Uterine sarcomas are relatively rare, accounting for only 3% to 8% of uterine cancers, and approximately 1 of every 800 smooth muscle tumors of the uterine is a leiomyosarcoma (LMS).1,2 Magnetic resonance imaging (MRI) is 1 of the most useful imaging modalities, but even with MRI, it is difficult to distinguish between malignant LMSs and benign leiomyomas. There are reports that the use of gray scale and Doppler sonographic findings can be helpful,3 with similar findings for the use of positron emission tomography using 2-[18F] fluoro-2-deoxy-D-glucose (PET-FDG)4; there is not at present an established practice for making an accurate presurgical differential diagnosis of these 2 tumor types. Recently, Yoshida et al5 reported that 16α[18F] fluoro-17β-estradiol PET was effective for differentiating uterine sarcomas from benign leiomyoma; however, in the 24 cases they studied, they had 2 false positives (a leiomyoma with adenomyosis and a hemorrhagic cellular variant leiomyoma) and 2 false negatives (a low-grade endometrial stromal sarcoma and an LMS).
The most reliable preoperative diagnostic method has been found to be a biopsy of the tumor. Because an LMS arises within the uterine smooth muscle, a biopsy of the endometrial malignant tissue is difficult to perform, and in many cases, the tumor is found only at surgery. Various authors have reported that LMS may be present in the submucosa of the uterus in 30% to 50% of patients, but even in such cases, the biopsy diagnosis is not easy.1 Although past reports have demonstrated the usefulness of a needle biopsy for uterine tumors,6,7 the use of the biopsy option has generally not gained widespread acceptance. To support its wider use, we report our experience with the ultrasound-guided needle biopsy for uterine tumors.
MATERIALS AND METHODS
We enrolled patients who had unusual MRI findings, in particular, those with a higher intensity tumor image than normal myometrium (50% or higher on T2-weighted images). In addition, we considered whether the patient had evidence of hemorrhagic necrosis (indicated by high signal intensity on T1-weighed image and nonenhancement in the same area) or irregular margins. From January 2005 to August 2012, we obtained informed consent from 63 patients to perform an ultrasound-guided needle biopsy of their uterine mass suspected by MRI of being a malignant tumor. This is a retrospective study, and the exact number of uterine tumors during that period with high or undetermined T2-weighed MRI signal is unclear. However, about 10% of the patients with uterine tumors met our eligibility to perform needle biopsy. Some patients desire surgery without performing preoperative needle biopsy, and others desire regular follow-up without performing needle biopsy. Needle biopsy was mainly performed to the cases that did not desire the surgery immediately but were suspected malignancy rather than typical leiomyoma.
The median patient age was 47.0 years (range, 21–83 years). The solid part of the mass was characterized as having a high signal intensity on a T2-weighted MRI (ie, the mass had a high signal intensity on a diffusion-weighted image [DWI] and a low apparent diffusion coefficient [ADC] map) and was specified in 3 dimensions. We evaluated MRI with DWI and ADC in 18 patients.
The biopsy was performed, without general anesthesia, using a Bard Magnum biopsy system (C. R. Bard, Inc, Murray Hill, NJ), with an automatic cutting needle (30 cm long, 18 gauge, with a 19-mm notch). The biopsy needle was inserted through a sterile needle guide that was attached to the ultrasound transducer. The ultrasound transducer was manipulated to position the uterine tumor along a line on the screen. After the tip of the needle was within the uterine tumor, the biopsy gun was fired, and a core of tissue was obtained. We performed 3 biopsies to reduce sampling error. A 50-mg diclofenac sodium suppository was administered before conducting the needle biopsy. All patients took a cephem antibiotic for 3 to 5 days after the needle biopsy. There were no cases that required postbiopsy hospitalization.
We recommended surgery for all cases diagnosed by biopsy with malignant tumors. In the cases diagnosed with benign tumors, we either observed or operated, depending on any myoma-related symptoms, infertility, or the patient’s desire to undergo surgery. In the operative cases, we compared the results of the needle biopsy and the results of the surgical specimen. The patients who did not undergo surgery were regularly followed at the hospital and have had no outcome event suggestive of malignancy at 4 to 81 months of follow-up (median, 41 months) after the needle biopsy.
