miomatosis por laparoscopia

1. Original Article
Feasibility, Safety, and Prediction of Complications for Minimally
Invasive Myomectomy in Women With Large and Numerous
Myomata
Maria V. Vargas, MD*, Gaby N. Moawad, MD, Cem Sievers, PhD, Jessica Opoku-Anane, MD1
,
Cherie Q. Marfori, MD, Paul Tyan, MD, and James K. Robinson, MD2
From the Department of Obstetrics and Gynecology, George Washington University Medical Center, George Washington University School of Medicine and
Health Sciences, Washington, DC (Drs. Vargas, Moawad, Opoku-Anane, Marfori, Tyam, and Robinson), Department of Pathology and Center for Cancer
Research, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts (Dr. Sievers), and Broad Institute of MIT and Harvard,
Cambridge, Massachusetts (Dr. Sievers).
ABSTRACT Objective: To assess perioperative outcomes and identify predictors of complications for minimally invasive surgery (MIS)
myomectomy in a cohort of women with large and numerous myomata.
Design: Case-control study (Canadian Task Force classification II-2).
Setting: Academic tertiary care medical center.
Patients: Women undergoing MIS myomectomy performed by 3 high-volume surgeons between April 2011 and
December 2014.
Interventions: Characteristics were compared between women who experienced complications and those who did not. Fac-
tors predictive of complications were then identified.
Measurements and Main Results: A total of 221 women underwent an MIS myomectomy, 47.5% via a laparoscopic
approach and 52.5% via robotic surgery. The mean 6 SD specimen weight was 408.1 6 384.9 g, uterine volume was
586.1 6 534.1 cm3
, dominant myoma diameter was 9.6 6 5.1 cm, and number of myomata removed was 4.5 6 4.1. The
most common complications were hemorrhage .1000 mL (8.6%) and blood transfusion (4.1%). The conversion rate was
1.8%. A dominant myoma diameter of R12 cm and a uterine volume of R750 cm3
increased the odds of complications
(odds ratio [OR], 7.44; 95% confidence interval [CI], 2.03–31.84; p 5 .004 and OR, 6.15; 95% CI, 1.55–30.02; p 5 .014
respectively). A receiver operating characteristic curve considering dominant myoma diameter and uterine volume had an
area under the curve of 0.81. A combination of dominant myoma diameter of R10 cm and uterine volume of 600 cm3
predicted complications with 79% sensitivity and 79% specificity.
Conclusion: Our cohort had large and numerous myomata with high specimen weights, but complications were comparable
to those reported in previous studies of MIS myomectomy with less complex pathology. Hemorrhage and transfusion ac-
counted for the majority of complications, and a combination of dominant myoma diameter and uterine volume was predictive
of complications. Both factors can be easily defined before surgery and may be used to guide patient counseling, referrals, and
implementation of preventative measures for hemorrhage and transfusion. Journal of Minimally Invasive Gynecology (2017)
24, 315–322 Published by Elsevier Inc. on behalf of AAGL.
Keywords: Complications; Laparoscopy; Minimally invasive surgery; Myomata; Myomectomy; Predictors; Robotic surgery
Dr. Moawad is a paid consultant for Intuitive Surgical and Applied Medical.
Dr. Robinson has received compensation for speaking at symposia held by
Bayer. The other authors declare that they have no conflicts of interest.
Corresponding author: Maria V. Vargas, MD, Department of Obstetrics and
Gynecology, George Washington University Medical Faculty Associates,
George Washington University School of Medicine and Health Sciences,
2150 Pennsylvania Ave, NW Suite 6a-408, Washington, DC 20037.
E-mail: mvvargas@mfa.gwu.edu
Submitted October 4, 2016. Accepted for publication November 30, 2016.
1
Present address: Department of Obstetrics and Gynecology, Division of
Gynecology, University of California San Francisco, San Francisco,
California.2
Present address: Department of Obstetrics and Gynecology,
Division of Minimally Invasive Gynecologic Surgery, MedStar Washington
Hospital Center, Washington, DC.
Available at www.sciencedirect.com and www.jmig.org
1553-4650/$ - see front matter Published by Elsevier Inc. on behalf of AAGL.
http://dx.doi.org/10.1016/j.jmig.2016.11.014

2. Minimally invasive surgery (MIS) myomectomy has
proven clinical benefits over traditional laparotomy and
mini-laparotomy, including decreased postoperative pain,
fewer febrile episodes, and a shorter hospital stay [1,2],
along with comparable pregnancy and myoma recurrence
rates [3]. Hemorrhage requiring transfusion and conversion
to laparotomy have been reported, however, especially in
cases with large and numerous myomata [4]. Little data
are available on the limitations of laparoscopic myomec-
tomy for complex cases, such as in women with a large
myomata burden or a large dominant myoma diameter.
