This copy is for personal use only. To order printed copies, contact reprints@rsna.
org
Indeterminate Adnexal Cysts at
US: Prevalence and Characteristics of
Ovarian Cancer1
Elizabeth A. Sadowski, MD
Viktoriya Paroder, MD
Krupa Patel-Lippmann, MD
Jessica B. Robbins, MD
Lisa Barroilhet, MD
Elizabeth Maddox, MD
Timothy McMahon, MS
Emmanuel Sampene, PhD
Ashish P. Wasnik, MD
Alexander D. Blaty, BS
Katherine E. Maturen, MD, MS
1
 From the Departments of Radiology (E.A.S., J.B.R., E.M.),
Obstetrics and Gynecology (E.A.S., L.B.), and Biostatistics
and Medical Informatics (E.S.), University of Wisconsin
School of Medicine and Public Health, Clinical Science
Center, E3/372, 600 Highland Ave, Madison, WI 53792-
3252; Department of Radiology, Montefiore Medical Center,
Albert Einstein College of Medicine, Bronx, NY (V.P., T.M.);
Department of Radiology, Vanderbilt University Medical
Center, Nashville, Tenn (K.P.L.); Department of Radiology
(A.P.W., A.D.B., K.E.M.) and Department of Obstetrics and
Gynecology (K.E.M.), University of Michigan Health System,
Ann Arbor, Mich. Received September 24, 2017; revision
requested November 7; revision received December 6; final
version accepted December 21. Address correspondence
to E.A.S. (e-mail: esadowski@uwhealth.org).
q
 RSNA, 2018
Radiology: Volume 000: Number 0—   2018  n  radiology.rsna.org
Original Research  n  Ultrasonography
Purpose: To assess the prevalence of indeterminate adnexal cysts in
women presenting to academic medical centers for pelvic
ultrasonography (US), determine the incidence of malig-
nancy, and identify cyst and patient characteristics that
are predictive of malignancy.
Materials and A multicenter study of US-detected adnexal cysts with ap-
Methods: propriate follow-up (surgical pathologic examination, imag-
ing and/or clinical examination) was conducted from Janu-
ary 2008 to June 2012. Indeterminate cysts were classified
as category 1 (typical benign appearing cysts .5 cm) or
category 2 (cysts with avascular solid components) on the
basis of a combination of definitions in the existing liter-
ature. The incidence of neoplasms and malignant tumors
was calculated. Patient and cyst characteristics associated
with neoplasm and malignant tumors were evaluated with
the x2 test or Fisher exact test for categorical variables and
the t test for continuous variables. A backward stepwise
logistic regression model was performed for two outcomes:
(a) the presence of any neoplasm (benign or malignant)
and (b) the presence of a malignant tumor.
Results: There were 1637 women with an adnexal cyst at US; 391
(mean age = 41.8 years 6 13.5.1; range = 17–91 years)
had an indeterminate adnexal cyst at US. The prevalence of
indeterminate adnexal cysts was 23.9% (391 of 1637; 95%
confidence interval [CI]: 0.22, 0.26). Three hundred three
indeterminate cysts in 280 women (mean age = 42.9 years
6 14.1; range = 17–88 years) had adequate follow-up. The
incidence of ovarian neoplasms (benign and malignant) was
24.8% (75 of 303 cysts; 95% CI: 0.20, 0.30), and the inci-
dence of malignant tumors was 3.6% (11 of 303 cysts; 95%
CI: 0.02, 0.06). The proportion of ovarian neoplasms dif-
fered between category 1 and category 2 cysts (17.5% [25
of 143 cysts; 95% CI: 0.12, 0.25] vs 31.3% [50 of 160 cysts;
95% CI: 0.24, 0.39], respectively; P = .001). The propor-
tion of malignant tumors differed between categories 1 and
2 cysts (0% [0 of 143 cysts] vs 6.9% [11 of 160 cysts; 95%
CI: 0.03, 0.12]; P , .001). The presence of an avascular
nodular component was a significant predictor of malig-
nancy at stepwise logistic regression analysis (odds ratio =
2.83; P  .0001; 95% CI: 1.69, 4.70).
Conclusion: The presence of an avascular nodular component was the
most significant predictor of the presence of malignancy
in indeterminate adnexal cysts. The risk of malignancy is
higher with category 2 cysts than with category 1 cysts.
q
 RSNA, 2018
1
ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al
A
dnexal cysts are commonly en-
countered in daily radiology
practice and most cysts can be
accurately characterized with ultra-
sonography (US) as likely benign or
malignant, regardless of symptoms
(1–7). When an adnexal cyst has be-
nign features at US, the patient can be
reassured of a very low risk of cancer.
Typical appearing simple cysts have a
0%–1% risk of cancer, and typical hem-
orrhagic cysts, endometriomas, and
dermoids have a 1%–2% risk of cancer
(7–14). Conversely, adnexal cysts with
a vascularized soft-tissue component
are considered to have a high probabil-
ity of ovarian cancer (5,7,15–21). In be-
tween the two ends of the spectrum are
cysts that cannot be definitively charac-
terized as probably benign or possibly
malignant, and they are considered in-
determinate at US. These sonographi-
cally indeterminate cysts have avascular
internal components, such as internal
irregular thick septations or solid-ap-
pearing nodules without blood flow, or
they are otherwise benign-appearing
entities, such as hemorrhagic cyst, en-
dometrioma, or mature teratoma, that
cannot be completely evaluated with US
