Multiparametric magnetic resonance imaging in epithelial ovarian cancer

(1)

DISSERTATIONS | AUNI LINDGREN | MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING IN... | No 532

uef.fi

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND Dissertations in Health Sciences

ISBN 978-952-61-3197-9 ISSN 1798-5706

Dissertations in Health Sciences

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

AUNI LINDGREN

MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING IN EPITHELIAL OVARIAN CANCER

Targeted therapies are under development for epithelial ovarian cancer (EOC). The new treatment modalities demand that also

imaging methods need to keep pace with these rapid developments. The goal of this thesis was to improve our understanding of the diffusion weighted and dynamic contrast- enhanced imaging in EOC. The results indicate

that multiparametric magnetic resonance imaging variables are sufficient for use as surrogate markers permitting a more detailed

characterization of EOC.

AUNI LINDGREN

(2)

(3)

MULTIPARAMETRIC MAGNETIC RESONANCE

IMAGING IN EPITHELIAL OVARIAN CANCER

(4)

(5)

Auni Lindgren

MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING IN EPITHELIAL OVARIAN 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, Kuopio on Friday 25^thOctober 2019, at 12 o’clock

Publications of the University of Eastern Finland Dissertations in Health Sciences

No 532

Departments of Obstetrics and Gynecology and Clinical Radiology, Institute of Clinical Medicine, Faculty of Health Sciences,

University of Eastern Finland Kuopio

2019

(6)

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 Jyväkylä, 2019

ISBN: 978-952-61-3197-9 (print) ISBN: 978-952-61-3198-6 (PDF)

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

(7)

Author’s address: Department of Obstetrics and Gynecology Kuopio University Hospital

Institute of Clinical medicine

School of Medicine, Faculty of Health Sciences University of Eastern Finland

KUOPIO FINLAND

Doctoral programme: Clinical Medicine Doctoral programme Supervisors: Adjunct Professor Maarit Anttila, M.D., Ph.D.

Department of Obstetrics and Gynecology Kuopio University Hospital

Institute of Clinical Medicine

KUOPIO FINLAND

Professor Ritva Vanninen, M.D., Ph.D.

Department of Clinical Radiology Kuopio University Hospital Institute of Clinical Medicine

KUOPIO FINLAND

Hanna Sallinen, M.D., Ph.D.

Department of Obstetrics and Gynecology Kuopio University Hospital

KUOPIO FINLAND

Adjunct Professor Kirsi Hämäläinen, M.D., Ph.D.

Department of Pathology and Forensic Medicine Kuopio University Hospital

KUOPIO FINLAND

(8)

Reviewers: Adjunct Professor Annika Auranen, M.D., Ph.D.

Department of Obstetrics and Gynecology Tampere University Hospital

TAMPERE FINLAND

Professor Hannu Aronen, M.D., Ph.D., M.Sc (Eng) Department of Clinical Radiology

University of Turku TURKU

FINLAND

Opponent: Adjunct Professor Irina Rinta-Kiikka, M.D., Ph.D.

Department of Clinical Radiology University of Tampere

TAMPERE FINLAND

(9)

To Antti and Vilho

(10)

(11)

Lindgren Auni

Multiparametric magnetic resonance imaging in epithelial ovarian cancer Kuopio: University of Eastern Finland

Publications of the University of Eastern Finland Dissertations in Health Sciences 532. 2019, 134 p.

ISBN: 978-952-61-3197-9 (print) ISSNL: 1798-5706

ISSN: 1798-5706

ISBN: 978-952-61-3198-6 (PDF) ISSN: 1798-5714 (PDF)

ABSTRACT

Epithelial ovarian cancer (EOC) is the leading cause of death from gynecological malignancies. Often the disease has already spread widely at the time of diagnosis, because it is usually symptomless in its early stage. There is no screening method capable of detecting ovarian cancer.

Patient survival of ovarian cancer is rather poor and therefore new, targeted treatments are under active investigation. These new treatment modalities demand that also imaging methods need to advance to keep pace with these developments.

This thesis investigated new functional magnetic resonance imaging-based methods in ovarian cancer diagnosis. We were interested to see if it would be possible to obtain more detailed information of tumor characteristics already in preoperative imaging i.e. information that could support treatment decisions. The prospective study examined an overall cohort of 42 patients with primary ovarian, fallopian tube or primary peritoneal cancer treated in Kuopio University Hospital from 2011-2014.

We found that apparent diffusion coefficient (ADC) values measured in diffusion weighted imaging (DWI) correlated with a more aggressive tumor and worse survival (I). In a dynamic contrast-enhanced (DCE) imaging study, higher K^trans (a rate constant of transferred contrast agent from plasma to extravascular extracellular space (EES)) was associated with high-grade serous ovarian cancer and better results in cytoreductive surgery (II). DCE imaging parameters K^trans and Kep (a rate constant of transferred contrast agent from EES to plasma) were lower in patients with elevated expression levels of hypoxia-inducible factor 1 alpha (HIF-1), a specific marker known to be elevated in hypoxic conditions (III). This thesis demonstrates that new functional imaging methods may serve as surrogate markers allowing a more detailed characterization of ovarian cancer.

National Library of Medicine Classification: WP 322, WN 185, WO 179, WB 142 Medical Subject Headings: Carcinoma, Ovarian Epithelial; Magnetic Resonance Imaging;

Diffusion Magnetic Resonance Imaging; Preoperative Care; Prognosis; Prospective Studies

(12)

(13)

Lindgren, Auni

Multiparametrinen magneettikuvataminen epiteliaalisessa munasarjasyövässä Kuopio: Itä-Suomen yliopisto

Publications of the University of Eastern Finland Dissertations in Health Sciences 532. 2019, 134 s.

ISBN: 978-952-61-3197-9 (nid.) ISSNL: 1798-5706

ISSN: 1798-5706

ISBN: 978-952-61-3198-6 (PDF) ISSN: 1798-5714 (PDF)

TIIVISTELMÄ

Gynekologisista syövistä epiteliaalinen munasarjasyöpä on suurin kuolemien aiheuttaja. Yleensä se on laajalle levinnyt jo diagnosointivaiheessa, koska varhaisoireet puuttuvat. Varhaista diagnosointia helpottaisi, jos munasarjasyöpään olisi seulontamenetelmä, mutta sellaista ei ole.

Elossaololuku munasarjasyöpäpotilailla on huono. Uusia targetoituja täsmähoitoja on ryhdytty kehittämään, jotta hoitotuloksia saataisiin parannettua ja elinaikaa lisättyä. Nämä uudet hoitomuodot edellyttävät, että kuvantamismetodeja kehitetään syövän hoidon nopean kehittämisen ohella.