One group suggested that clinical parameters for an MIS
approach include a dominant myoma diameter ,8 cm and/
or fewer than 3 myomata to be removed [5], whereas others
have concluded that MIS myomectomy is safe even in
women with numerous and very large myomata [6]. To
date, no guidelines have been established, and predictors
of complications remain poorly defined.
The aim of the present study was to assess perioperative
outcomes and identify predictors of complications in a pop-
ulation of women with large and numerous myomata under-
going MIS myomectomy performed by high-volume
fellowship-trained surgeons.
Materials and Methods
Study Population and Data Collection
This retrospective case-control studywas initiated byiden-
tifying cases of myomectomy. Allpatientsundergoing laparo-
scopic, robotic, or abdominal myomectomy between April
13, 2011, and December 22, 2014, performed by 1 of 3
fellowship-trained minimally invasive gynecologic surgeons
at George Washington University Hospital and Sibley Memo-
rial Hospital were included. This study was approved by
George Washington University’s Institutional Review Board.
We reviewed charts for clinical characteristics, including
age, race, gravidity, parity, payment method, body mass
index (BMI), indication for surgery, dominant myoma diam-
eter on preoperative imaging, uterine volume on preoperative
imaging (ellipsoid formula) [7], and history of previous
abdominal surgery. We then gathered data on characteristics
of the surgeries, including surgeon, surgical approach,
concomitant hysteroscopic myomectomy, total operative
time from skin incision to closure, estimated blood loss
(EBL), number of myomata removed, and intraoperative
complications. Intraoperative complications were defined as
conversion, EBL R1000 mL, blood transfusion, and visceral
injury. Data on perioperative outcomes, including patholog-
ical diagnosis and specimen weight from pathology reports,
hospital length of stay, readmission within 30 days (defined
as emergency room visit or readmission to the hospital), and
postoperative complications were collected as well. Postoper-
ative complications included reoperation, intensive care unit
admission, deep vein thrombosis or pulmonary embolism,
transfusion, bowel obstruction, ileus, and sepsis.
Surgical Technique
All surgeries were performed by 1 of 3 fellowship-trained
minimally invasive gynecologic surgeons (Drs. Moawad,
Marfori, and Robinson). Surgeon preference determined
the approach used. Dr. Moawad performed robotic myomec-
tomy exclusively, whereas Drs. Marfori and Robinson per-
formed predominantly conventional laparoscopic
myomectomy. When Drs. Marfori and Robinson used a ro-
botic approach, the decision was influenced by the presence
of obesity and the number of myomata that needed to be
removed by separate incisions (and thus the amount of sutur-
ing that would be necessary during the procedure). When
number of myomata was considered, determination was
made with the assistance of preoperative magnetic reso-
nance imaging. Typically, Drs. Marfori and Robinson used
R10 myomata with the largest myoma %7 cm as a cutoff
for use of the robotic approach. Conventional laparoscopic
myomectomy was completed using a multiport approach.
Pneumoperitoneum was achieved with the Veress needle.
Before the Food and Drug Administration’s (FDA) safety
warning on power morcellation, all 3 surgeons used a 5-mm
umbilical laparoscope port with two 5-mm low contralateral
ports and a 12-mm suprapubic port. Tissue extraction for
myomata ,10 cm in diameter was completed with power
morcellation through the suprapubic incision. Larger speci-
mens were extracted through a 2.5-cm suprapubic incision
using a scalpel for tissue extraction.
After the FDA safety warning, the port configuration
changed. A 12-mm umbilical port was used, and the incision
was extended to 2.5 cm for bagged scalpel extraction using a
‘‘paper roll’’ technique [8]. In addition, all 3 accessory ports
were placed laterally (rather than using a suprapubic port).
One 5-mm lateral port was placed approximately 8 cm ceph-
alad to a 12-mm low lateral port, typically on the right side,
and a low 5-mm port was placed, typically on the left side.
Dilute vasopressin (20 U of vasopressin in 200 mL of saline
solution) was used for every case. The seromuscular layer
was incised transversely using a Harmonic scalpel (Ethicon
EndoSurgery, Cincinnati, OH). The myoma was grasped
with a 10-mm laparoscopic tenaculum and then enucleated
with traction, blunt dissection, and the Harmonic scalpel
when needed.