owing to their large size and/or atypical
features (1,2,12,21–26).
Implications for Patient Care
nn The incidence of neoplasm in
sonographically indeterminate
adnexal cysts was strongly associ-
ated with the cyst characteristics,
in addition to patient age and
menstrual status, when all covari-
ables were considered.
nn The incidence of malignant tumor
in sonographically indeterminate
adnexal cysts was strongly associ-
ated with the presence of avas-
cular nodules within the cyst
when all covariables were
considered.
nn Sonographically indeterminate
adnexal cysts should be followed
up or further evaluated to deter-
mine their benign or malignant
nature and to ensure that the
patient is treated appropriately.
2 radiology.rsna.org  n Radiology: Volume 000: Number 0—   2018
Indeterminate adnexal cysts can-
not be preoperatively classified as be-
nign or malignant and pose a dilemma
for radiologists and clinicians. If these
cysts are referred for gynecologic or
surgical evaluation, there is a risk of
overtreatment. If imaging surveillance
is suggested, there is a risk of losing
the “window” for catching early stage
disease. There is currently a paucity of
data with regard to the risk of ovarian
cancer in sonographically indetermi-
nate cysts. Previous studies that in-
clude sonographically indeterminate
cysts were not designed to study this
question or have a small number of pa-
tients. Furthermore, these studies re-
ported variable frequencies of ovarian
cancer within sonographically indeter-
minate cysts ranging from 5% to 40%
(7,27,28). The goals of our study were
to assess the prevalence of indetermi-
nate adnexal cysts in a general popula-
tion of women presenting to academic
medical centers for pelvic US, to deter-
mine the prevalence of ovarian cancer
within these cysts, and to identify pa-
tient and cyst characteristics that could
be used to predict ovarian cancer.
Materials and Methods
Subjects and Inclusion
Our Health Insurance Portability and
Accountability Act–compliant, retro-
spective, and multicenter study received
institutional review board approval from
four participating academic institutions.
The requirement to obtain informed
consent was waived. Female subjects
at least 17 years of age were included
in the study. The images from pelvic
US examinations from January 2008 to
June 2012 were reviewed by abdominal
radiologists (E.A.S., V.P., K.P., J.R.,
E.M., A.P.W., K.E.M., with 1–16 years
of postresidency experience) for adnexal
cysts that met the following cyst inclusion
criteria based on a combination of exist-
ing literature and depicted in Figure 1
(2,7,29): (a) typical benign-appearing
cysts, including simple-appearing cysts,
hemorrhagic cysts, endometriomas,
and dermoids, measuring at least 5 cm
in greatest diameter (category 1), and
(b) cysts with internal solid component,
including thickened septations, nodules,
and papillary excrescences, without
blood flow (category 2).
The methods for identifying studies
that contained adnexal cysts, including
follicles and nonphysiologic adnexal
cysts, varied among institutions. At
two institutions, abdominal radiologists
(E.A.S., K.P., J.B.R., E.M.) reviewed
all nonobstetric pelvic US examinations
in female patients in consecutive fash-
ion and recorded all patients with ad-
nexal cysts at imaging that met the in-
clusion criteria outlined earlier. At the
other two institutions, the investigators
first searched the text of all nonobstet-
ric pelvic US reports by means of the
radiology information network using
the keywords “ovarian cyst,” “complex
cyst,” “cystic mass,” “adnexal cyst,” “cys-
tic neoplasm,” “ovarian mass,” “ovarian
cancer,” and “ovarian neoplasm.” The
investigators (K.E.M., A.P.W., V.P.)
subsequently reviewed the images from
these examinations to identify adnexal
cysts that met the inclusion criteria
above. A patient inclusion flow diagram
is depicted in Figure 2.
Only patients examined with both
transvaginal and transabdominal tech-
nique, which included the addition of
Doppler evaluation, were evaluated.
Medical chart review was performed
https://doi.org/10.1148/radiol.2018172271
Content codes:
Radiology 2018; 000:1–9
Abbreviations:
CI = confidence interval
OR = odds ratio
Author contributions:
Guarantors of integrity of entire study, E.A.S., V.P., E.M.,
E.S.; study concepts/study design or data acquisition
or data analysis/interpretation, all authors; manuscript
drafting or manuscript revision for important intellectual
content, all authors; approval of final version of submitted
manuscript, all authors; agrees to ensure any questions
related to the work are appropriately resolved, all authors;
literature research, E.A.S., V.P., L.B., E.M., A.P.W., A.D.B.,
K.E.M.; clinical studies, E.A.S., V.P., K.P.L., J.B.R., L.B., E.M.,
T.M., A.P.W., A.D.B., K.E.M.; statistical analysis, E.A.S., E.M.,
E.S., A.D.B., K.E.M.; and manuscript editing, E.A.S., V.P.,
K.P.L., J.B.R., E.M., T.M., E.S., A.P.W., A.D.B., K.E.M.
Conflicts of interest are listed at the end of this article.
ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al

