2.3 Applications of DWI in oncological imaging ... 53

2.4.2.4 DWI in breast imaging ... 53

2.4.2.4.1 Lesion diagnostics ... 53

2.4.2.4.2 Lesion characterization ... 57

2.4.2.4.3 Axillary lymph node staging ... 57

2.4.2.4.4 Neoadjuvant chemotherapy assessment ... 57

2.5 COMPUTER-AIDED DIAGNOSTICS AND ARTIFICIAL INTELLIGENCE ... 58

3 AIMS OF THE STUDY ... 59

4 DIFFUSION-WEIGHTED IMAGING IN 3.0 TESLA BREAST MRI:

DIAGNOSTIC PERFORMANCE AND TUMOR

CHARACTERIZATION USING SMALL SUBREGIONS VS.

WHOLE TUMOR REGIONS OF INTEREST ... 61

4.1 ABSTRACT ... 61

4.1.1 Introduction ... 61

4.1.2 Methods ... 61

4.1.3 Results ... 61

4.1.4 Conclusion ... 62

4.2 INTRODUCTION ... 62

4.3 MATERIALS AND METHODS ... 63

4.3.1 Patients and study design ... 63

4.3.2 Breast MRI protocol ... 66

4.3.3 DW image analysis ... 66

4.3.4 Histopathological analysis ... 67

4.3.5 Statistical analysis ... 68

4.4 RESULTS ... 69

4.4.1 Diagnostic performance of ADC values ... 69

4.4.2 Comparison of WL-ROIs with S-ROIs ... 70

4.4.3 Association between ADC values and prognostic factors ... 70

4.5 DISCUSSION ... 71

5 INCIDENTALLY DETECTED ENHANCING LESIONS FOUND IN BREAST MRI: ANALYSIS OF APPARENT DIFFUSION COEFFICIENT AND T2 SIGNAL INTENSITY SIGNIFICANTLY IMPROVES SPECIFICITY ... 79

5.1 ABSTRACT ... 79

5.1.1 Objectives ... 79

5.1.2 Methods ... 79

5.1.3 Results ... 79

5.1.4 Conclusions ... 79

5.2 KEY POINTS ... 80

5.3 INTRODUCTION ... 80

5.4 MATERIALS AND METHODS ... 81

5.4.1 Study design and patients ... 81

5.4.2 Breast MRI protocol ... 81

5.4.3 MR image analysis ... 82

20

5.4.4 Histopathological analysis ... 83 5.4.5 Statistical analysis ... 83 5.5 RESULTS ... 84 5.5.1 Diagnostic performance of the methods ... 86 5.6 DISCUSSION ... 91 6 GADOTERATE MEGLUMINE DECREASES ADC VALUES OF

BREAST LESIONS DEPENDING ON THE B VALUE

COMBINATION ... 95 6.1 ABSTRACT ... 95 6.2 INTRODUCTION ... 95 6.3 MATERIALS AND METHODS ... 96 6.3.1 Study design and patients ... 96 6.3.2 Breast MRI protocol ... 97 6.3.3 Generation of pre- and post-contrast ADC maps with

different b value combinations and the ROI placement .. 98 6.3.4 Histopathological evaluation of the lesions ... 99 6.3.5 Statistical analysis ... 99 6.4 RESULTS ... 99 6.5 DISCUSSION ... 101

7 DISCUSSION ... 111 7.1 STUDY I ... 111 7.2 STUDY II ... 112 7.3 STUDY III ... 112 7.4 GENERAL DISCUSSION, RELEVANCE, FUTURE

PERSPECTIVES ... 113

8 CONCLUSIONS ... 115

REFERENCES ... 117

22

ABBREVIATIONS

ADC apparent diffusion coefficient

AJCC American Joint Committee on Cancer

ALND axillary lymph node dissection

B0 external magnetization field BCS breast-conserving surgery BI-RADS Breast Imaging Reporting

and Data System

BPE background parenchymal enhancement

BRCA1 breast cancer type 1 susceptibility protein BRCA2 breast cancer type 2

susceptibility protein CBE clinical breast examination ChT chemotherapy CIS carcinoma in situ CIS ductal carcinoma in situ CNR contrast-to-noise ratio DCE dynamic contrast-enhanced

DWI diffusion-weighted imaging

EES extravascular extracellular space

ER estrogen receptor ESMO European Society for

Medical Oncology ET endocrine therapy

EUSOBI European Society of Breast Imaging

EUSOMA European Society of Breast Cancer Specialists

G-CA gadolinium-based contrast agent

HER2 human epidermal growth factor receptor 2

LABC locally advanced breast cancer

NACT neoadjuvant chemotherapy NCCN National Comprehensive

Cancer Network NMLE non-mass lesion

NMV net magnetization vector ML mass lesion

MR magnetic resonance

MRI magnetic resonance imaging

24

PALB2 partner and localizer of BRCA2

PR progesterone receptor ROI region of interest RT radiation therapy Tis in situ tumor

TNM Classification of Malignant Tumors (tumor, lymph nodes, metastasis) UICC Union for International

Cancer Control US ultrasonography

26

1 INTRODUCTION

Breast cancer is the most frequently diagnosed new cancer in women all around the world and a major cause of cancer-related deaths. Magnetic resonance imaging (MRI) is increasingly being utilized in breast imaging for a variety of indications, e.g.

disease staging and neoadjuvant therapy response evaluation.

