MINISTRY OF EDUCATION
Hanoi medical university
dOAN TIEN LUU
STUDY THE VALUE OF MR IMAGING
IN OVARIAN CANCER DIAGNOSIS
Specilised : Radiology
Code : 62720166
Summarise of thesis of Philosophy doctor
Hanoi - 2019
Thesis made in hannoi medical university
1. Ass Prof. BUI VAN LENH
2. Ass Prof. VU BA QUYET
Ass Prof. Thai Khac Chau
Ass Prof. Nguyen Dinh Tuan
Ass Prof. Lam Khanh
Thesis will be protected in congress university level
of Hanoi Medical University
Thesis will be found in:
- National library
- Library of Hanoi medical university
Researchs publised concerning
to the thesis
Đoàn Tiến Lưu, Bùi Văn Lệnh, Vũ Bá Quyết (2019), “Nghiên cứu áp
dụng thang điểm cộng hưởng từ trong chẩn đoán khả năng ác tính của
u buồng trứng”, Tạp chí Y học thực hành, số tháng 1 năm 2019
(1089), trang 86 - 90.
Đoàn Tiến Lưu, Bùi Văn Lệnh, Vũ Bá Quyết (2019), “Nghiên cứu
giá trị của xung cộng hưởng từ động học sau tiêm thuốc đối quang từ
trong chẩn đoán phân biệt u buồng trứng ác tính với u buồng trứng lành
tính”, Tạp chí Y học thực hành, số tháng 3 năm 2019 (1098), trang 23 –
Ovarian cancer is the third most common cancer, after cervical
cancer and endometrial cancer, but is the leading cause of death among
female genital cancers. The disease has a poor prognosis because the
majority of cases are detected late. Most cases have pelvic invasion and
peritoneal metastases when detected. To change the prognosis, early
detection and proper treatment should be made.
Recently, with the development of new magnetic resonance pulse
chains (CHT), it has faster shooting time, better resolution of images,
analysis of many characteristics of tumor tissue, differentiating cancer
tissue from benign tissue. At the same time, it is easier to surveyed the
entire abdominal cavity to detect peritoneal metastases, lymph node
metastasis. Magnetic resonance (MR) imaging is increasingly showing
advantages in definitive diagnosis and diagnosis of ovarian cancer stage.
There have been studies of the value of MR imaging in the diagnosis
of ovarian cancer in developed countries. However, in Vietnam, there are
no studies, especially studies on MR machines with advanced software
for diagnosing ovarian cancer. So we conducted a research on the topic
"Study the value of MR imaging in diagnosing ovarian cancers" with the
following three research objectives:
1. Describe the MR imaging characteristics of ovarian cancer.
2. Evaluate the value of MR imaging in the differential diagnosis of
ovarian cancer with benign ovarian tumors.
3. Evaluate the value of MR imaging in the diagnosis of ovarian
* Urgency of the project: Finding out a medical imaging modality
which has hight value in diagnosis of ovarian cancers and staging ovarian
cancers. MR imaging with 1,5T machines, many new sequences, has
hight potential in accurately differentiating between ovarian cancers and
benign ovarian tumors, in staging ovarian cancers.
* New contributions of the thesis: This thesis is the first
Vietnamese research of values of MR imaging in diagnosis of ovarian
cancers. The study outcomes showed efficacy of MR imaging in
diffentiating between ovarian cancers and benign ovarian tumors, and in
staging ovarian cancer. Ovarian cancer’s tissues are always restricted
diffusion (hight intensity on DW-b1000). Cutt-off of ADC value ≤
1,26x10-3 mm2/s has moderate value in diffentiating between ovarian
cancers and benign ovarian tumors. Ovarian tumor’s tissue with
enhanced curve type II or type III is the most important characteristics in
diagnosis of ovarian cancers. Diffusion imaging (DW-b1000) has hight
sensitive in discovering peritoneal metastases.
This thesis covers 131 pages, including: preamle (2 pages), chapter
1: The Overview (37 pages), chapter 2: Material and method (19 pages),
chapter 3: Study outcomes (28 pages), chapter 4: Discussion (42 pages),
Conclusions (2 pages), Recommendation (1 page). The thesis consists of
35 tables, 3 charts, 3 diagram, 29 figures. There are 105 references, of
which 5 in Vietnamese and 100 in English.