Figure 1

Figure 1:  Chart of the US criteria used to define a category 1 and category 2 sonographically indeterminate cyst. ∗ 5 Criteria
per Levine et al (29). ∗∗ 5 Criteria per Ekerhovd et al (2).

by the investigators (E.A.S., V.P.,
K.P., J.B.R., L.B., E.M., T.M., A.P.W.,
A.D.B., K.E.M.), and only patients with
adequate follow-up, defined as surgical-
pathologic diagnosis, imaging evidence
of cyst resolution or decrease in size,
cyst stability for more than 2 years,
or clinical follow-up with documented
pelvic examination more than 2 years
after initial pelvic US, were included in
our final study population.
Adnexal Cyst Evaluation and Image Data
Collection
US examinations were performed by
using GE Logiq 9 or E9 (GE Healthcare,
Waukesha, Wis) and Phillips iU22
(Philips Healthcare, Andover, Mass)
machines. Syngo Dynamics (Siemens
Healthineers, Mountain View, Calif),
McKesson (McKesson Radiology, San
Francisco, Calif), or GE Centricity (GE
Healthcare, Waukesha, Wis) picture
archiving and communication systems

Radiology: Volume 000: Number 0—   2018  n  radiology.rsna.org 3

ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al
Figure 2
Figure 2:  Flow diagram of final study population.
were used by the abdominal radiologists
(E.A.S., K.P., J.B.R., E.M., K.E.M.,
A.P.W., V.P.) to review images dur-
ing the study. Research image review
was performed from December 2014
to June 2017. All cysts that met the
inclusion criteria were recorded with
documentation of cyst characteristics,
including largest cyst diameter, side of
the pelvis, and presence or absence of
solid components (ie, number of septa-
tions, number and greatest diameter of
nodules and/or papillary excrescences).
Data and Statistical Analysis
Counts and percentage of subjects
were used to summarize categorical
and ordinal data. Means and standard
deviations were used to summarize
continuous data for descriptive statis-
tics purposes. Differences in baseline
characteristics were compared with the
x2 or Fisher exact test for categorical
4 radiology.rsna.org  n Radiology: Volume 000: Number 0—   2018
variables and with the t test for con-
tinuous variables. Evaluation of differ-
ences in age, menopausal status, great-
est diameter of the cyst, number of
neoplasms, and number of malignant
tumors for the various patient outcome
measurements was performed. The
prevalence of sonographically indeter-
minate adnexal cysts in our population
was calculated. Ovarian neoplasms in-
cluded both benign (eg, dermoids, cyst-
adenomas, cystadenofibromas) and ma-
lignant (eg, borderline tumors, serous
carcinomas) tumors. The incidence of
benign and malignant tumors was cal-
culated in the entire cyst population
and in category 1 and category 2 cysts.
We performed a backward step-
wise logistic regression model for the
two outcomes, where the probability to
remove a variable from the model was
set at PR = 0.2. The primary end points
were (a) the presence of any neoplasm
(benign or malignant) and (b) the pres-
ence of a malignant tumor. Covariables
considered for both end points included
age, menopausal status, greatest diam-
eter of the cyst, number of septations,
and number and greatest diameter of
nodular components. Furthermore, we
assessed the goodness-of-fit measures
of the model by reporting the area un-
der the curve. P , .05 was considered
indicative of a statistically significance
difference. Statistical analysis was per-
formed with software (STATA, version
14; Stata, College Station, Tex).
Results
The patient demographic characteris-
tics and cyst dimensions are presented
in Table 1 for all adnexal cysts, category
1 cysts (typical benign appearing cysts
.5 cm), and category 2 cysts (cysts
with avascular internal components).
In our study population, most patients
were premenopausal (71.6%). Of the
303 patients, 145 (47.8%) were in-
cluded on the basis of surgical resection
of the cyst, 149 (49.2%) were included
on the basis of follow-up imaging, and
nine (3%) were included on the basis
of clinical follow-up more than 2 years
after baseline US.
Prevalence and Incidence of
Indeterminate Cysts, Neoplasms, and
Malignant Tumors
The overall prevalence of indeterminate
adnexal cysts in our entire population
was 23.9% (391 of 1637 cysts; 95%
confidence interval [CI]: 0.22, 0.26).
Figure 3 illustrates the number of neo-
plasms and malignant tumors in our
population of indeterminate adnexal
cysts. Among all indeterminate cysts,
the incidence of ovarian neoplasms
was 24.8% (75 of 303 cysts; 95% CI:
0.20, 0.30) and the incidence of malig-
nant tumors was 3.6% (11 of 303 cysts;
95% CI: 0.02, 0.06). The incidence of
ovarian neoplasms differed between
category 1 cysts (typical benign appear-
ing cysts .5 cm) and category 2 cysts
(cysts with avascular internal compo-
nents) (17.5% [25 of 143 cysts; 95%
CI: 0.12, 0.25] vs 31.3% [50 of 160
cysts; 95% CI: 0.24, 0.39], respectively;
ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al
Table 1
Patient Demographic Characteristics and Cyst Size of the Indeterminate Adnexal
Cysts Included in the Final Study Population
Parameter
All adnexal cysts (n = 303)
  Mean age (y)
  Age range (y)
  Premenopausal women
  Postmenopausal women
  Mean greatest diameter (cm)
  Greatest diameter range (cm)
Category 1 cysts (typical benign appearing cysts .5 cm) (n = 143)
  Mean age (y)
  Age range (y)
  Premenopausal women
  Postmenopausal women
  Mean greatest diameter (cm)
  Greatest diameter range (cm)
Category 2 cysts (cysts with avascular internal components) (n = 160)
  Mean age (y)
  Age range (y)
  Premenopausal women
  Postmenopausal women
  Mean greatest diameter (cm)
  Greatest diameter range (cm)
Note.—Numbers in parentheses are percentages.
P = .001). The incidence of malignant
tumors also differed between categories
1 and 2 cysts (0% [0 of 143 cysts] vs
6.9% [11 of 160 cysts; 95% CI: 0.03,
0.12]; P , .001).
Analysis of the Patient and Cyst
Characteristics: Nonneoplastic Cyst,
Neoplastic Cyst, and Malignant Tumors
Univariable comparison between non-
neoplastic and neoplastic cysts and be-
nign versus malignant diagnoses was
performed, and results are presented
in Table 2. Those patients who ulti-
mately were diagnosed with an ovar-
ian neoplasm were older (P , .0001)
and had adnexal cysts that were sig-
nificantly larger in greatest diameter
(P , .0001). These adnexal cysts had
more septations (P = .0003) and nodu-
lar components (P = .0139) than non-
neoplastic cysts. Malignant tumors had
more nodular components than benign
tumors (P , .0001). All other charac-
teristics were similar between the two
groups.
Radiology: Volume 000: Number 0—   2018  n  radiology.rsna.org 5
Value
42.9 6 14.1
17–88
217 (71.6)
86 (28.4)
5.9 6 3.2
1.0–20.7
40.7 6 13.9
17–88
107 (74.8)
36 (25.2)
7.0 6 2.5
5–20.7
44.8 6 14.2
17–85
110 (68.8)
50 (31.2)
4.9 6 3.4
1.0–20
Backward Stepwise Regression Analysis
of Patient and Cyst Characteristics:
Neoplasms
Stepwise logistic regression modeling
revealed that the number of nodular
components was the strongest predictor
of neoplasm accounting for all other co-
variables, with a 95% increased odds for
neoplasm for every 1-unit increase in the
number of nodular components (odds
ratio [OR] = 1.95, P < .0001) (Table 3).
Other significant predictors were age,
greatest cyst diameter (in centimeters),
and number of septations, with ORs of
1.03 (P = .037), 1.25 (P , .0001), and
1.37 (P = .001), respectively. Menstrual
status was not a statistically significant
predictor in this analysis, with an OR
of 2.19 (P = .057). The area under the
curve for this model is 80.5%.
Backward Stepwise Analysis of Patient
and Cyst Characteristics: Malignant
Tumors
Stepwise logistic regression modeling
revealed that the number of nodular
components was the single strongest in-
dependent predictor of malignancy, ac-
counting for all other covariables (Table
3). There is a 183% increased odds for
malignancy for every 1-unit increase
in the number of nodular components
(OR = 2.83, P , .0001). Cyst diameter
(OR = 1.16; P = .067) and menstrual
status (OR = 3.01; P = .152) were not
statistically significant predictors of ma-
lignancy. In addition, the area under
the curve for this model is reported as
85%.
Benign Tumors: Patient and Cyst
Characteristics
There were 64 benign tumors at final
pathologic examination. The average
age (6standard deviation) of women
with benign tumors was 49.8 years 6
16.1 (range, 19–85 years). The major-
ity of the benign tumors (42 of 64 cysts,
65.6%) were serous cystadenomas. The
remaining cysts were mature teratomas
(n = 8), serous cystadenofibromas
(n = 7), mucinous cystadenomas (n = 6),
and Brenner tumor (n = 1).
Malignant Tumors: Patient and Cyst
Characteristics
Table 4 lists the age, cyst character-
istics, and final pathologic finding for
each patient in whom a malignant tu-
mor was diagnosed. The majority of the
malignant tumors (seven of 11) were
borderline and low-grade serous tumor
types. Four of the 11 malignant tumors
were more aggressive cell types of ovar-
ian cancer.
Discussion
Grayscale and Doppler US are usu-
ally the first imaging modality by
which an adnexal cyst is evaluated,
as US can be used to characterize
the vast majority of adnexal cysts
(1–4,6,7,9,11,12,20,21,29,30). To our
knowledge, there is no universally ac-
cepted definition of a sonographically
“indeterminate” adnexal cyst; however,
many authors have described an inde-
terminate adnexal cyst as a cyst with in-
ternal avascular nodularity or thickened
septations, nontypical-appearing endo-
metriomas, dermoids and hemorrhagic
ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al