In cases of robotic myomectomy, a Veress needle was
used to achieve pneumoperitoneum and a 12-mm optical
midline port, usually at the umbilicus, was placed under
visualization. Initially, all 3 robotic arms were used; howev-
er, since 2013, the majority of robotic myomectomies have
been completed with 2 operative arms. When necessary, a
5- or 12-mm assistant port was used. Sutures were preloaded
when necessary. The ports for the first and second robotic
arms were placed approximately 8 cm laterally and slightly
caudad on either side of the port accommodating the laparo-
scope. When used, the third arm was set 8 cm to the left of
the port for second robotic arm. A side docking technique
was used exclusively. Dilute vasopressin was used for every
316 Journal of Minimally Invasive Gynecology, Vol 24, No 2, February 2017

3. case. The robotic monopolar scissors were used to incise the
seromuscular layer transversely. Enucleation was performed
using the robotic tenaculum and robotic monopolar scissors.
Tissue extraction techniques were the same as in the conven-
tional laparoscopic cases.
In some patients with very large dominant myoma or
multiple myomata that required removal, we used a hybrid
robot-assisted laparoscopic myomectomy technique as
described previously [9]. This approach involved enucle-
ation of myomata using a conventional laparoscopic tech-
nique, followed by patient-side cart docking and closure of
the hysterotomy using a robotic approach. These cases
were categorized as robotic myomectomies for our analysis.
In both surgical approaches, suturing was completed with
0 V-Loc 180 unidirectional barbed suture in multiple layers.
When the endometrial cavity was encountered, either 3-0
polyglecaprone 25 was used to close the defect or imbrica-
tion of the overlying myometrium with 0 V-Lock 180 suture
was used to reapproximate the defect.
Statistical Analysis
Patient demographic, clinical, and surgical characteristics
were summarized using descriptive statistics. A bivariate
analysis was completed for associations between patient
characteristics and surgical complications using the c2
and
Fisher’s exact tests for categorical variables, t tests for nor-
mally distributed continuous variables, and Wilcoxon’s
rank-sum test for non-normally distributed continuous vari-
ables. Variables that could be potentially identified preoper-
atively meeting an a value of 0.1 on bivariate analysis were
included in a multivariable regression model for prediction
of complications in MIS cases. The model was adjusted
for age, race, BMI, previous laparotomy, and surgical
approach (robot-assisted or conventional laparoscopy) as
potential confounders. A receiver operating characteristic
(ROC) curve was computed using a leave-one-out cross-
validation to evaluate the performance of selected preopera-
tive variables.
Statistical analyses were performed using R version 3.2.2
(R Foundation for Statistical Computing, Vienna, Austria).
ROC curves were computed using the R package ROCR [10].
Results
The cohort comprised 250 patients who underwent myo-
mectomy between April 2011 and December 2014. The ma-
jority of women underwent MIS myomectomy (42%
robotic, 46.4% laparoscopic, and 11.6% abdominal), and
the rate of MIS myomectomy increased during the study
period (Fig. 1). The characteristics of the MIS (n 5 221)
and abdominal myomectomy (n 5 29) groups are summa-
rized in Table 1. The women in the abdominal myomectomy
group were older (mean age 41.1 years vs 37.3 years;
p 5 .029), more likely to be black (93.1% vs 59.7%;
p 5 .014), and more likely to have Medicaid insurance
(31.0% vs 13.6%; p 5 .018). BMI and parity did not differ
between the 2 groups. Women in the abdominal myomec-
tomy group were more likely to have undergone previous
abdominal surgery (44.8% vs 25.7%; p 5 .046), specifically,
previous laparotomy (42.4% vs 22.1%; p 5 .036); more
likely to have infertility as an indication for surgery
(31.0% vs 20.4%; p 5 .005); and had more myomata
removed during surgery (mean, 18.3 6 12.1 vs 4.5 6 4.1;
p , .001), a larger uterine volume on preoperative imaging
(mean, 1442.2 6 680.0 cm3
vs 683.7 6 658.7 cm3
;
p , .001), and a greater specimen weight on pathology re-
ports (mean, 908.8 6 470.4 g vs 410.6 6 384.8 g;
p , .001). There were no between-group differences in
dominant myoma diameter, type of myomata removed,
EBL, or operative time. There was no difference in the
rate of readmission, but women in the abdominal myomec-
tomy group were more likely to have a length of stay
exceeding 1 night (100% vs 10.6%; p , .001). There were
trends toward higher rates of any major complication
(20.6% vs 10.4%; p 5 .121), major intraoperative complica-
tions (13.8% vs 8.6%; p 5 .320), and major postoperative
complications (10.3% vs 2.3%; p 5 .053) in the abdominal
myomectomy group, but the differences were not statisti-
cally significant. Women in the abdominal myomectomy
group were significantly more likely to receive a postopera-
tive blood transfusion (10.3% vs 1.4%; p 5 .022).
In the MIS group, 4 women (1.8%) underwent conversion
to laparotomy secondary to high blood loss at a point remote
from case completion, 19 (8.6%) had an EBL R1000 mL, 2
(0.9%) had an unanticipated ICU admission secondary to
concern for airway edema in the setting of high-volume
blood transfusion, and 13 (5.8%) received a blood transfu-
sion. There were no cases of reoperation, visceral injury,
deep vein thrombosis or pulmonary embolism, ileus, bowel
obstruction, or sepsis (Table 1).