The evaluation of breast lesions by MRI is based on both lesion morphology and contrast enhancement kinetics. Novel MRI parameters are emerging in attempts not only to reduce false-positive findings but also to preoperatively characterize tumors. Diffusion-weighted imaging (DWI) is based on the movement of water molecules in tissues; the diffusion of water inversely correlates with tissue cellularity and membrane integrity. Restricted diffusion, which is reflected in low apparent diffusion coefficient (ADC) values, is a recognized biomarker for malignancy.

DWI could also have profound benefits on diagnostic imaging of the breast as a problem-solving tool. This doctoral dissertation focuses on DWI and ADC values and their implementation into clinical practice.

28

2 REVIEW OF THE LITERATURE

2.1 BREAST DISEASES

2.1.1 Subtypes and prevalence of benign breast lesion

Breast diseases are categorized into benign and malignant tumors. Benign breast lesions are derived from different breast components (epithelial, stromal, adipocyteal, or vascular) ¹. Benign breast lesions are common. The actual incidence of benign breast conditions remains undetermined. The majority of symptomatic patients have benign tumors ²; approximately 75 % of breast lesions biopsied are benign ³. Nonetheless, it has been proposed that only 10-20 % of all benign breast conditions are verified histopathologically ^1,4.

The fibrocystic changes occurring in the breast are of interest, because they are the most frequently biopsy-confirmed benign breast conditions ². Three subgroups of fibrocystic changes have been identified; non-proliferative lesions, proliferative lesions without atypia, and lesions with atypical hyperplasia ^1,2. Lesions with a proliferative tendency, especially atypical hyperplasia, have an increased relative risk for breast cancer development ^5,6. In a large study sample of over 10 000 breast biopsies, 69 % of samples represented non-proliferative changes and only 3.6 % were atypical hyperplasia ⁷. A review of post-mortem studies indicated that nearly every other woman may develop a fibrocystic breast change ^1,4.

2.1.2 Malignant breast diseases 2.1.2.1 Breast cancer subtypes

Malignant breast tumors are categorized as either non-invasive (in situ) or invasive carcinomas ^8,9. Approximately 20 % of breast carcinomas are non-invasive ¹⁰. Non- invasive and invasive carcinomas are divided into several subtypes based on their histopathology ¹¹. The share of these subtypes according to the National Cancer Institute’s Survey, Epidemiology, and End Results Program is presented in Table 1.

2.1.2.2 Incidence, survival and mortality rates

Breast cancer is the most common cancer in women (1.67 million new cancers worldwide in 2012) and accounts for 25 % of all female cancers ¹³. The incidence of breast cancer varies by region, with the highest rates reported in Western Europe and the United States with the lowest rates in Africa and Asia ¹⁴. The global incidence of breast cancer is increasing ¹⁵.

Five-year survival rates of breast cancer vary regionally and are 89.7 %, 81.8 % and 90.6 % for the United States, the European Union and Finland, respectively ^11,16,17. In the developed regions, breast cancer is the the most frequent cause of cancer deaths among women ¹⁸, even though the mortality has declined due to the implementation of screening and adjuvant therapies ^19–21. Estimated breast cancer incidence and mortality rates per 100 000 women at a global level, in the United States, the European Union and Finland are presented in Table 2.

CIS = carcinoma in situ. ¹ Lobular CIS is considered a benign tumor in the American Joint Committee for Cancer (AJCC) staging. Other cancer types occurring in breasts, i.e. sarcomas, phyllodes tumors and breast lymphomas, are not staged as breast cancers in UICC or AJCC staging manuals. ^8,9,12

Table 1. The percentage share of invasive and non-invasive breast cancer subtypes according to their histology in 2010-2014 in the United States ¹¹.

Invasive Non-Invasive

Cancer subtype Share (%) Cancer subtype Share (%)

Ductal carcinoma 73.5 Ductal CIS 85.0

Lobular carcinoma 9.3 Lobular CIS¹ 11.6

Mixed ductal and

lobular carcinoma 9.1 Mixed ductal and

lobular CIS 2.7

Mucinous carcinoma 1.9 Other CIS 0.7

Papillary carcinoma 0.6 Tubular carcinoma 0.5

Other 5.1

Table 2. Estimated age-standardized incidence and mortality rates of breast cancer in 2012 ¹³.

