Magnetic Resonance Imaging to Rule out Leiomyosarcoma in Patients Undergoing Surgery for Leiomyomas: A Real World Experience in an Unenhanced Patient Population

Abstract

Objective:

Surgery for leiomyomas is common; yet, no reliable test can help distinguish a benign leiomyoma and malignant leiomyosarcoma (LMS).

Materials and Methods:

This retrospective observational cohort study evaluated patients before and after implementation of a protocol to identify LMS, which included magnetic resonance imaging (MRI) with diffusion-weighted imaging.

Results:

This study revealed the incidence of uterine pathology, as well as MRI, lactate dehydrogenase (LDH), and pathology results, in 1085 patients—479 before and 606 after implementation of the protocol. Two cases of LMS were identified in the postprotocol cohort, and 70% of the patients underwent MRI. Test statistics for MRI to detect LMS in this cohort were: sensitivity of 100%; specificity of 67%; positive predictive value of 1%; negative predictive value of 100%; false–positive rate of 33%; and false–negative rate of 0%. For patients with both MRI and LDH results (358/606, 59%), 56.7% had normal MRI and LDH, 9.8% had negative MRI but high LDH, 6.4% had abnormal MRI and high LDH, and 27.1% had abnormal MRI and normal LDH.

Conclusion:

Preoperative MRI for detecting LMS had high a sensitivity and a high false–positive rate, which warrants caution in interpreting MRI results, particularly in women of childbearing age.

Introduction

Leiomyoma is a common condition, with 70%–80% of women receiving the diagnosis during their lifetimes, and hysterectomy and myomectomy are among the most-common procedures performed in the United States.^1,2 Unfortunately, there is significant difficulty in discriminating between a benign leiomyoma and a malignant leiomyosarcoma (LMS); when LMS is present, power morcellation can lead to an inadvertent spread of malignant cells throughout the abdominal cavity.

In November 2014, the U.S. Food and Drug Administration issued a safety communication cautioning providers regarding these risks.³ The American Association of Gynecologic Laparoscopists responded shortly thereafter with a practice report⁴ and the American College of Obstetricians and Gynecologists recently published a committee opinion to address recent updates in the literature.⁵ Reports have subsequently been published on rising laparotomy rates, which, in turn, are associated with increased surgical complications.^6–8 Although many hospitals either banned or restricted the use of power morcellation, insurance carriers and device companies limited the availability of the technique further, and contained (“in-bag”) tissue extraction products became commercially available, the problem of whether or not to perform manual tissue extraction (either contained or uncontained) is still highly relevant to surgical decision-making.

In order to address the concerns regarding morcellation while continuing to offer minimally invasive options to patients, a protocol was instituted at the New York University (NYU) Langone Health hospital in New York City, based on previously published work, which required preoperative magnetic resonance imaging (MRI) to identify patients at increased risk for LMS.⁹

This current study's objectives were to: (1)

Evaluate the effectiveness of this protocol in identifying LMS

(2)

Determine the test characteristics of MRI for identifying LMS.

Materials and Methods

This study was approved by the NYU School of Medicine's institutional review board prior to data collection. All data and materials are available for review.

A retrospective chart review of patients at NYU Langone Health who underwent surgery for leiomyomas from April 23, 2013, through April 23, 2015, was performed. The protocol was instituted on April 23, 2014, and patients' charts were separated into 2 cohorts: 1 year before (preprotocol) and 1 year after (postprotocol) implementation.

For surgeons utilizing uncontained power morcellation, the protocol required the following pre-operative testing: MRI with diffusion-weighted imaging, total serum lactate dehydrogenase (LDH), and LDH isoenzyme 3 (LH3). In the preprotocol group, MRIs were done at the surgeon's discretion; the same was the case for those surgeons who chose not to comply with the protocol in the postprotocol cohort.

