Materials and Methods
The subjects included 5 patients confirmed to have a total of 10 leiomyosarcoma lesions (5 primary lesions, one 2-cm-diameter small-intestine recurrence, one 3.9-cm-diameter sacral recurrence, one 12-cm-diameter omentum metastasis, and 2 other disseminated Douglas pouch–lesion cases of 12.7 cm and 2.7 cm) by postoperative pathological examination (4 cases) or autopsy (1 case), and 76 patients were confirmed to have a total of 83 leiomyoma nodules by postoperative pathological examination. All patients underwent diffusion-weighted magnetic resonance (MR) imaging (MRI) preoperatively from April 2008 through December 2011. Clinical findings are shown in Tables 1 and 2 .
Variable | Leiomyomas (76 cases, 83 lesions) | Leiomyosarcomas (5 cases, 10 lesions) | Significance ( t test or Welch, Fisher) |
---|---|---|---|
Age, y (n = 76, 5) | 44.6 ± 8.1 (29-75) Pm: 5/76 (6.6%) | 55 ± 4.6 (50-62) Pm: 2/5 (40.0%) | P < .01 P < .01 |
No. of lesions (n = 76, 5) | 1.1 ± 0.3 (1-3) Median: 1 | 2 ± 0 (1-3) Median: 2 | |
Lesion size, cm (n = 83, 10) | 8.7 ± 3.9 (2.8-21.5) Median: 8.2 | 10.7 ± 6.0 (2-19.5) Median: 12 | P = .323 |
Uterine weight, g (n = 76, 5) | 745.8 ± 654.5 (94-3110) Median: 583.5 | 1640 ± 770.6 (590-2260) Median: 2050 | P < .01 |
No. of pregnancies (n = 76, 5) | 1.3 ± 1.4 (0-6) Median: 1 Nongravida: 32/76 (42.1%) | 1.8 ± 1.3 (0-3) Median: 2 Nongravida: 1/5 (20%) | P = .407 P = .330 |
Parity (n = 76, 5) | 1.0 ± 0.9 (0-4) Median: 1 Nonparous: 37/76 (48.7%) | 1.6 ± 1.1 (0-3) Median: 2 Nonparous: 1/5 (20%) | P = .259 P = .213 |
Total LDH, IU/L (n = 76, 5) | 179.1 ± 28.6 (114-233) Median: 176 Positive cases: 1/76 (1.3%) | 297.4 ± 97.9 (180-419) Median: 271 Positive cases: 4/5(80%) | P = .054 P < .001 |
Chief symptom | Leiomyoma (76 cases) | Leiomyosarcoma (5 cases) |
---|---|---|
Hypermenorrhea | 32 | 1 |
Anemia | 23 | 2 |
Noted in medical screening | 17 | |
Abnormal genital bleeding | 14 | 2 |
Dysmenorrhea | 13 | |
Palpable abdominal mass | 12 | 2 |
Lower abdominal pain | 7 | 2 |
Abdominal dilatation | 7 | 2 |
Lumbago | 3 | |
Prolonged menses | 3 | |
Pollakiuria | 2 | 1 |
Sterility | 2 | |
Abnormal discharge | 2 | |
Dysuria | 1 | |
Others | 7 |
All MR examinations were performed with a 1.5-T superconductive MR unit (Excelart Vantage version 8.02; Toshiba Medical Systems Corp, Ootawara city, Tochigi Prefecture) using an 8-channel phased-array coil. Fast-spin-echo T1-weighted images (repetition time/echo time [TR/TE]: 520/10 milliseconds) in the axial plane and fast-spin-echo T2-weighted images (TR/TE: 4086-4264/90 milliseconds) in the sagittal and axial plane were obtained for all subjects. Dynamic MRI was performed for 2 leiomyosarcomas and 15 leiomyomas that were obtained at 30-40 (first phase), 60-70 (second phase), and 180 (third phase) seconds after the start of rapid injection (3 mL/s) with 0.2 mL/kg (=0.1 mmol/kg) of gadolinium diethylenetriamine pentaacetic acid. All these images were obtained prior to acquiring the DWI. DWI were obtained in the sagittal and axial planes by using a single-shot echo-planar imaging sequence (TR/TE: 5915-100 milliseconds, flip angle: 90 degrees, excitations: 3, matrix: 128 × 96, bandwidth: 1302 Hz/pixel) with inversion recovery pulse and a parallel imaging technique. The b values corresponding to the diffusion sensitizing gradient were 0 and 1000 s/mm 2 . All sagittal images, including DWI, were uniform with a thickness of 5.0 mm, an intersection gap of 1.5 mm, and a field of view of 300 mm. ADC maps were derived automatically on a pixel-by-pixel basis from the DWI according to the following equation: diffusion coefficiency = –[natural logarithm(S1) – natural logarithm(S0)]/(b1 – b0), where b0 and b1 represent lower and higher b values, respectively, and S0 and S1 are the signal intensities for these b values.