Incidence (per 100 000 females) /

Total number Mortality (per 100 000 females) / Total number

World 43.1 / 1 700 000 12.9 / 520 000

The United States 92.9 / 230 000 14.9 / 44 000

The European Union 80.3 / 360 000 15.5 / 92 000

Finland 89.4 / 4500¹ 13.6 / 860

1 According to the Finnish Cancer Registry, there were 5161 female breast cancers diagnosed in 2015 ¹⁷.

30

2.1.3 Female breast cancer risk factors

There are non-modifiable and modifiable lifestyle-related risks factors for breast cancer ²². Non-modifiable factors are characteristics that predispose a person to an increased risk of breast cancer and cannot be altered ²². Lifestyle-related factors are less significant for breast cancer development than non-modifiable risk factors ¹⁰.

2.1.3.1 Non-modifiable breast cancer risk factors

Breast cancer is a hundred times more likely to affect women than men ^14,18. The risk of being diagnosed with a breast cancer increases with age ²³. Having a personal history of a non-invasive or an early invasive breast cancer or a high risk lesion predisposes a woman to breast cancer ^6,10. A history of other cancers (endometrium, ovary and colon) has been associated with an increased breast cancer risk ²⁴.

A family history of breast and / or ovarian cancers increases an individual’s risk of breast cancer; the number of first-degree relatives and the age of the affected relative(s) at the time of breast cancer diagnosis also plays a role ^10,25. It has been estimated that 10-30 % of breast cancer cases are attributed to hereditary factors, but only 5-10 % of breast cancer cases have a strong inherited background ²⁶. Women with a rare high-prevalence genetic predisposition are more likely to be diagnosed with breast cancer. It has been estimated that 35-65 % of women having mutations in BRCA1, BRCA2 or PALB2 cancer susceptibility genes will develop a breast cancer by the age of 70 years ^24,27–30.

Higher mammographic breast tissue density, i.e. the relative amount of breast’s glandular and connective tissues to the portion of fatty tissue, has been shown to increase the risk of breast cancer. An increased mammographic breast density is an established risk factor for the development of female breast cancer, only less significant than the risks posed by the patient’s age, existing breast cancer and BRCA mutations ^31,32. Other known risk factors include high systemic estrogen levels, early menarche and late menopause ³³.

2.1.3.2 Lifestyle-related breast cancer risk factors

The higher breast cancer incidence in high-income countries is partly attributable to a higher prevalence of lifestyle-related breast cancer risk factors ³⁴. Lifestyle-related risk factors include adult weight gain, excess body weight, use of menopausal hormone therapy or oral contraceptives, physical inactivity, alcohol consumption, high-dose radiation to breast and reproductive and hormonal factors ²⁴.

2.1.4 Breast cancer prognosis

Clinical and biological prognostic factors provide information on the overall cancer outcome in untreated individuals ^35,36. Tumor size, axillary nodal status and the presence of metastasis are well-defined anatomical prognostic factors. It has been

32

proposed that biological prognostic factors should also be incorporated into risk models in order to obtain a more refined assessment of prognosis ^8,12,37. Tumor histology, nuclear grade, estrogen / progesterone receptor (ER / PR) statuses, human epidermal growth factor 2 (HER2) overexpression / amplification, vascular invasion and the expression of proliferation markers (e.g. Ki-67) are the most often used prognostic factors in clinical practice for prognosis estimation and treatment tailoring

38–40. In addition, during the past few years, gene expression panels have been

introduced into clinical practise to aid in prognosis determination 12,38,41,42.

2.1.4.1 Anatomical staging of breast cancer

The TNM classification system is the most widely used system for staging breast cancer. Staging of invasive breast cancers using a four-level system (stages I-IV) is based on the largest diameter of the primary tumor (T), regional lymph node involvement (N) and distant metastasis (M); non-invasive (in situ) breast cancers are referred to as Tis ^8,9,43. A larger tumor size and an increased number of involved lymph nodes have been associated with decreased breast cancer survival ⁴⁴. The spread of cancer cells from the primary tumor to distant organs, i.e. distant metastasis (stage IV disease), is a determinant of poor prognosis and a contributor to disease related mortality ^45–47. The relative 5-year survival rates in the US between the years 2007 and 2013 were 100.0 %, 92.0 %, 73.2 % and 26.5 % for stages I to IV, respectively

48.

2.1.4.2 Prognostic biological factors in early breast cancer

Histopathological analysis and grading are routinely performed. Grading is performed using the Nottingham Grading System (NGS) ⁴⁹ that relies on the evaluation of the degree of tubule or gland formation, nuclear pleomorphism and mitotic count ⁵⁰. Well-differentiated low-grade tumors have a better prognosis than poorly differentiated high-grade tumors ⁴⁹. Immunohistochemistry has been used to determine the expression of intracellular estrogen and progesterone receptors; 70 % of invasive cancers are ER positive and 50 % are PR positive ⁵¹. ER positivity has been correlated to prolonged disease-free survival ^51,52; the role of PR as an independent factor remains unclear ^53,54. Overexpression / over-amplification of the transmembrane receptor HER2 detected either by immunohistochemistry or in situ hybridization is present in 15 to 30 % of invasive breast carcinomas ^38,55,56, and it has been associated with higher mortality, greater recurrence rates and more aggressive diseases ⁵⁷. Nonetheless, in the era of HER2-targeted therapies, patients with HER2 positive breast cancers have similar 5-year overall and relapse free survival rates as patients with HER2 negative, hormone receptor positive diseases ⁵⁸. Increases in the levels of a proliferation marker (e.g. nuclear protein Ki-67) expression have been associated with poorer overall disease-free survival ^59,60.

