YGYNO-975582; No.

of pages: 5; 4C: 4
Gynecologic Oncology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Gynecologic Oncology
journal homepage: www.elsevier.com/locate/ygyno

Serial ultrasonographic evaluation of ovarian abnormalities with a

morphology index
Jeffrey W. Elder a,⁎, Edward J. Pavlik a, Ashleigh Long a, Rachel W. Miller a, Christopher P. DeSimone a,
John T. Hoff a, Walker R. Ueland a, Richard J. Kryscio b, John R. van Nagell Jr. a, Frederick R. Ueland a
a
Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Kentucky Chandler Medical Center-Markey Cancer Center, Lexington, KY, United States
b
Department of Statistics, University of Kentucky Chandler Medical Center-Markey Cancer Center, Lexington, KY, United States

H I G H L I G H T S

• Serial evaluation of tumor morphology index (MI) differentiates between malignant and benign ovarian tumors.
• Utilization of delta MI allows clinicians to avoid unnecessary surgical evaluations of ovarian tumors.

a r t i c l e i n f o a b s t r a c t

Available online xxxx Objective. Transvaginal ultrasonography with tumor morphology index (MI) has been used to predict the risk
of ovarian malignancy. Our objective was to analyze changes in serial MI scores for malignant and non-malignant
Keywords: ovarian tumors in a large and asymptomatic population.
Ovary
Methods. Eligible subjects participated in the University of Kentucky Ovarian Cancer Screening Program and
Tumor
had abnormalities that included cysts, cysts with septations, complex cysts with solid areas, and solid masses.
Ultrasound
Screening
Analysis included: MI, change in MI (delta MI), delta MI per scan and per month, number and duration of scans.
Morphology Results. From 1987 to 2012, 38,983 women received 218,445 scans. Of the 7104 eligible subjects, 6758 tumors
were observed without surgery and 472 were surgically removed. Eighty-six percent (5811) of observed tumors
were resolved. There were 74 malignant and 272 non-malignant tumors. Eighty-five percent of malignancies had
MI ≥5 at decision for surgery. The risk of malignancy based on MI was: MI = 5 (3%), MI = 6 (3.7%), MI = 7
(12.6%), MI = 8 (26.7%), MI = 9 (27.8%), MI = 10 (33.3%). The mean delta MI per month decreased for tumors
that resolved (delta MI −1.0, p b 0.001) or persisted without surgery (delta MI −0.7, p b 0.001). For abnormal-
ities surgically removed, the mean delta MI per month increased significantly more for malignancies than for
benign tumors (delta MI +1.6 vs. +0.3, p b 0.001).
Conclusions. The mean MI for malignant ovarian tumors increases over time, while non-malignant tumors
have a decreasing or stable MI. Serial MI analysis can improve the prediction of ovarian malignancy by reducing
false-positive results, thereby decreasing the number of operations performed for benign abnormalities.
© 2014 Elsevier Inc. All rights reserved.

Introduction
Prior to the widespread availability of transvaginal ultrasonography,
any palpable ovarian tumor in a postmenopausal woman was an indica-
tion for surgical removal [1]. This paradigm began to change in the late
1980s when a conservative approach for some cystic ovarian tumors
was suggested [2]. Physical examination of the ovary is easy to perform
but has not been an effective way to monitor ovarian health, and the
⁎ Corresponding author at: Division of Gynecologic Oncology, Greenville Health System
Cancer Institute, 900 W Faris Rd, Greenville, SC 29605, United States. Fax: +1 864 404
2011.
E-mail address: jelder2@ghs.org (J.W. Elder).
superiority of ultrasound has been widely accepted [3–5]. Today, sono-
graphic observation is a standard practice for asymptomatic women
with either unilocular or septate ovarian cysts because they are at
very low risk for malignancy [6,7].
A morphology index (MI), developed at the University of Kentucky
(Fig. 1) is an ultrasound-based evaluation of an ovarian tumor's volume
and morphologic complexity. The MI score ranges from 0 to 10, and the
risk of malignancy rises with increasing MI score. Although both compo-
nents are independently predictive of cancer, tumor structure is twice
as important as volume [8]. Ovarian tumors at highest risk for malignan-
cy have a MI score ≥ 5 (range 0–10); however, higher MI scores may
also occur with benign abnormalities which lower the test's positive
predictive value.