In the MIS group, women who experienced 1 or more
complications and those who did not experience complica-
tions were compared (Table 2). In characteristics that could
be identified preoperatively, the 2 subgroups did not differ in
Fig. 1
Rates of surgical approaches to myomectomy by year.
Vargas et al. Minimally Invasive Myomectomy for Large and Numerous Myomata 317

4. Table 1
Patient and surgical characteristics
Characteristic
MIS
myomectomy
(n 5 221)
Abdominal
myomectomy
(n 5 29) p value
Preoperative characteristics
Age, yr, mean 6 SD 37.3 6 7.0 41.1 6 6.0 .029
Race, n (%)
White 39 (17.6) 0 (0) .011
Black 132 (59.7) 27 (93.1)
Hispanic 1 (0.4) 0 (0)
Asian/Pacific
Islander
14 (6.3) 0 (0)
Native American 0 (0) 0 (0)
Other 35 (15.8) 2 (6.9)
Payment method, n (%)
Private 180 (81.4) 19 (65.5) .018
Medicaid 30 (13.6) 9 (31.0)
Medicare 0 (0) 0 (0)
Self-pay 0 (0) 0 (0)
Unknown 11 (5.0) 1 (3.4)
BMI, mean 6 SD* 28.0 6 6.5 28.8 6 6.4 .892
Parity, median
(IQR)
0 (0–0) 0 (0–0) .932
Previous abdominal surgery, n (%)y
Any abdominal
surgery
57 (25.7) 13 (44.8) .046
Laparotomy 49 (22.1) 12 (41.4) .036
Myomectomy 20 (9.0) 5 (17.2) .185
Procedure, n (%)
Laparoscopic 105 (47.5) d
Robotic 116 (52.5) d
Combined
hysteroscopic
procedure, n (%)
24 (10.9) 2 (6.8) .749
Indication for surgery, n (%)y
Abnormal
bleeding
95 (43.0) 10 (34.5) .429
Pelvic pain/
pressure
30 (13.6) 3 (10.3) .777
Infertility 23 (10.4) 9 (31.0) .005
Other 79 (35.7) 13 (44.8) .482
Uterine volume,
cm3
,
mean 6 SD*
683.7 6 658.7 1442.1 6 680.0 ,.001
Dominant myoma
diameter, cm,
mean 6 SD*
9.6 6 5.1 9.7 6 5.4 .783
Number of
myomata,
mean 6 SD*
4.5 6 4.1 18.3 6 12.1 ,.001
Type of myomata, n (%)y
Intramural 149 (67.4) 19 (65.5) .833
Subserosal 95 (43.0) 18 (62.1) .073
Broad ligament 10 (4.5) 1 (3.5) 1.000
Pedunculated 16 (7.2) 1 (3.5) .702
(Continued)
Table 1
Continued
Characteristic
MIS
myomectomy
(n 5 221)
Abdominal
myomectomy
(n 5 29) p value
Intraoperative characteristics
Operating time,
min,
mean 6 SD*
188.1 6 77.2 169.0 6 53.9 .111
Estimated blood
loss, mL,
mean 6 SD
341.1 6 460.5 464.7 6 410.8 .142
Specimen weight,
g, mean 6 SD*
410.6 6 384.8 908.9 6 470.4 ,.001
Postoperative characteristics
Length of stay
.1 day, n (%)
20 (10.6) 29 (100) ,.001
Readmission within
30 days, n (%)*
10 (4.5) 2 (6.9) .233
Complications
Any major
complication, n
(%)
23 (10.4) 6 (20.7) .121
Major
intraoperative
complication, n
(%)y
21 (9.5) 4 (13.8) .507
Conversion to
laparotomy
4 (1.8) d d
EBL .1000 mL 19 (8.6) 4 (13.8) .320
Blood
transfusion
11 (4.9) 1 (3.4) 1.000
Visceral injury 0 (0) 0 (0) d
Major postoperative
complications, n
(%)y
5 (2.3) 3 (10.3) .053
Reoperation 0 (0) 0 (0) d
ICU admission 2 (0.9) 0 (0) 1.000
DVT/PE 0 (0) 0 (0)
Blood
transfusion
3 (1.4) 3 (10.3) .022
Bowel
obstruction/
ileus
0 (0) 0 (0) d
Sepsis 0 (0) 0 (0) d
Any transfusion, n
(%)
13 (5.8) 4 (13.8) .119
BMI 5 body mass index; DVT 5 deep vein thrombosis; EBL 5 estimated blood
loss; ICU 5 intensive care unit; IQR 5 interquartile range; MIS 5 minimally
invasive surgery; PE 5 pulmonary embolism.