Tumors may be grouped into intrinsic subtypes according to their biological prognostic factors to allow a more detailed prognostic assessment (Table 3). The

luminal A subtype is associated with a highly favorable prognosis; the luminal B subtype has a worse prognosis in comparison to the luminal A subtype ⁶¹. The introduction of the HER2-targeted therapies has markedly improved the survival of patients with HER2 positive tumors while the triple negative tumors have an unfavourable prognosis in comparison to other tumor subtypes ⁶¹.

Recently, gene expression panels have been introduced to complement the pathological assessment in the prognostic evaluation 12,38,41,42. Genomic tests, MammaPrint, Oncotype DX, EndoPredict, Breast Cancer Index and Prosigna Breast Cancer Prognostic Gene Signature Assay with varying eligibility criteria have been validated and / or are under evaluation for the estimation of the breast cancer recurrence risk ⁶². These genomic tests may help in determining breast cancer recurrence and the benefit of adjuvant treatments ^63,64.

2.2 BREAST CANCER TREATMENTS

The tumor stage dictates the goal of the treatments. Curative surgical intervention aims at complete resection of the disease with negative resection margins. Adjuvant therapies (chemotherapy, radiotherapy (RT), endocrine therapy (ET), targeted therapies) in local diseases may be used to reduce local / distant recurrence and to increase disease-free and overall survival ³⁸. Neoadjuvant therapies i.e. treatments preceding surgery, are prescribed with the intent of reducing the extent of the surgically removed area in local diseases or to achieve complete surgical resection by downstaging the disease in locally advanced breast cancers (LABCs) ⁴⁷. Adjuvant treatments may be provided after neoadjuvant therapies ⁴⁷. Although metastatic breast cancer is treatable, it remains virtually incurable; therefore the rationale is to

Table 3. Intrinsic subtypes of breast cancer ^38,41,61.

Luminal

A Luminal B

(HER2 -) Luminal B (HER2 +)

luminal Non- (HER2 +)

Triple- negative and basal⁵ Estrogen

receptor status + + + - -

Progesterone

receptor status +¹ + / low^1,4 High or low - - Human

epidermal growth factor 2 receptor (HER2)

- - + + -

Proliferation

marker (Ki-67) Low² High or low^2,4 High or low² n / a n / a Molecular risk

type

Low-risk molecular signature³

High-risk molecular

signature³ n / a n / a n / a - = negative, + = positive, n / a = not applicable. ¹ Suggested cut-off value is 20 %, ² scores should be interpreted using the local laboratory reference values ³ defined, if available, using gene expression panels, ⁴ either progesterone receptor status low or Ki-67 high, ⁵ approximately 80 % overlap between the triple negative and the basal subtypes.

34

prolong survival and improve the quality of life through systemic and local therapies and supportive care ^45,65.

2.2.1 Surgical management of the tumor and the axilla

In local diseases and LABCs responding to neoadjuvant therapies, breast surgery is the mainstay of treatment. There are two main types of surgery: breast-conserving surgery (BCS) aims at the removal of the tumor with pathologically sufficient cancer- free margins with a cosmetically more comparable outcome whereas mastectomy refers to surgical removal of the whole breast ^38,47. There is a slightly higher risk of local recurrence after BCS with postoperative RT as compared to mastectomy ⁶⁶. Women undergoing mastectomy may consider primary or secondary reconstructive surgery.

The axilla is routinely assessed preoperatively by ultrasound (US). Suspicious US-visible axillary lymph nodes undergo an US-guided core biopsy, or in some centers, fine needle aspiration biopsy sampling. Lymph node metastasis may also be assessed by intraoperative sentinel lymph node biopsy guided by injected blue dye color and / or an isotope substance that allows the surgeon to locate the lymph nodes.

If a lymph node macrometastasis is confirmed in the pathological analysis, axillary lymph node dissection (ALND) may be performed ⁶⁷. The role of ALND has been recently challenged. A prospective randomized trial reported that ALND did not significantly affect either the overall or disease-free survival of patients with SLN- positive, clinically T1-T2 breast cancers treated with BCS, adjuvant systemic therapy and whole-breast RT ⁶⁸. Consequently, the AMAROS trial showed that ALND and axillary RT both provided excellent and comparable axillary control for patients with SNL-positive, T1-T2 primary breast cancers; axillary radiotherapy caused significantly less morbidity ⁶⁹.