http://dx.doi.org/10.1016/j.ygyno.2014.07.091
0090-8258/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Elder JW, et al, Serial ultrasonographic evaluation of ovarian abnormalities with a morphology index, Gynecol Oncol
(2014), http://dx.doi.org/10.1016/j.ygyno.2014.07.091
2 J.W. Elder et al. / Gynecologic Oncology xxx (2014) xxx–xxx
Fig. 1. The University of Kentucky morphology index for ovarian tumors. Tumors are assigned a volume score from 0 to 5 and a structural score from 0 to 5; the total MI score is the sum of
the volume and structure score and ranges from 0 to 10.
Skates and Jacobs report that a longitudinal evaluation of CA125 can
improve early detection of ovarian cancer, and their Risk of Ovarian
Cancer Algorithm (ROCA) monitors changes in serum biomarkers in
the same subject over time [9]. Ultrasound is also a biomarker of ovarian
disease, and can be applied sequentially to monitor the malignant risk of
an ovarian tumor. The study objective was to examine the serial use of
the morphology index in women with asymptomatic ovarian tumors.
Materials and methods
Women enrolled in the University of Kentucky Ovarian Cancer
Screening Program (OCSP) from January 1987 to June 2012 were eligi-
ble for this investigation. Study methods and eligibility have been
previously reported [10]. In brief, criteria for eligibility include: asymp-
tomatic women aged 50 years or older, or at least 25 years with a
documented family history of ovarian cancer. Women with a known
ovarian tumor or a personal history of ovarian cancer were excluded.
This investigation was approved by the University of Kentucky Institu-
tional Review Board.
Transvaginal ultrasonography
Since 2000, transvaginal ultrasonography has been performed using
General Electric Logiq 400 units with a 5-mHz vaginal probe or a Gener-
al Electric Voluson 730 ProV unit with a 5- to 9-mHz vaginal probe.
Transvaginal scans were performed in low lithotomy position with an
empty bladder. If the ovary was not visualized or the tumor was too
large to be completely imaged, an abdominal transducer was used
with a full urinary bladder. Tumor measurements included length (L),
width (W), and height (H) and the calculated volume (prolate ellipsoid
formula for ovarian volume = L × W x H × 0.523). All ultrasound abnor-
malities were reviewed by at least one of the authors. Clinical and path-
ological data were entered into a Medlog database.
Please cite this article as: Elder JW, et al, Serial ultrasonographic evaluation of ovarian abnormalities with a morphology index, Gynecol Oncol
(2014), http://dx.doi.org/10.1016/j.ygyno.2014.07.091
The MI score (Fig. 1) was determined using the published index [8].
The criteria for abnormality included an ovarian volume greater than 20
ml for premenopausal women and greater than 10 ml for postmeno-
pausal women [11], or any cystic ovarian tumor with a solid component
or papillary projection. Women with a normal sonogram were sched-
uled to return in 12 months for a repeat ultrasound, while those
with an abnormality typically underwent a confirmatory scan in 4 to
6 weeks. Based on the sonographic characteristics and the MI score,
subjects either continued with interval sonographic surveillance or
were recommended for surgical removal.
For this investigation, the non-surgery group included women with
an ovarian abnormality for which surgery was not recommended. Ovar-
ian tumors that were persistent or spontaneously resolved without sur-
gery were included in the observation analysis. The malignant cohort
included all primary epithelial ovarian cancers, tumors of borderline
malignancy, and granulosa cell tumors identified at surgery. Surgically
removed non-malignant tumors were classified as a false positive result.
All cancers were staged according to the International Federation of
Gynecology and Obstetrics system and histologically classified using
the World Health Organization system.
Change in MI score
The change in MI (delta MI) was calculated by comparing the final
MI score prior to surgery to the initial MI score. Abnormalities of oppo-
site ovaries in the same woman were considered independent events. If
an abnormality was detected and then resolved, any new abnormality
that developed later was recorded as a new event. Mean scores were
used for statistical comparison. The mean delta MI per scan was deter-
mined by dividing the mean delta MI by the mean number of scans per-
formed. The delta MI per month was calculated using the total number
of months scanned as the denominator.
Statistical methods included comparison of means using one-factor
ANOVA t-tests and log rank testing. Linear regression was used to
 J.W. Elder et al. / Gynecologic Oncology xxx (2014) xxx–xxx 3