Väitöskirjatyössäni on tutkittu multiparametrisiä

magneettikuvantamismenetelmiä munasarjasyöpäpotilailla. Halusimme nähdä, onko magneettikuvantamisella mahdollista saada yksityiskohtaisempaa tietoa tuumorien ominaisuuksista jo preoperatiivisesti, jotta voitaisiin saada tukea hoitotavan valintaan. Tässä prospektiivisessa tutkimuksessa oli mukana 42 primaaria munasarjasyöpää, munanjohdinsyöpää tai peritoneaalista karsinoosia sairastavaa potilasta, joiden hoidot aloitettiin Kuopion yliopistollisessa sairaalassa 2011-2014 välisenä aikana. Saimme selville, että matalammat diffuusiokerroin ADC:n arvot (apparent diffusion coefficient), jotka on mitattu diffuusiokuvantamisella, korreloivat aggressiivisen syöpätyypin kanssa ja olivat yhteydessä huonompaan selviytymiseen (I). Dynaamista kontrastikuvantamista DCE (dynamic contrast enhanced) tutkivassa työssä korkeampi K^trans (varjoaineen siirtymistä verisuonista solun ulkoiseen tilaan kuvaava suure) oli yhteydessä high grade seroosiin munasarjasyöpätyyppiin ja parempaan leikkaustulokseen syöpäleikkauksessa (II). Dynaamisessa kuvantamisessa K^trans ja Kep (varjoaineen siirtymistä solun ulkoisesta tilasta plasmaan kuvaava suure) arvot olivat matalampia potilailla, joiden hapen puutetta kuvaava muuttuja eli hapenpuutteen indusoima tekijä HIF-1α (hypoxia-inducible factor 1 alpha) oli voimakkaammin expressoitunut (III). HIF-1α on hapenpuutteelle herkkä muuttuja, joka nousee, kun kudoksissa vallitsee hapenpuute. Tämä väitöskirjatyö todistaa, että uudet funktionaaliset

(14)

magneettikuvantamismenetelmät voivat toimia surrogaattimuuttujina yksityiskohtaisempaa munasarjasyövän luonnehdintaa varten.

Luokitus: WP 322, WN 185, WO 179, WB 142

Yleinen suomalainen ontologia: munasarjasyöpä; diffuusiokuvaus; magneettikuvaus;

hoito; ennusteet; seurantatutkimus

(15)

ACKNOWLEDGEMENTS

This study was carried out in the Department of Gynecology and Obstetrics and Department of Clinical Radiology, University of Eastern Finland and Kuopio University Hospital during the years 2014-2019. I am grateful to have had the possibility to be a member of the Graduate School of Clinical Research, Kuopio University Hospital and University of Eastern Finland. I am sincerely thankful to all of you who have participated or otherwise contributed to my research work during these years.

I owe my deepest feelings of respect and gratitude to my principle supervisor Adjunct Professor Maarit Anttila. Your brilliance, enthusiasm, never-ending ideas have been inspirational. From the first day when I visited KUH to meet you and we discussed the possibility of participating in a research project, you have demonstrated real expertise in medical research. Although you have many ongoing projects, you have always been helpful to me. Another brilliant person to whom I want to extend my deepest gratitude is Professor Ritva Vanninen. I really appreciate that I have had the good fortune to have had you as my supervisor. Your knowledge, enthusiasm, guidance and humanity have been a driving force for me, especially during the tough times. You have always found time to go through my texts and answer my questions. Your scientific expertise and encouragement at every step has been really important for me.

I would also like to extent my gratitude to my other supervisor, Hanna Sallinen. Your work ethic, guidance, humility and encouraging words have been a source of inspiration to me. Your friendship has been so important to me. I am thankful that I have had the opportunity to know you - your warm attitude and support throughout these years have meant that I can always call you and ask for your opinion. It has been a priviledge to get to know you.

Likewise, I would like to thank my supervisor Adjunct Professor Kirsi Hämäläinen.

Your kindness, patient and encouraging attitude made me believe that I can learn to analyse pathological samples. You have been always ready to help me.

To all of my four supervisors, without your constructive, sometimes challenging, but invariably brilliant input, I wouldn’t be where I am now.

I am honored to have Adjunct Professor Irina Rinta-Kiikka from the University of Tampere as my opponent.

I want to express my sincere gratitude to the official reviewers of this thesis Adjunct Professor Annika Auranen and Professor Hannu Aronen. Your comments and concerns have improved this thesis substantially.

I wish to thank my fellow co-authors; Mervi Könönen, Suvi Rautiainen, Otso Arponen, Annika Kivelä, Petri Mäkinen, Kirsi Härmä, Veli-Matti Kosma and Seppo

(16)

Ylä-Herttuala – you all have been important in the completion of this project. My deepest thanks belong to Mervi - you have been so patient and helpful in guiding me through the strange world of physics which underpins magnetic resonance imaging.

It has been a real pleasure to know you. Suvi - you helped me a lot when I was taking my first steps. Without your experience, my start with different MRI sequences would have been even much harder. I am so happy that I have had the opportunity to work together with Otso in that small chamber in the Radiology Department. Your enthusiasm for research was infectious. When I had a lot of work and other things on my mind, it helped so much simply to come to our chamber to witness your enthusiasm with research.

Likewise, I wish to express my warm gratitude also to my colleagues in the Department of Obstetrics and Gynecology in Kuopio University Hospital for working together and supporting me.

I owe my sincere thanks to my dear friends Annu, Hanna, Jaana, Johanna, Melisa, Mirka and Paula: you have been such an important part of my life and I thank you for the support during all this years. I thank also all my other friends who have helped me to forget about medical research every now and then during the completion of this thesis.

Last, but not least, I would like to thank my family for all their love and encouragement. To my parents, Arja and Heimo and my sisters, Tiina and Maiju, thank you for showing me love and commitment. One couldn’t have wished for a better family and for a better upbringing, all that I have accomplished is built on the support and belief that you have instilled into me.

To Antti my love. You are my home harbour, my safe place and together with Vilho you are the lights that brighten my days. Thank you for your everlasting support during all this years. Thank you for believing in me even when I lost belief in myself, thank you for being there for me and Vilho whenever we needed it. And most importantly, thank you for being you.

Kuopio, October 2019 Auni Lindgren

This work was supported by grants from the Kuopio University Hospital-VTR funds, EVO funding, University of Eastern Finland, the Instrumentarium Foundation, the Finnish Cultural Foundation of Northern Savo, the Paavo Koistinen Foundation, the Kuopio University Hospital Research Foundation, the Cancer Society of Finland, the Mauri and Sirkka Wiljasalo Grant, the Finnish Medical Foundation.

(17)

LIST OF ORIGINAL PUBLICATIONS

This dissertation is based on the following original publications:

I Lindgren Auni, Anttila Maarit, Rautiainen Suvi, Arponen Otso, Kivelä Annukka, Mäkinen Petri, Härmä Kirsi, Hämäläinen Kirsi, Kosma Veli-Matti, Ylä-Herttuala Seppo, Vanninen Ritva, Sallinen Hanna. Primary and metastatic ovarian cancer: Characterization by 3.0T diffusion weighted MRI. Eur Radiol.

2017 Sep; 27(9):4002-4012.

II Lindgren Auni, Anttila Maarit, Arponen Otso, Rautiainen Suvi, Könönen Mervi, Vanninen Ritva, Sallinen Hanna. Prognostic Value of Preoperative Dynamic Contrast-Enhanced Magnetic Resonance Imaging in Epithelial Ovarian Cancer. Eur J Radiol 2019; 115:66-73.

III Lindgren Auni, Anttila Maarit, Rautiainen Suvi, Arponen Otso, Hämäläinen Kirsi, Könönen Mervi, Vanninen Ritva, Sallinen Hanna. Dynamic contrast- enhanced perfusion parameters in ovarian cancer: good accuracy in identifying high HIF-1α expression. PLoS One 2019; 14(8):e0221340.

The publications were adapted with the permission of the copyright owners.