CHAPTER 1: OVERVIEW
1.1. GENERAL ON OVARIAN CANCERS
1.1.1. Epidemiology of ovarian cancers
Ovarian cancer represents the sixth most commonly diagnosed
cancer among women in the world, and causes more deaths per year than
any other cancer of the female reproductive system. it accounts for about
4% of all cancers in women. It is the third common after cervical cancer
and endometrial cancer.
An estimated 1 in 70 women in the United States will develop
ovarian cancer in their lifetime. On a worldwide basis, an estimated
204,000 new cases are diagnosed and 125,000 women die of ovarian
cancer annually. In 2007, approximately 22,430 new cases of ovarian
cancer will be diagnosed and 15,280 ovarian cancer-related deaths are
expected in the United States. Mortality is high because women typically
present with late-stage disease when the overall 5-year relative survival
rate is 45%. Thus, the public health burden is significant.
Ovarian cancer affects women in the age 20 - 80 years and older
more frequently than younger women. More than 80% of all ovarian
cancers occur in women in the age more than 40 years old.
1.1.2. Hispathology of ovarian cancers
- Over 90% of ovarian neoplasms arise from the epithelial surface of
the ovary, the rest from germ cells or stromal cells. The epithelial
neoplasms are classified as serous (30–70%), endometrioid (10–20%),
mucinous (5–20%), clear cell (3–10%), and undifferentiated (1%).
- 5% - 10% of ovarian cancers are of germ cell origin, included
dysgerminoma, endodermal sinus tumor, embryonal carcinoma,
choriocarcinoma, malignant teratoma.
- Sex cord-stromal are rare, <5% of ovarian cancers, included
malignant granulosa cell tumor, Sertoli-Leydig cell tumor, fibrosarcoma.
1.1.3. Ovarian cancer staging
Stage I: Tumor confined to ovaries
- IA Tumor limited to 1 ovary, capsule intact, no tumor on surface,
- IB Tumor involves both ovaries otherwise like IA.
- IC Tumor limited to 1 or both ovaries IC1 Surgical spill IC2
Capsule rupture before surgery or tumor on ovarian surface. IC3
Malignant cells in the ascites or peritoneal washings.
Stage II: Tumor involves 1 or both ovaries with pelvic extension
(below the pelvic brim) or primary peritoneal cancer
- IIA Extension and/or implant on uterus and/or Fallopian tubes
- IIB Extension to other pelvic intraperitoneal tissues
Stage III: Tumor involves 1 or both ovaries with cytologically or
histologically confirmed spread to the peritoneum outside the pelvis
and/or metastasis to the retroperitoneal lymph nodes
- IIIA ( Positive retroperitoneal lymph nodes and /or microscopic
metastasis beyond the pelvis) IIIA1 Positive retroperitoneal lymph nodes
only IIIA1(i) Metastasis ≤ 10 mm IIIA1(ii) Metastasis > 10 mm IIIA2
Microscopic, extrapelvic (above the brim) peritoneal involvement ±
positive retroperitoneal lymph nodes
- IIIB Macroscopic, extrapelvic, peritoneal metastasis ≤ 2 cm ±
positive retroperitoneal lymph nodes. Includes extension to capsule of
- IIIC Macroscopic, extrapelvic, peritoneal metastasis > 2 cm ±
positive retroperitoneal lymph nodes. Includes extension to capsule of
Stage IV: Distant metastasis excluding peritoneal metastasis
- IVA Pleural effusion with positive cytology
- IVB Hepatic and/or splenic parenchymal metastasis, metastasis to
extraabdominal organs (including inguinal lymph nodes and lymph nodes
outside of the abdominal cavity)
1.2. MR PROTOCOL
MR imaging is performed with a closed-configuration
superconducting 1.5-T system (Signa HDxT; GE Healthcare).
MR imaging is performed with the patient lying in the supine
position (feet first).
Localizer sequence in the three spatial planes;
Axial T2-weighted single-shot fast spin-echo (SSFSE) sequence,
section thickness 6 mm, interslice gap 0.6 mm, used as second
localiser to identify the longitudinal axis of the uterus in the case
of laterally deviated uterus.
- Sagittal T2-weighted fast spin-echo (FSE) sequence parallel to
the longitudinal axis of the uterus (identified on the previous
SSFSE sequence), section thickness 4 mm, interslice gap 1 mm.