predict malignancy based on initial MI and delta MI for the clinical deci- Table 2
sion guide. Statistical significance was determined at the p b 0.05 level. Evolution of ultrasound screening performance.

n TP % TP FP % FP PPV
Results Study 1987–2012 39,337 85 0.22% 472⁎ 1.20% 15.30%
Period A 1987–1/2000 15,049 20 0.13% 226⁎ 1.50% 8.10%
There were 38,983 women enrolled in the University of Kentucky Period B 2000–1/2008 24,695 41 0.17% 173⁎ 0.70% 19.20%
Ovarian Cancer Screening Program (OCSP) from January 1987 to June Period C 1/2008–2012 23,450 24 0.10% 73⁎ 0.31% 24.70%
2012. Clinical characteristics of the study population are detailed in N = number of subjects receiving a scan during the period identified. TP, true positive; FP,
Table 1. A total of 218,445 scans were performed (mean, 5.5 scans per false positive; PPV, positive predictive value. PPV = TP/(TP + FP). *P b 0.05 Study vs A or
subject). The non-surgery group included 6758 subjects. Four hundred B or C p b 0.05 using TP,TN,FP,FN (2 × 4) B vs C, A vs C p b 0.05 using TP,TN,FP,FN (2 × 4)
Study vs A or B or C p b 0.05 using FP & [TP+TN+FN] (2 × 2) B vs C, A vs C p b 0.05 using
seventy-two surgeries identified 74 primary ovarian malignancies. FP & [TP+TN+FN] (2 × 2).
Twenty-four subjects with a malignancy had only one scan before sur-
gery, leaving 50 women for the delta MI analysis. In subjects monitored
without surgery, ovarian abnormalities resolved in 86% (5811/6758) of removal. Non-malignant tumors and resolved ovarian cysts averaged
the non-surgery group. There were 53 primary epithelial ovarian can- 3.6 and 3.0 scans over mean periods of 9.6 and 11.5 months, respective-
cers (EOC), 5 granulosa cell tumors, and 16 ovarian tumors of borderline ly. Both number of scans (p b 0.001) and time to removal (p b 0.001)
malignancy. The stage distribution for EOC was as follows: stage I, 24/53 were significantly lower for malignant compared to non-malignant
(45%); stage II, 12/53 (23%); and stage III, 17/53 (32%). No stage IV cases and resolved tumors.
were observed. For the borderline ovarian tumors, 15 were stage I and Fig. 2 displays the magnitude of the mean delta MI graphed over
one was stage III. All five granulosa cell tumors were stage I (Supple- time. Ovarian tumors observed without surgery (resolved, persistent)
mental Table 1). exhibited a consistent decline in mean MI score over time with a
The ultrasound algorithm for the OCSP has been modified over time mean delta MI = − 2.7 (− 1.0 per month resolved cysts; − 0.7 per
to reflect the evolving knowledge of ovarian sonography. We analyzed month persistent cysts). For all tumors surgically removed (benign
the study population (N = 38,983) in three distinct time periods that and malignant), the mean delta MI was 1.2 (0.7 per scan). When ana-
correspond to changes in the ultrasound screening algorithms lyzed separately, the mean delta MI for malignant tumors was 1.9 com-
(Table 2). Period A included a time when all confirmed ovarian abnor- pared to 0.7 for non-malignant tumors (p b 0.001). Likewise, the mean
malities were recommended for surgery. Period B corresponded to a delta MI per month was statistically greater for malignant than benign
time when unilocular cysts were observed without surgery. Lastly, Peri- tumors (1.6 vs. 0.3, p b 0.001).
od C included the time interval when expanded serial ultrasound mon- The time interval between scans is influenced by clinical factors
itoring included septate and more complex cystic abnormalities. The including the initial MI or the change in MI over time. The delta MI be-
result of a more selective, serial ultrasound approach resulted in an tween temporally proximate scans may have a smaller absolute change.
increase in positive predictive value (PPV) in the screening population Therefore, it was important to determine an absolute mean delta MI
from 8.1% to 24.7% over the study period. which is relatively time independent, allowing changes in the MI score
An increasing preoperative MI score correlated with a rising risk of to be meaningful regardless of the time interval between scans. Thus,
ovarian malignancy; furthermore, 85% of all malignancies had a MI delta MI per month was calculated, and linear regression analysis was
score ≥5 (Table 3). Over the history of the screening trial, not all sub- used to calculate the time independent mean delta MI from the slope
jects with ovarian abnormalities had more than one MI score recorded; of the regression line.
therefore, we report only the data for 50 malignant and 272 non-
malignant subjects with multiple MI values. When analyzing the change Discussion
in MI over time, three distinct zones were apparent (Table 4). The mean
delta MI for malignant tumors was 1.9 (or 1.6 per month) indicating a Many publications and professional societies emphasize the value of
significant increase in MI over time (p b 0.001) when compared to sur- identifying high risk ovarian tumors prior to surgery, to promote proper
gically evaluated benign tumors with a mean delta MI of 0.7 (or 0.3 per counseling, surgical planning, and consultation with a gynecologic
month). Ovarian tumors which ultimately resolved had a negative oncologist [12–18]. The International Ovarian Tumor Analysis (IOTA)
mean delta MI indicating a notable decrease in MI with each visit group has stated the importance of expert sonography, providing clear
(p b 0.001) compared to tumors requiring surgery. On average, malig- definitions of sonographic features [19], simple rules for evaluating
nant tumors received 2.1 scans over a mean 2.3 months prior to surgical an adnexal mass [20], and data suggesting that predictions are not