* Some patients had missing data. The total missing for each variable: BMI,
n 5 2; uterine volume, n 5 10; dominant myoma diameter, n 5 15; number
of myomata removed, n 5 7; operating time, n 5 20; specimen weight,
n 5 9; and readmission, n 5 30.
y
Categories are not mutually exclusive.
318 Journal of Minimally Invasive Gynecology, Vol 24, No 2, February 2017

5. terms of age, race, payment method, BMI, parity, previous
abdominal surgery, surgical approach, undergoing a com-
bined hysteroscopic procedure, or indication for surgery.
However, the women with complications had a greater
mean uterine volume on preoperative imaging
(1464.3 6 986.9 cm3
vs 586.1 6 534.1 cm3
; p , .001),
mean dominant myoma diameter (15.2 6 6.4 cm vs
9.5 6 4.49 cm; p 5 .002), and mean number of myomata
removed (6.7 6 6.3 vs 4.3 6 3.7; p 5 .031). Regarding intra-
operative characteristics, women with complications had
longer mean operating times (271.4 6 73.3 minutes vs
178.9 6 72.1 minutes; p , .001), greater mean EBL
(1282.7 6 823.8 vs 231.7 6 213.3 mL; p , .001), and
greater mean specimen weight (990.8 6 622.8 g vs
343.0 6 279.5 g; p , .001). Regarding postoperative char-
acteristics, women with complications were more likely to
have a length of stay exceeding 1 day (6.5% vs 42.9%;
p , .001). Readmission was also more common in the sub-
group with complications (13.6% vs 3.5%; p 5 .073),
although the difference did not reach statistical significance.
Only uterine volume, number of myomata removed, and
dominant myoma diameter met the criteria for inclusion in
the logistic regression model. In the unadjusted model, a
dominant myoma diameter of R12 cm significantly
increased the odds of complications compared with a domi-
nant myoma diameter of ,12 cm (odds ratio [OR], 6.45;
95% confidence interval [CI]. 2.47–18.26; p , .001). Uter-
ine volume R750 cm3
on preoperative imaging also signif-
icantly increased the odds of complications compared with
uterine volume ,750 cm3
(OR, 5.64; 95% CI, 2.28–15.43;
p , .001).
After adjusting for age, race, BMI, previous laparotomy,
and surgical approach, both dominant myoma diameter
R12 cm and uterine volume R750 cm3
remained predictive
of complications (OR, 7.44; 95% CI, 2.03–31.84; p 5 .004
and OR, 6.15; 95% CI, 1.55–30.02; p 5 .014, respectively).
Table 2
Patient characteristics by the presence of complications
Characteristic
No
complications
(n 5 198)
One or more
complications
(n 5 23) p value
Preoperative characteristics
Age, yr, mean 6 SD 37.2 6 7.1 38.5 6 6.0 .327
Race, n (%)
White 37 (18.6) 2 (8.7) .711
Black 117 (59.0) 15 (65.2)
Hispanic 1 (0.5) 0 (0.0)
Asian/Pacific
Islander
13 (6.7) 1 (4.3)
Native American 0 (0) 0 (0)
Other 30 (15.2) 5 (21.7)
Payment method, n (%)
Private 162 (81.8) 18 (78.2) .277
Medicaid 25 (12.6) 5 (21.7)
Medicare 0 (0) 0 (0)
Self-pay 0 (0) 0 (0)
Unknown 11 (5.6) 0 (0.0)
BMI, mean 6 SD* 28.0 6 6.5 28.4 6 7.0 .760
Parity, median (IQR) 0 (0–0) 0 (0–0) .503
Previous abdominal surgery, n (%)y
Any abdominal
surgery
51 (25.8) 6 (26.1) 1.0
Laparotomy 44 (22.2) 5 (21.7) 1.0
Myomectomy 20 (10.1) 0 (0) .238
Procedure, n (%)
Laparoscopic 91 (46.1) 14 (69.5) .192
Robotic 107 (54.0) 9 (39.1)
Combined
hysteroscopic
procedure, n (%)
23 (11.6) 1 (4.3) .482
Indication for surgery, n (%)y
Abnormal bleeding 83 (41.9) 12 (52.1) .379
Pelvic pain/pressure 28 (14.1) 2 (8.7) .747
Infertility 20 (10.1) 3 (13.0) .716
Other 42 (36.4) 7 (30.4) .300
Uterine volume, cm3
,
mean 6 SD*
586.1 6 534.1 1464.3 6 986.9 ,.001
Dominant myoma
diameter, cm,
mean 6 SD*
9.5 6 4.49 15.2 6 6.4 .002
Number of myomata,
mean 6 SD*
4.3 6 3.7 6.7 6 6.3 .0314
Type of myomata, n (%)y
Intramural 130 (65.7) 19 (82.6) .151
Subserosal 83 (41.9) 12 (52.1) .379
Broad ligament 9 (4.5) 1 (4.3) 1.0
Pedunculated 15 (7.6) 1 (4.3) 1.0
Intraoperative characteristics
Operative time, min,
mean 6 SD*
178.9 6 72.1 271.4 6 73.3 ,.001
EBL, mL, mean 6 SD 231.7 6 213.3 1282.7 6 823.8 ,.001
Specimen weight, g,
mean 6 SD*
343.0 6 279.5 990.8 6 622.8 ,.001
(Continued)
Table 2
Continued
Characteristic
No
complications
(n 5 198)
One or more
complications
(n 5 23) p value
Postoperative characteristics
Length of stay .1 days,
n (%)
11 (6.5) 9 (42.9) ,.001
Readmission in
30 days, n (%)*
7 (3.5) 3 (13.0) .073
BMI 5 body mass index; IQR 5 interquartile range; EBL 5 estimated blood
loss.