All myoma nodules that could be identified clearly by MRI and measured 2.8-21.5 cm were included in the study. All leiomyosarcoma lesions were the same as above and measured 2-19.5 cm. Lesions measuring <2 cm were not included as they may not be detectable by DWI.
The signal intensity of DWI was based on a 3-tier scale: low (below the intensity of skeletal muscle; eg, iliopsoas), intermediate, and high (above the intensity of uterine endometrium) as determined by b values at 1000 s/mm 2 ( Figure 1 ).
Two persons measured the ADC values as follows: 10 leiomyosarcoma lesions, 33 DWI low-intensity leiomyoma nodules, 41 DWI intermediate-intensity leiomyoma nodules, and 9 DWI high-intensity leiomyoma nodules. The circular region of interest used to measure the ADC values was placed such that it would be as large as possible within the solid tumoral component, not contain necrotic cysts as much as possible by referring to the T2-weighted images, and not contain high-intensity areas on the T1-weighted images, thus avoiding the susceptibility effect of hemorrhagic contents. If several ADC values were measured, the lowest value was adopted.
Pathological diagnoses were performed by a single pathologist, and the diagnosis of each histological type of uterine sarcoma was established according to routine guidelines. The staging system applied was the new International Federation of Gynecology and Obstetrics system for uterine sarcoma.
This was a retrospective study, and Microsoft Excel for Macintosh (Microsoft Corp, Redmond, WA) and StatMate III for Macintosh, version 3.14 (ATMS Corp, Hongo, Bunkyo-ku, Tokyo, Japan) were used to assess statistical differences via t tests or Welch tests and Fisher exact test. A value of P < .05 was considered statistically significant. We have informed consent and ethics committee approval.
Results
Comparison between clinical images of myoma cases and leiomyosarcoma cases
The data described below are summarized in Tables 1 and 2 . The 76 myoma patients had an average age of 44.6 ± 8.1 years (range, 29–75 years), while that of the 5 uterine sarcoma patients was 55 ± 4.6 years (range, 50–62 years), with a significant difference ( P < .01). In addition, 5 (6.6%) of the 76 myoma patients and 2 (40%) of the 5 uterine sarcoma patients were menopausal women. In the myoma patients, the most common symptom was excessive menstruation, which was evident in 32 cases, followed by 23 anemic cases and 13 dysmenorrhea cases. The main symptom among the leiomyosarcoma cases was presence of an abdominal mass, and 2 patients reported a feeling of abdominal distension, lower abdominal pain, unexplained bleeding from the sexual organs, and anemia.
The mean number of prior pregnancies and births was insignificant between myomas and leiomyosarcomas. The number of nongravida and nonparous women was insignificant between myomas and leiomyosarcomas.