Table 1
Clinical characteristics of the study population.

Non-surgery group Surgery group

Persistent cohort Resolved cohort Benign cohort Malignant cohort

Subjects, n 38,983 947 5811 272 74

Total screens 218,445 21,915 28,960 2013 371
Abnormal screens (%) 16,499 (7.6) 6456 (29.5) 8921 (30.8) 965 (47.9) 157 (42.3)
Mean age, year 56.8 58 54.5 58.6 63.5
Median age, year 56 57 53 57.5 64
Weight, lbs. 161.2 163.7 161 166.1 160.3
Parity, mean 2.1 2.3 2.2 2.4 2.0
Nulliparity 5851 (15) 151 (15.9) 937 (16.1) 26 (9.8) 16 (21.6)
FH ovary (%) 8766 (22.5) 275 (29) 1871 (32.2) 68 (25) 15 (20.2)
FH breast (%) 16,268 (41.7) 429 (45.3) 2779 (47.8) 135 (49.6) 39 (52.7)
No HRT 8013 (20.6) 208 (22) 1465 (25.2) 33 (12.1) 5 (6.8)
HRT on last visit 5140 (13.2) 85 (9) 552 (9.5) 25 (9.2) 9 (12.2)
Surgery (%) 472 (1.2) 0 0 272 (0.7) 74 (0.2)
Resolved cysts (%) 6915 (17.7) 0 6915 0 0
Persisting cysts (%) 2006 (5.1) 2006 0 0 0

Please cite this article as: Elder JW, et al, Serial ultrasonographic evaluation of ovarian abnormalities with a morphology index, Gynecol Oncol
(2014), http://dx.doi.org/10.1016/j.ygyno.2014.07.091
4 J.W. Elder et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Table 3
Risk of malignancy for each morphology index score.