Oblique coronal T2-weighted FSE sequence parallel to the
longitudinal axis of the uterus, section thickness 4 mm, interslice
gap 1 mm.
- Oblique axial T2-weighted FSE sequence perpendicular to the
longitudinal axis of the uterus, section thickness 4 mm, interslice
gap 1 mm.
- Axial oblique fat suppressed T2-weighted FSE sequence, section
thickness 4 mm; interslice gap 1 mm.
- Axial T1-weighted gradient-echo (GRE) sequence in-out (chemicalshift imaging), section thickness 6 mm; interslice gap 0,6 mm.
- Axial DWI SE EPI (TR/TE 3000/74,1; flip angle 90°; section
thickness 5 mm; interslice gap 1 mm.
- Axial oblique T1-weighted 3D gradient-echo liver acquisition
with volume acquisition (LAVA) sequence with fat suppression,
section thickness 3.4 mm; overlap locations −1.7 mm.
After i.v. administration of 0.1 mmol/kg paramagnetic contrast
agent (Dotarem) at a flow rate of 2 ml/s, followed by 20 ml of saline
solution at the same flow rate, the following sequences are acquired:
- Dynamic axial T1-weighted 3D gradient-echo LAVA with fat
suppression, section thickness 3.4 mm, overlap locations
−1.7 mm, 10 sequences acquired in 6 minutes after contrast
OBJECTS AND METHOD
Objects included 184 patients with ovarian tumors (93 patients with
ovarian cancers and 91 patients with benign ovarian tumors).
All the patients got pelvis and abdominal MR imaging to diagnose
ovarian cancers at Radiology department of Hanoi University Hospital,
operated at Oncology department of Hanoi University Hospital and
National hospital of obstetrics and gynecology, from November 2013 to
2.2. Study methods
Prospective, descriptive cross-sectional study.
2.2.1. Study equipments
- MR system 1.5T Signa X (GE Healthcare).
- Paramagnetic contrast agent (Dotarem) 0,5mmol/1ml.
2.2.2. Study design
- Select eligible patients.
- Various MR criterias were evaluated on the basis of several
previously published terms:
+ A purely cystic lesion was defined by the absence of solid tissue
and the absence of internal enhancement after injection and corresponded
to a unilocular cyst or hydrosalpinx, both of which have low T1-weighted
and high T2-weighted MR signal intensities.
+ A purely endometriotic mass was defined as a lesion that displayed
high T1- weighted signal intensity that was greater than or equal to that
of subcutaneous fat, shading on T2-weighted MR images, and no solid
+ A purely fatty mass was defined as a lesion that displayed high T1weighted signal intensity that decreased after fat saturation and that
displayed no solid tissue.
+ Readers also recorded enhancement of the cyst wall, bi- or
multilocularity, and the presence of thickened regular septa or grouped septa.
+ The presence of a solid tissue and its morphology (solid portion,
vegetation, thickened irregular septa) were also evaluated. Then, T2weighted signal intensity within the solid tissue (low or intermediate
compared with that of the outer myometrium) and b = 1000 sec/mm2–
weighted signal intensity within the solid tissue (high b = 1000 sec/mm2–
weighted signal intensity compared with that of serous fluid [urine in the
bladder or cerebrospinal fluid]) were analyzed. As previously
demonstrated, we described lesions that displayed both low T2 and b =
1000 sec/mm2–weighted signal intensity within the solid tissue.
+ Finally, readers analyzed the perfusion-weighted images at a
standard workstation by using the breast or prostate perfusion tool and
selecting two regions of interest—one in the external myometrium and one
in the most enhancing part of any solid tissue. We classified the
enhancement of the solid tissue by using a previously published time–
signal intensity curve classification. A gradual increase in the signal
intensity of the solid tissue, without a well-defined “shoulder,” was defined
as curve type 1. A moderate initial increase in the signal intensity of solid
tissue relative to that of myometrium, followed by a plateau, was defined
as curve type 2. An initial increase in the signal intensity of solid tissue that
was steeper than that of myometrium was defined as curve type 3.
+ The presence of free fluid in peritoneal cavity.
+ Peritoneal implants was also noted: Nodular thickening of the
peritoneum that is restricted diffusion (hight intensisty on DW-b1000) and
enhances after gadolinium chelate injection.