The MRIs were performed on either a 1.5T or 3T clinical system using a torso-phased array coil. Imaging sequences obtained included: multiplanar turbo-spin-echo T2-weighed imaging (T2WI), axial in-and-opposed-phase gradient-echo T1-weighted imaging (T1WI), axial fat-suppressed single-shot echo-planar diffusion-weighted imaging (DWI) at multiple B-values ranging from B0 to B800, and dynamic three-dimensional fat-suppressed spoiled gradient-echo T1WI at multiple timepoints following administration of intravenous contrast (0.1 mmol/kg of gadopentetate dimeglumine or gadobutrol).

An algorithm was devised by a senior radiology attending, using specific characteristics to evaluate risks for LMS. These were: leiomyoma size(s) and number(s); presence of a large dominant mass occupying the majority of the uterus; atypical signal-intensity characteristics (including hyperintense and heterogeneous signals on T2-weighted images and increased signals on T1-weighted images compatible with hemorrhage); ill-defined, irregular, nodular, or infiltrative lesion borders; avid enhancement on postcontrast images with areas of necrosis; diffusion restriction within the lesion(s); extrauterine extension; rapid growth in comparison to prior examinations; and presence of ascites, implants, and/or abnormal lymph nodes. These criteria were chosen based on prior published reports describing the MRI features of LMS.^9–14

After evaluating all of these MRI features, the likelihood of leiomyosarcoma was determined by the interpreting radiologist based on the number of atypical features. While there have been several studies of various scoring schemes and MRI features, none of them have been widely accepted in clinical use.^15–20 Therefore, no formal scoring system was used and it was left up to the radiologist to make a clinical judgment with regard to possibility of LMS and to state what the likelihood of it might be. All radiologists reading MRIs at this institution are fellowship-trained and specialize in reading abdominal and pelvic MRIs.

A patient was eligible for uncontained laparoscopic power morcellation if all three results (MRI, total LDH, and LH3) were normal. If one of the results was abnormal, the surgeon was restricted from using the power morcellator; clinical management, as well as the resultant mode of surgery and method of specimen removal, was left to the surgeon's discretion.

Women between ages 18 and 99 with symptomatic fibroid uteri who underwent intra-abdominal surgery for leiomyomas were included. This included myomectomies (laparoscopic, robotic, and open approaches) and hysterectomies (both total and supracervical, with laparoscopic, robotic, vaginal, laparoscopically assisted vaginal, and open approaches). Hysteroscopic myomectomies were excluded. Patients were screened by a review of the operating room schedule during the study period (Fig. 1). Eligible patients' charts were reviewed to extract relevant data including: patient demographics, preoperative workup, type of surgery, and method of tissue extraction. The internal institutional gynecologic cancer database (created and maintained for tumor-board treatment planning) was reviewed to confirm that all cases of LMS were captured in the study. All data were entered into a protected database by 1 of the current authors and reviewed by several of the other current authors to ensure accuracy.

FIG. 1.

Study flowchart. This figure identifies the process of inclusion in the final analysis.

The outcome metrics were: (1)

Correlation between MRI results and final pathology

(2)

Test characteristics (true positives, true negatives, false–positives, false–negatives, sensitivity, specificity, positive predictive value [PPV] and negative predictive value [NPV] and accuracy) of MRI testing for detection of LMS and other uterine pathologies.

Patient demographic data were compared, using exact Mann–Whitney-U tests, and binary demographic parameters (menopausal status, history of cancer, tamoxifen use, and pelvic radiation) were compared using Fisher's exact tests. The incidence of uterine pathology was compared using Fisher's exact tests. MRI and LDH results in the pre- and postprotocol groups stratified by pathology were compared with Fisher's exact tests. All statistical tests were conducted using SAS software version 9.3 (SAS Institute, Cary, NC). Test performance characteristics were calculated using 2 × 2 tables of the raw data.

Results

A total of 1678 patients undergoing intra-abdominal surgery (myomectomies and hysterectomies) during the study period were identified. Of these, a total of 593 patients were excluded for having benign non-leiomyoma indications for surgery or for histories of prior uterine malignant or premalignant conditions (i.e., endometrial hyperplasias or endometrial cancers). A total of 1085 patients were included in the study, with 479 patients in the preprotocol group and 606 patients in the postprotocol group.

