Uterine Arterial Embolization for Leiomyomas: Perfusion and Volume Changes at MR Imaging and Relation to Clinical Outcome1
1 From the Robert Steiner Magnetic Resonance Unit, Department of Imaging, Faculty of Medicine at Imperial College, Hammersmith Hospital, DuCane Rd, London W12 0HS, England. Received March 9, 2001; revision requested April 11; revision received May 25; accepted July 5.
ABSTRACT
PURPOSE: To monitor changes in perfusion and volume of uterus and leiomyomas after bilateral uterine artery embolization (UAE) and to correlate immediate perfusion changes with subsequent reduction in leiomyoma volume and clinical outcome.
MATERIALS AND METHODS: Eleven consecutive women underwent magnetic resonance (MR) imaging before UAE, immediately after, and at 1 and 4 months. Reduction in maximal enhancement above baseline at 90 seconds (ME90) after injection of the dominant leiomyoma immediately after embolization was correlated with its volume reduction at 4 months and with clinical response at 12 months.
RESULTS: Forty-five leiomyomas were noted (mean, four per patient). Myometrium enhanced briskly (ME90 of 110%), with a reduction in ME90 to 26% immediately after embolization. Initial leiomyoma ME90 was lower (P < .001), but it suppressed to baseline levels immediately after embolization. At 1 and 4 months, myometrial perfusion returned to normal, but leiomyoma perfusion remained suppressed (P < .001). Immediate reduction in leiomyoma ME90 correlated with clinical response (Spearman {rho} = 0.64). Leiomyomas initially high in SI on T2-weighted images showed significantly greater volume reduction than those low in SI (P = .006). Well-perfused leiomyomas did not show greater volume reduction than those that were poorly perfused. Volume reduction did not correlate with improvement in clinical symptom score.
CONCLUSION: Immediate reduction in leiomyoma perfusion after bilateral UAE correlates with clinical response, whereas leiomyomas initially high in SI on T2-weighted images indicate a likely greater volume reduction.
Index terms: Arteries, therapeutic embolization, 969.1264 • Arteries, uterine, 969.12 • Leiomyoma, 854.318 • Uterine neoplasms, MR, 854.121411, 854.121412, 854.12143 • Uterine neoplasms, therapy, 854.1264
INTRODUCTION
Uterine arterial embolization (UAE) is gaining popularity as a safe and cost-effective way of treating leiomyomas that produce symptoms of menorrhagia, sensation of pressure, or a mass (1–4). Although the procedure obliterates flow in both uterine arteries, findings from clinical follow-up suggest that perfusion of the myometrium recovers, with reports of successful full-term pregnancies (5). However, documentation of immediate or late changes in perfusion of the leiomyoma compared with the surrounding uterine muscle is poor, and the mechanism of selective infarction of leiomyomas is poorly understood. The purpose of this study was to use magnetic resonance (MR) imaging to monitor changes in uterine and leiomyoma perfusion and volume immediately and 1 and 4 months after embolization of both uterine arteries and to correlate immediate perfusion changes with subsequent reduction in leiomyoma volume and with clinical outcome.
MATERIALS AND METHODS
Eleven consecutive women (age range, 29–48 years; mean age, 40 years ± 6.6 [SD]) who were to undergo bilateral UAE of leiomyomas, which were causing symptoms of menorrhagia, were recruited. The research ethics committee of the hospital granted approval for MR imaging, and informed consent was obtained after the MR imaging procedure was explained. All patients had menorrhagia, whereas abdominal distension (feeling of a mass) was present in eight cases. In one patient, images were obtained with a 1.5-T system (Eclipse; Marconi Medical Systems, Cleveland, Ohio) and a body coil and in the other 10 patients with a 0.5-T system (Apollo; Marconi Medical Systems) and a pelvic phased-array coil.
Imaging was performed before and within 30 minutes of the completion of embolization and 1 and 4 months after the procedure. MR imaging consisted of coronal and sagittal T1-weighted spin-echo sequences (496/20 repetition time msec/echo time msec) obtained before and after contrast material administration, sagittal T2-weighted fast spin-echo sequences (4,000/84 [effective], echo train length of 16), and single-section dynamic gadolinium-enhanced (Omniscan [0.1 mmol per kilogram of body weight of gadopentetate dimeglumine]; Nycomed Amersham, Oslo, Norway) sequences (31/13; 90° flip angle) obtained with a temporal resolution of 4 seconds. In the patient who underwent MR imaging with the 1.5-T system, an echo time of 8 msec was used for dynamic contrast-enhanced sequences (192 x 256 matrix, two signal averages, 35-cm field of view, and 8-mm section thickness). At least 5 hours had elapsed between the administration of the preembolization dose of gadolinium-enhanced contrast material and the immediate acquisition of postembolization images. In eight cases, preembolization imaging was performed the day prior to the procedure.
Embolization was performed with a femoral approach by using a 5-F cobra (Cordis, Miami, Fla) catheter. Both uterine arteries were selectively catheterized in all cases. Embolization was achieved by injecting 355–500-µm polyvinyl alcohol particles (Contour Emboli, St Albans, England) into each uterine artery until the flow had ceased. Pain was managed with intravenously administered opiates, either as a bolus or with a patient-controlled pump. In two cases, no intravenous postprocedural analgesia was required.