MI Total Malignant ROM (%)

0 28,615
1 2349
2 2365
3 2635
4 1579
5 1061
6 241
7 87
8 30
9 18
10 3 1
Total 38,983
0 0.00
1 0.04
0 0.00
3 0.11
7 0.44
29 2.73
9 3.73
11 12.64
8 26.67
5 27.78
33.33
74 0.71

improved by the addition of CA125 alone [21,22] or in combination

Fig. 2. Change in mean MI per time for the study population.
algorithms like the Risk of Malignancy Index [23]. However, the multi-
institutional IOTA investigations have mostly been performed in popu-
lations of high cancer prevalence, rather than primary care or screening
for epithelial malignancies with the majority of malignancies diagnosed
populations of low prevalence.
as stage I or II (Supplemental Table 1). Used appropriately, serial
It is well-accepted that unilocular ovarian cysts and septate tumors
sonography can reduce the number of operations performed for asymp-
without complexities frequently resolve without surgical intervention,
tomatic ovarian tumors. The authors believe that the Kentucky MI is
even in postmenopausal women [2–4]. We now appreciate that certain
both simple and objective and can be effectively used by devoted gyne-
complex ovarian abnormalities will improve or spontaneously normal-
cologic sonographers even if they are not specialty trained [26].
ize over time. Yet, reliable predictions regarding the malignant risk of
The longitudinal use of CA125 in the ROCA algorithm improves pre-
complex tumors is still challenging [24], even for logistic regression
clinical ovarian cancer detection [9]; likewise, serial ultrasonography en-
models [25]. Ultimately, the clinician must determine whether patient
hances test performance by sequentially evaluating tumor size and struc-
referral or surgery is required and this judgment is traditionally made
ture. In our experience, longitudinal surveillance does not result in a
during a single patient visit, with a single imaging encounter.
clinically significant delay in time to surgery since the mean time to sur-
Serial ultrasonography is a safe, easy and clinically useful way to
gery for malignancies was only 2.3 months compared to 9.6 months for
evaluate ovarian tumors. The delta MI provides an objective diagnostic
benign tumors and 11.5 months for cysts that spontaneously resolved.
measure for monitoring the behavior of ovarian abnormalities. This
For persistent, stable tumors at low risk for malignancy, ultrasound eval-
strategy can help provide quantitative evidence that a complex tumor
uation at six to twelve month intervals is prudent and inexpensive. Con-
is at low risk for malignancy (in the process of resolving) or high risk
comitant use of serum biomarkers, color Doppler, or diagnostic surgical
for malignancy (increasing in complexity). In this investigation, the
procedures were not evaluated as part of this algorithm; additional re-
magnitude of change in MI was significantly different for cancers and
search is needed to determine whether these adjuvant tests will enhance
non-cancers. Malignant tumors had an increase in the mean delta MI
the diagnostic accuracy and cost-effectiveness of serial ultrasound.
per month of 1.6, while non-malignant tumors showed a negligible
A notable strength of this investigation is that it presents data from a
change in delta MI per month. This significant increment in MI score
large, well-defined, 25 year prospective study with carefully monitored
over time for malignancies reflects morphological changes in both
quality control. The trial involved experienced sonographers trained in
tumor volume and structure not demonstrated by the non-malignant
consistent methods for recording the MI score. In addition, the results
group.
were obtained in a challenging study population of low cancer prevalence
Not all ovarian tumors are appropriate for serial evaluation. Symp-
(b 0.2%), compared to patients referred for a known or symptomatic ovar-
toms or other clinical indicators of malignancy may necessitate prompt
ian tumor, where cancer prevalence ranges from 20% to 30% [20–25]. It is
surgery. Also, a very high initial MI may warrant surgery rather than a
important to acknowledge that serial sonography cannot reliably evaluate
repeat ultrasound. The authors define the MI score ≥ 5 as high risk,
non-ovarian gynecological malignancies (peritoneal and other non-
and recommend short interval monitoring (first scan 4–6 weeks, then
adnexal cancers). A potential limitation of this study is that all ultrasound
2–3 months) to avoid a delay in cancer diagnosis. For tumors under
procedures were performed at a single institution with expertise in ovar-
serial sonographic surveillance, these data suggest that a per month
ian sonography, thus external validation is encouraged.
MI score increase of ≥ 1 is concerning and surgical removal is
Prompt operative removal is recommended for symptomatic ovari-
recommended.
an tumors or for overt signs of malignancy, including ascites or metasta-
Over the past 25 years of ultrasound screening, we have gradually
tic disease on examination or imaging. When immediate surgery is not
introduced serial monitoring and have documented a statistically signif-
imperative, serial ultrasound may help characterize tumor biology by
icant decrease in false positive results and an improvement in PPV from
providing insight into the pattern of tumor growth. This investigation
8.1% to 24.7% (Table 2). We have also noted a shift in stage at detection
demonstrates that malignant tumors have an increasing MI score on
serial sonography, while non-malignant tumors have a score that
Table 4 decreases or remains unchanged. Thus, serial MI analysis can improve
Mean change in morphology index score. the accuracy of detecting ovarian malignancy by reducing false-
n ΔMI p-Value ΔMI per ΔMI per p-Value
positive results, decreasing the number of operations performed for
scan month benign abnormalities.
Supplementary data to this article can be found online at http://dx.
Primary ovarian 50 1.9 p b 0.001 0.9 1.6 p b 0.001
malignancy doi.org/10.1016/j.ygyno.2014.07.091.
Non-malignancy 272 0.7 p b 0.001 0.2 0.3 p b 0.001
Persistent cysts 947 −2.7 p b 0.001 −1.3 −0.7 p b 0.001 Conflict of interest statement
Resolved cysts 5811 −2.7 p b 0.001 −1.1 −1.0 p b 0.001
The authors have no conflicts of interest to report.

