Manuscript Title:

Typical Differentiation of Serous and Non-serous Epithelial Ovarian Cancer by Use of Pelvic Magnetic
ResonanceRadiological Imaging

Running Title:

Distinguish epithelial lesions

Abstract

Introduction: Ovarian cancer has a high incidence and fatality rate, with most prevalent type is
epithelial origin. Serous subtype of epithelial cancer predominates. To choose the therapy, it is
important to differentiate this subtype. Due to the fact that some subtypes, such as high-grade
serous and clear cell carcinoma, are still chemosensitive. While other subtypes become more
resistant. The gold standard is histopathology. This examination calls for careful patient preparation.
Patients are typically diagnosed with comorbidities when they are elderly. Thus, that not every
patient can have a biopsy. Radiological modalities, particularly an Magnetic Resonance Imaging
(MRI), are superior in discriminating tissue compared to CT or ultrasound. The use of the T1
weighted imaging (T1WI) and T2WI sequences can best differentiate between a cystic and solid
lesion. The goal of this study was to use radiological examination to assist in the identification of
ovarian cancer.

Methods: This study used histology and MRI data from patients with ovarian cancer as secondary
data for its cross-sectional design. Data was gathered utilizing electronic medical records in Dr.
Sardjito Hospital between January 2017 and May 2022. The following MRI characteristics are
evaluated: ascites, papillary projection, solid nodule, signal intensity of solid and cystic components,
size, configuration, enhancement of contrast and bilateral of the lesion.

Results: Thirty-eight participants made up the study's sample, and 63% of them had serous subtype.
Bilateral lesion suggested a three times greater likelihood that it was serous ovarian cancer (p 0.02;
binary logistic regression). Age >50 years old and strong enhancement on contrast were also
relevant for separating the serous subtype from other subtypes (enhancement p 0.02; age p 0.044)

Conclusions: A bilateral lesion with a significantly enhanced pattern can be seen on the MR imaging
of a serous subtype of epithelial ovarian cancer. The elderly are also more likely to develop this
cancer.

Keywords: cancer, ovarian, epithelial, serous

INTRODUCTION

Ovarian cancer is the eighth most frequent cancer in women and the 18th most frequent
cancer overall. In 2020, there were around 313,000 new cases of ovarian cancer [1]. After cervical
cancer, ovarian cancer ranks as the second most frequent gynecological malignancy and has the
highest fatality rate in the US [2]. There are different forms of ovarian cancer, with the epithelial
type accounting for between 70 and 90 percent of cases [3,4]. There are many subtypes of the
epithelial type exist, including the endometrioid, serous, mucinous, clear cell, and Brenner's tumor.
The management of cases in patients depends on determining the type of tumor. Some types of
epithelial ovarian cancer are still sensitive of chemotherapy regimen, such as high-grade serous
carcinoma and clear cell carcinoma. While low-grade serous, endometrioid and mucinous carcinoma
may become chemo-resistant [5]. Among other ovarian epithelial cancers, serous
cystadenocarcinoma is the most prevalent type of malignancy with 40% prevalence [6]. This is what
also drives how the samples in this study were grouped. Due to the asymptomatic nature of ovarian
epithelial carcinoma, 75% of patients have advanced disease when they are finally detected.
Although ovarian cancer can develop at any age, the majority of cases are detected in people over
the age of 50 [7].
Each subtype shows characteristicsThis is related to each tumor subtype's characteristics.
According to studies, the mucinous subtype typically manifests as big cystic lesions that are more
likely to metastasize to the abdominal regions, whereas the serous subtype frequently exhibits
several tiny lesions [84]. Of course, this stage cannot be separated from the radiological imaging
strategy. Imaging techniques like MRI, CT scans of the pelvis, and ultrasonography are frequently
used to diagnose gynaecological organ cancers. Unlike other imaging modalities, such as
ultrasonography and CT, some physical characteristics of borderline lesions can be picked up by MRI.
For depicting anatomical structures and tissue characteristics, T1WI and T2WI sequences work well.
A malignant lesion is visible on the DWI sequence which shows strong signal, meanwhile. Due to the
fact that the research samples employed in this study were all malignant tumors, it was obvious that
it was not helpful [6].
Histopathological analysis remains to be the gold standard for ovarian cancer diagnosis. The
preparation of the patient's condition must be of utmost importance in this regard because this
examination necessitates an aggressive technique in efforts to retrieve the tissue. In patients who
are elderly or have comorbidities such heart disease, diabetes mellitus, kidney failure, and other
illnesses, it may be challenging to identify the perfect patient condition. Meanwhile, radiological
examinations often do not require invasive procedures, only in contrast injection procedures,
patients usually receive intervention. The peripheral veins need to be accessed for this procedure.
The components of each subtype of this epithelial type of cancer could be thoroughly described by
MRI. Thus, this study is expected to be able to determine the characteristic differences in MRI
images based on the type of ovarian epithelial malignancy.