+ Pelvis invasion: Normally, there is a hight intensity interface on
T2W between ovarian tumors and rectum, uterus, blader and pelvis wall.
When we can not see this interface or the interface is irregular or retracted,
it is said that there is pelvis invasion.
+ Metastases lymph nodes: Oval or round in shape, transverse
diameter over 8mm, hyperintensity on DW-b1000, enhances after
gadolinium chelate injection, beside to pelvis vessels and aorta, vena cava
and superior mesenteric vessels.
+ Diagnosis of malignant ovarian tumors based on MR criterias,
comparing MR imaging diagnosis with surgery diagnosis to asses the
values of MR imaging in diagnosis malignant ovarian tumors.
+ Staging ovarian cancers based on MR imaging diagnosis of
peritoneal implants, metastases lymph nodes, pelvis invasion.
+ Calculating the malignancy ratio of tumors according to MR
characteristics, following malignancy ratio to rank ovarian tumors in
Thomassin's ADNEX MR scoring system, the ADNEX MR scoring
system has 5 scores (score 1: no tumors; score 2: benign tumors; score 3:
tumors with low malignancy ratio; score 4: tumors with high malignancy
ratio; score 5: almost certainly malignant).
- Based on MR imaging’s diagnosis of invasion of malignant ovarian
tumors, diagnosis of peritoneal metastases, visceral metastases, lymph
node metastases, to diagnose of pre-operative disease stage. Compare the
results of MR imaging’s diagnosis of ovarian cancer stage with staging
ovarian cancer after surgery and hispathologic results. Calculate the
accuracy of MR imaging in diagnosing ovarian cancer stage.
3.1. General characteristics of the research object
184 cases of ovarian tumors in the study, there were 93 malignant
cases (50.5%), 91 cases of benign tumors (49.5%). Malignant epithelial
neoplasms account for the majority (84.6%), germ cell malignancy 8.3%,
and genital cord cell cancers 7.1%.
Patients with ovarian cancer had an average age of 49.5 ± 14.8 years,
72.1% at age 40 - 69 years.
Clinical signs of malignant ovarian tumors were nonspecific, each of
which showed signs of <50% of cases, but gynecological examination
palpable the masses in 88.7% of cases.
Level of serum CA15 has the sensitivity of 94.9% with stage III, IV
ovarian cancers, 78.9% with stage II ovarian cancers, 57.1% with stage I
3.2. MR imaging’s characteristics of ovarian cancer
Malignant ovarian tumors have an average size of 93.9 ± 37.8mm.
93.6% cases of ovarian cancers had complex structure (tissue and
cystic structure), 3.2% completed tissue structure, 3.2% purely cystic
Malignant ovarian tumors had serous fluid (increased signal on
T2W, decreased signal on T1W) in 84.4% of cases; mosaic fluid,
increased signal on T1W and T2W compartments and decreased signal
on T1W, increased signal on T2W compartments (mucus, protein-rich
fluid) in 10% of cases; chronic bleeding fluid, increased signal on T1W,
T1W fat-saturation, decreased T2W signal in 5,6% of cases.
100% of malignant tumors’ tissues increased signal on DW-b1000,
94.4% of soft tissues had time–signal intensity curve type 2 or type 3,
5.6% of soft tissues had time–signal intensity curve type 1.
3.3. The value of MR in differential diagnosis of malignant ovarian
tumors with benign ovarian tumors
3.3.1. Analysis of MR characteristics to diagnose of malignant
Table 3.1: Multivariate analysis of MR characteristics for diagnosing
malignant ovarian tumors
Size of tumor ≥
tissue on DW
Curve type 2
aOR (CI 95%)
When analyzing univariate, the MR characteristics of ovarian
tumors, size of tumor ≥ 80mm, multilocularity, wall thickness ≥ 3mm,
vegetations, tissues increase signal on DW-b1000, tissue’s ADC value ≤
1,26x10-3 mm2 / s, enhancement with curve type 2 or type 3, peritoneal
metastases, peritoneal cavity’s fluid, all have value in diagnosis of
malignant ovarian tumors.The malignant ratio of tumors with one of the
above characteristics is higher than the malignant ratio of tumors with no
corresponding characteristics, the difference is statistically significant p
<0.05, malignancy odd ratio OR > 1.