Patient demographics are shown in Table 1. The preprotocol group had more white women, more postmenopausal women, and more women taking tamoxifen, compared to the women in the postprotocol group. The postprotocol group had more Hispanic women. No differences were noted in ages, body mass indices (BMIs), cancer histories, or histories of pelvic radiation.

Table 1.

Selected Demographic and Clinical Characteristics

Characteristics	Preprotocol group n = 479	Postprotocol group n = 606	p-Value
Age (years)	42 (23–70)	43 (22–76)	0.114
BMI (kg/m²)	27.4 (16–52)	27.5 (17–53)	0.893
Race
White	154 (32)	161 (27)	0.044
Black	193 (40)	231 (38)	0.466
Hispanic	21 (4)	47 (8)	0.021
Other	50 (10)	79 (13)	0.153
Unstated	3 (1)	10 (2)	0.124
Menopausal status
Premenopausal	442 (92)	559 (92)	1.000
Postmenopausal	19 (4)	10 (2)	0.022
Unknown	18 (4)	37 (6)	0.095
Family history of cancer	142 (30)	162 (27)	0.307
Personal history of cancer	25 (2)	18 (3)	0.062
Tamoxifen use	10 (2)	3 (0)	0.022
Pelvic radiation	0 (0)	1 (0)	1.000

Notes: Age and BMI are listed as average (range). Other characteristics are listed as n (%).

p-Values were determined by exact Mann–Whitney-U-tests (for BMI and age) and by Fisher's exact tests (for the rest of the characteristics listed here). Bolded p-values are significant.

BMI, body mass index.

Given that the 2 cohorts were based on the timing of protocol initiation, the prevalence of uterine pathology were compared in both time periods/cohorts (Table 2). There were no differences between the pre- and postprotocol groups in incidence of leiomyomas, uterine malignancies, or atypical uterine tumors. Only benign non-leiomyoma diagnoses (including adenomyosis, endometrial polyps, and others) were different between the 2 groups: 22% versus 30% in the pre- and postprotocol groups, respectively.

Table 2.

Final Pathology Reported in Pre- and Post-Protocol Groups

Pathology	Preprotocol group n = 479	Postprotocol group n = 606	p-Value
Benign
Benign leiomyoma	448 (94)	570 (94)	0.80
Benign non-leiomyoma^a	106 (22)	182 (30)	0.004
No uterine pathology reported	3 (0.6)	0 (0)	0.09
Benign total^b	451 (94)	587 (97)	0.04
Endometrial hyperplasia	5 (1)	5 (0.8)	0.76
Cancer
Gynecologic cancer total^c	6 (1)	4 (0.7)	0.35
Leiomyosarcoma	0 (0)	2 (0.3)	0.51
Endometrial adenocarcinoma	4 (0.8)	1 (0.2)	0.18
Other gynecologic cancer^d	2 (0.4)	1 (0.2)	0.59
Nongynecologic cancer^e	1 (0.2)	0 (0)	0.44
Smooth-muscle tumor variant^f	13 (3)	10 (12)	0.29
Any smooth-muscle abnormality^g	15 (3)	13 (2)	0.34

Notes: Data are n (%).

p-Values were determined by Fisher's exact tests. Bolded p-values are significant.

Column totals do not equal group totals and percentages do not add up to 100% because more than one pathology was often found within a single specimen (for example, the same specimen might have leiomyomas and adenomyosis).

“Benign non-leiomyoma” includes adenomyosis, adenomyoma, endometrial polyp, chronic endometritis, denuded endometrium, polypoid adenomyoma, endometriosis, fibrosis, infarctive endometrium, endometrium with cystic degeneration, atrophic endometrium, focal compression atrophy, tubal metaplasia, glandular crowding, proliferative endometrium, foreign-body giant-cell reaction, & chronic serosal inflammation.