METHODS
This study is cross-sectional and uses secondary data from ovarian cancer patients'
electronic medical records in the form of imaging information and histological findings. Patients with
a history of ovarian cancer who had at least one MRI of the pelvic region performed in Dr. Sardjito
General Hospital comprised the study population. Ascites and an ovarian mass/tumor were
discovered during the examination. Ascites, papillary projection, solid nodules, signal intensities of
solid and cystic components, size, configuration, contrast enhancement, and side(s) of the lesion
were the MRI parameters assessed in this study.
These patients' histopathological analyses revealed ovarian epithelial carcinoma with serous,
mucinous, endometrioid, or clear cell subtypes. Thereafter, patients will be grouped into serous and
non-serous due to the uneven number of samples. The MRI examination time frame was from
January 2017 to May 2022. Without randomization, samples were collected one after the other until
the required minimum sample size of 38 subjects was reached. These factors served as the study's
inclusion criteria as well. Exclusion criteria for patients included histopathological diagnosis obtained
by cytological examination, nodal metastatic examination and histopathological examination other
than large surgical excision of tissue and frozen sections of tissue, as well as histopathological
examination results showing non-epithelial malignancy. Based on the inclusion and exclusion
criteria, there were 25 patients with a total sample of 38 research subjects because a total of 1 53
patients showed bilateral lesions on both ovaries. Some of these bilateral lesions exhibit different
subtypes in the same patient.
When interpreting MRI scans of samples, it is important to consider the age factor as well as
the bilaterality sides of lesions, size, configuration, papillary projection, solid nodules, intensity of
solid and cystic components, pattern of enhancement, and presence of ascites.
A radiologist served as an intraobserver for this study's re-reading of the MRI image
interpretation in order to evaluate reliability. Chi-square tests and binary logistic regression were
used for the descriptive and inferential data analysis. A relationship between two or more
dependent variables and the independent variables can be seen using the chi-square test. Nominal
data is what was used.
Multivariate analysis aims to determine the relationship between one or more independent
variables with one or more dependent variables. Binary logistic regression, meanwhile, can evaluate
the strength of the link between a predictor (independent) and a response (dependent) variable
with the dependent variable was a dichotomous categorical variable. From the two inferential tests,
values will be obtained in testing the hypothesis. The smaller the value of p, the stronger the
evidence to reject the null hypothesis.

RESULTS

The kappa score of 0.78 from the reliability test on ten of the 38 subjects in this study
indicates that the instruments used in this investigation had good reliability in assessing research
variables.
By analyzing the tumor's properties based on the acquired MRI images, the fundamental
characteristics of the study sample were ascertained. Following data collection and descriptive
analysis, absolute and relative frequency distribution data (in percentage form) are produced, and
Table 1 shows these results.