However, when analyzing multi-variable Regression Multinomial
Logistic, there are only two characteristics, contrast-enhanced tissue with
curve type 2 or 3 and peritoneal metastases, that are valuable for
diagnosis of malignant tumor, malignancy aOR very high, dynamic range
of odds ratio (CI 95%) always > 1, with p <0.001, other characteristics
have no value to diagnose malignant ovarian tumors, aOR malignancy
rate of these characteristics <1 or dynamic range (CI 95%) of aOR with
oscillation <1, with p> 0.05.
This is explained by the MR characteristics after multivariate
analysis with p> 0.05, the ratio of aOR <1, are all characteristics when
univariate analysis have low sensitivity (19.4% - 77.4%), specificity is
not high (48.4% - 92.3%), low accuracy (55.4% - 71.2%) , the negative
predictive value and the positive predictive value are also not high.
On the contrary, contrast-enhanced tissue with curve type 2 or 3 is
the MR characteristics with high diagnostic value, high sensitivity
91.4%, high specificity 90.1%, when tumors has contrast-enhanced tissue
with curve type 2 or 3, they will often include many other malignant
diagnostic characteristics. Peritoneal metastases is MR characteristics
with low sensitivity but high specificity (reaches 100%), when we
observe peritoneal metastases on MR images the tumors are almost
The results of our research are similar to the results of some other
studies. Thomassin et al. (2013) researched to construct of MR scores in
differential diagnosis of ovarian malignant tumors to ovarian benign
tumors, multivariate analysis also identified only two characteristics of
malignant diagnosis had high value in diagnosis of ovarian malignant
tumors that are contrast-enhanced tissue with curve type 2 or 3 and
peritoneal metastases, other MR characteristic are not valuable when
Other tudies of applying Adnex MR score in the diagnosis of ovarian
cancers, such as Pereira’s study (2018), Sadowski's study (2018), both
studies had results which are similar to our research result, only two
characteristics, contrast-enhanced tissue with curve type 2 or 3 and
peritoneal metastases, are applied into malignant diagnostic criteria.
Figure 3.1: Cancerous tissue with contrast-enhanced curve type 3/Left
ovarian mucinous adenocarcinoma. Cancerous tissue’s curve is purple,
uterus’curve is green, the purle one if steeper than the green one .
Patient: Vu Hong Ch., 44 years old, patient code: 178/13.
3.3.2. Analysis of MR characteristics for diagnosis of benign ovarian
Table 3.2: MR characteristics to diagnose benign ovarian tumors
Benign MR characteristics
Cyst with thick wall, without
Cyst with thin wall, without
Endometriotic cyst, without
Fatty cyst, without contrastenhanced tissue
Tissue’s low intensity on DW
Tissue with curve type 1
Benign ovarian cysts often do not have cell proliferation inside so
they have MR images of pure cysts without tissue, whereas ovarian
malignant cysts have malignant cell proliferation inside, often have MR
images of mix cysts (solid portions inside the cysts). Therefore, ovarian
cysts without tissue are more likely to be benign tumors.
Thin-walled cysts are common, are benign serous cysts, benign
mucous cysts, cysts of ovary, paraovarian cyst, peritoneal inclusion cyst,
these benign cysts are without cell proliferation in the cyst, so they have
Endometriotic cysts are the most common benign cysts in the ovary,
these cyst are high intensity on T1W and T1W-fatsaturate images, low
signal on T2W images (chronic bleeding), thick wall or thin wall, but no
solid portion inside.
In some cases of endometrial carcinoma, clear cell carcinoma also
has chronic internal bleeding, so they are high intensity on T1W and
T1W fat-saturation images, low signal on T2W images, but these cancers
often have solid portions, the solid portions are strongly contrastenhanced.
Pure fatty cystic tumors (high signal on T1W, low signal on T1W
fat-saturation), without contrast-enhanced tissue inside, are benign
teratomas or dermoid cysts. In cases of teratomas with contrast-enhanced
tissue inside may be benign or malignant teratomas, not absolutely
The tissue, which is low signal on T2W and DW, is benign, rich in
fiber, poor in cells, not restricted on DW image, so the tumors with this
tissue are also usually benign.
The tissue, with contrast-enhanced curve type 1, is hypovascular, so this
tissue also has high benign ratio, low malignancy ratio.
Therefore, in our study, thin-walled cyst, endometriotic cyst, fatty
cyst, tissue with low signal on DW were definitely benign with
malignancy ratio of 0%.