(18)

(19)

ABBREVIATIONS

ADC Apparent diffusion coefficient AIF Arterial input function

AUC Area under the curve BRCA Breast cancer-associated gene CA-125 Cancer antigen 125

CD34 Cluster of differentiation 34- protein

CT Computed tomography DCE Dynamic contrast-enhanced DWI Diffusion-weighted imaging EES Extravascular extracellular space

EOC Epithelial ovarian cancer ESUR European Society of Urogenital Radiology

FIGO International Federation of Gynecology and Obstetrics HE4 Human epididymis protein 4 HIF-1α Hypoxia-inducible factor 1 alpha

HR Homologous recombination IAUGC Integrated area under the concentration time

ICC Interclass correlation coefficient IOTA International Ovarian Tumor Analysis

IVIM Intravoxel incoherent motion K^trans Rate constant for transfer of contrast agent from plasma to EES Kep Rate constant for transfer of contrast agent from EES to plasma L-ROI Large region of interest

MRI Magnetic resonance imaging MVD Micro-vessel density NACT Neoadjuvant chemotherapy OC Ovarian cancer

OS Overall survival

PARP Poly adenosine diphosphate- ribose polymerase

(24)

PD-L1 Programmed cell death ligand 1

Peak Maximal enhancement

PET Positron emission tomography PET/CT Positron emission tomography /Computed tomography PS Permeability surface area product

qRT-PCR Quantitative transcription polymerase chain reaction RFS Recurrence-free survival RECIST Response Evaluation Criteria in Solid Tumors

RMI Risk of malignancy ROI Region of interest S-ROI Small region of interest SPAIR Spectral attenuated inversion recovery

T Tesla

TATI Tumor associated trypsin inhibitor

TE Echo time

Time-To-Peak Time when contrast agent reaches the peak volume TR Repetition time

TVU Transvaginal ultrasound Ve Contrast agent distribution volume EES volume fraction VEGF Vascular endothelial growth factor

VEGFR Vascular endothelial growth factor receptor

V^p Plasma volume

WashIn Initial up-slope of the DCE curve

WashOut Initial down-slope of the DCE curve

WHO World Health Organization WLsp-ROI Whole lesion single plane region of interest

WB-DWI Whole body diffusion weighted imaging

(25)

1 INTRODUCTION

Epithelial ovarian cancer (EOC) is one of the leading causes of death in gynecological malignancies. Commonly, EOC has already widely spread at the time of diagnosis, because early symptoms are missing, and there is no screening system for this cancer.

EOC may cause abdominal pain, changes in bowel motility, and a palpable mass in pelvis, weight loss and occasionally further swelling of the abdomen, breathing problems and pain if there is the presence of ascites.

Survival of EOC is poor in advanced stages of the disease. The golden standard for treatment is debulking surgery where all visible tumor material is removed, followed by platinum-based chemotherapy. A residual tumor after surgery is one of the strongest prognostic factors (Bois et al., 2009). Other prognostic factors are International Federation of Gynecology and Obstetrics (FIGO) stage, age and patient’s condition at the time of diagnosis (Davidson and Tropé, 2014; Winter et al., 2007). Studies have been conducted to determine whether there is a difference in survival if the patient receives neoadjuvant chemotherapy before the operation in comparison to surgery without neoadjuvant therapy. Results from neoadjuvant treatment are conflicting and the issue is under debate (Chang et al., 2015; Wright et al., 2016).

It is difficult to know in advance which patients will benefit from chemotherapy.

EOC is frequently chemo-sensitive, patients may benefit from treatment but the high recurrence rate is a problem since as many as 70% of patients suffer a relapse (Ledermann et al., 2013). For this reason, treatment is under continuous research.

Platinum-based treatment has been used for decades but lately an anti-vascular endothelial growth factor (VEGF) antibody, bevacizumab, has been combined in the treatment as it is known that angiogenesis is essential for tumor growth and metastasis. It has been indicated that bevacizumab, which causes changes in neoangiogenesis, can prolong disease-free time (Aghajanian et al., 2012; Burger et al., 2012; Perren et al., 2011). In addition, other new, targeted treatments, such as agents affecting hypoxia-inducible-factor 1 alpha (HIF-1α) are under development. Hypoxia has an important role in tumor growth and treatment resistance. HIF-1α expression increases in hypoxic conditions and modulates expression of many genes that are involved in crucial cellular processes such as metabolism, angiogenesis and proliferation (Jiang and Feng, 2006; Semenza, 1999; Wang et al., 1995).

These new treatment modalities can lead to changes in tumor consistency not only in the size of the tumor. Follow-up is based on response evaluation criteria in solid tumors (RECIST) (Eisenhauer et al., 2009). However, in the future, it will be important to devise different and more accurate methods for follow-up.

Furthermore, in diagnostics, alternative imaging methods other than computed tomography (CT) may provide more information on tumor behavior. Vaginal ultrasound is the first step in the diagnosis of OC. If ultrasound suggests a strong

(26)

possibility for a malignant tumor, then CT is the second imaging modality.

Conversely, if there is an indeterminate tumor observed in ultrasound, magnetic resonance imaging (MRI) is recommended. The European Society of Urogenital Imaging also recommends the use of the following techniques, diffusion weighted imaging (DWI) and dynamic contrast-enhanced (DCE) imaging, analysis of the enhancement curve shape, when conducting a detailed analysis of the tumor (Bazot et al., 2017).

DWI and DCE imaging are functional MRI techniques that provide more information about the tumor than normal conventional structural MRI. DWI has been shown to give additional information and improve conventional MRI sensitivity and specificity to differentiate between benign and malignant tumors (Bollineni et al., 2015; Thomassin-Naggara et al., 2011). DWI can help also to detect the peritoneal spread of OC (Fujii et al., 2008; Kyriazi et al., 2010). DCE imaging is based on the movement of contrast agent from blood to tissue and then back again.

Semi-quantitative measurements and contrast enhancement curve shape analysis provide important qualitative information. Pharmacokinetic perfusion parameters represent quantitative measurements from contrast agent flow, diffusion and distribution in tissue. In many clinical studies, dynamic contrast MRI has been widely exploited for noninvasive detection, characterization, and therapy monitoring (Ho et al., 2007; Li et al., 2017; Ren et al., 2008; Thomassin-Naggara et al., 2017; Zahra et al., 2007).

This thesis was intended to contribute to our understanding of the diffusion weighted and dynamic contrast-enhanced imaging in EOC. It reveals new information on the opportunities of different magnetic resonance imaging sequences.

This study supports the hypothesis that these functional imaging sequences are useful supplements to traditional conventional magnetic resonance imaging.

(27)

2 REVIEW OF THE LITERATURE

2.1 EPITHELIAL OVARIAN CANCER

2.1.1 Evolution

When one considers the origin of epithelial ovarian carcinomas, it has been traditionally understood that the many different ovarian tumors are all derived from the ovarian surface epithelium; this forms an inclusion cyst after invagination in the underlying stroma where it undergoes a malignant transformation. Metaplastic changes during the process have explained the different cell types that are encountered as tumors. (Kurman and Shih, 2010) According to an alternative theory of the origin of EOC, it has been postulated that EOC could arise from tissue that is derived embryologically from the mullerian duct. (Dubeau, 2008)

2.1.2 Epidemiology

Epithelial ovarian cancer (EOC) is one of the most lethal malignancies of women.

While it is only the sixth most common cancer in women, it is the gynecological malignancy that carries the greatest risk of death. Nonetheless, it is a rare disease: in 2016 in Finland, the incidence was 14.61/100 000 women. In Europe, the incidence varies between 5-15/100 000, being highest in Lithuania, Latvia, Ireland, Slovakia and Czech Republic and lowest in Portugal and Cyprus. During 2016 in Finland, 437 new patients were diagnosed with ovarian cancer (OC) and there were 357 deaths due to OC (Institute for Statistical and Epidemiological Cancer Research, 2016). The incidence of EOC is highest in Europe, the USA and Israel and lowest in the developing countries and in Japan. The incidence of OC increases with age; it is highest in women aged 60 to 69 years (Institute for Statistical and Epidemiological Cancer Research, 2016).