2
 Table 1. Baseline characteristics

Serous Mucinous Endometrioid Clear Cell Mix type

n(%) n(%) n(%) n(%) n(%)
Number of subjects 24(63.1%) 3(7.9%) 4(10.5%) 5(13.2%) 2(5.3%)
Age
 ≤ 50 y o 9(37.5%) 2(66.7%) 2(50%) 4(80%) 2(100%)
> 50 y.o 15(62.5%) 1(33.3%) 2(50%) 1(20%) 0(0%)
BilateralitySide(s) of
lesion
Bilateral 22(91.7%) 2(66.7%) 0(0%) 4(80%) 2(100%)
Unilateral 2(8.3%) 1(33.3%) 4(100%) 1(20%) 0(0%)
Tumor size
< 6 cms 6(25%) 0(0%) 3(75%) 1(20%) 1(50%)
>= 6 cms 18(75%) 3(100%) 1(25%) 4(80%) 1(50%)
Tumor configuration
Solid 7(29.2%) 2(66.7%) 1(25%) 1(20%) 0(0%)
Solid-cystic 11(45.8%) 0(0%) 3(75%) 1(20%) 2(100%)
Uniloculated cyst 2(8.3%) 0(0%) 0(0%) 1(20%) 0(0%)
Multiloculated cyst 4(16.7) 1(33.3%) 0(0%) 2(40%) 0(0%)
Papillary projection
Yes 15(62.5%) 3(100%) 3(75%) 2(40%) 2(100%)
No 9(37.5%) 0(0%) 1(25%) 3(60%) 0(0%)
Solid nodule
Yes 18(75%) 2(66.7%) 4(100%) 2(40%) 2(100%)
No 6(25%) 1(33.3%) 0(0%) 3(60%) 0(0%)
Signal in solid
component
T1 hypo-T2 hyper 1(4.2%) 0(0%) 0(0%) 0(0%) 0(0%)
T1 iso-T2 hyper 19(79.2%) 2(66.7%) 3(75%) 2(40%) 2(100%)
T1 iso-T2 hypo 0(0%) 1(33.3%) 0(0%) 1(20%) 0(0%)
T1 hyper-T2 iso 0(0%) 0(0%) 0(0%) 1(20%) 0(0%)
Inhomogen 1(4.2%) 0(0%) 1(25%) 0(0%) 0(0%)
T1 & T2 iso 3(12.5%) 0(0%) 0(0%) 1(20%) 0(0%)
Signal in cystic
component
T1 hypo-T2 hyper 14(58.3%) 3(100%) 2(50%) 2(40%) 1(50%)
T1 iso-T2 hyper 6(25%) 0(0%) 0(0%) 0(0%) 0(0%)
T1 iso-T2 hypo 1(4.2%) 0(0%) 1(25%) 0(0%) 0(0%)
T1 hyper-T2 iso 1(4.2%) 0(0%) 1(25% 1(20%) 1(50%)
Inhomogen 1(4.2%) 0(0%) 0(0%) 1(20%) 0(0%)
T1 & T2 iso 1(4.2%) 0(0%) 0(0%) 1(20%) 0(0%)
Contrast enhancement
Non-contrast 1(4.2%) 0(0%) 0(0%) 3(60%) 0(0%)
No enhancement 2(8.3%) 0(0%) 0(0%) 0(0%) 0(0%)
Mild 1(4.2%) 1(33.3%) 0(0%) 1(20%) 0(0%)
Moderate 0(0%) 0(0%) 3(75%) 0(0%) 0(0%)
Strong 20(83.3%) 2(66.7%) 1(25%) 1(20%) 2(5.3%)
Ascites
Yes 11(45.8%) 0(0%) 1(25%) 2(40%) 2(100%)
No 13(54.2%) 3(100%) 3(75%) 3(60%) 0(0%)
 According to the table, a total of 24 subjects, or 63.2% of all samples, demonstrated the
serous tumor type. Sixty two point five percents of the serous type samples in the age category had
an age more than 50. More than 50% of the tumor samples for both mucinous and clear cell types
were between the ages of 31 and 50. Between the age groups of 31 to 50 years and above 50 years,
the endometrioid type exhibits the same number.
The bilaterity variable showed that 91.7%, 66.7%, and 80% of serous, mucinous, and clear
cell tumor types were bilateral lesions, respectively. The three different tumor types also
demonstrated that 75%, 100%, and 80% of lesions were 6 cm or more.
Seventy five percents of endometrioid types had solid cystic shape, while 66.7% had
mucinous with solid cystic predominance. Serous tumors reveal that 45.8% of all tumor types are
solid-cystic lesions, whereas the remaining 29.2% are solid-dominant lesions.
Serous Non-serous
p
(n=24) (n=14)
Signal in solid
component
T1 hypo-T2 hyper 1 0
T1 iso-T2 hyper 19 9
0.271***
T1 iso-T2 hypo 0 2
T1 hyper-T2 iso 0 1
Inhomogen 1 1
T1 & T2 iso 3 1
Signal in cystic
component
T1 hypo-T2 hyper 14 8
T1 iso-T2 hyper 6 0
0.265***
T1 iso-T2 hypo 1 1
T1 hyper-T2 iso 1 3
Inhomogen 1 1
T1 & T2 iso 1 1
Contrast
enhancement
Non-contrast 1 3
No enhancement 2 0
0.02**
Mild 1 2
Moderate 0 3
Strong 20 6
Ascites
Yes 11 5
0.735*
No 13 9
Most serous, mucinous, and endometrioid cancers have papillary projections and solid
nodules. Projections for papillary growth are 62.5%, 100%, and 75%. Nodules that are 100%, 66.7%,
and 75% solid.
Table 2. Characteristics of MR images; serous vs non serous lesions