Thick-walled cysts, tissue portion with contrast-enhanced curve type
1 are likely benign, with low malignant ratio, 5.3% and 15.2,
respectively. Our research results are similar to other studies. The study
of Thomassin (2013), also identified merely thin-walled cysts,
endometriotic cysts, simple cyst, tumors with tissue of low signal on DW
have malignant ratio of 0%, thick-walled cysts and tissue portion with
contrast-enhanced curve type 1 have low malignant ratios of 4.3 - 4.5%.
Pereira's study (2018), thin-walled cysts, endometriotic cysts, simple
fatty cyst, tumors with tissue of low signal on DW have malignant ratio
of 0%, thick-walled cysts and tissue portion with contrast-enhanced
curve type 1 have low malignant ratios of 5.1%.
Table 3.3: Ovarian malignancy ratio according to the Adnex MR scores
Low possibility of
High possibility of
MR characteristics or ovarian
No ovarian masse
Tissues are low intensity on T2W,
Solid portions with curve type 1
Solid portions with curve type 2
Solid portions with curve type 3
Score 1 is without ovarian masses, in our study all patients have
tumors so there is no case of score 1. Score 2 is definitely benign,
malignancy ratio is 0%. Score 3 is ovarian tumors with low possibility of
malignancy, malignant ratio is 13.0%. Score 4 is tumors with high
malignant possibility, malignant ratio is 82.2%. Score 5 is almost
certainly malignant, malignancy ratio is 100%.
Table 3.4: Value of magnetic resonance in the differential diagnosis of
malignant ovarian tumors with benign ovarian tumors
Kết quả mô bệnh học
We diagnose malignant ovarian tumors if the tumors have solid
portions with contrast-enhanced curve type 2,3 or peritoneal (Adnex MR
score of 4 or 5).
The results of our study showed that MR has high value in the
diagnosis of malignant ovarian tumors, with high sensitivity of 93.5%,
high specificity of 92.3%, high positive predictive value 92.6%, the
negative predictive value is 93.3%, the high accuracy is 92.9%.
Studies of Thomassin (2013) and Pereira (2018) also diagnose
malignant tumors if the tumors have a Adnex MR score of 4 or 5, the
results of these studies are similar to our results, MR is highly valuable in
diagnosing malignant ovarian tumors.
Research result of Thomassin has 93.5% sensitivity, 96.6%
specificity, 96.0% accuracy. Pereira's research result (2018) has 94.9%
sensitivity, 97.5% specificity, 96.6% accuracy.
3.4. The value of magnetic resonance in the diagnosis of ovarian cancer
3.4.1. The value of magnetic resonance in diagnosis of pelvic invasion
We diagnose ovarian cancer’s invasion to the uterine, rectal, bladder
invasion if we find to lost the high signal on T2W planes, which separate
the tumor and the wall of the uterus, bladder wall, rectum wall, or these
planes are irregular or shrinkage.
We diagnose ovarian cancer’s invasion to pelvic wall if we find to
lost the high signal on T2W plane between tumors and muscles, bones of
the pelvic wall or tumor’tissue surrounds more than half of the perineal
pelvic vassel’s perimeter. These characteristics are the standards applied
in some studies of the values of MR in diagnosing ovarian cancer stage.
When these standards were used to diagnose pelvic invasion of ovarian
cancer, we found that MR has low sensitive 56.9%, low specificity
77.1%, quite high false positive rate 19.5%.
The results of our study are similar to other research results, Forstner
et al. (1995) studied the value of MR to diagnose ability of radical
surgery for 50 cases of invasive ovarian cancer. MR diagnosis of pelvic
invasion has low sensitivity 60%, not very high positive predictive value
69%, high false positives rate 31%. Kurst et al. (1999), investigating the
value of MR in the diagnosis of ovarian cancer stage, MR diagnosis of
pelvic invasion with low sensitivity of 58.3%, positive predictive value
of 71.4% , high false positive rate is 28.6%. In many cases, the fat layer
of pelvic wall is thick, so when the tumor has already invaded into pelvic
wall but not yet invaded all the thickness of the fat layer, so on MR
images we still see the fat layer between tumors and muscles and bones.