“Benign total” includes benign leiomyomas, benign non-leiomyoma, & no uterine pathology categories.

“Gynecologic cancer total” includes leiomyosarcoma, endometrial adenocarcinoma, & other gynecologic cancers.

“Other gynecologic cancer” includes aggressive angiomyxoma, Müllerian adenosarcoma, & endometrial stromal sarcoma.

Nongynecologic cancer” includes metastatic breast adenocarcinoma.

“Smooth-muscle tumor variant” includes adenostromyoma, smooth-muscle tumor of uncertain malignant potential, symplastic leiomyoma, cellular leiomyoma, adenomatoid tumor, mitotically active leiomyoma, lipoleiomyoma, leiomyoma with atypia, & mitotically active smooth-muscle tumor with atypia

“Any smooth-muscle abnormality” includes leiomyosarcoma, other gynecologic cancer, & smooth-muscle tumor variants.

There were no differences in the rates of LMS, other gynecologic cancers, or smooth-muscle tumor variants between the 2 groups. There were only 2 cases of LMS in the entire study (2/1085, or 0.2%), both of which occurred in the postprotocol cohort. There were 5 cases of endometrial adenocarcinoma: 4 cases in the preprotocol group and 1 case in the postprotocol group. There were 23 cases of smooth-muscle tumor variants identified: 13 in the preprotocol group and 10 in the postprotocol group. There were 3 cases classified as other gynecologic cancers: 2 in the preprotocol group and 1 in the postprotocol group. These included an aggressive angiomyxoma, a Müllerian adenosarcoma, and an endometrial stromal sarcoma.

Tables 3 and 4 list MRI results by final pathologic diagnoses. In both the pre- and postprotocol groups, the majority of patients with abnormal MRI results had benign pathologies. Of the 479 patients in the preprotocol group, 176/479 (37%) had MRI. Of these, 92/176 (52%) were read as normal and 84/176 (48%) were read as abnormal. Of those read as normal, 1 was found to have nongynecologic cancer (metastatic breast cancer) and 3 were found to have smooth-muscle tumor variants; the majority (92%) had benign leiomyomas. Of the 84/176 MRIs read as abnormal, 93% were benign leiomyomas. Of the remainder, 1 gynecologic cancer (other than LMS) and 3 smooth-muscle tumor variants were in this group. In the postprotocol group, 426/606 patients (70%) had MRI (Table 4). Of these 426 patients, 284 had normal MRI results (67%) and 142 (33%) had abnormal MRI results. Both cases of LMS were identified as abnormal on MRI; however, most of the MRIs classified as abnormal had normal pathology (94% had leiomyomas and 18% had benign non-leiomyoma pathologies). There were three STUMP (Smooth Muscle Tumor of Undetermined Significance) tumors and one other gynecologic cancer that were identified as abnormal by MRI.

Table 3

MRI Results by Final Pathology in Preprotocol Group

Uterine pathology	Normal MRI n = 92	Abnormal MRI n = 84
Benign
Benign leiomyoma	85 (92)	78 (93)
Benign non-leiomyoma^a	11 (12)	14 (17)
No uterine pathology reported	1 (1)	0 (0)
Benign total^b	82 (89)	76 (90)
Endometrial hyperplasia	0 (0)	0 (0)
Cancer
Gynecologic cancer total^c	6 (1)	4 (0.7)
Leiomyosarcoma	0 (0)	0 (0)
Endometrial adenocarcinoma	0 (0)	0 (0)
Other gynecologic cancer^d	0 (0)	1 (1)
Nongynecologic cancer^e	1 (1)	0 (0)
Smooth-muscle tumor variant^f	3 (2)	3 (4)
Any smooth-muscle abnormality^g	3 (3)	4 (5)

Notes: Data are n (%).

Column totals do not equal group total and percentages do not add up to 100% because more than one pathology was often found within a single specimen (for example, the same specimen might have leiomyomas and adenomyosis).

“Benign total” includes benign leiomyomas, benign non-leiomyoma, & no uterine pathology categories.