Figure 3 cysts, and cysts that cannot be com-

pletely characterized with US owing
to their large size (7,27,29–34). In our
study, we defined an indeterminate ad-
nexal cyst as a typical benign-appearing
simple cyst, hemorrhagic cyst, endome-
trioma or dermoid larger than 5 cm, or
a cyst that contains internal avascular
solid components (nodules, papillary
excrescences, or septations) (2,7,29).
The percentage of sonographically in-
determinate adnexal cysts is potentially
dependent on the imaging practice, in-
cluding the expertise of the interpret-
ing physician and baseline prevalence
of ovarian cancer. Our study demon-
strated a 23.9% (391 of 1637 cysts;
95% CI: 0.22, 0.26) prevalence of inde-
terminate adnexal cysts among all ad-
nexal cysts detected in a general popu-
lation undergoing pelvic US at academic
institutions. This is comparable to that
of other studies, which have found up
to 20% of cysts that are not able to be
characterized as benign or likely ma-
lignant at the initial US examination
(4,7,19,27,31–33,35,36).
When an indeterminate adnexal
cyst was discovered at US in our study,
the incidence of neoplasm was 24.8%
Figure 3:  Bar graph depicts percentage of nonneoplastic cysts and benign and malignant tumors in all (75 of 303 cysts; 95% CI: 0.20, 0.30)
cysts, in category 1 cysts (typical benign appearing cysts .5 cm), and in category 2 cysts (cysts with avas- and the incidence of malignant tumors
cular internal components). Percentages refer to fraction of each column.
was 3.6% (11 of 303 cysts; 95% CI:
0.02, 0.06). The cysts were further sub-
divided into two categories for further