The plane of separation between the tumor and the rectum wall,
bladder, uterus and ovary is the fat and serosa layer of the organs, in
some cases the invasion does not occupy all the thickness of these layers,
so it is still still observed on MR images. These cases lead to false
negative diagnosis, so MR is not highly sensitive. In contrast, in some
cases, the fat layer is very thin, so when the tumor is close to the muscle,
the bone we can not see this layer but in fact that there is not invasion.
At the same time, in many cases the serous fat layer of the organs in
the pelvis is very thin, although the tumor has not invaded the pelvic
organs but with the resolution of 6mm thickness T2W images we still can
not observe the plane separating between the tumor and the organs. These
cases lead to MR diagnosis with significant false positive rate, low
Figure 3.2: Images of pelvic invasion. Ovarian serous adenocarcinoma
invades the uterus and rectum, lost the fat layer between the tumor and
the uterine wall (white arrow), between the tumor and the rectum wall
(yellow arrow), on T2W image horizontal plane (left), vertical vertical
plane (right image). Patients: Pham Thi T., patient’code: 405/31.
4.4.2. The value of magnetic resonance in large lymph nodes detection
According to the study of Takeshima et al. (2005), most metastatic
lymph nodes of the ovarian cancers are beside the abdominal aorta, and
next to the pelvic arteries, the rate of ovarian cancer with metastatic
lymph nodes accounts for the proportion 13-60% of cases, including
many cases of metastatic lymph nodes but small in size.
However, according to research by Maggioni et al. (2006),
comparing the group of patients with surgical ovarian cancer, having
systemic lymph node surgery with the group of ovarian cancer patients
who had surgery but not systemic lymph node surgery, only large lymph
nodes are taken as a attempt of maximal cutting of tumor, the study
results show that there is no difference in the overall life time of these
two groups of patients.
In our study, surgeons only performed lymph node removal when
there was a large size lymph node, so we only studied the role of MR in
detecting enlarged lymph nodes, the horizontal diameter of lymph nodes
are 8mm or more.
The detection of enlarged lymph nodes before surgery helps
surgeons to find large lymph nodes during surgery to remove them. There
were 10 patients in our study, that surgeons discovered and removed
large lymph nodes, MR detected large lym nodes in 8 cases, sensitivity
reached 80%, specificity was 97.6%.
There were 2 cases who had enlarged lymph nodes but MR did not
detect enlarged lymph nodes. At the same time, there were two cases that
MR diagnosed with enlarged lymph nodes but surgeons did not find in
surgery, the rate of false positive was 20%.
The sensitivity of MR in detection of lymphadenopathy in our study
was lower than some other research results.
Research by Kim J.K et al (2008), 125 cases of cervical cancer, 30
metastatic lymph nodes size ≥8mm, MR detected these nodes with 87%
sensitivity. Harisinghani et al. (2003), studied the value of MR in
detecting metastatic lymph nodes size ≥8mm in patients with prostate
cancer, MR has high sensitivity of 91%.
In our study, two false positive cases are due to the misdiagnosis of
peritoneal metastatic nodules in pelvic wall, near the pelvic arteries as
lymph nodes. In contrast, two false negative cases we misdiagnosed
lymph nodes as peritoneal metastatic nodules. Kim J.K's research
subjects were cervical cancer cases, Harisinghani's research subjects
were prostate cancers. These types of cancers often have metastatic
lymph nodes, but peritoneal metastasis is less common than ovarian
cancer, so it is seldom confused between pelvic lymph nodes and
peritoneal metastases of pelvic wall. So the sensitivity and specificity of
these studies are higher than our study.
3.4.3. The value of MR in diagnosis of visible peritoneal metastasis
Peritoneal metastatic lesions are nodules in the parietal peritoneum
(peritoneum of the diaphragm, colon recess, abdominal wall, pelvic
wall), visceral peritoneum (on the surface of the liver, spleen, intestinal
wall), greater omentum, mesentery.
Peritoneal metastatic lesions are characterized by restricted diffusion
on DWI, the signal of lesions is higher than the signals of abdominal
organs on DW-b1000 image, low signal compared to peritoneal fat on
T1W and T2W images, enhanced on T1W fatsat after contrast injection.