2.1.3 Classification

The histogenic classification of ovarian cancer includes coelomic epithelium, mesenchymal (sex cord stroma) and germ cells (DiSaia and Creasman, 2012), see Table 1. Approximately 95% of ovarian carcinomas originate from epithelium. The World Health Organization (WHO) updated the classification of EOC 2014 to include serous high- and low-grade carcinoma, mucinous carcinoma, endometrioid carcinoma, clear cell carcinoma, malignant Brenner tumor, seromucinous carcinoma and undifferentiated carcinoma. Most of the tumor types have also their own benign and borderline malignant counterparts. (Meinhold-Heerlein et al., 2015).

(28)

Table 1. The histogenic classification of ovarian malignancies. (DiSaia 2012)

Origin Tumor type

Epithelium 90-95% Serous

Mucinous Endometrioid Clear-cell

Undifferentiated carcinoma Carcinosarcoma

Germ cells 3-5% Dysgerminoma

Immature teratoma

Secondary neoplasm from mature cystic teratoma

Chorioncarcinoma

Gonadal stroma 5-10% Granulosa cell tumor

Non-specific mesenchyme Mixed mesodermal sarcoma

Lymphoma

Metastatic GI tract

Breast Endometrium Lymphoma

The cell types of EOC resemble normal non-neoplastic cells situated in other organs of the female genital tract. Serous, endometrioid, clear cell, mucinous and transitional cell (Brenner) carcinomas morphologically resemble the epithelia of the fallopian tube, endometrium, gastrointestinal tract or endocervix and urinary bladder, respectively. These cell types are not present in normal ovarian tissue.

(Cheng et al., 2005; Kim et al., 2012). It has been postulated that most of the serous tumors develop from fimbria, endometrioid and clear cell tumors from endometrial tissue via retrograded menstruation and transitional (Brenner) tumors may have originated from transitional-type epithelium located at the tubal-mesothelial junction, where the fimbria makes contact with the peritoneum. (Dubeau, 2008;

Kurman and Shih, 2010)

Even though the etiology and pathology behind EOC are not fully understood, it has been claimed that EOC is a group of diseases with different morphologies, immunohistochemistries, molecular and biologic behaviors rather than a single homogenous disease. Instead, the different histological subtypes of EOC are indicative of five different diseases: high-grade serous carcinoma (HGSC), low-grade serous carcinoma (LGSC), mucinous carcinoma (MC), endometrioid carcinoma (EC) and clear cell carcinoma (CCC). Each of these diseases has its own different epidemiological and genetic risk factors, precursor lesions, patterns of dissemination, molecular genetic alterations, response to chemotherapy and prognosis. (Kurman and Shih, 2016; Prat et al., 2018) Mixed forms, malignant Brenner tumors, undifferentiated carcinomas and malignant mixed tumors (carcinocarcomas) are

(29)

additional entities. The five main types of OC and their behaviors are presented in Table 2.

Table 2. Main types of ovarian cancer and their morphology and behaviors. (Prat 2018;

Kurman 2016)

High-grade

serous Low-

grade serous

Mucinous Endometrioid Clear cell

Percentage (%) 70-80 <5 3 10 5-10

Stage Advanced Early or

advanced

Early Early Early

Origin/

precursor lesion Tubal

neometaplasia in inclusions of ovarian surface epithelium

Serous borderline tumor

Adenoma- borderline- carcinoma sequence;

teratoma

Endometriosis,

adenofibroma Endometriosis, adenofibroma

Genetic BRCA1/2 HNPCC

Molecular

abnormalities TP53, BRCA B-RAF or K-RAS

K-RAS, ERBB2

PTEN, CTNNB1, ARID1A, PIK3CA, K- RAS, MI

HNF-1β, ARID1A, PTEN, PIK3CA

Proliferation High Low Intermediat

e

Low Low

Response to primary chemotherapy

80% 26%-28% 15% 15%

Prognosis Poor Favorable Favorable Favorable Intermediate

2.1.4 Prognosis and Survival Rates

EOC has a relatively poor prognosis. Most often, EOC is diagnosed only at an advanced stage, because there are few or no early symptoms and furthermore, it is not uncommon for EOC to metastasize early in its development. Low-grade tumors do not metastasize as quickly as high-grade tumors. Although the mortality rate declined by 33% between 1976 and 2015, the five-year survival rate is still low, 47%

(Torre et al., 2018). Survival in early stage disease is good, the five-year survival rate is as high as 90%, but it drops dramatically in advanced stages of the disease. Table 3 illustrates survival rates in OC in the Finnish population. Although treatment has developed during recent decades and the response rate to primary treatment is good,

(30)

a high recurrence and the development of resistance to chemotherapy are problems in EOC treatment (Ledermann et al., 2013). If it were possible to diagnose EOC earlier, then the patient’s survival possibilities would increase, but unfortunately, at present there is no screening system for EOC.

Table 3. Survival rates in ovarian cancer for all patients and for different age groups, in the Finnish population during the years 2014-2016. (Finnish Cancer Registry, 2018)

1y 2y 3y 4y 5y

all 73.17 60.57 51.38 44.63 41.40

age group

0-54 90.72 82.91 77.16 71.38 68.44

55-74 83.36 71.09 59.17 50.54 46.09

75+ 46.66 30.63 23.96 19.74 18.32

2.1.5 Risk factors

Family history is the strongest known risk factor for EOC. The average lifetime risk of developing EOC is 1.4 percent, the risk increase 1.5 times among women who have a first degree relative with EOC and 1.1 times among those with a relative who has had breast cancer (Wentzensen et al., 2016). A clear identifiable genetic predisposition is present in only 10-15% of patients; in most of these cases, there are germline BRCA1 and BRCA2 mutations. The average cumulative risk of EOC in a 70y woman with a BRCA1 mutation is 43-76% and 7.5-34% if she carries a BRCA2 mutation (Mavaddat et al., 2013). Other known risk factors are nulliparity and infertility, early menarche, late menopause and use of hormone replacement therapy (Gong et al., 2013; John et al., 1993; Mørch et al., 2009). Women who have ever used hormone replacement therapy (estrogen alone or estrogen combined with progesterone) have a 20% higher risk of developing OC in comparison with those who have never used these preparations. The risk is highest with current users and those who have stopped within 5 years, where it may be elevated as much as 40%

(Collaborative Group on Epidemiological Studies of Ovarian cancer, 2015).

Tubal ligation, a greater number of pregnancies, a longer duration of breast feeding and the use of oral contraceptives reduce the risk of EOC (Danforth et al., 2007; Hankinson et al., 1993; John et al., 1993). The extent of the reduction is about 35% in women who have used contraceptives from 5 to 9 years (Beral et al., 2008).

2.1.6 Symptoms and Diagnosis

Commonly, EOC has already spread widely during the time of diagnosis, because early symptoms are missing. OC may cause abdominal swelling, abdominal pain, dyspepsia, change in bowel motility, a palpable mass in pelvis, weight loss and changes in urinary frequency (Goff et al., 2004; Olson et al., 2001). If the cancer starts

(31)

to form ascites, this may cause abdominal bloating and a loss of appetite. Some patients may have breathing problems due to pleural effusions.