Table 2
Univariable Analysis of Patient and Cyst Characteristics Associated with Nonneoplastic versus Neoplastic Cysts and Benign versus
Malignant Tumors
Parameter Nonneoplastic Cysts Neoplasms P Value Benign Tumors Malignant Tumors P Value

No. of total cysts 228 75 64 11

Average age (y) 40.8 6 12.8 (17–88) 49.7 6 15.6 (17–85) ,.0001* 50.0 6 16.1 (17–85) 47.2 6 14.5 (25–74) .6195
Menopausal status† ,.0001* .708
 Premenopausal 180 (78.9) 37 (49.3) 31 (48.4) 6 (54.5)
 Postmenopausal 48 (21.1) 38 (50.7) 33 (51.6) 5 (45.5)
Average greatest diameter (cm) 5.4 6 2.6 (1–20.7) 7.3 6 4.3 (2–20) ,.0001* 7.1 6 4.0 (2–18.5) 8.5 6 5.8 (2.5–20) .3377
Average no. of septations 0.6 6 1.4 (0 to 5) 1.4 6 2.0 (0 to 5) .0003* 1.4 6 2.0 (0 to 5) 1.0 6 1.8 (0 to 5) .5296
Average no. of nodular components 0.3 6 0.6 (0–4) 0.6 6 1.2 (0–5) .0139* 0.3 6 0.6 (0–3) 2.1 6 2.2 (0–5) ,.0001*
Average greatest diameter of nodular 2.9 6 8.0 (0–5.0) 5.2 6 14.4 (0–10.4) .0868 4.0 6 14.3 (0–10.4) 11.6 6 14.0 (0–3.7) .1064
components (mm)

Note.—Except where indicated, numbers in parentheses are the range.

* Statistically significant.
†
Numbers in parentheses are percentages.

6 radiology.rsna.org  n Radiology: Volume 000: Number 0—   2018

ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al

analysis. Category 1 included cysts that malignant tumors was 6.9% (11 of 160
were typical benign-appearing cysts cysts). The incidence of both neoplasms
(simple or hemorrhagic cysts, endo- and malignant tumors was significantly
metriomas, or dermoids) at US and greater in category 2 cysts (P , .001).
measured more than 5 cm in diameter. In our study, we focused on identi-
Within category 1 cysts, the incidence of fying characteristics within an indeter-
neoplasm was 17.5% (25 of 143 cysts), minate adnexal cyst that may suggest a
and there were no malignant tumors. neoplasm or malignant tumor is pre-
Category 2 included cysts with avascu- sent. We found that category 2 cysts,
lar internal solid components, including adnexal cysts with avascular internal
septations, papillary excrescences, and/ solid components (ie, septations, pap-
or nodules. Within category 2 cysts, illary excrescences, or nodules), have a
the incidence of neoplasms was 31.3% higher incidence of neoplasms and ma-
(50 of 160 cysts) and the incidence of lignant tumors compared with category
Table 3
Parameter Estimates from Backward Stepwise Logistic Regression Analysis
Evaluating Predictors of Neoplasm and Malignancy
Parameter Odds Ratio 95% Confidence Interval
Outcome = neoplasm
 Age 1.03 1.00, 1.06
  Menstrual status 2.19 0.97, 4.91
  Greatest cyst diameter 1.25 1.13, 1.38
  No. of septations 1.37 1.14, 1.64
  No. of nodular components 1.95 1.36, 2.80
Outcome = malignancy
  Menstrual status 3.01 0.67, 13.63
  Greatest cyst diameter 1.16 0.99, 1.37
  No. of nodular components 2.83 1.70, 4.70
* Statistically significant.
1 cysts, large adnexal cysts with a typi-
cal benign appearance (simple or hem-
orrhagic cysts, endometriomas, or der-
moids). The stepwise logistic regression
demonstrated that the size of the cyst
(OR = 1.25, P , .0001) and the pres-
ence of internal septations (OR = 1.37,
P = .001) and avascular nodules (OR
= 1.95, P < .0001), in addition to age
(OR = 1.03, P = .037), were significant
predictors of an ovarian neoplasm. The
presence of internal avascular nodular
components had the strongest asso-
ciation with ovarian malignancy (OR
= 2.83, P , .0001). Menstrual status
and age were not significant predictors
of ovarian malignancy when accounting
for all other covariables. This suggests
that the morphologic appearance of the
P Value
indeterminate cyst should guide the
level of concern for malignancy and fol-
low-up of the patient, even among pre-
.037* menopausal women.
.057 Currently, both American and Euro-
,.0001* pean imaging societies have recommen-
.001* dations for the follow-up of sonographi-
,.0001* cally indeterminate adnexal cysts. The
Society of Radiologists in Ultrasound
.152
guidelines recommend either referral
.067
to magnetic resonance (MR) imaging
,.0001*
or surgical evaluation in all indeter-
minate cysts, except for large typical
benign-appearing hemorrhagic cysts,