“Gynecologic cancer total” includes leiomyosarcoma, endometrial adenocarcinoma, & other gynecologic cancers.

“Other gynecologic cancer” includes aggressive angiomyxoma, Müllerian adenosarcoma, & endometrial stromal sarcoma.

“Nongynecologic cancer” includes metastatic breast adenocarcinoma.

“Any smooth-muscle abnormality” includes leiomyosarcoma, other gynecologic cancer, & smooth-muscle tumor variants.

MRI, magnetic resonance imaging.

Table 4.

MRI Results by Final Pathology in Postprotocol Group

Uterine pathology	Normal MRI n = 284	Abnormal MRI n = 142
Benign
Benign leiomyoma	275 (97)	133 (94)
Benign non-leiomyoma^a	72 (25)	26 (18)
No uterine pathology reported	1 (1)	0 (0)
Benign total^b	275 (97)	134 (94)
Endometrial hyperplasia	1 (0)	2 (1)
Cancer
Gynecologic cancer total^c	6 (1)	4 (0.7)
Leiomyosarcoma	0 (0)	2 (1)
Endometrial adenocarcinoma	0 (0)	1 (0)
Other gynecologic cancer^d	0 (0)	1 (0)
Nongynecologic cancer^e	0 (0)	0 (0)
Smooth-muscle tumor variant^f	2 (1)	2 (1)
Any smooth-muscle abnormality^g	2 (1)	4 (3)

Notes: Data are n (%).

“Benign total” includes benign leiomyoma, benign non-leiomyoma, & no uterine pathology categories.

“Gynecologic cancer total” includes leiomyosarcoma, endometrial adenocarcinoma, & other gynecologic cancers.

“Other gynecologic cancer” includes aggressive angiomyxoma, Müllerian adenosarcoma, & endometrial stromal sarcoma.

“Nongynecologic cancer” includes metastatic breast adenocarcinoma.

“Any smooth-muscle abnormality” includes leiomyosarcoma, other gynecologic cancer, & smooth-muscle tumor variants.

MRI, magnetic resonance imaging.

The performance of MRI was examined in the preoperative detection of LMS in the postprotocol study timeperiod. Of 426 MRIs, 142/426 (33%) were abnormal (i.e., test-positive). Of these 142, only 2 were true positives. In the same cohort, 284 MRIs were read as negative; all of these were confirmed as true negatives by pathologic review; no false–negative MRIs were noted. The test statistics for MRI to detect LMS in this postprotocol cohort were: sensitivity of 100% (95% confidence interval [CI]: 15.8%–100.0%), specificity of 67% (95% CI: 62.0%–71.1%), a PPV of 1% (95% CI: 0.2%–5.0%), an NPV of 100% (95% CI: 98.1%–100.0%), a false–positive rate of 33% (95% CI: 29.9%–38.0%), and false–negative rate of 0% (95% CI: 0%–84.2 0%).

In the postprotocol group with false–positive MRI (n = 140), the most-common abnormal features were “heterogeneous slightly high signal on T2, necrosis,” and “degeneration,” followed by “ascites,” and “lesion borders not well-circumscribed.” Detailed analyses were not performed for each variable used, given the relatively small numbers in each group.

It was not possible to perform similar analyses on total LDH and LH3 with and without MRI results because too few patients had all three tests performed to make these analyses meaningful. In the postprotocol group, the mean LDH and LH3 values were 284 and 25, respectively. For the 2 LMS cases identified in this group, 1 patient had an elevated total LDH of 2648 (no LH3 test was performed), and the second patient had an LDH of 343 (high), and an LH3 of 23 (normal). Test characteristics were also examined using a combination of MRI and LDH in preoperative detection of LMS in the postprotocol group. In the preprotocol group, only 10 of 479 (2.1%) patients had LDH testing, and only 4 of 479 patients (0.83%) had both MRI and LDH. Therefore, patients were not used from the preprotocol group in this analysis. A total of 358/606 patients (59.1%) had both tests performed. Of these, 23/358 (6.4%) had abnormal results for both tests (i.e., test-positive); 2 out 23 were true positives. Negative results were found in 203/358 patients for both tests (56.7%), all of which were true negatives. In addition, discordance between MRI and LDH testing was noted. Finally, 97/358 patients (27.1%) had abnormal MRI and normal LDH and 35/358 (9.8%) had normal MRI and high LDH.