The first steps in the diagnostic phase are the anamnesis, clinical examination and rectovaginal palpation. Vaginal ultrasound is the golden standard in the differential diagnosis. Diagnostic tools such as the risk of malignancy index (RMI), which combines multiple risk factors (menopausal status, CA-125 level, ultrasound findings) (Tingulstad et al., 1996) have been developed to help physicians to differentiate benign tumors from their malignant counterparts. The International Ovarian Tumor Association (IOTA) group has published guidelines to distinguish benign from malignant tumors; these take into account tumor size, solid proportions, tumor shape, ascites, acoustic shadows and vascularity (Timmerman et al., 2008). If there is a strong possibility for a malignant tumor, computed tomography (CT) is usually the next imaging modality to be performed. However, if there is an indeterminate tumor in ultrasound, magnetic resonance imaging is recommended.

The European Society of Urogenital Imaging recommends the inclusion of diffusion weighted imaging (DWI) and dynamic contrast-enhanced (DCE) imaging enhancement curve shape analysis to detailed analysis (Forstner et al., 2017). Tumor markers like cancer antigen 125 (CA-125), tumor associated trypsin inhibitor (TATI), human epididymis protein 4 (HE4) may also help in the differential diagnosis. CA- 125 has been the most widely used marker, but it is not specific for EOC, also benign conditions like endometriosis, as well as pregnancy enhance its levels (Einhorn et al., 2005; Miralles et al., 2003). TATI does provide extra information in mucinous cell types, but its concentration is not elevated in other cell types of EOC (Stenman, 2002).

HE4 is a newer marker; its levels are elevated especially in serous and endometrioid EOC (Drapkin et al., 2005). At present, there is no practical screening modality for OC. In the primary situation, it is important to define as precisely as possible the stage of disease as well as a differential diagnosis, since these are crucial for treatment planning. The final diagnosis is based on an evaluation of surgical findings and a histopathological assessment of tissue samples (Benedet et al., 2000).

2.1.7 Staging

Staging is very important in EOC treatment. Staging defines the tumor extension at the primary presentation. EOC staging is based on surgical findings in laparotomy with a midline incision, but in the early stage (stage I), laparoscopy is an alternative method (Park et al., 2013). The International Federation of Gynecology and Obstetrics (FIGO) staging system has been the most widely used staging system in EOC; FIGO was updated in 2014 (Mutch and Prat, 2014). The American Joint Committee on Cancer (AJCC) TNM staging system is another widely used. Both of these staging systems use three variables to classify the cancer. Table 4 illustrates the FIGO and TNM staging system.

(32)

Table 4. Figo and TNM staging.

I Tumor confined to ovaries or fallopian tube(s) T1 IA Tumor limited to one ovary, capsule intact no tumor on

surface, negative washing/ascites

T1a N0M0 IB Tumor involves both ovaries otherwise like IA T1b N0M0 IC Tumor limited to one or both ovaries

IC1 Surgical spill

IC2 Capsula rupture before surgery of tumor on ovarian surface

IC3 Malignant cells in the ascites or peritoneal washing

T1c N0M0

II Tumor involved below the pelvic brim or primary peritoneal cancer

T2 IIA Extension and/or implant on uterus and/or fallopian

tubes

T2aN0M0

IIB Extension to other pelvic IP tissues T2b N0M0

III Confirmed spread outside the pelvis and/or metastasis to the retroperitoneal lymph nodes

T3 IIIA Positive peritoneal lymph nodes and/or microscopic

metastasis beyond the pelvis

IIIA1 Positive peritoneal lymph nodes only IIIA1(i) Metastasis ≤ 10mm

IIIA1(ii) Metastasis > 10mm

IIIA2 Microscopic, extrapelvic (above the brim) peritoneal involvement ± positive retroperitoneal lymph nodes. Includes extension to capsule of liver/spleen

T3aN1M0 T3a/T3aN1

T3a/T3aN1

IIIB Macroscopic, extrapelvic, peritoneal metastasis ≤ 2cm ± positive retroperitoneal lymph nodes. Includes extension to capsule of liver/spleen

T3b/T3bN1M0

IIIC Macroscopic, extrapelvic, peritoneal metastasis > 2cm ± positive retroperitoneal lymph nodes. Includes extension to capsule of liver/spleen

T3c/T3cN1M0

IV Distant metastasis excluding peritoneal metastasis AnyT,anyN,M1 IVA

IVB Pleural effusion with positive cytology

Hepatic and/or splenic parenchymal metastasis, metastasis to extra-abdominal organs (including inguinal lymph nodes and lymph nodes outside of the abdominal cavity)

AnyT,anyN,M1

It is important that the surgical procedure is conducted adequately to ensure a reliable staging, see Table 5. Before staging surgery, the patient undergoes staging imaging with computed tomography (CT) and chest x-ray, for disease evaluation and screening of pleural metastases (Benedet et al., 2000). If there is a suspicion of a pleural effusion, then cytology must be examined, because this is the way in which the disease is upgraded to stage IV. Histological type and grade should be designated at staging. In addition, the tumor’s primary site, i.e. ovary, fallopian tube or peritoneum, should always be characterized whenever possible.

(33)

Table 5. Adequate operation procedures for correct FIGO staging. (Benedet 2000).

Guidelines for surgical staging according to FIGO

Four peritoneal washings (diaphragm, right and left abdomen, pelvis) Careful inspection and palpation of all peritoneal surfaces

Biopsy of all suspicious lesions Infracolic omentectomy

Biopsy or resection of any adhesions In the absence of obvious implants:

Random biopsies of normal peritoneum of undersurface of right hemidiaphragm, bladder reflection, cul-de-sac,

right and left paracolic recesses, and both pelvic sidewalls Lymphadenectomy of pelvic and aortic nodes

Total abdominal hysterectomy, bilateral salpingo-oophorectomy, and excision of masses when considered prudent

Appendectomy for mucinous tumors

2.1.8 Prognostic Factors and Current Treatments

The presence of residual tumor after debulking surgery is one of the strongest prognostic factors in EOC (Bois et al., 2009). Other prognostic factors are FIGO stage, grade, age and general condition at diagnosis (Davidson and Tropé, 2014; Winter et al., 2007). In addition, low CA125 levels and the presence of the BRCA2 mutation seem to be protective factors. Because surgery plays such an important role in prognosis, studies have been conducted to determine whether there is a difference if the patient receives neoadjuvant chemotherapy (NACT) before the operation (Vergote et al., 2010). It has been claimed that survival of these patient is not worse and the surgical operation is easier to perform because of the lower tumor burden.

This issue has been a subject of ongoing debate in recent years (Chang et al., 2015;

Kehoe et al., 2015; Wright et al., 2016).

The golden standard for primary treatment is debulking surgery where all visible tumor tissue is removed, followed by platinum-based chemotherapy. The two most commonly used cytostatic agents are carboplatin and paclitaxel. Patients with early stage disease IA or IB, non clear cell histology and well-differentiated (G1) tumors do not seem to benefit from adjuvant chemotherapy after optimal debulking surgery (ICON1 and EORTC-ACTION, 2003) but for all other patients, adjuvant chemotherapy is indicated (Ledermann et al., 2013).

EOC is frequently chemotherapy sensitive, and patients benefit from treatment, but the high recurrence rate is a major problem. As many as 70% of patients relapse within the first 18 months (Ledermann et al., 2013). A relapse is treated usually with chemotherapy. If the patient is platinum sensitive and the first relapse occurs after 12 months of treatment, secondary chemotherapy is most often a combination of platinum and some other cytostatic agent, such as paclitaxel, docetaxel or pegylated

(34)

liposomal doxorubicin. If the time for first relapse is 6-12 months, then the patient is defined as partially sensitive, and if relapse is within 6 months, they are platinum resistant. It has been shown that if the relapse is only local and it is possible to remove recurrent tumor totally by surgery, then a new operation will be beneficial (Du Bois et al., 2018). Usually, the second relapse occurs more quickly than the first one.