Table 4
Patient and Cyst Characteristics of the Sonographically Indeterminate Adnexal Cysts That Were Malignant Tumors at Final Pathologic
Examination
Cyst Greatest No. of Nodular Nodular Component
Patient No./Age (y) Diameter (cm) No. of Septations Components Greatest Diameter (cm) Cyst Category Final Pathologic Diagnosis

1/52 4.1 0 1 0.5 2 Low-grade serous tumor

2/63 20 0 5 3.7 2 Low-grade serous tumor
3/74 6.8 5* 0 0 2 High-grade serous with clear cell features
4/35 7 0 1 0.2 2 Borderline serous tumor
5/46 17 0 1 0.7 2 Clear cell carcinoma
6/42† 2.8 (right) 0 5 1.4 2 Borderline serous tumor
6/42† 5.6 (left) 0 5 2.4 2 Borderline serous tumor
7/59 2.5 0 1 0.4 2 Endometrioid carcinoma
8/25 4.1 1‡ 4 0.9 2 Borderline serous tumor
9/31 5.3 3‡ 0 0 2 Metastasis (colon)
10/50 9.9 3* 4 3.5 2 Borderline serous tumor

* Thick (3 mm) irregular septations.

†
Same patient.
‡
Thin (,3 mm) regular septations.