‍
Age 0.044***
≤ 50 y.o. 9 10
> 50 y.o. 15 4
Side(s) of lesion 0.034*
Bilateral 22 8
 Unilateral 2 6
Tumor size 0.712*
< 6 cms 6 5
>= 6 cms 18 9
Tumor configuration 0.986***
Solid 7 4
Solid-cystic 11 6
Uniloculated cyst 2 1
Multiloculated cyst 4 3
Papillary projection 0.728*
Yes 15 10
No 9 4
Solid nodule 1.000*
Yes 18 10
No 6 4
Signal in solid component 0.271***
T1 hypo-T2 hyper 1 0
T1 iso-T2 hyper 19 9
T1 iso-T2 hypo 0 2
T1 hyper-T2 iso 0 1
Inhomogen 1 1
T1 & T2 iso 3 1
Signal in cystic component 0.265***
T1 hypo-T2 hyper 14 8
T1 iso-T2 hyper 6 0
T1 iso-T2 hypo 1 1
T1 hyper-T2 iso 1 3
Inhomogen 1 1
T1 & T2 iso 1 1
Contrast enhancement 0.02**
Non-contrast 1 3
No enhancement 2 0
Mild 1 2
Moderate 0 3
Strong 20 6
Ascites
Yes 11 5 0.735*
No 13 9

Note: *Fisher Exact test, ** Mann-Whitne test, *** Chi-square test with expected cells < 5 no more than 20%