Discussion

This study was conducted to examine the performance of a preoperative protocol (MRI and LDH testing) for triaging patients when planning minimally invasive surgery for leiomyomas. This study showed that a negative MRI result is very effective for ruling out LMS, as MRI alone had a very high NPV. However, the high false–positive rate and the low PPV are concerning. In general, false–positive rates are reliant on the prevalence of the condition under investigation. Although the rate of LMS was low (2/1085, or 0.2%), it was within the range of expected, based on the literature. The addition of total LDH to MRI led to a marked decrease in the false–positive rate in concordance with a prior study by Goto et al.⁹; however, having an abnormal MRI alone with normal LDH may be worrisome enough to steer patients and providers alike into unnecessary interventions.

The current study's results differed from previous reports. Goto et al. performed a retrospective review of a prospectively collected cohort of patients undergoing surgery for leiomyomas.⁹ In their study, the use of “dynamic” or diffusion-weighted imaging was associated with a 0% false–positive rate. Other retrospective studies have similarly recorded very favorable test characteristics for MRI utilizing a variety of grading systems.^10–14 In the Goto et al. series, there were 10 patients with LMS over a 10-year period in a cohort of 227 patients (4.4%)⁹; the other series shared an enhanced number of cases with LMS and this high prevalence of the disease may have aided their overall test statistics.^10–14 The current study showed that, in a normal unenhanced population that clinicans take care of in the “real world,” it remains difficult to identify the rare cases of LMS without counting a large number of benign cases as “abnormal.”

With respect to the current study's MRI protocol, imaging sequences used closely paralleled those used in prior studies.^{9,10,15,20,21} Widely used standard MRI imaging sequences included T1-weighted, T2-weighted, diffusion-weighted, and pre- and post-gadolinium contrast-enhanced imaging with slight modifications, as follows. In the study by Goto et al.⁹ post-contrast images were acquired dynamically every 12 seconds for 3 minutes. In the current study's protocol, postcontrast images were obtained at 60 seconds and at 180 seconds only; however, in the Goto et al. study,⁹ those researchers found that it was most important to obtain images at 40–80 seconds after contrast injection as sarcomas showed rapid enhancement during this phase. In addition, Suzuki et al.¹⁰ used T2-weighted and diffusion-weighted sequences similar to those used in the current study's protocol to devise a grading system for leiomyomas and uterine smooth-muscle tumors. In the current study, apparent diffusion coefficient (ADC) mapping was used to quantitate diffusion-restriction rather than a calculation of ADC values.

The current study has several limitations. It was a retrospective study from a single institution, and, as such, generalizability could be limited. There were some demographic differences between the 2 cohorts, but the current authors do not believe that these were clinically important. Patient cohorts were utilized based on the timing of initiation of a protocol and compared patients before and after protocol initiation. However, no difference was found between the 2 groups with regard to the incidence of uterine pathology, which indicated that there was likely no change in the incidence of LMS. A small number of uterine malignancies were detected during the study time period and no uterine LMS's were identified in the year prior to the protocol. As has been pointed out in the prior literature, it is difficult to conduct a study that would be powered to detect differences in LMS rates, as the best estimates for the rate of LMS place a range of 1 in 770 and 1 in 10,000 surgeries for presumed symptomatic leiomyomas.^5,22,23

In keeping with the relative incidence, more cases of undiagnosed endometrial adenocarcinomas were found than cases of uterine leiomyosarcomas. Although the protocol was intended to identify sarcomas, the results reinforced the need for general thoroughness in preoperative planning, including timely endometrial biopsy for patients who are at increased risk for endometrial cancer. Other strengths of the current study include a large and “real-life” study population that was unenhanced for uterine sarcoma pathology. In addition, the radiologists reading the MRIs were all specialists in gynecologic imaging, allowing for high-quality reads.