Radiology: Volume 000: Number 0—   2018  n  radiology.rsna.org 7

ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al
endometriomas, and dermoids in pre-
menopausal women, for which 6–12-
week follow-up US is recommended
(29). The European Society of Urogen-
ital Radiology recommends MR imaging
for the further evaluation of indetermi-
nate adnexal cysts on the basis of the
accuracy of excluding a malignant tumor
at MR imaging (34). In a study of 497
adnexal masses, which incorporated an-
atomic (T1- and T2-weighted) and func-
tional (diffusion-weighted and perfusion
series) MR imaging, the sensitivity and
specificity for the diagnosis of a malig-
nant tumor were 94% and 97%, respec-
tively (32). Furthermore, the lack of any
enhancement of the solid components
of a cyst yielded a positive predictive
value for malignancy of 0%, which is
highly reassuring for benignity. Given
the 6.9% incidence of malignancy in cat-
egory 2 sonographically indeterminate
adnexal cysts in our study, we favor the
use of MR imaging, when feasible, for
further characterization of such cysts.
If malignant features are present, the
women can be expeditiously referred
to a gynecologic oncologic surgeon for
appropriate staging and treatment,
whereas if the patient has a benign cyst,
more conservative management can be
contemplated.
Previous studies have addressed ma-
lignancy in sonographically indetermi-
nate adnexal cysts. Adusumilli et al (27)
analyzed a group of 95 patients with so-
nographically indeterminate masses and
found a 5% incidence of malignant tu-
mors in their population, similar to our
study. Their criteria for a sonographi-
cally indeterminate adnexal mass were
suboptimal image quality, large mass (no
defined diameter), or an adnexal mass
in which the radiologist could not render
a specific diagnosis (27). These investi-
gators included patients who underwent
resection of the adnexal mass and those
with appropriate clinical follow-up. Two
other studies in the literature, those by
Timmerman et al (7) and Zhang et al
(28), found an incidence of malignant
tumor in their sonographically indeter-
minate adnexal mass population of up
to 40%. Neither study was primarily
designed to determine the incidence
of malignant tumors in sonographically
8 radiology.rsna.org  n Radiology: Volume 000: Number 0—   2018
indeterminate adnexal masses, but both
included indeterminate adnexal masses
in a subanalysis within the study. Tim-
merman et al (7) investigated the diag-
nostic performance of the International
Ovarian Tumor Analysis simple rules,
and the criterion for a sonographically
indeterminate mass was a cyst that did
not fit into their malignant or benign cat-
egory. Zhang et al (28) studied the di-
agnosis of malignancy with intravenous
contrast material–enhanced US in inde-
terminate adnexal masses, and their def-
inition of an indeterminate mass was an
adnexal mass with internal solid parts
(not further defined), a multilocular
cyst, and a solid mass. In addition to dif-
ferent criteria for a sonographically in-
determinate adnexal mass, both of these
studies included only patients with re-
sected masses. This requirement would
tend to overestimate the prevalence of
malignancy due to the exclusion of more
benign-appearing yet sonographically in-
determinate cysts and may account for
the difference from our results.
There are some limitations of our
study. First, the population studied is
a general population presenting to ac-
ademic institutions for pelvic US, and
these results may not be applicable to
a population at high risk for ovarian
cancer. The majority of our patients
were premenopausal (71.6%), and this
likely accounted for the high number of
physiologic and benign-appearing cysts;
in a population with a greater number
of postmenopausal women, the rates
of neoplasm and malignancy might be
different. The small number of ma-
lignant tumors (n = 11) does limit the
analysis, and it is possible that other
cyst characteristics would become sig-
nificant predictors of malignancy with a
larger number of malignant cysts. Fur-
thermore, we only included cases with
2-year follow-up or pathologic diagno-
sis, and this limits our ability to study
the remaining cases. Our study is ret-
rospective in nature and not all factors
affecting the frequency of ovarian cancer
could be evaluated, such as high-risk
genetic mutation status, family history,
and hormone status. However, in every-
day practice, radiologists often lack such
clinical information and, therefore, we
believe our study reflects the real-world
clinical setting. Last, we have no repro-
ducibility data on our methods for the
diagnosis of an indeterminate adnexal
lesion at US or on the variability be-
tween readers in this study.
In conclusion, when evaluating sono-
graphically indeterminate adnexal cysts,
cyst features, patient age, and menstrual
status are predictive of the presence of
an ovarian neoplasm, whereas the pres-
ence of avascular nodules was the only
predictor of ovarian malignancy in our
study, when all covariates were consid-
ered. Category 2 indeterminate cysts (in-
ternal avascular solid components) con-
vey a higher risk of malignancy than do
category 1 indeterminate cysts (typical
benign appearing cysts). In women with
sonographically indeterminate cysts,
further evaluation is prudent given that
the incidence of malignancy was up to
6.9%. Further imaging characterization
or follow-up is particularly important for
cysts with avascular solid components to
ensure that women with malignant cysts
receive expedited care and women with
benign cysts can consider conservative
management.
Acknowledgment: This research was performed
by members (E.A.S., K.E.M., V.P., K.P., J.R.,
A.P.W.) of the Society of Abdominal Radiology
Uterine and Ovarian Cancer Diseased Focused
Panel.
Disclosures of Conflicts of Interest: E.A.S. dis-
closed no relevant relationships. V.P. disclosed
no relevant relationships. K.P. disclosed no rel-
evant relationships. J.B.R. disclosed no relevant
relationships. L.B. Activities related to the pre-
sent article: disclosed no relevant relationships.
Activities not related to the present article: in-
stitution received money for expert testimony.
Other relationships: disclosed no relevant rela-
tionships. E.M. disclosed no relevant relation-
ships. T.M. disclosed no relevant relationships.
E.S. disclosed no relevant relationships. A.P.W.
disclosed no relevant relationships. A.D.B. dis-
closed no relevant relationships. K.E.M. dis-
closed no relevant relationships.
References
1. Brown DL, Doubilet PM, Miller FH, et al.
Benign and malignant ovarian masses: se-
lection of the most discriminating gray-scale
and Doppler sonographic features. Radiol-
ogy 1998;208(1):103–110.
2. Ekerhovd E, Wienerroith H, Staudach A,
Granberg S. Preoperative assessment of
ULTRASONOGRAPHY: Indeterminate Adnexal Cysts at US Sadowski et al
unilocular adnexal cysts by transvaginal ul-
trasonography: a comparison between ultra-
sonographic morphologic imaging and histo-
pathologic diagnosis. Am J Obstet Gynecol
2001;184(2):48–54.