2
 Figure 1. Bilateral lesion from 2 ovaries. Both showed histopathologically as serous ovarian cancer
(arrow)

Table 2 displays the results of the Chi-square test for each variable. The Fisher Exact's test
was employed as an alternative for the variables for bilaterality, tumor size, papillary projection,
solid nodules, and ascites. The bilaterality variable revealed, which was statistically significant at p
0.034. An ordinal variable is the pattern of contrast enhancement. The Mann-Whitney test was used
as an alternative Chi-square test results. This enhancement variable has a p-value of 0.02. The age
variable has a value of 0.044 by Chi-square, making both variables significant.
Test with binary logistic regression using a dichotomous Y response variable. In this study
categorized into serous and non-serous. Meanwhile, the predictor variable X was defined as
bilaterality, enhancement pattern and age. The first step in the binary logistic regression test is to
perform a model fit test. This is useful to know a model without insignificant variables is the best
model. This test uses the Hosmer and Lemeshow test to assess suitability, where p 0.666, indicating
greater than  so that it is concluded that the model is hypothesized according to the observation
data.
The next step is to determine the estimated parameters. After the parameter estimation
results are obtained, the significance of the response variable is tested. Testing can be carried out
simultaneously or partially/individually. The statistical test used to test the significance of the
simultaneous regression parameters is the X2 test by looking at the Model Coefficient Omnibus
table, where H0 is rejected if the value p <  (for  = 0.05), which shows X2 = 14.159 and the p-value
0.003 means that there is one variable predictor (X) that statistically significantly affects the
response variable.
Partial test of logistic regression parameters using Chi-square, H0 is rejected if the value p <
. In this study there was only 2 significant parameters, these are the bilaterality of the lesions with
p 0.02 and age of subjects with p 0.036.

3
 Table 3. Variable Significance Test in Binary Logistic Regression

Variable B p
Age -2,398 0,036
Pattern of tumor
-0,167 0,565
augmentation
Bilaterality Side(s)
3,071 0,020
of the lesion
Constant -0,379 0,822

Based on Table 3 above, the regression equation can be formed as follows:

Y = A + B1X1 + B2X2 + B3X3
JT = -0,379 – 2,398U – 0,167PP + 3,071BL
A constant of -0,379 means that if there is no difference in age, pattern of tumor
augmentation and bilaterality of the lesion, there will be no difference in the type of serous and non-
serous malignancy. A negative constant means that the probability is considered 0 or there is no
chance. The regression coefficient of the bilaterality variable is 3.0 which means that whenever there
is bilaterality of the lesion, there is a possibility of a serous malignancy of more than 3 times.
At the end of the analysis using this regression method, it was also assessed to what extent
the differences in the Y variable could be explained by the X predictor variable by looking at the
Nagelkerke R-square. In this study, a value of 0.425 was obtained, which means that 42.5% of the
differentiation of variable Y in the form of serous or non-serous epithelial malignancy subtypes can
be explained by predictor variables.

A B Courtesy of Dr. Sardjito Hospital

Figure 2. (A) There is a blurry nodule at the edge of the cystic wall (B) Lesion is clearly seen with
strong enhancement in contrast addition (arrow)

DISCUSSION

Data from MRI image results dating back to 2017, the first year of Dr. Sardjito General
Hospital, up until May 2022, were used in this research. The results of exams conducted by the
Department of Anatomical Pathology FKKMK UGM, as well as results of examinations conducted
outside the hospital, were used to acquire histopathological data. Without randomization, samples
were collected one after the other until the required minimum sample size of 38 subjects was

4
reached. Serous subtypes made up the 24 subjects. This is in line with data on the incidence of
epithelial tumors, the majority of which are high grade serous epithelial tumors (high grade serous
carcinoma), accounting for up to 70% of cases, then endometrioid by 10%, clear cell 10%, mucinous
3% and low grade serous less than 5% [95].

The majority of the samples in the age group were discovered to be older than 50. This is
also consistent with the findings of earlier studies [76], which found that while ovarian cancer can
develop at any age, the majority of cases occur in people who are older than 50. Early age at
menarche and late age at menopause increases risk by increasing the number of ovulatory cycles [9]

In the Chi-square test, the variables of bilateralityside(s) of lesion, pattern of

enhancementcontrast enhancement and age showed were considered statistically significant.
Similarly, in the multivariate test, in which tumor bilateral tumor bilaterality was a strong predictor
that a lesion was a serous malignancytumor. This is consistent with other studies showing that
serous and mucinous epithelial tumors malignancy are bilateral lesions with thin relatively small
wallslesion [107]. Different gene expression is seen during the developing phase in an ovarian cancer
with bilateral lesions, according to chromosomal analysis. While it is claimed that bilateral ovarian
cancer has the same genesis based on the prior understanding. This further demonstrates the
likelihood that the types of tumors on either side of the ovary can vary [118].