Treatment for EOC is under continuous development. Although platinum-based treatment has been used for decades, lately a human monoclonal vascular endothelial growth factor (VEGF) antibody, bevacizumab, has been combined with these more traditional drugs in EOC treatment. There are other new drugs being investigated e.g. orally delivered poly (ADP-ribose) polymerase (PARP) inhibitors, like olaparib, niraparib and rucaparib.

2.1.9 Strategies for targeted therapies 2.1.9.1 Bevacizumab and PARP inhibitors

Angiogenesis is essential for tumor growth and metastasis. An increased expression of pro-angiogenic factors and increased vascularization have been associated with advanced stage and poor prognosis in many tumors (Hicklin and Ellis, 2005). VEGF proteins and their receptors play a pivotal role in tumor angiogenesis and the pathogenesis of EOC as is the case in a wide range of other human tumors (Carmeliet, 2005). There are two strategies that can be used to inhibit the VEGF pathway:

inhibition of VEGF ligand with antibodies or soluble receptors, and inhibition of the VEGF receptors (VEGFRs) with tyrosine kinase inhibitors or receptor antibodies. The US Food and Drug Administration (FDA) has approved bevacizumab, nintedamib, pazopanib and cediranib for clinical use. Bevacizumab is the most widely used of these drugs and it has been claimed in many studies that bevacizumab can prolong disease-free time (Aghajanian et al., 2012; Burger et al., 2012; Perren et al., 2011).

Unfortunately, resistance to antiangiogenic therapy is not uncommon and not all patients gain any benefits from this treatment (Welti et al., 2013). Furthermore, the side effects of the therapy may be problematic; these include high blood pressure, proteinuria and bowel perforations.

The poly (ADP-ribose) polymerase (PARP) inhibitors are novel treatment options for those EOC patients who carry a BRCA1 or BRCA2 mutation. BRCA1 and BRCA2 proteins are required for homologous recombination (HR) to suppress genetic instability, which can lead to cancer (Venkitaraman, 2002). The genetic interaction between PARP and BRCA has been described as synthetic lethal. The individual loss of either gene is compatible with life, but the simultaneous loss of both genes results in cell death. Hartwell et al. suggested already 1997 that a synthetic lethal approach could be used in cancer treatment. The PARP-BRCA interaction is the first example of this kind of approach that has achieved good results both in vivo (Bryant et al., 2005) and in clinical (Fong et al., 2009; Mirza et al., 2017; Moore et al., 2019) studies.

In a study with olaparib maintenance therapy, a patient with newly diagnosed

(35)

advanced ovarian cancer and a BRCA1/2 mutation who was treated with olaparib had a 70% lower risk of disease progression and death in comparison to a patient receiving only placebo (Moore et al., 2019). PARPs are a large family of multifunctional enzymes. The PARP-1 isoform is the form intended to be blocked by PARP inhibitors but most of these compounds also inhibit PARP-2 and in some cases, PARP-3. It seems that PARP inhibitors are well tolerated and associated with only minor side effects, because they are selectively targeting BRCA defective cells (Bryant et al., 2005; Fong et al., 2009; Moore et al., 2019).

2.1.9.2 Immunological treatments

Immunogenic pathways and immune systems are activated during cancer development (Zhang et al., 2003). Immunotherapy is based on the target immunogenic antigens expressed on the surface of tumor cells. (Drerup et al., 2015;

W. Wang et al., 2019; Zhang et al., 2003) The four main categories for immunotherapy are antibodies, immune checkpoint inhibitors, vaccines and adoptive cell therapy.

Many novel immunological therapeutic agents are now undergoing preclinical studies, with several already in clinical trials. Results from the trials in ovarian cancer were not as positive as expected but investigations are continuing. Nowadays, it is common to give a combination of different targeted treatments simultaneously (Z.

Wang et al., 2019; Zeng et al., 2019). There is no reliable biomarker for immunological treatments. In clinical drug trials, PD-L1 1% positivity has been used as an inclusion criterion.

2.1.9.3 New approaches, novel drug delivery systems

The aims of drug delivery systems are to encapsulate or embed a chemotherapeutic agent to achieve enhanced cancer treatment. Multiple research groups have been investigating ways to enhance cisplatin delivery to the tumor and to reduce the drug’s kidney toxicity (Paraskar et al., 2010); this has also been attempted for enhancing paclitaxel delivery (Devalapally et al., 2007; Lu et al., 2007). Single agent delivery systems based on surface-modification as well as engineering the size and shape of the nanoparticles have been investigated. Zeng et al. has developed a doxorubicin-releasing nanoparticle with the cucumber mosaic virus vector which enhanced in vivo treatment efficacy and reduced cardiotoxicity (Zeng et al., 2013).

There are also some co-delivery systems based on different nanoparticle combinations. For example, co-delivery of paclitaxel and ceramide using poly- modified poly-nanoparticles was able to enhance tumor suppression. Ceramide that accumulates in cancer cells would be one way to promote cancer cell apoptosis, and thus overcome drug-resistance and enhance the outcome of chemotherapy.

(Devalapally et al., 2007) Another group has developed a magnetic iron oxide nanoparticle for doxorubicin delivery, to be combined with hyperthermia and

(36)

conventional chemotherapy (Taratula et al., 2013). Drug release was triggered by an alternating magnetic field; this combination achieved a superior therapeutic outcome with fewer side effects.

2.1.10 Hypoxia and angiogenesis

2.1.10.1 Significance

Hypoxia has an important role in tumor development, it has a significant influence on the treatment response and on the clinical outcome in solid tumors (Daponte et al., 2008; Osada et al., 2007; Wilson and Hay, 2011). When a tumor grows, it requires more oxygen and this initiates neoangiogenesis in order to enhance the tumor’s blood supply. Neoangiogenesis is a complex process where many different growth factors and their receptors are present, studied already in 1970s (Folkman, 1971). The VEGF family has been the most widely studied; it exerts angiogenic, mitogenic and vascular hyperpermeability effects in tumors. Four different subtypes, VEGF-A, -B, - C and -D have been detected and they signal through three tyrosine kinase receptors VEGFR-1 (flt-1), VEGFR-2 (kdr/flt-1) and VEGFR-3 (flt-4) (Claesson-Welsh and Welsh, 2013). Cross-signaling between different ligands and receptors occurs but VEGF-A and -B mainly signal through VEGFR-1 and -2. They are considered as the major receptors of angiogenesis, vasculogenesis and vascular permeability (Horowitz and Seerapu, 2012; Nishida et al., 2004; shibuya, 2006). VEGF-C and -D act via VEGFR-3 which is the most important pathway in lymphangiogenesis (Achen et al., 1998; Joukov et al., 1996).