3. Fishman DA, Cohen L, Blank SV, et al. The
role of ultrasound evaluation in the detec-
tion of early-stage epithelial ovarian cancer.
Am J Obstet Gynecol 2005;192(4):1214–
1221; discussion 1221–1222.
4. Mohaghegh P, Rockall AG. Imaging strategy
for early ovarian cancer: characterization of
adnexal masses with conventional and ad-
vanced imaging techniques. RadioGraphics
2012;32(6):1751–1773.
5. Sayasneh A, Ekechi C, Ferrara L, et al. The
characteristic ultrasound features of specific
types of ovarian pathology (review). Int J
Oncol 2015;46(2):445–458.
6. Sharma A, Apostolidou S, Burnell M, et al.
Risk of epithelial ovarian cancer in asymptom-
atic women with ultrasound-detected ovarian
masses: a prospective cohort study within
the UK collaborative trial of ovarian cancer
screening (UKCTOCS). Ultrasound Obstet
Gynecol 2012;40(3):338–344.
7. Timmerman D, Ameye L, Fischerova D, et
al. Simple ultrasound rules to distinguish
between benign and malignant adnexal
masses before surgery: prospective valida-
tion by IOTA group. BMJ 2010;341:c6839.
8. Castillo G, Alcázar JL, Jurado M. Natural
history of sonographically detected simple
unilocular adnexal cysts in asymptomatic
postmenopausal women. Gynecol Oncol
2004;92(3):965–969.
9. Greenlee RT, Kessel B, Williams CR, et al.
Prevalence, incidence, and natural history of
simple ovarian cysts among women .55 years
old in a large cancer screening trial. Am J Ob-
stet Gynecol 2010;202(4):373.e1–373.e9.
10. Levine D, Gosink BB, Wolf SI, Feldesman
MR, Pretorius DH. Simple adnexal cysts:
the natural history in postmenopausal
women. Radiology 1992;184(3):653–659.
11. Modesitt SC, Pavlik EJ, Ueland FR, DePriest
PD, Kryscio RJ, van Nagell JR Jr. Risk of
malignancy in unilocular ovarian cystic tu-
mors less than 10 centimeters in diameter.
Obstet Gynecol 2003;102(3):594–599.
12. Valentin L, Ameye L, Franchi D, et al. Risk
of malignancy in unilocular cysts: a study of
1148 adnexal masses classified as unilocular
cysts at transvaginal ultrasound and review
of the literature. Ultrasound Obstet Gynecol
2013;41(1):80–89.
13. Kido A, Togashi K, Konishi I, et al. Dermoid
cysts of the ovary with malignant transfor-
mation: MR appearance. AJR Am J Roent-
genol 1999;172(2):445–449.
Radiology: Volume 000: Number 0—   2018  n  radiology.rsna.org 9
14. Munksgaard PS, Blaakaer J. The associa-
tion between endometriosis and gyneco-
logical cancers and breast cancer: a review
of epidemiological data. Gynecol Oncol
2011;123(1):157–163.
15. Nunes N, Ambler G, Foo X, Naftalin J,
Widschwendter M, Jurkovic D. Use of IOTA
simple rules for diagnosis of ovarian cancer:
meta-analysis. Ultrasound Obstet Gynecol
2014;44(5):503–514.
16. Van Calster B, Van Hoorde K, Froyman W, et
al. Practical guidance for applying the ADNEX
model from the IOTA group to discriminate
between different subtypes of adnexal tumors.
Facts Views Vis ObGyn 2015;7(1):32–41.
17. Van Calster B, Van Hoorde K, Valentin L, et
al. Evaluating the risk of ovarian cancer be-
fore surgery using the ADNEX model to dif-
ferentiate between benign, borderline, early
and advanced stage invasive, and secondary
metastatic tumours: prospective multicentre
diagnostic study. BMJ 2014;349:g5920.
18.
Stukan M, Dudziak M, Ratajczak K,
Grabowski JP. Usefulness of diagnostic in-
dices comprising clinical, sonographic, and
biomarker data for discriminating benign
from malignant ovarian masses. J Ultra-
sound Med 2015;34(2):207–217.
19. Timor-Tritsch IE, Goldstein SR. The com-
plexity of a “complex mass” and the sim-
plicity of a “simple cyst”. J Ultrasound Med
2005;24(3):255–258.
20. Alcázar JL, Guerriero S, Laparte C, Ajossa
S, Ruiz-Zambrana A, Melis GB. Diagnos-
tic performance of transvaginal gray-scale
ultrasound for specific diagnosis of benign
ovarian cysts in relation to menopausal
status. Maturitas 2011;68(2):182–188.
21. Sokalska A, Timmerman D, Testa AC, et al.
Diagnostic accuracy of transvaginal ultra-
sound examination for assigning a specific
diagnosis to adnexal masses. Ultrasound
Obstet Gynecol 2009;34(4):462–470.
22. Buy JN, Ghossain MA, Hugol D, et al. Char-
acterization of adnexal masses: combination
of color Doppler and conventional sonogra-
phy compared with spectral Doppler analysis
alone and conventional sonography alone.
AJR Am J Roentgenol 1996;166(2):385–393.
23. Guerriero S, Alcazar JL, Coccia ME, et al.
Complex pelvic mass as a target of evaluation
of vessel distribution by color Doppler sonogra-
phy for the diagnosis of adnexal malignancies:
results of a multicenter European study. J Ul-
trasound Med 2002;21(10):1105–1111.
24. Medeiros LR, Rosa DD, da Rosa MI, Boz-
zetti MC. Accuracy of ultrasonography with
color Doppler in ovarian tumor: a systematic
quantitative review. Int J Gynecol Cancer
2009;19(7):1214–1220.
25. Stein SM, Laifer-Narin S, Johnson MB, et
al. Differentiation of benign and malignant
adnexal masses: relative value of gray-scale,
color Doppler, and spectral Doppler sonogra-
phy. AJR Am J Roentgenol 1995;164(2):381–
386.
26. Valentin L. Prospective cross-validation of
Doppler ultrasound examination and gray-
scale ultrasound imaging for discrimination
of benign and malignant pelvic masses. Ultra-
sound Obstet Gynecol 1999;14(4):273–283.
27. Adusumilli S, Hussain HK, Caoili EM, et al.
MRI of sonographically indeterminate ad-
nexal masses. AJR Am J Roentgenol 2006;187
(3):732–740.
28. Zhang X, Mao Y, Zheng R, et al. The contri-
bution of qualitative CEUS to the determina-
tion of malignancy in adnexal masses, inde-
terminate on conventional US: a multicenter
study. PLoS One 2014;9(4):e93843.
29. Levine D, Brown DL, Andreotti RF, et al. Man-
agement of asymptomatic ovarian and other
adnexal cysts imaged at US Society of Radi-
ologists in Ultrasound consensus conference
statement. Ultrasound Q 2010;26(3):121–131.
30. Buys SS, Partridge E, Greene MH, et al.
Ovarian cancer screening in the Prostate,
Lung, Colorectal and Ovarian (PLCO) can-
cer screening trial: findings from the initial
screen of a randomized trial. Am J Obstet
Gynecol 2005;193(5):1630–1639.
31. Spencer JA, Ghattamaneni S. MR imaging of
the sonographically indeterminate adnexal
mass. Radiology 2010;256(3):677–694.
32. Thomassin-Naggara I, Aubert E, Rockall A,
et al. Adnexal masses: development and pre-
liminary validation of an MR imaging scoring
system. Radiology 2013;267(2):432–443.
33. Kinkel K, Lu Y, Mehdizade A, Pelte MF, Hri-
cak H. Indeterminate ovarian mass at US:
incremental value of second imaging test for
characterization—meta-analysis and Bayes-
ian analysis. Radiology 2005;236(1):85–94.
34. Forstner R, Thomassin-Naggara I, Cunha
TM, et al. ESUR recommendations for MR
imaging of the sonographically indetermi-
nate adnexal mass: an update. Eur Radiol
2017;27(6):2248–2257. [Published correc-
tion appears in Eur Radiol 2017;27(6):2258.]
35. Sohaib SA, Sahdev A, Van Trappen P, Ja-
cobs IJ, Reznek RH. Characterization of ad-
nexal mass lesions on MR imaging. AJR Am
J Roentgenol 2003;180(5):1297–1304.
36. Kinkel K, Hricak H, Lu Y, Tsuda K, Filly RA.
US characterization of ovarian masses: a meta-
analysis. Radiology 2000;217(3):803–811.