The variables of age and contrast enhancement pattern also have significant values. Lesions
with bilateral tumor features and show strong augmentation, and occur in patients over 50 years of
age suggest a serous epithelial malignancy.

Delivery and retention of the contrast by the lesion are necessary for ovarian mass
enhancement. The accumulation of contrast within the bulk and its increased amplification are
caused by the vascular supply, capillary threads, and leakage of contrast into the interstitial
(extravascular) area. Technical limitations like the procedure for contrast injection and delays in
taking scan images might also have an impact on variations. It is well known that angiogenesis
frequently occurs in cancer situations and aids in the growth of tumors. Vascular endothelial growth
factor (VEGFR-2) regulates angiogenesis and vascular permeability, and invasive malignant lesions
are thought to express more of it than benign lesions [129].

Rare reports of ovarian cancer imaging characteristics exist. The histopathological picture is
examined by the anatomical pathology (PA) department after the morphological picture has been
reported from the results of surgeries. Rarely is the macroscopic appearance of the lesion discussed
in reports from the PA. The specimen that was sent to PA could, however, be a lesion that is
incomplete or merely a small portion of the entire lesion, thus this report could simply be
hypothetical. The position and specifics of the lesion can be seen clearly with an MRI scan, though.

Serous epithelial carcinoma is the most prevalent lesion of an ovarian malignancy, and other
primary cancers of the ovarian epithelium include mucinous, endometrioid, and clear cell subtypes.
Debulking surgery normally comes first in these patients' primary treatment plans, which are then
followed by chemotherapy with cytotoxic drugs. Frequently, the findings indicate that the majority
of patients will fare well. The terms low-grade serous carcinoma (LGSC) and high-grade serous
carcinoma are widely used to describe these serous cancers (HGSC) [510]. A borderline lesion that
frequently has papillary projections gives rise to several LGSCs. HGSC, however, is typically found as

5
a solid nodule. Additionally, compared to the mucinous subtype, the serous epithelium malignancy is
said to have a smaller look. This subtype often exhibits large multiloculated lesions [1110].

Figure 3. Big cystic lesion showed small part of solid lesion at its wall. Histologically showed high-grade serous carcinoma

Many have discovered that endometrioid and clear cell carcinomas are intimately associated
to the malignant transformation of endometriosis in subsequent advances to date. Clear cell and
mucinous cell subtypes are believed to be resistant to chemotherapy, although endometrioid
malignancy itself is thought to be a susceptible subtype [105].

Research on ovarian cancer using radiological modalities, particularly MRI has not been done
much in Indonesia. This study can certainly be a starting point to further test the validity of MRI’s
ability to determine ovarian epithelial cancer subtypes. So that later it can help in the approach to
the diagnosis of ovarian cancer.

There are undoubtedly some shortcomings in this research. Researchers find it challenging
to manage the homogeneity of the MRI modalities utilized, such as the type of MRI localization
employed, pelvic, abdominal, or lumbar MRI which also displays intrapelvic organs, thanks to data
obtained with secondary data. It is also challenging to distinguish between the 1.5 Tesla and 3 Tesla
MRI devices that are employed. Even though the research sample exceeded the required minimum
sample size, the findings were unbalanced since there were substantial numbers of differences
across each subtype.

CONCLUSIONS

The typical bilaterality of the pelvic MRI data indicates that the lesion is three times more
likely to be a serous ovarian epithelium cancer than a non-serous ovarian epithelium cancer. The
pattern of significant contrast enhancement on the pelvic area MRI images and the patient's age of

6
above 50 years can further help distinguish a serous epithelial ovarian cancer from a non-serous
tumor.

DECLARATIONS

Competing interest

The authors declare no competing interest in this study.