2.1.10.2 Hypoxia markers

Hypoxia-inducible factor 1 alpha (HIF-1α) is one of the most widely used biomarkers for detecting hypoxia. HIF-1α is a heterodimer transcription factor, which is formed from alpha and beta subunits. The β unit is a stable, constitutively expressed structure whereas the α unit is excessively secreted in hypoxic conditions (Wang et al., 1995) Thus, HIF-1α has been used as a molecular marker of hypoxia (Desir et al., 2014). Subsequently, three different alpha subunits 1α, 2α and 3α have been discovered which all have slightly different actions; two β isoforms, HIF-1β and HIF- 2β have been detected (Pezzuto and Carico, 2018). HIF-1α works as a key regulator of the cellular response to hypoxia; it modulates the expression of several genes involved in important cellular processes such as glucose uptake, metabolism, proliferation, angiogenesis, erythropoiesis and apoptosis (Jiang and Feng, 2006;

Semenza, 1999; Wang et al., 1995). In addition, it has been shown that HIFs control the proliferation of cancer stem cells (CSC) as well as their differentiation and pluripotency through specific signaling pathways such as Oct4, Wnt, Notch (Majmundar et al., 2015; Mathieu et al., 2011; Risbud et al., 2010).

(37)

Eppendorfs PO2 histograph is an older method for hypoxia detection but after the identification of HIF-1α, it is no longer used extensively. There are some other hypoxia markers in use e.g. carbonic anhydrase IX (CA-IX) and glucose-transporter- 1 (Glut-1).

2.1.10.3 Targeted treatment

Target treatments for HIF-1α and its pathway are under development. It has been shown that one mechanism of tumor resistance to antiangiogenic treatment is up- regulation of the transcription factor HIF-1α (Blagosklonny, 2004; Hicklin and Ellis, 2005). Thus, a combination of antiangiogenic therapy and HIF-1α inhibition seems to be a promising strategy for targeted cancer treatment (Moroney et al., 2012; Powis and Kirkpatrick, 2004).

2.2 IMAGING OF EOC

2.2.1 Transvaginal ultrasound

In the diagnostic imaging of ovarian tumors, transvaginal ultrasound (TVU) is the golden standard. TVU is the first modality for imaging if there is a suspicion of ovarian tumor since this device is widely available and inexpensive to use. Doppler technology makes ultrasound even more sensitive for tumor imaging. Certain findings are indicative of malignancy e.g. solid components, irregular shape, highly vascularized solid parts, ascites and changes in both ovaries. A two-dimensional technique combined with Doppler technology can be used to differentiate between benign and malignant tumors with a 91.0% sensitivity and a 91.7% specificity (Dodge et al., 2012). Sokalska et al. found that Doppler ultrasound monitoring for diagnosing extra uterine mass possessed a sensitivity of 77-86% and a specificity of 94-100%

(Sokalska et al., 2009). Various different prediction models for TVU have been developed to help to optimize the diagnosis. The conclusion of a large meta-analysis (195 studies with 19 different prediction models) was that IOTA logistic regression mathematical models 2 (LR2) and simple rules would be the best ultrasound based models and strategies for use in clinical practice, with sensitivities of 92% and 93%

and specificities of 83% and 81%, respectively (Kaijser et al., 2014). IOTA simple rules are described in Table 6 and Figure 1.

(38)

Table 6. International Ovarian Tumor Analysis (IOTA) simple rules. (Timmerman 2008) Rules for predicting a malignant tumor

(M-rules)

Rules for predicting a benign tumor (B- rules)

M1 Irregular solid tumor B1 Unilocular

M2 Presence of ascites B2 Presence of solid components where the largest solid component has a largest diameter < 7mm

M3 At least four papillary structures B3 Presence of acoustic shadows M4 Irregular multilocular solid tumor with

largest diameter ≥ 100mm

B4 Smooth multilocular tumor with largest diameter < 100mm

M5 Very strong blood flow (color score 4) B5 No blood flow (color score 1)

If one or more M-rules apply without a B-rule, the tumor is classified as malignant and if one or more B-rules apply without an M-rule, then it is classified as benign. A tumor cannot be classified if both M- and B-rules apply or if no rules apply.

(39)

Figure 1. Illustration of the IOTA simple rules in TVU. (Copyright by IOTA)

In addition, a 3-dimensional technique has been developed to enable a closer evaluation of the structure and vascularity of the tumor. In fertile aged women, ovarian tumors are most often benign although they may display a mixed shadow or solid portion, commonly these findings are attributable to a hemorrhoid cyst, myoma or dermoid cyst. The risk for malignancy increases if these findings are observed in prepuberty or postmenopausal women (Kuivasaari-Pirinen and Anttila, 2011). Even when examined by an experienced physician, approximately 20% of the ovarian tumors are indeterminate in TVU. For these tumors, MRI with DWI and DCE sequences are recommended for further imaging (Forstner et al., 2017).

(40)

2.2.2 Computed tomography

CT is the first modality for staging imaging in OC (Forstner et al., 2016). It is based on different absorption of radiation in different tissues and organs. Single detector helical CT technique was first developed in 1990, subsequently (late 1990s) it was superseded by a multidetector CT (MDCT) technique allowing multiple slice acquisition during a single tube rotation. CT equipment has developed rapidly; in 1998, the first four-detector row CT equipment was introduced but already in 2010, a 320-row system was produced (Rubin, 2014). CT is not the most accurate imaging modality for pelvic tumors, although intravenous contrast agents can increase morphologic information. CT is readily available, less expensive and faster than MR imaging and therefore it is still the first modality in ovarian cancer imaging, although some studies have demonstrated that MRI and DWI- and DCE-MRI sequences are more accurate than CT in detecting the primary tumor, tumor metastases, predicting incomplete resection and recurrence (Fan et al., 2015; Low et al., 2015; Low and Barone, 2012; Michielsen et al., 2017; Vargas et al., 2013). CT is used also in the follow- up of ovarian cancer. If there are contraindications to CT, such as renal insufficiency or allergy to the contrast agent, then MRI is recommended for staging imaging.

2.2.2.1 Imaging technique

CT with MDCT technology allows high-resolution imaging in the coronal, sagittal and oblique planes, see Figure 2. An unenhanced imaging protocol can be used but intravenous administration of a contrast agent improves the morphologic information (Mohaghegh and Rockall, 2012). A monophasic CT protocol is usually used in the portal venous phase. In ovarian cancer imaging, it is worthwhile utilizing oral contrast media to aid in the differentiation of peritoneal tumor metastases from bowel loops.

The use of CT is increasing steadily which has raised concerns about the cancer risk related to diagnostic x-ray imaging. Factors that modify an individual’s risk of cancer include the radiation dose, age and sex. Children and young adults are at a higher risk than in elderly patients. Dose reduction has been achieved with advances in CT dose optimization technology without compromising image quality. In addition, doses have decreased continuously with the newer CT devices.

(41)

Figure 2. A 56-year-old woman with high-grade endometrioid ovarian cancer stage IC. A) CT image in the axial plane, B) T2W-MRI image in the axial plane, C) CT image in the sagittal plane, D) T2W-MRI image in the sagittal plane, E) CT image in the coronal plane and F) T2W- MRI image in the coronal plane.