Table 1 shows the MRI appearance of the 63 patients who underwent a needle biopsy. Among the 46 patients who had solid parts with high signal intensity on T2-weighted MRI, there were 2 atypical malignant cases. One case had a well-defined cystic area that was suspected of being a leiomyoma with degenerative liquefaction; the other had a nodule with a high-signal T2-weighed image on MRI typical of a leiomyoma. Among 16 patients who had a mass with an MRI pattern of hemorrhagic necrosis and/or an irregular margin, there were 10 malignant tumors by biopsy. We could not determine the T2-weighted image signal ratio in 1 case because we could not find any normal myometrium for comparison.
Figure 1 shows the outcomes of 63 patients with uterine tumors, with a high signal intensity or an undetermined ratio (1 patient), on T2-weighted imaging, which we preoperatively performed an ultrasound-guided needle biopsy on. Among the 63 patients, malignant tumors were identified by biopsy in 12 cases (19%), and benign tumors were diagnosed in 51 cases (81%). Table 2 shows the characteristics of patients with a malignant tumor.
Of 12 patients with the diagnosis of a malignant tumor by needle biopsy, 11 had surgery performed, and 1 was treated with chemotherapy because it was an advanced case. Two advanced cases were treated with surgery after chemotherapy. Surgery specimens were classified as LMS (n = 10) and endometrial stromal sarcoma (n = 1). Of 51 patients with the diagnosis of benign tumors, 27 underwent a hysterectomy or myomectomy. After the surgery, the condition of 1 of the 27 patients was diagnosed with a low-grade endometrial stromal sarcoma, and the other 26 were diagnosed with a leiomyoma.
Twenty-four patients were managed conservatively, being observed every 3 months according to their symptoms. Tumor and uterine sizes were measured by ultrasonography at each visit. These 24 patients lacked any evidence of uterine malignancy during subsequent follow-ups.
In the 38 cases where surgery was performed, the sensitivity and specificity of the biopsy diagnosis was 91.7% and 100%, respectively. The positive predictive value of the biopsy was 100%, and the negative predictive value was 96.2%. We had 2 cases of infection that needed surgery to control. Both cases were LMS with widespread necrosis. Other than these 2 cases, no major complications, such as intraperitoneal hemorrhage or injury to adjacent structures that required surgery, were observed.
To treat a uterine tumor properly without doing unnecessary surgery, differentiating uterine sarcomas from benign leiomyomas is very important. The MRI is an important tool for diagnosing uterine tumors. The typical MRI appearance of an LMS reveals a heterogeneous appearance with intermediate to high signal intensity on T2-weigted images and enhances well. In a typical case, MRI reveals high intensity on T1-weigted images with coagulated tumor necrosis.8 Confeld et al9 reported finding a distribution of imaging criteria between leiomyoma and other mesenchymal neoplasms. Objective criteria included T1 and T2 signal characteristics, enhancement pattern, the presence of cystic changes, and an ill-defined margin.9 However, in a previous report of patients with findings of coagulative tumor cell necrosis, the incidence of LMS was limited to 68%, and in the 5% of the patients without findings of coagulative tumor cell necrosis, LMS was noted.4 There are many case reports that had few of these typical imaging characteristics. On the other hand, there is also 1 report claiming that malignant tumors were rare in cases that were operated on solely because they seemed to be rapidly growing.10
Recently, by using diffusion-weighted MRI, it became possible to perform diagnoses that are more detailed.11 We performed our very first uterine DWI in September 2008. From January 2010 onward, we performed DWI for all such cases. Namimoto et al12 reported that ADC combined with T2-weigted imaging is significantly better than ADC or T2-weighted imaging alone at differentiating between sarcomas and leiomyomas. However, there are limitations to their study; the number of sarcomas was relatively small, and some leiomyomas were not proven diagnostically.12
In 2002, Kawamura et al reported obtaining good diagnostic results using a transcervical needle biopsy for the differential diagnosis between uterine sarcoma and leiomyoma.7 They reported that a transcervical needle biopsy, using a histopathological scoring system that was established based on the criteria proposed by Bell et al,8 is highly precise, with an especially high negative predictive value.
By performing a needle biopsy, we can diagnose whether the uterine mass is a benign or malignant tumor. As a result, we can avoid unneeded surgery in cases with a poor general status or distant metastasis, or do only limited procedures, such as fertility-sparing surgery, laparoscopic surgery, or transvaginal surgery.
In addition to it, DWI reduced the cases that needed needle biopsy. So we have performed needle biopsy with limited cases in recent years.