Please cite this article as: Elder JW, et al, Serial ultrasonographic evaluation of ovarian abnormalities with a morphology index, Gynecol Oncol
(2014), http://dx.doi.org/10.1016/j.ygyno.2014.07.091
 J.W. Elder et al. / Gynecologic Oncology xxx (2014) xxx–xxx
References
[1] Barber H, Graber E. The PMPO syndrome (postmenopausal ovary syndrome). Obstet
Gynecol 1971;38:921–3.
[2] Goldstein S, Subramanyam B, Snyder J, Beller U, Raghavendra N, Beckman M. The
postmenopausal cystic adnexal mass: the potential role of ultrasound in conserva-
tive management. Obstet Gynecol 1989;73:8–10.
[3] Padilla LA, Radosevich DM, Milad Magdy P. Accuracy of the pelvic examination in
detecting adnexal masses. Obstet Gynecol 2000;96:593–8.
[4] Ueland F, DePriest P, DeSimone C, Pavlik EJ, Lele S, Kryscio R, et al. The accuracy of
examination under anesthesia and transvaginal sonography in evaluating ovarian
size. Gynecol Oncol 2005;99:400–3.
[5] American College of Obstetricians and Gynecologists. ACOG Practice Bulletin.
Management of adnexal masses. Obstet Gynecol 2007;110:201–14.
[6] Modesitt S, Pavlik E, Ueland F, DePriest P, Kryscio R, van Nagell Jr J. Risk of malignan-
cy in unilocular ovarian cystic tumors less than 10 cm in diameter. Obstet Gynecol
2003;102:594–9.
[7] Saunders B, Podzielinski I, Ware R, Goodrich S, DeSimone C, Ueland F, et al. Risk of
malignancy in sonographically confirmed septated cystic ovarian tumors. Gynecol
Oncol 2010;188:278–82.
[8] Ueland F, DePriest P, Pavlik E, Kryscio R, van Nagell Jr J. Preoperative differentiation
of malignant from benign ovarian tumors: the efficacy of morphology indexing and
Doppler flow sonography. Gynecol Oncol 2003;91:46–50.
[9] Skates S, Menon U, MacDonald N, Rosenthal A, Oram D, Knapp R, et al. Calculation of
the risk of ovarian cancer from serial CA-125 values for preclinical detection in post-
menopausal women. J Clin Oncol 2003;21:206s–10s.
[10] van Nagell J, DePriest P, Ueland F, DeSimone C, Cooper A, Pavlik E, et al. Ovarian
cancer screening with annual transvaginal sonography: findings of 25,000 women
screened. Cancer 2007;109:1887–96.
[11] Pavlik E, DePriest P, Gallion H, Ueland F, Reedy M, Kryscio R, et al. Ovarian volume
related to age. Gynecol Oncol 2000;77:410–2.
[12] NIH consensus conference. Ovarian cancer. Screening, treatment, and follow-up. NIH
Consensus Development Panel on Ovarian Cancer. JAMA 1995;273:491–7.
[13] Ozols R. Update of the NCCN ovarian cancer practice guidelines. Oncology 1997;
11:95–105.
[14] Elit L, Plante M, Bessette P, DePetrillo A. Surgical management of an adnexal mass
suspicious for malignancy. SOGC Clin Pract Guidel 2000;97:1–5.
[15] ACOG Committee opinion: number 280, December 2002. The role of the generalist
obstetrician–gynecologist in the early detection of ovarian cancer. Obstet Gynecol
2002;100:1413–6.