The mean myoma weight was 745.8 ± 654.5 g (range, 94–3110 g) (n = 76). The mean leiomyosarcoma weight was 1640 ± 770.6 g (range, 590–2260 g) (n = 5) ( P < .01, regarding myoma vs leiomyosarcoma). The mean greatest dimension of the 83 myoma nodules was 8.7 ± 3.9 cm (range, 2.8–21.5 cm), insignificant from that of the 10 leiomyosarcoma lesions, which were 10.7 ± 6.0 cm (range, 2–19.5 cm). The mean total lactate dehydrogenase (LDH) of the 76 myomas was 179.1 ± 28.6 IU/L (range, 114–233 IU/L), insignificant ( P = .054) from that of the 5 leiomyosarcomas (297.4 ± 97.9 IU/L; range, 180–419 IU/L). However, an elevated total LDH positive rate for the 5 leiomyosarcomas was 80% (4 cases), which differed significantly from that of the 76 myomas with 1.3% (1 case) ( P < .001).
Examination of DWI signal intensity
The data described below are summarized in Figure 2 . Of the 83 myoma nodules, low DWI signals were evident in 33 cases (39.8%), intermediate signals in 41 (49.4%), and high signals in 9 (10.8%). Of the 10 leiomyosarcoma lesions, 1 (10%) exhibited an intermediate signal and 9 (90%) exhibited high signals, no low-signal lesions were detected.
Considering the assumption that intermediate- or high-intensity DWI-showing lesions indicated the presence of a leiomyosarcoma (DWI intermediate- or high-intensity group) and that low-signal lesions indicated a myoma (DWI low-intensity group), the technique yielded the following regarding the diagnosis of leiomyosarcoma lesions: sensitivity, 100%; specificity, 39.8%; positive predictive value (PPV), 16.7%; negative predictive value, 100%; and accuracy, 46.2%.
The data described below are summarized in Table 3 . Of 9 high-intensity myoma nodules on DWI, only 1 (11.1%) was dark on a T2 image, and of 33 low-intensity myoma nodules on DWI, 28 (84.8%) were dark on T2 images. Nine high-intensity leiomyosarcoma lesions showed the following characteristics on T2 images: 4 heterogeneous, 2 bright, and 3 dark. The 3 dark lesions on T2 images were 2 small recurrent lesions (2 cm; 3.9 cm) and a disseminated omentum lesion (12 cm). One intermediate, primary leiomyosarcoma lesion was heterogeneous on a T2 image. High-intensity areas on T1 images demonstrated 5 myoma nodules (6%) and 4 leiomyosarcoma lesions (40%). Border irregularity was seen with 2 myoma nodules (2.4%) and 2 leiomyosarcoma lesions (20%), but 4 recurrent, metastatic, and disseminated lesions and 1 intermediate primary lesion on DWI had no border irregularity. With dynamic study, we found no correlation with DWI for either myomas or leiomyosarcomas.
Variable | T2 image | High-intensity area on T1 image | Border irregularity | Dynamic study | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Dark | Heterogeneous | Bright | Total | – | + | Total | Regular | Irregular | Unclear | Total | Early phase | Later phase | Total | ||
Leiomyoma | |||||||||||||||
DWI | High | 1 | 5 | 3 | 9 | 8 | 1 | 9 | 7 | 2 | 0 | 9 | 2 | 1 | 3 |
Intermediate | 13 | 18 | 10 | 41 | 38 | 3 | 41 | 41 | 0 | 0 | 41 | 4 | 3 | 7 | |
Low | 28 | 3 | 2 | 33 | 32 | 1 | 33 | 33 | 0 | 0 | 33 | 5 | 0 | 5 | |
Total | 42 | 26 | 15 | 83 | 78 | 5 | 83 | 81 | 2 | 0 | 83 | 11 | 4 | 15 | |
Leiomyosarcoma | |||||||||||||||
DWI | High | 3 a | 4 | 2 a | 9 | 5 a | 4 | 9 | 4 a | 1+1 a | 3 | 9 | 1 | 0 | 1 |
Intermediate | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | |
Low | |||||||||||||||
Total | 3 | 5 | 2 | 10 | 6 | 4 | 10 | 5 | 2 | 3 | 10 | 1 | 1 | 2 |