* Some patients had missing data. The total missing for each variable was as
follows: BMI, n 5 2; uterine volume, n 5 8; dominant myoma diameter,
n 5 14; number of myomata removed, n 5 5; operating time, n 5 18; specimen
weight, n 5 8; readmission, n 5 29.
y
Categories are not mutually exclusive.
Vargas et al. Minimally Invasive Myomectomy for Large and Numerous Myomata 319

6. Having 2 to 4 or R5 myomata removed did not increase the
odds of complication compared with having 1 myoma
removed in either the unadjusted or the adjusted logistic
regression model.
To assess the sensitivity and specificity of uterine volume
and dominant myoma diameter for predicting complica-
tions, we developed a ROC curve (Fig. 2). Uterine volume
was more predictive of complications than dominant myoma
diameter (area under the curve [AUC], 0.79 vs 0.74); howev-
er, combining both uterine volume and diameter in an ROC
curve conferred the best predictive performance (AUC,
0.81). Uterine volume R750 cm3
had a sensitivity of
73.7% and a specificity of 70.7% for any complication.
Dominant myoma diameter R12 cm had a sensitivity of
68.4% and a specificity of 79.3% for any complication. A
combination of dominant myoma diameter R10 cm and a
uterine volume R600 g had a sensitivity of 78.9% and a
specificity of 78.7% for any complication.
Discussion
The results of the present study support the conclusions of
previous reports showing that MIS invasive myomectomy is
a safe option for patients with myomata who desire uterine
preservation [1]. This study stands apart, however, in that
our patients had on average greater specimen weights,
more myomata removed, and larger diameters of dominant
myomas than previous reports of MIS myomectomy
[2,9,11–13]. For example, our mean specimen weight in
grams and number of myomata removed were
408.1 6 384.9 g and 4.5 6 4.1, respectively, compared
with 263.4 6 286.2 g and 3.54 6 4.10 in the largest
reported US-based cohort [12]. The largest European cohort
reported a mean dominant myoma diameter of
6.4 6 2.61 cm, compared with the 9.6 6 5.1 cm seen in
our cohort, and a mean number of myomas removed of
2.26 6 1.8 [6]. Our overall complication rate was compara-
ble to that in previous reports of MIS myomectomy
[6,13,14]. Our rates of transfusion were slightly higher
than the rates in some reports (6% vs 0.14% [6], 2.2%
[15], and 0% [2]), but comparable to other reports of cohorts
with a large myomata burden (18% [16], 8.1% [17], and
5.7% [18]). Our operative times were longer, likely influ-
enced by the greater surgical complexity of our cases
compared with most previous reports of MIS myomectomy.
In addition, it is possible that our operative times are influ-
enced by a practice of performing combined hysteroscopic
and MIS procedures, resident and fellow training, and a
recent change in our practice of tissue extraction owing to
the FDA safety warning on power morcellation. Notably,
the efficiency of the bagged paper roll technique [8]
currently used for tissue extraction in our practice has
increased with experience [19].
In our population, the odds of complications significantly
increased with a uterine volume R750 cm3
or a dominant
myoma diameter R12 cm. In a statistically validated ROC
curve, uterine volume predicted complications more accu-
rately than dominant myoma diameter; however, combining
the 2 characteristics increased the AUC to 0.81, signifying a
maximum specificity of approximately 80% and a maximum
sensitivity of approximately 80%. The combination of domi-
nant myoma diameter R10 cm and a uterine volume R600 g
came the closest to attaining maximum sensitivity and spec-
ificity in predicting complications.