The current study protocol required an MRI prior to the performance of surgery if power morcellation was under consideration. The protocol was only enforced for providers interested in using power morcellation. The current authors are not suggesting that MRI should be a prerequisite to all surgeries involving fibroid uteri. In many cases, an ultrasound should be the imaging study of choice, given the decreased expense and time involved. MRIs are widely used in clinical practice during preoperative planning for “surgical mapping” of leiomyomas, particularly prior to laparoscopic myomectomy. It is critical that surgeons recognize the high false–positive rate associated with MRI when utilized to rule out LMS. These abnormal results can have a significant negative impact, irrespective of the decision to use power morcellation. False–positive results can lead to additional testing, unnecessary oncology referrals, and undue patient distress.

Moreover, in patients who are interested in future fertility, counseling regarding the type of procedure (myomectomy versus hysterectomy) could be tainted by false–positive MRI results with significant implications. Surgeons should be aware of these limitations and review the possibility of a false–positive test result with patients as part of preoperative counseling. The decision to proceed with myomectomy in these cases requires an extensive informed-consent process as a myomectomy may be associated with a worse prognosis in patients ultimately diagnosed with LMS.^24,25

Future directions for this research include examining the effects of this protocol on downstream factors, such as laparoscopy rates, use of contained morcellation, patient acceptance of the potential harm associated with the high false–positive rate for MRI, and overall costs.

Conclusions

MRI has a high false–positive rate when used as a screening test for LMS in women undergoing surgery for presumed leiomyomas. Prospective studies are warranted, given that the current study showed a significant discrepancy between false–positive rates, compared to existing literature.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was obtained for this research.

References

Baird

, Dunson

, Hill

, Cousins

, Schectman

. High cumulative incidence of uterine leiomyoma in black and white women: Ultrasound evidence. Am J Obstet Gynecol, 2003; 188:100.

Wright

, Herzog

, Tsui

, et al. Nationwide trends in the performance of inpatient hysterectomy in the United States. Obstet Gynecol, 2013; 122(2[pt1]):233.

U.S. Food and Drug Administration (FDA). Laparoscopic Uterine Power Morcellation in Hysterectomy and Myomectomy: FDA Safety Communication. April 17, 2014. Online document at: www.chartrrg.com/wp-content/uploads/2014/07/FDA-Warning.pdf Accessed August 12, 2019.

AAGL [American Association of Gynecologic Laparoscopists] Advancing Minimally Invasive Gynecology Worldwide. AAGL practice report: Morcellation during uterine tissue extraction. J Minim Invasive Gynecol, 2014; 21:517.

American College of Obstetricians and Gynecologists (ACOG). ACOG Committee Opinion No. 770: Uterine morcellation for presumed leiomyomas. Obstet Gynecol, 2019; 133:e238.

Barron

, Richard

, Robinson

, Lamvu

. Association of the U.S. Food and Drug Administration morcellation warning with rates of minimally invasive hysterectomy and myomectomy. Obstet Gynecol, 2015; 126:1174.

Harris

, Swenson

, Uppal

, Kamdar

, Mahnert

, As-Sanie

, Morgan

. Practice patterns and postoperative complications before and after US Food and Drug Administration safety communication on power morcellation. Am J Obstet Gynecol, 2016; 214:98.e1.

Wright

, Chen

, Burke

, Hou

, Tergas

, Ananth

, Hershman

. Trends in use and outcomes of women undergoing hysterectomy with electric power morcellation. JAMA, 2016; 316:877.

Goto

, Takeuchi

, Sugimura

, Maruo

. Usefulness of Gd-DTPA contrast-enhanced dynamic MRI and serum determination of LDH and its isozymes in the differential diagnosis of leiomyosarcoma from degenerated leiomyoma of the uterus. Int J Gynecol Cancer, 2002; 12:354.