Diffusion weighted (DW) is the basic sequence to detect peritoneal
metastases, recent studies on the value of MR in the diagnosis of
peritoneal metastases are based on DW, diagnosis of peritoneal
metastasis based on high signal characteristics on DW with high b values
(b800, b1000, b1200). Therefore, in the study, we diagnosed peritoneal
metastasis based on high signal of the lesions on DW-b1000. There were
33 cases which had visible peritoneal metastases, MR detected visible
peritoneal metastasis with relatively high sensitivity of 87.9%, accuracy
The results of our research are similar to the results of other
researchs. Low R. et al. (2009), study theo value of DW to diagnose
visible peritoneal metastasis, the sensitivity of MR reaches 90%,
accuracy of 91%.
Tempany et al. (2000), studying comparing the value of diagnostic
methods in ovarian cancer staging, MR has higher value than CT scanner
and ultrasound, MR's sensitivity in diagnosis of visible peritoneum
Figure 3.3: Images of metastases of visceral peritoneum and parietal
peritoneum. Peritoneal metastases have high signal on DW-b1000.
Metastatic nodules of the liver surface (yellow arrows), metastatic
nodules of the peritoneum of abdominal wall (white arrows), metastasis
of greater omentum (red arrow). Patient: Do Thi N., patient’code:
Our study also found that MR sensitivity depends on the size of
There were 5 cases of peritoneal metastasis at microscopic level that
were not clearly observed by eye when surgery, MR did not detect any
cases, the sensitivity was 0%; 17 cases of peritoneal metastasis, observed
in surgery with size ≤10mm, MR detected 13/17 cases, sensitivity
reached 76.5%; 16 cases of peritoneal metastasis with size > 10mm, MR
detected 16/16 cases, sensitivity of 100%.
MR sequences investigated all abdominals with slice thickness of
6mm, which had certain resolution, and there were some image noises
(noise due to movement of the respiratory, peristalsis). So small lesions,
which may have the signal combined with the surrounding structure, will
be difficult to distinguish from the surrounding structures, so the
sensitivity of MR will be lower. In contrast, large-sized lesions are easier
to observe, the signal distinction is clearer, so MR has higher sensitivity
with larger lesions.
3.4.4. The value of MR in the diagnosis of ovarian cancer stage
Table 3.5: MR's value in ovarian cancer staging
Staging by MR
Staging by surgery
MR diagnoses the stage of ovarian cancer based on diagnosis of
pelvic invasion, diagnosis of lymphadenopathy, diagnosis of peritoneal
metastases. The results of our study, MR diagnosed each stage with low
accuracy of 68.8%. The reason that MR diagnosed pelvic invasion with
low accuracy of 64.5%, MR could not diagnose microscopic peritoneal
metastasis, MR only had high accuracy of 95.6% with macroscopic
Other research results in the world, MR diagnoses each stage with
low accuracy, fluctuating 53% - 88%, our research results are also in this
range. Kurtz et al. (1999) compared MR with other diagnostic imaging
methods in diagnosing stages of ovarian cancer, MR had an accuracy of
72.9%. Forstner et al. (2004) studied the value of imaging diagnostic
methods in diagnosing stages of ovarian cancer, MR had accuracy of
78%. The authors all have identified that 1.5 Tesla MR machine and
multi-sequence CT scanner nearly equal value in diagnosis of ovarian
cancer staging, but MR is a method that requires more complex
techniques, to convey the entire abdomen will take longer time. Highermagnetic MR machines (3 Tesla) with new hardware and software for
better image processing can improve these problems, with one sequence
can convey the entire abdomen, faster speed, DW and new other higher
resolution sequences enable to increase accuracy of MR in the diagnosis
of peritoneal metastases and metastatic lymph nodes.
Table 3.6: MR differential diagnosis of stage I, II with stage III, IV
Stage I, II
Stage III, IV
Stage I, II
Stage III, IV
In our study, MR diagnosis of each stage has not high accuracy, but
the differential diagnosis late stages (stage III and IV) with the earlier
stage (phase I, II), MR has a high accuracy of 90.3%.
Kurtz's study and Forstner's study also identified that MR had highvalue in distinguishing ovarian cancer stage I and II from stage III, IV.
Kurtz et al. (1999), MR differentially diagnosed the early stages to the
late stages with accuracy of 90%. Forstner et al. (2004), MR differential
diagnosis of early-stages with late stages with accuracy of 95%. The
differential diagnosis of the late stages with the early stages helps to get
prognosis before treatment.
1. MR imaging’s characteristics of ovarian cancer
Malignant ovarian tumors have an average size of 93.9 ± 37.8mm.