Previous reports also have assessed clinical predictors of
complications for minimally invasive myomectomy. Sac-
cardi et al [5] reported an increased operative time and esti-
mated blood loss with intramural myomata .8 cm and
subserosal myomata .12 cm in patients undergoing MIS
myomectomy for a single myoma. Overall complication
rates were low (1.35% conversion to laparotomy, 0.45%
transfusion), and the authors concluded that MIS myomec-
tomy is safe even in patients with very large myomata. Sizzi
et al [6] reported results from a multicenter analysis of
approximately 2000 patients who underwent myomectomy
and found that a dominant myoma diameter .5 cm, removal
of 3 or more myomata, and removal of an intraligamentous
myoma increased the odds of major complications. Unlike
the aforementioned studies, our present study did not find
an association of myoma classification with complications,
but it is possible that the study was underpowered to detect
this difference. Sandberg et al [12] identified the number
of myomata and specimen weight .500 g as factors signif-
icantly increasing the odds of conversion to laparotomy, but
myoma classification was not associated with conversion in
their cohort.
Fig. 2
ROC curve showing the true-positive (sensitivity) and false-positive
rates (1-specificity) obtained for the logistic regression classifier that
predicts the occurrence of complications based on the diameter of the
dominant myoma and uterine volume. The ROC was determined using
leave-one-out cross-validation.
320 Journal of Minimally Invasive Gynecology, Vol 24, No 2, February 2017

7. The present study has several limitations. First, the data
collection was retrospective, which inherently limits the
quality of data and possible comparisons. Second, given
that the standard practice at our institution is to offer an
MIS approach to most patients requesting myomectomy,
only very complex cases are completed abdominally; thus,
we cannot compare outcomes between abdominal and MIS
myomectomy with equivalent dominant myoma diameters
and numbers of myomata removed. A previous randomized
trial comparing mini-laparotomy and conventional laparos-
copy for myomectomy in women with relatively high
complexity (mean specimen weights just under 400 g) found
no difference in complications between the 2 groups but
decreased blood loss, less postoperative pain, shorter length
of stay, and faster recovery in the laparoscopic group [2].
Another analysis of patients undergoing robotic myomec-
tomy matched to abdominal myomectomy controls by uter-
ine size found no difference in complication rates between
the 2 groups, even in cases with uterine size .16 weeks
[20]. The robotic group had higher associated costs, but
less intravenous narcotic use and shorter length of stay.
Another limitation of the present study is its small sample
size, which limited the ability to detect rare complications.
Finally, the generalizability of the predictive characteristics
must be tested in diverse patient populations and varying
levels of surgeon expertise to truly assess performance in
clinical practice.
A strength of this study is the relatively complex pathol-
ogy in our cohort, allowing for assessment of the MIS myo-
mectomy technique in a unique patient population. Our
results indicate that a minimally invasive approach is safe
even for these complex cases, although a more conservative
diameter and uterine volume cutoff for an MIS approach
may be necessary for safe completion, depending on surgeon
expertise. Moreover, 2 clinically relevant patient characteris-
tics that are typically seen on preoperative imaging were
identified as predictors of complications. After further
testing, such a predictive model may enhance preoperative
counseling and surgical planning, and may aid in deter-
mining the appropriate route of surgery. Another important
aspect of our findings is that although a large proportion of
the procedures in our cohort were performed after a nation-
wide change in tissue extraction practice triggered by the
FDA safety warning on power morcellation, these proced-
ures were safely completed in patients with relatively serious
pathology using a bagged scalpel tissue extraction technique.
In conclusion, our results suggest that when completed by
an experienced surgeon, MIS myomectomy is feasible and
safe in patients with large and numerous myomata. Hemor-
rhage and transfusion are the most common complications,
but can be predicted using a combination of easily defined
clinical characteristics. In our patient population, complica-
tions were most frequent in women with a dominant myoma
diameter R12 cm or uterine volume R750 cm3
. Further
study, preferably in a randomized controlled design, is needed
to identify other associated clinical characteristics and
specific cutoffs that are clinically relevant and reproducible
across diverse populations. Identifying patients at risk for
complications of surgery can facilitate appropriate referral,
guidepreoperative counseling,helpdetermine the appropriate
surgical approach, and allow for the timely implementation of
preventative measures for potential complications.
Acknowledgments
We thank Lindsey Powers Happ, MPH (George Washing-
ton University’s Milken Institute of Public Health) and
Richard Amdur, PhD (George Washington University Med-
ical Faculty Associates) for their assistance with manuscript
review, and Thalia Grant-Wisdom (George Washington Uni-
versity Medical Faculty Associates) and Ariundari Tsogoo
(George Washington University) for their assistance with
data collection.
References
1. Bhave Chittawar P, Franik S, Pouwer AW, Farquhar C. Minimally inva-
sive surgical techniques versus open myomectomy for uterine fibroids.