10.

Suzuki

, Wada

, Nakajima

, et al. Magnetic resonance imaging grading system for preoperative diagnosis of leiomyomas and uterine smooth muscle tumors. J Minim Invasive Gynecol, 2018; 25:507.

11.

, Liu

, Qiang

, Zhang

, Ma

. Diffusion-weighted imaging for differentiating uterine leiomyosarcoma from degenerated leiomyoma. J Comput Assist Tomogr, 2017; 41:599.

12.

Lakhman

, Veeraraghavan

, Chaim

, et al. Differentiation of uterine leiomyosarcoma from atypical leiomyoma: Diagnostic accuracy of qualitative MR imaging features and feasibility of texture analysis. Eur Radiol, 2017; 27:2903.

13.

Lin

, Yang

, Huang

, et al. Comparison of the diagnostic accuracy of contrast-enhanced MRI and diffusion-weighted MRI in the differentiation between uterine leiomyosarcoma/smooth muscle tumor with uncertain malignant potential and benign leiomyoma. J Magn Reson Imaging, 2016; 43:333.

14.

Sato

, Yuasa

, Fujita

, Fukushima

. Clinical application of diffusion-weighted imaging for preoperative differentiation between uterine leiomyoma and leiomyosarcoma. Am J Obstet Gynecol, 2014; 210:368e1.

15.

Tanaka

, Nishida

, Tsunoda

, Okamoto

, Yoshikawa

. Smooth muscle tumors of uncertain malignant potential and leiomyosarcomas of the uterus: MR findings. J Magn Reson Imaging, 2004; 20:998.

16.

Laughlin-Tommaso

, Stewart

. Moving toward individualized medicine for uterine leiomyomas. Obstet Gynecol, 2018; 132:961.

17.

Thomassin-Naggara

, Dechoux

, Bonneau

, et al. How to differentiate benign from malignant myometrial tumours using MR imaging. Eur Radiol, 2013; 23:2306.

18.

Nagai

, Takai

, Akahori

, et al. Novel uterine sarcoma preoperative diagnosis score predicts the need for surgery in patients presenting with a uterine mass. Springerplus, 2014; 3:678.

19.

Nagai

, Takai

, Akahori

, et al. Highly improved accuracy of the revised PREoperative sarcoma score (rPRESS) in the decision of performing surgery for patients presenting with a uterine mass. Springerplus, 2015; 4:520.

20.

Barral

, Placé

, Dautry

, et al. Magnetic resonance imaging features of uterine sarcoma and mimickers. Abdom Radiol (NY), 2017; 42:1762.

21.

Tamai

, Koyama

, Saga

, Morisawa

, Fujimoto

, Mikami

, Togashi

. The utility of diffusion-weighted MR imaging for differentiating uterine sarcomas from benign leiomyomas. Eur Radiol, 2008; 18:723.

22.

Pritts

, Vanness

, Berek

, Parker

, Feinberg

, Olive

. The prevalence of occult leiomyosarcoma at surgery for presumed uterine fibroids: A meta-analysis. Gynecol Surg, 2015; 12:165.

23.

Parker

, Kaunitz

, Pritts

, Olive

, Chalas

, Clarke-Pearson

, Berek

; Leiomyoma Morcellation Review Group. U.S. Food and Drug Administration's guidance regarding morcellation of leiomyomas: Well-intentioned, but is it harmful for women?. Obstet Gynecol, 2016; 127:18.

24.

Perri

, Korach

, Sadetzki

, Oberman

, Fridman

, Ben-Baruch

. Uterine leiomyosarcoma: Does the primary surgical procedure matter?. Int J Gynecol Cancer, 2009; 19:257.

25.

Morice

, Rodriguez

, Rey

, et al. Prognostic value of initial surgical procedure for patients with uterine sarcoma: Analysis of 123 patients. Eur J Gynaecol Oncol, 2003; 24(3–4):237.