Patients with inoperable LABCs or metastatic diseases may undergo surgical treatments. In such cases, decisions are jointly made in a multi-disciplinary meeting

45,47.

2.2.2 Radiation therapy of the breast and the lymph node regions

RT is performed with external beams and is delivered after surgery and after chemotherapy, if the latter has been administered. There are several factors that determine the need for adjuvant RT, e.g. the features of the tumor (risk of recurrence) and the type of surgery ⁷⁰. Women who undergo BCS for DCIS / invasive carcinomas often undergo whole-breast irradiation with a booster dose to the tumor bed if unfavorable risk factors exist for local disease control. RT is performed in traditional fractioning or in a hypofractioned manner; partial breast irradiation is under investigation ^71,72. After mastectomy, RT is given only to women with invasive carcinomas who have a high risk for local recurrence.

Adjuvant RT approximately halves the recurrence rate and reduces the breast cancer death rate by approximately a sixth in women who undergo BCS ⁷¹. Breast-

conserving surgery with postoperative adjuvant whole-breast irradiation has been shown to be comparable in terms of survival as well as with the risk of developing metastasis in comparison to mastectomy ^66,73.

Lymph node areas may be irradiated if the patient has axillary metastasis ³⁸. Delineation of different lymph node regions is out of the scope of this work. As discussed in chapter 2.2.1, the role of axillary lymph node RT may increase in the future. Furthermore, RT may be given with a palliative intent to women with inoperable LABCs or metastatic breast cancers.

2.2.3 Systemic therapies

There are three main types of systemic therapies in breast cancer; ET, chemotherapy (ChT) and targeted therapies ^74,75. They can be used in the neoadjuvant setting preceding surgery either to decrease the extent of surgery of large operable breast cancers or used to treat locally advanced breast cancers in an attempt to downstage the tumor and to reduce the risk of local and distant recurrence ^38,47,76. After radical surgery of the primary tumor, ET, ChT, and targeted therapies may be used in an adjuvant setting to reduce the risk of local and distant recurrence. In metastatic breast cancers, systemic therapies are provided to prolong survival and to maintain local disease control.

The choice of systemic therapy regimes is based on the assumed sensitivity of tumor cells to the treatment. The benefit-harm ratio of adjuvant systemic therapies is assessed based on the patient’s risk of relapse / tumor progression and the patient’s biological age, health status, co-morbidities, preferences and preceding treatments

38,42,45,47. When one considers adjuvant therapy regimes, there are assessment tools

(e.g. the PREDICT tool ⁷⁷, the Nottingham Prognostic Index ⁷⁸) estimating the recurrence / survival on the basis of prognostic biological factors; the PREDICT tool

77 also estimates the incremental value of adjuvant treatments. Additionally, breast cancer genome test panels (reviewed in chapter 2.1.4.2) may be used when considering the recurrence risk and balancing the benefits and harms of adjuvant treatments. In metastatic breast cancers, the choice of systemic therapies is highly tailored.

It is likely that hormone receptor positive (luminal) breast cancers respond to ETs that can prevent estrogen-induced breast cancer growth either by suppressing estrogen levels (aromatase inhibitors) or by exerting antiestrogenic effects (for example, selective estrogen-receptor modulators) ^51,52,79. Chemotherapy i.e. anticancer drug treatments, target dividing cells ⁷⁵. Chemotherapy is especially important in the treatment of non-luminal and triple negative breast cancers as these cancer types are more aggressive and do not respond to ETs ^38,42,80. There are targeted therapies, e.g.

HER2-directed therapies that are used for women with HER2 positive cancers 38,42,47,80. In addition to HER2-directed therapies, numerous other targeted treatments are used or are under investigation ^47,74 in the (neo)adjuvant setting and in the treatment of metastatic breast cancer.

36

2.3 BREAST DIAGNOSTICS

Breast cancer screening is performed on asymptomatic women to detect breast cancer earlier. Currently, mammography is the only method appropriate for mass screening of breast cancer ⁸¹. Tomosynthesis is a promising screening method, although the scientific evidence supporting its use in mass screening is still accumulating ⁸². Magnetic resonance imaging (MRI) screening is indicated for women at a high risk of developing breast cancer.

The main rationale for diagnostic breast imaging is to exclude or confirm malignancy; triple assessment i.e. clinical breast examination (CBE), radiological and pathological evaluation are the cornerstones in the assessment of any suspicious breast symptoms or findings ^83–85. Conventional breast imaging of breast tumors comprises mammography and breast and axillary ultrasonography (US), and in some cases, MRI ⁸³. Illustrative images obtained with commonly applied imaging modalities are shown in Figure 1. Lesions that are considered suspicious of malignancy undergo imaging-guided biopsy and tissue sampling ⁸⁶. Furthermore, neoadjuvant therapies demand imaging-based response assessment methods for post-neoadjuvant therapy staging prior to surgical treatment.