Cochrane Database Syst Rev. 2014;10:CD004638.
2. Alessandri F, Lijoi D, Mistrangelo E, Ferrero S, Ragni N. Randomized
study of laparoscopic versus minilaparotomic myomectomy for uterine
myomas. J Minim Invasive Gynecol. 2006;13:92–97.
3. Frishman GN, Jurema MW. Myomas and myomectomy. J Minim Inva-
sive Gynecol. 2005;12:443–456. quiz 457–458.
4. Falcone T, Parker WH. Surgical management of leiomyomas for
fertility or uterine preservation. Obstet Gynecol. 2013;121:856–868.
5. Saccardi C, Gizzo S, Noventa M, Ancona E, Borghero A, Litta PS.
Limits and complications of laparoscopic myomectomy: which are
the best predictors? A large cohort single-center experience. Arch Gy-
necol Obstet. 2014;290:951–956.
6. Sizzi O, Rossetti A, Malzoni M, et al. Italian multicenter study on com-
plications of laparoscopic myomectomy. J Minim Invasive Gynecol.
2007;14:453–462.
7. Goldstein SR, Horii SC, Snyder JR, Raghavendra BN, Subramanyam B.
Estimation of nongravid uterine volume based on a nomogram of gravid
uterine volume: its value in gynecologic uterine abnormalities. Obstet
Gynecol. 1988;72:86–90.
8. Wong WS, Lee TC, Lim CE. Novel vaginal ‘‘paper roll’’ uterine mor-
cellation technique for removal of large (.500 g) uterus. J Minim Inva-
sive Gynecol. 2010;17:374–378.
9. Gargiulo AR, Srouji SS, Missmer SA, Correia KF, Vellinga TT,
Einarsson JI. Robot-assisted laparoscopic myomectomy compared
with standard laparoscopic myomectomy. Obstet Gynecol. 2012;
120(2 Pt 1):284–291.
10. Sing T, Sander O, Beerenwinkel N, Lengauer T. ROCR: visualizing
classifier performance in R. Bioinformatics. 2005;21:3940–3941.
11. Sinha R, Hegde A, Mahajan C, Dubey N, Sundaram M. Laparoscopic
myomectomy: do size, number, and location of the myomas form
limiting factors for laparoscopic myomectomy? J Minim Invasive Gy-
necol. 2008;15:292–300.
12. Sandberg EM, Cohen SL, Jansen FW, Einarsson JI. Analysis of risk fac-
tors for intraoperative conversion of laparoscopic myomectomy. J
Minim Invasive Gynecol. 2016;23:352–357.
13. Gobern JM, Rosemeyer CJ, Barter JF, Steren AJ. Comparison of ro-
botic, laparoscopic, and abdominal myomectomy in a community hos-
pital. JSLS. 2013;17:116–120.
14. Iavazzo C, Mamais I, Gkegkes ID. Robotic assisted vs laparoscopic
and/or open myomectomy: systematic review and meta-analysis of
the clinical evidence. Arch Gynecol Obstet. 2016;294:5–17.
Vargas et al. Minimally Invasive Myomectomy for Large and Numerous Myomata 321

8. 15. Barakat EE, Bedaiwy MA, Zimberg S, Nutter B, Nosseir M, Falcone T.
Robotic-assisted, laparoscopic, and abdominal myomectomy: a compar-
ison of surgical outcomes. Obstet Gynecol. 2011;117(2 Pt 1):256–265.
16. Sinha R, Hegde A, Warty N, Mahajan C. Laparoscopic myomectomy:
enucleation of the myoma by morcellation while it is attached to the
uterus. J Minim Invasive Gynecol. 2005;12:284–289.
17. Chang WC, Huang PS, Wang PH, et al. Comparison of laparoscopic
myomectomy using in situ morcellation with and without uterine artery
ligation for treatment of symptomatic myomas. J Minim Invasive Gyne-
col. 2012;19:715–721.
18. Zhao F, Jiao Y, Guo Z, Hou R, Wang M. Evaluation of loop ligation of
larger myoma pseudocapsule combined with vasopressin on laparo-
scopic myomectomy. Fertil Steril. 2011;95:762–766.
19. Moawad GN, Abi Khalil ED, Opoku-Anane J, et al. Comparison of
methods of morcellation: manual versus power. Acta Obstet Gynecol
Scand. 2016;95:52–54.
20. Nash K, Feinglass J, Zei C, et al. Robotic-assisted laparoscopic myo-
mectomy versus abdominal myomectomy: a comparative analysis of
surgical outcomes and costs. Arch Gynecol Obstet. 2012;285:
435–440.
322 Journal of Minimally Invasive Gynecology, Vol 24, No 2, February 2017