In our report, although the exact number of uterine tumors with high or undetermined T2-weighed MRI signal between January 2005 and August 2012 is unclear, we could recruit 63 cases for needle biopsy. Of those 63 cases, 13 cases turned out to be sarcomas. On the other hand, a total of 41 uterine sarcomas were treated in our hospital during that period. Therefore 13 (31%) of 41 uterine sarcomas were biopsied as a consequence. Because conducting the needle biopsy provided a correct differential diagnosis, 24 of the 63 cases were able to avoid an unneeded operation altogether, 7 had laparoscopic surgery, and 4 had transvaginal surgery.
We only targeted tumors with high intensity T2 signal on MRI. On the other hand, there were no malignant cases among 51 patients with a low intensity T2 signal. We had suspected that these were malignant tumors because of their rapid growth or because they had hemorrhagic necrosis. Among the 46 patients who were not strongly suspected of having a malignant tumor (because they were only high signal intensity on T2-weighted images), there were 2 atypical cases of malignant tumor that were then diagnosed as malignant tumor by needle biopsy.
We evaluated 18 patients with DWI; in the group of 11 patients who had increases in DWI and decreases in ADC, there were 4 malignant tumors; in the 7 patients who had an increase DWI but no decrease in ADC, there were no malignant tumors. There is now the strong possibility that we can perform a needle biopsy more efficiently by restricting its use to patients with tumors that have a high signal with T2-weighted DWI imaging and a low signal with ADC.
One of the possible critical drawbacks of doing a needle biopsy is the possibility of spreading cancer cells. In the 12 tumor cases where we performed surgery after the needle biopsy, the median number of days from biopsy to surgery was 37.5 days (range, 10–113 days); in this cohort, there was no evidence that conducting the biopsy had spread the cancer. In cases where we strongly suspected a malignancy by MRI, we first evaluated with computed tomography whether there were distant metastases before conducting the needle biopsy. In 5 advanced cases, the computed tomography pointed out the presence of metastatic sites before the needle biopsy.
When performing a needle biopsy, there can be problems with accurate sampling, such as difficulties with the specimens, the diagnostic accuracy of any microsample, and the frequency of postbiopsy complications. In our report, 3 of the 68 specimens (4.8%) were difficult to diagnose; 1 was too small, and 2 consisted of only necrotic material. One patient underwent repeated ultrasound-guided needle biopsies and was with diagnosis of a malignant tumor. In another patient, repeating the biopsy was difficult because of an infection, and her condition was diagnosed after surgery with an LMS. Two other patients who underwent repeated biopsies did not undergo surgery because a malignant tumor was not strongly suspected.
In our research, the negative predictive value of the biopsy was 96.2%. There was 1 case in which the postsurgical diagnosis of a low-grade endometrioid stromal sarcoma (ESS) was different than by needle biopsy, which had suggested it was a leiomyoma. The existence of invasion is important for the distinction of an ESS from a benign tumor, and such invasion may be difficult to determine in a biopsy specimen.
There are numerous reports of cases simultaneously having leiomyoma and LMS. Mittal and Joutovsky reported that a spectrum of morphologic and immunohistochemical changes, from benign to malignant, is seen in 50% of LMSs, indicative of the progression of some leiomyoma to LMS.13 Thus, when an unambiguous nodule is present within a myoma, we should bear in mind the possibility of the presence of a malignancy.
In our report, there were 2 cases of infection associated with an LMS with widespread necrosis. When we performed the needle biopsy on these cases, we had to consider carefully which area to perform the biopsy on and which type of patient management to conduct afterward.
Except in cases of low-grade malignancy, a presurgical guided needle biopsy for difficult cases may generally be as useful as a postsurgical histological examination of the tumor. Thus, we have found that in cases where a uterine malignancy is suspected by MRI, a preoperative needle biopsy is a highly effective tool for determining that most such cases are benign, a result which greatly expands our ability to use more conservative and appropriate treatment options.
On the other hand, there are complications, such as infection, that can occur from conducting a needle biopsy. It is also important to take into consideration that there will be rare cases that are just too difficult to diagnose by needle biopsy, that there may be the possibility of cancer spread after biopsy, and that malignancy can never be 100% ruled out by this means. We are thus ethically required to be selective of the cases we chose to diagnose in this manner.
There are now several other presurgical treatment options that can greatly benefit the patient. Examples are administering chemotherapy or molecular-targeted drugs, conducting radiotherapy, or induction of arterial embolism in the tumor. Going forward, a guided needle biopsy will help us choose which of these nonsurgical treatment plans is most appropriate for our patients.
© 2014 by the International Gynecologic Cancer Society and the European Society of Gynaecological Oncology.