38

Figure 1. An illustrative figure. A female patient in her 50’s was diagnosed with an invasive lobular cancer in her right breast on mammography (A, circle). The contralateral left breast was re-evaluated because of an asymmetry in the posterior third of the breast (B, dotted arrow).

Tomosynthesis revealed a spiculated mass at the same site (C, between dotted arrows) behind a large cyst (C, arrowheads). A second smaller spiculated mass was also seen in the caudal part of the left breast (C, arrow). MRI showed a larger round mass with a non-circumscribed irregular contour (D, dotted arrow) and an anterior lesion with non-mass enhancement (D, arrowheads). Two other masses were visible in the caudal (D, short arrow) and cranial (D, long arrow) areas of the left breast. Restricted diffusion (mean ADC value 0.8 x 10^-3 mm² s^-1) was seen in the large posterior mass (E, short arrow) on the ADC map. The large cyst (E, arrowhead) showed no diffusion restriction (mean ADC value 2.9 x 10^-3 mm² s^-1); the mean ADC value of the non-mass enhancement area around the cyst was 1.6 x 10^-3 mm² s^-1. The cranial mass (F, arrow) showed no diffusion restriction (mean ADC value 1.5 x 10^-3 mm² s^-1).

The caudal mass showed diffusion restriction (mean ADC value 0.9 x 10^-3 mm² s^-1). In ultrasound, the posterior mass was poorly-defineable with heterogeneous echogeneity (H, arrows) and the caudal mass was round and hypoechogenic (I, arrow). The following histopathological diagnoses were confirmed with ultrasound-guided core biopsies: the posterior tumor was an invasive lobular carcinoma (J) and the non-mass enhancing lesion was a lobular carcinoma in situ (K) with some areas of grade 3 ductal carcinoma in situ. The cranial tumor was a fibroadenoma and the caudal lesion was an invasive ductal carcinoma.

2.3.1 Diagnostic breast imaging 2.3.1.1 Mammography

Mammographic imaging is based on low-energy X-ray radiation directed at the breast using a dedicated X-ray unit ⁸⁷; the visibility of breast tumors depends on differences in the attenuation of the X-ray beam in normal breast parenchyma and in breast pathologies ^88,89. Full-field digital mammography requires a lower breast radiation dose ⁸⁸ yet increases contrast and accuracy ⁹⁰ and has largely replaced screen-film mammography ⁹¹. Full-field digital mammography is also superior to screen-film mammography in cancer detection in the screening of women with dense breast tissue or women who are under the age of 50 years ⁹⁰.

Diagnostic mammography is performed on symptomatic patients, or to further evaluate a finding detected on screening mammography or other imaging studies ⁹². When there is a suspicion of a breast malignancy, mammography is the first-line imaging modality for women over the age of 30 to 40 years ^80,83,93 and is performed preoperatively for all women with a breast cancer ^38,80. Bilateral whole- breast craniocaudal, mediolateral oblique and varied angle views and supplemental spot compression and magnification views from the clinically suspicious areas may be obtained during the diagnostic mammographic imaging ⁸³.

The interpretation of the diagnostic mammographies is structured according to the 5^th edition of the Breast Imaging Reporting and Data System (BI-RADS^®) lexicon. Breast composition in regard to the parenchymal distribution is first evaluated using a four-level scale. ⁸⁶ Increased mammographic breast density has been shown to reduce mammography’s sensitivity and specificity in cancer detection

94 and to increase the risk of breast cancer ³². The breast findings are then categorized into masses, calcifications, architectural distortions or asymmetries. Malignancy-

associated features, such as skin thickening or skin or nipple retraction in mammographic images may also necessitate further evaluations. ⁹⁵ Based on the mammographic features, breast findings are classified according to a numeric scale and managed accordingly (see chapter 2.3.1.4.7).

2.3.1.2 Tomosynthesis

Digital breast tomosynthesis is based on low-energy X-ray radiation; pseudo-three- dimensional images are reconstructed from several projection images spaced typically at 1 mm intervals and acquired using an X-ray tube moving in an arc ⁸². The main rationale for the use of tomosynthesis over mammography is the improved detection and characterization of breast lesions, especially in dense breasts where overlapping fibroglandular tissue may hinder lesion detectability ^82,96. Tomosynthesis further improves the detection of architectural distortions and spiculated masses. There is growing evidence to support the use of tomosynthesis in breast diagnostics ⁹⁷. However, the widespread use of digital breast tomosynthesis in both breast cancer screening and diagnostic imaging has been hindered by the concerns over increased radiation dose, increased reading times and cost / benefit justification ^82,98.