Ethics approval and consent to participate

This research approved by Medical and health Research Ethics Committee (MHREC), Faculty of
Medicine, Public Health and Nursing, Universitas Gadjah Mada – Dr. Sardjito General Hospital on
April 28th 2022 with reference number KE/FK/0538/EC/2022

Funding
The authors have no financial or personal relationship with any third party whose interests could be
positively or negatively influenced by the article’s content. This research did not receive any specific
grant from funding agencies in the public, commercial, or not-for-profit sectors.

Acknowledgment

The Authors wish to thank dr. Retno Sutomo, dr. Yana Supriatna, dr. Bambang Supriyadi for their
supports and reviews to this research.

7
REFERENCES
1. International WCRF. Ovarian cancer statistics | World Cancer Research Fund International
[Internet]. 2022 [cited 2023 Feb 21]. Available from:
https://www.wcrf.org/cancer-trends/ovarian-cancer-statistics/

2. Desai A, Xu J, Aysola K, Qin Y, Okoli C, Hariprasad R, et al. Epithelial ovarian cancer: An

overview. World J Transl Med [Internet]. 2014 [cited 2021 Dec 27];3(1):1. Available from:
/pmc/articles/PMC4267287/

3. Momenimovahed Z, Tiznobaik A, Taheri S, Salehiniya H. Ovarian cancer in the world:

Epidemiology and risk factors. Int J Womens Health [Internet]. 2019;11:287–99. Available
from: http://doi.org/10.2147/IJWH.S197604

4. Andrew E Green M. Ovarian Cancer: Practice Essentials, Background, Pathophysiology

[Internet]. Medscape. 2021 [cited 2022 Sep 17]. Available from:
https://emedicine.medscape.com/article/255771-overview

5. Matulonis UA, Sood AK, Fallowfield L, Howitt BE, Sehouli J, Karlan BY. Ovarian cancer. Nat
Rev Dis Prim. 2016;2:1–22.

6. Foti PV, Attinà G, Spadola S, Caltabiano R, Farina R, Palmucci S, et al. MR imaging of ovarian
masses: classification and differential diagnosis. Insights Imaging. 2016;7(1):21–41.

7. Aluloski I, Tanturovski M, Jovanovic R, Kostadinova-Kunovska S, Petrusevska G, Stojkovski I,

et al. Survival of Advanced Stage High-Grade Serous Ovarian Cancer Patients in the Republic
of Macedonia. Open Access Maced J Med Sci [Internet]. 2017 Dec 15 [cited 2021 Dec
27];5(7):904. Available from: /pmc/articles/PMC5771292/

8. Aslam Sohaib SA, Sahdev A, Van Trappen P, Jacobs IJ, Reznek RH. Characterization of adnexal
mass lesions on MR imaging [Internet]. Vol. 180, American Journal of Roentgenology. 2003.
Available from: www.ajronline.org

9. Reid BM, Permuth JB, Sellers TA. Epidemiology of ovarian cancer: a review. Cancer Biol Med
[Internet]. 2017;14(1):9–32. Available from: www.cancerbiomed.org

10. Tanaka YO, Okada S, Satoh T, Matsumoto K, Oki A, Saida T, et al. Differentiation of epithelial
ovarian cancer subtypes by use of imaging and clinical data: a detailed analysis. Cancer
Imaging. 2016;16(1):1–9.

11. Smebye ML, Haugom L, Davidson B, Trope CG, Heim S, Skotheim RI, et al. Bilateral ovarian
carcinomas differ in the expression of metastasis-related genes. Oncol Lett [Internet].
2017;13(1):184–90. Available from: http://cancer.sanger.ac.uk/census/,

12. Pannu HK, Ma W, Zabor EC, Moskowitz CS, Barakat RR, Hricak H. Enhancement of Ovarian
Malignancy on Clinical Contrast Enhanced MRI Studies. ISRN Obstet Gynecol. 2013 Feb
13;2013:1–8.