Please cite this article as: Elder JW, et al, Serial ultrasonographic evaluation of ovarian abnormalities with a morphology index, Gynecol Oncol
(2014), http://dx.doi.org/10.1016/j.ygyno.2014.07.091
5
[16] Myers E, Bastian L, Havrilesky L, Kulasingam S, Terplan M, Cline K, et al. Management
of adnexal mass. Rockville, MD: US Department of Health and Human Services; 2006
[DHHS Publication No. 06-E004].
[17] Engelen M, Kos H, Willemse P, Aalders J, de Vries E, Schaapveld M, et al. Surgery by
consultant gynecologic oncologists improves survival in patients with ovarian carci-
noma. Cancer 2006;106:589–98.
[18] Hillner B, Smith T, Desch C. Hospital and physician volume or specialization and
outcomes in cancer treatment: importance in quality of cancer care. J Clin Oncol
2000;18:2327–40.
[19] Timmerman D, Valentin L, Bourne T, Collins W, Verrelst H, Vergote I. Terms defini-
tions and measurements to describe the sonographic features of adnexal tumors: a
consensus opinion from the International Ovarian Tumor Analysis (IOTA) group.
Ultrasound Obstet Gynecol 2000;16:500–5.
[20] Timmerman D, Ameye L, Fischerova D, Epstein E, Melis G, Guerriero S, et al. Simple
ultrasound rules to distinguish between benign and malignant adnexal masses be-
fore surgery: preoperative validation by IOTA group. Br Med J 2010;341:1–8.
[21] Valentin L, Jurkovic D, van Calster B, Testa A, van Holsbeke C, Bourne T, et al. Adding
a single CA125 measurement to ultrasound imaging performed by an experienced
examiner does not improve preoperative discrimination between benign and malig-
nant adnexal masses. Ultrasound Obstet Gynecol 2009;34:345–54.
[22] Van Calster B, Timmerman D, Bourne T, Testa A, van Holsbeke C, Domali E, et al. Dis-
crimination between benign and malignant adnexal masses by specialist ultrasound
examination versus serum CA125. J Natl Canc Inst 2007;99:1706–14.
[23] Timmerman D, Van Calster B, Jurkovic D, Valentin L, Testa A, Bernard J, et al. Inclu-
sion of CA-125 does not improve mathematical models developed to distinguish
between benign and malignant adnexal tumors. J Clin Oncol 2007;25:4194–200.
[24] Valentin L, Hagen B, Tingulstad S, Eik-nes S. Comparison of ‘pattern recognition’ and
logistic regression models for discrimination between benign and malignant pelvic
masses: a prospective cross validation. Ultrasound Obstet Gynecol 2001;18:357–65.
[25] Valentin L, Ameye L, Jurkovic D, Metzger U, Lecuru F, van Huffel S, et al. Which ex-
trauterine pelvic masses are difficult to correctly classify as benign or malignant
on the basis of ultrasound findings and is there a way of making a correct diagnosis?
Ultrasound Obstet Gynecol 2006;27:438–44.
[26] Barnsfather K, Fitzpatrick B, Wilson J, Linn C, Brizendine E, Schilder J. The morphol-
ogy index: predictive value of malignancy among clinicians at various levels of
training. Gynecol Oncol 2012;127:94–7.