2.3.1.3 Ultrasonography

US imaging is based on reflected ultrasound echoes; different breast tissue types have different mechanical and acoustic properties that translate into differences in brightness. As compared to breast parenchyma, breast lesions usually appear hypoechoic. ⁹⁹ Ultrasonography is the first-line imaging modality for women who are younger than 30 to 40 years or who are pregnant and have palpable masses in the breast ^80,100. In the setting of breast diagnostics, other applications of breast ultrasonography include the evaluation of palpable masses and breast-related symptoms, suspected abnormalities detected in other imaging modalities, and breast implant problems ⁹². The second-look ultrasonography, after the detection of primarily MRI-detected lesions, is widely performed as it can be applied to target the biopsy of the suspicious lesion ^101–103. Contrast-enhanced US, i.e. the “third-look US”, may be a feasible method to identify occult MRI-visible lesions that have no correlate on the targeted second-look US ¹⁰⁴.

The BI-RADS^® lexicon describes mass lesions with five morphological US features; shape, orientation, margin, lesion boundary, internal echo pattern, and posterior acoustic feature type. In addition, US can be used to evaluate calcifications and malignancy-associated features (cysts, architectural distortion, duct and skin changes, edema, vascularity, elasticity and lymph nodes) ¹⁰⁵. Based on the sonographic features, breast findings are classified on a numeric scale and managed accordingly (see chapter 2.3.1.4.7).

40

2.3.1.4 Magnetic resonance imaging 2.3.1.4.1 Physical basis

In simple terms, magnetic resonance imaging is based on the observation of an intrinsic atomic property called spin; the spin system of an atom’s nucleus in an external magnetic is excited by a radiofrequency (RF) pulse. Clinical MRI most commonly focuses on hydrogen atoms since their spin is ½ and they are abundant in the human body ¹⁰⁶. A spin ½ particle can be thought of as a bar magnet which is spinning around its axis in a disoriented manner until an external magnetization field (B⁰) is applied, leading to the alignment of the spin ½ particles according to the external field (either in parallel to or opposing the field direction) ¹⁰⁷. The net cumulative effect of magnetic moments, termed the net magnetization vector (NMV), is flipped to rotate at an angle (a) due to the application of a RF pulse. There are two magnetization vector components in the rotating NMV, the longitudinal and the transverse magnetization components ¹⁰⁷. Simply put, the transverse magnetization is responsible for the induction of the current that translates into the MR signal; as the external RF pulse is discontinued, the NMV starts to realign with the B0 field (in a process called T1 recovery) and the transverse magnetization decreases through additional mechanisms termed as T2* and T2 decays ¹⁰⁷. Differences in T1 and T2 processes and the number of hydrogen nuclei (i.e. proton density) contribute to differences in tissue contrasts ^106–108 and this permits the diagnostic work-up of MR images.

2.3.1.4.2 Clinical imaging technique and sequences

The European Society of Breast Cancer Specialists (EUSOMA) recommends the use of MRI scanners with a magnetic field strength ≥ 1.0 tesla with bilateral dedicated multichannel coils ¹⁰⁹. In clinical practice, the field strength is often either 1.5 or 3 tesla

110. EUSOMA considers a minimal bilateral protocol to contain at least one unenhanced high-contrast sequence (such as T2 weighted fast / turbo spin-echo or short tau inversion recovery (STIR) sequence), and a bilateral T1 weighted dynamic sequence with a contrast agent. Temporal subtracted images are recommended ¹⁰⁹.

2.4.1.4.3 Imaging procedure

If clinically feasible, pre- / perimenopausal women should undergo MRI during the second week of their menstrual cycle for optimal results in most clinical indications;

postmenopausal women may undergo MRI at any time ^110,111. Women undergoing MRI lie in a prone position on the table with their breasts placed in dedicated coils

110; supine positioning may be beneficial for preoperative localization of the MRI- detected lesions ¹¹². When properly positioned, the table on which the patient is lying moves into the magnet. While undergoing imaging, the contrast agent is injected into the cubital vein. A routine imaging protocol lasts 15 to 30 minutes. ¹¹⁰

2.3.1.4.4 Indications and contraindications for breast MRI

Although bilateral mammography and ultrasonography are the traditional primary preoperative imaging modalities ^38,113,114, MRI is increasingly being utilized in preoperative breast imaging ¹¹⁵. Preoperative MRI imaging is not recommended for all patients diagnosed with breast cancer, but should be considered in specific situations as a part of the diagnostic work-up ^38,42.

Indications for diagnostic MRI vary in the different recommendations ^109–111. The European Society of Breast Imaging (EUSOBI) and The American College of Radiology (ACR) agree that diagnostic breast MRI is indicated for preoperative staging as well as for the assessment of the neoadjuvant chemotherapy response, for the evaluation of women with breast implants and when mammography- and ultrasonography-negative occult tumors are being evaluated when there is a suspicion of a primary breast cancer based on (axillary) metastasis ^110,111. MRI may also be performed when there is a suspicion of local recurrence and for problem solving when prior imaging examinations have been inconclusive ^110,111 especially when needle biopsy cannot be performed ¹¹⁰. According to the EUSOMA guidelines, MRI should also be considered whenever there is an existing discrepancy in tumor size > 1 cm between mammography and US with an expected impact on treatment decision in patients < 60 years of age, in patients with lobular carcinoma and in selected cases where partial breast irradiation treatment is considered ¹⁰⁹.