2.2.3 Positron emission tomography

In positron emission tomography (PET), the patient is given a radioactive tracer agent, with a short half-life. The radioactive tracer emits positrons to its surroundings; when positrons collide with electrons, then a process called annihilation occurs, and gamma radiation is emitted. Two different gamma- quantums are emitted in opposite directions, which can be captured with a gamma camera. This collision point can be localized and the computer forms a three dimensional picture of the distribution of the radioactive tracer in the body. The fluorine isotope ¹⁸F incorporated into fluorine deoxyglucose (FDG) has been the most commonly used radioactive tracer. FDG is concentrated in organs and cells that use large amounts of glucose such as brain, liver and malignant tumors. However, inflammation affects its accumulation and this has to be considered in the differential diagnosis. Different tracers can be used, e.g. [¹⁸F] EF5 (2-(2-nitro-1H-imidazol-1-yl)- N-(2,2,3,3,3-pentafluoropropyl)-acetamide) has been used to label hypoxia in tumors (Silvoniemi et al., 2018). CT is commonly combined with PET so that anatomical differentiation is more robust. PET/CT does not give any better results in staging imaging of ovarian cancer and is not used in primary staging imaging (Iyer and Lee, 2010; Lee et al., 2015; Michielsen et al., 2014). It is helpful in situations when there is

(42)

a suspicion of recurrence, CA-125 is elevated and CT and MRI have been unsuccessful in determining recurrence (Gu et al., 2009; Hauth et al., 2005; Sebastian et al., 2008). PET/CT might be the best option for imaging if there is a plan to reoperate due to recurrent disease because it is important that all of the tumor will be removed when a second staging operation is planned. PET/CT can indicate if the tumor has an extensive distribution; interestingly, Michielsen et al. found that whole body DWI-MRI was capable of achieving the same outcome (Michielsen et al., 2014).

Figure 3. ¹⁸F-FDG PET/CT imaging, a 53-year-old woman with high-grade serous ovarian cancer stage IV. A) PET/CT image in the axial plane, B) PET/CT image in the coronal plane, C) still images from three-dimensional PET/CT illustrations.

2.2.4 Magnetic resonance imaging

MRI is a non-invasive imaging technology that produces two- or three-dimensional detailed anatomical images. In an external magnetic field, some atomic nuclei including hydrogen (¹H) nuclei are able to absorb and emit radio-frequency type

(43)

energy. MRI sensors receive the detectable radio-frequency signal, which is generated by the hydrogen nuclei. Hydrogen atoms are present throughout the human body particularly in water and fat tissues. For this reason, conventional magnetic resonance imaging has good soft tissue contrast and hydrogen nuclei are the most often used nuclei in clinical MRI. The contrast between different tissues depends on their relaxation times and proton density. (Klein et al. 2019; Lin and Brown, 2007; McRobbie et al., 2006; Moser et al., 2009) It is very important to determine whether an adnexal mass is benign or malignant, especially in patients of reproductive age, and thus MRI is recommended for patients with an indeterminate ovarian tumor in TVU (Forstner et al., 2017). It has been reported that MRI may serve as a more accurate imaging method than TVU to differentiate pelvic inflammatory disease (Tukeva et al., 1999). MRI is also recommended when there is a suspected ovarian malignancy, if there is a contraindication to CT. An obvious advantage of MRI is the avoidance of ionizing radiation. MRI is reported to have 91.9% sensitivity and 88.4% specificity in ovarian cancer staging (Dodge et al., 2012). In the detection of peritoneal dissemination, gadolinium enhanced MRI was better than CT with overall sensitivity 95% and 92%, respectively (Tempany et al., 2013). Findings that predict malignancy in ovarian tumors are described in Table 7. Covering the entire abdomen with high resolution images by MRI may be difficult, which is a limitation of this modality. Other limitations of MRI include longer examination and interpretation times, higher costs and limited availability.

Table 7. Imaging criteria for malignancy. (Forstner 2016; Mohaghegh 2012)

Size >4cm

Morphology Complex solid and cystic

Solid enhancing component Thick septations >3mm Papillary projections Central necrosis

Vascularization Type 3 dynamic contrast curve

Additional findings Ascites

Lymph node enlargement Peritoneal carcinomatosis Organ invasion

2.2.4.1 Imaging protocol

The MRI protocol in ovarian cancer imaging typically includes rapid multiplanar (axial, sagittal, coronal) T2-weighted (T2W) images and fluid sensitive T2W images with fat suppression, and pre- and post-contrast T1-weighted images. Field strengths of 1.5 T and 3 T are frequently used in ovarian cancer studies. Recently, concerns have been raised about the accumulation of gadolinium contrast agent inside the brain and there is an ongoing debate about its safety (Kanda et al., 2014; Pasquini et al., 2018). The addition of diffusion weighted sequences to conventional MRI imaging

(44)

increases the discrimination without the need for administration of a contrast agent.

Figure 4 illustrates conventional MRI in ovarian cancer.

Figure 4. An 81-year-old woman with high-grade ovarian cancer stage IV. A) T2-weighted image in the axial plane, B) T1-weighted image in the axial plane, C) T2-weighted image in the sagittalplane and D) T2-weighted image in the coronal plane.

2.2.5 Diffusion weighted magnetic resonance imaging

Diffusion weighted magnetic resonance imaging (DWI-MRI) is based on the movement of water molecules in tissues, so-called Brownian motion. It offers improved contrast in comparison to traditional MRI with T1W and T2W sequences.

(Patterson et al., 2008). The different gradient amplitudes, duration, time intervals, b- value measured as sec/mm² between diffusion gradients and even different scanners are all factors that influence the intensity of the detected signal in DWI-MRI.

Parametric apparent diffusion coefficient (ADC) maps are reconstructed with different b-values and this enables the quantitative measurement of ADC values.

(Chenevert et al., 2014; Koh and Collins, 2007). ADC maps can be reconstructed with either mono-, bi- or multi-exponential models.

(45)

ADC is affected by both molecular diffusion and blood flow: low b-values reflect fast diffusion (ie. motion of water molecules within the tissue and also the capillary perfusion in the tissue). High b-values reflect slow diffusion that is not influenced by capillary perfusion. (Le Bihan et al., 1986) Intravoxel incoherent motion (IVIM) imaging is based on multiple b-values and models DW signal as the sum of two exponential curves. It enables the quantification of the relative contribution of perfusion and true molecular diffusion to the random motion of water molecules within biological tissue (Dixon et al., 1988; Le Bihan et al., 1988).

Areas that are highly cellular, such as many malignant tumors, give rise to low ADC values because of restricted water movements; while high ADC values are typical to benign lesions that contain low cell density. Other factors that influence ADC values are fluid viscosity, membrane permeability, macromolecular structures and blood flow (Chenevert et al., 2014; Le Bihan et al., 1988). DWI-MRI has been utilized at first in neurological imaging (Hagmann et al., 2007, Warach et al., 1992) and later in the diagnosis and characterization of different cancers (Gu et al., 2019; Le Bihan, 2019; Pandey et al., 2019; Partridge et al., 2017; Sumi and Nakamura, 2018;

Tamada et al., 2017; Toivonen et al., 2019). Figure 5 demonstrates DWI-MRI features in ovarian cancer.

Figure 5. An 86-year-old woman with serous papillary ovarian adenocarcinoma stage IIIC, high-grade. A) T2-weighted image in the axial plane, B) T2W image with fat suppression in the axial plane, C) Diffusion weighted image where the tumor is bright and D) ADC map where the solid tumor is shown with a dark color.

Multiparametric magnetic resonance imaging in epithelial ovarian cancer

uef.fi

Dissertations in Health Sciences

AUNI LINDGREN

MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING IN EPITHELIAL OVARIAN CANCER

AUNI LINDGREN

MULTIPARAMETRIC MAGNETIC RESONANCE

IMAGING IN EPITHELIAL OVARIAN CANCER

Auni Lindgren

MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING IN EPITHELIAL OVARIAN CANCER

ABSTRACT

TIIVISTELMÄ

ACKNOWLEDGEMENTS

LIST OF ORIGINAL PUBLICATIONS

CONTENTS

ABBREVIATIONS

1 INTRODUCTION

2 REVIEW OF THE LITERATURE

2.1 EPITHELIAL OVARIAN CANCER

2.2 IMAGING OF EOC