Absolute contraindications for MRI are non-MRI compatible intracranial aneurysm clips and iron splinters in the eyes. Some implanted (electronic) devices (e.g. non-MR-compatible pacemakers, stents, screws) and iron-pigment-containing tattoos are additional contraindications ¹¹⁰. Allergic reactions to contrast media and poor kidney function (glomerular filtration rate < 30 ml / min × 1.73 m²) should be taken into consideration ¹¹⁰. Pregnancy is a relative contraindication to contrast- enhanced MRI ^110,116. Severe claustrophobia may prevent the patient from undergoing MRI ^110,116.

2.3.1.4.5 Rationale for preoperative MRI staging of the breast

MRI was shown to be superior to CBE, mammography and ultrasonography in the detection of additional disease foci (multifocal and / or multicentric diseases in ipsilateral and contralateral breasts) and disease extent evaluation 109,117–119. Furthermore, it has been reported that additional incidental MRI-detected lesions are found in 5.3 % to 48 % of studies ^84,120–127. It has been hypothesized that accurate preoperative imaging could reduce re-excision, mastectomy and recurrence rates while improving the long-term outcomes.

In a study of 111 prospectively enrolled patients with a known or suspected invasive cancer (177 malignant foci, 258 lesions) who underwent bilateral CBE, mammography, ultrasonography and MRI for the assessment of the local extent, the sensitivities and accuracies were 50.3 % / 63.6 %, 67.8 % / 70.2 %, 83.0 % / 67.8 % and 94.4 % / 72.9 %, respectively ¹¹⁷. Furthermore, studies on high risk women reported

42

superior sensitivity and specificity (90.7-92.6 % / 97.2-98.4 %) for MRI over mammography (32.6-33.3 % / 96.8-99.1 %) and ultrasonography (37.0-39.5 % / 90.5- 98 %) or the combination of mammography and ultrasonography (48.1-48.8 % / 89.0- 98.3 %) ^128,129.

There are three ^84,122,123 randomized prospective controlled multicenter trials which investigated the preoperative utility of MRI (summarized in Table 4). The clinical primary endpoints of all of the three studies were changes in the numbers of surgical procedures (re-operations rates). In contrast to the two larger studies, paradoxically the MONET study detected higher re-operation rates in the MRI group. The COMICE trial revealed similar re-operation rates in the MRI and non- MRI groups. Women who had undergone MRI also underwent re-excisions more seldom in the POMB. The COMICE and the MONET trials observed no higher overall conversion rate to mastectomy from the BCS; this kind of trend was seen in the POMB trial, but there was a lower initially planned mastectomy rate in the MRI-group. The POMB trial showed that MRI led to a change in the treatment plan in 18 % of patients and reduced the number of re-operations. The POMB trial did not follow the patients and the COMICE and MONET trials had a follow-up time of one year. Disease-free survival and disease-specific survival benefits of breast MRI were not studied.

MRI has been associated with an increased number of incremental investigations, mastectomies, contralateral prophylactic mastectomies and longer wait times to surgery in a large population-based retrospective study ¹¹⁵. The level of evidence and existing results regarding the possible disadvantages underline the importance of rigorous patient selection and adherence to the indications of preoperative MRI.

43

Table 4. Summary of the three prospective controlled randomized studies on preoperative breast magnetic resonance imaging (MRI) of women with newly diagnosed breast cancer. Trial (Year) Primary endpoint(s) N in MR / non-MR groupsPatient characteristicsResults COMICE (2010) 84Re-operation rates 816 / 807

Patients were newly diagnosed with malignant lesions that were scheduled for wide local excision operations.

There was no difference in the re-operation rates. MONET (2011) 122Re-operation rates 207 / 2111All patients had non- palpable lesions (the BI- RADS® categories from 3 to 5) detected on mammo- graphy or ultrasound.

The re-operation rate was higher in the MRI group vs. the non-MRI-group. POMB (2014) 123Primary surgical management, re- operation rates, NACT rates 220 / 220Patients (aged < 56 years) were newly diagnosed with malignant lesions.

The re-operation rate was lower in the MRI group vs. the non-MRI group (5 % vs. 15 %). MRI-detected findings in the ipsi- or the contralateral breast and in the axilla changed the surgical management in 18 % of patients. An increased number of mastectomies was performed in the MRI-group. NACT rates were similar in both groups.

N 1= number of patients, MRI = magnetic resonance imaging, MR = magnetic resonance,NACT = neoadjuvant chemotherapy. In the study sample, there were 74 and 75 women with malignant (non-invasive / invasive) lesions in the MR-imaged and the non-MR groups, respectively.