A postal questionnaire on clinical symptoms was completed by each patient before embolization and at 12 months after embolization. All recorded data were related to the symptoms of menses: namely length, blood loss, and associated pain. Each symptom was scored from 1 to 3 to produce a total final clinical response score of 9. Menses length of less than 3 days was scored 1; 4–5 days, 2; and 6 or more days, 3. Light menstrual loss (spotting) was scored 1; loss requiring one to two changes of sanitary napkins per 24 hours, 2; and heavy loss with or without clots and requiring more than two changes of sanitary napkins per 24 hours, 3. Occasional twinges of pain were scored 1; cramps, 2; and cramps requiring analgesia, 3. The change in the clinical score at 12 months from that at preembolization levels was calculated as a percentage with the following formula: difference in score/preembolization score x 100.
Serum follicle-stimulating hormone (FSH) levels were measured by using a microparticle enzyme immunoassay before embolization and at the same stage of the menstrual cycle 4 months after the procedure.
Image Analyses
In each patient, the number of leiomyomas and the individual signal intensities (SIs) relative to the myometrium were recorded on initial, immediate, and 1- and 4-month postembolization T2-weighted images as hypointense, isointense, or hyperintense. All leiomyoma-to–adjacent myometrium SI ratios were also recorded on T1-weighted images obtained before embolization, immediately after, and at 1 and 4 months. Leiomyoma locations (submucous, myometrial, and subserosal) were also recorded.
In every patient, perfusion data were obtained from the dynamic enhanced series before embolization, immediately after, and at 1 and 4 months. Regions of interest were drawn over an area of maximal enhancement within the dominant leiomyoma and an adjacent area of myometrium in the same z plane (head to foot) to avoid SI variations due to the pelvic phased-array coil. Regions of interest were placed by one author (N.M.d.S.) and were between 0.3 and 1.5 cm2 in area (median, 0.5 cm2). SIs were recorded with an imaging perfusion profile software (Marconi Medical Systems, Highland Heights, Ohio). In addition, for every leiomyoma depicted on each single section of the dynamic series, the enhancement above baseline was noted. Leiomyomas were well perfused when the plateau enhancement (after 90 sec) was equivalent to or greater than that of the adjacent myometrium, or poorly perfused, when the plateau enhancement was less than that of adjacent myometrium.
Volume measurements of the dominant leiomyoma and the entire uterus (including all leiomyomas) were made by one author (N.M.d.S.) by drawing around the area with an electronic caliper and by using standard scanner software to compute the area per section. The sum of the areas was multiplied by the section thickness. Measurements were made before embolization, immediately after, and at 1 and 4 months. Individual volume measurements of all leiomyomas were made similarly on preembolization and 4-month postembolization images. Measurements obtained before and after embolization were made while viewing corresponding images simultaneously so that individual leiomyomas could be clearly identified and compared.
Statistical Analyses
Statistical analyses were done by using software (Unistat Ltd, London, England). Data were tested for normality by using normality profiles and a Shapiro-Wilk normality test to determine subsequent use of appropriate tests for statistical comparison. For each time point of the dynamic series for both myometrium and dominant leiomyoma, the percentage enhancement above baseline was calculated by using the equation (SIpost - SIpre)/SIpre, where post is after enhancement and pre is before enhancement. The calculated change in SI from that at baseline would eliminate any T2* effects in the absolute values of SI. The mean and SD of the pooled patient data were plotted before embolization, immediately after, and at 1 and 4 months. The maximal enhancement above baseline at 90 seconds (ME90) of myometrium and that of leiomyoma before embolization was compared by using a paired t test. The ME90 of myometrium and leiomyoma before embolization was similarly compared with the values obtained immediately after embolization and at 1 and 4 months.
All distributions tested normal. The percentage change in volume at 4 months between areas of high SI and those of low SI on T2-weighted images of leiomyomas or between well- and poorly perfused leiomyomas was compared by using a nonpaired two-sample t test with unequal variance. Changes in individual leiomyomas were considered independently of each other. High SI well- versus poorly perfused leiomyomas were compared as were low SI well- versus poorly perfused leiomyomas, large (">="
50 cm3) versus small (<50 cm3) volume leiomyomas, and submucosal versus myometrial leiomyomas.
In dominant leiomyomas, the immediate reduction in ME90 (reduction in a percentage enhancement between preembolization and immediate postembolization images) was correlated with their reduction in volume at 4 months by using Spearman rank correlation coefficient. Similar correlation coefficients were obtained between immediate reduction in ME90 of myometrium and leiomyoma and clinical response (percentage change in symptom scores at 12 months) and between total volume reduction of all leiomyomas in each patient at 4 months and clinical response at 12 months.
The significance of the difference in serum FSH levels at 4 months from those obtained before embolization was assessed by using a paired t test.
RESULTS
Patients had multiple leiomyomas (range, one to eight leiomyomas; mean, four; median, four). Forty-five leiomyomas were noted, and it was possible to obtain volume measurements from all of them. Leiomyoma volumes ranged from 0.6 to 434 cm3 (mean, 89.9 cm3 ± 110.5; median, 41.2 cm3). Although 11 leiomyomas were isointense or hyperintense on T2-weighted images, compared with the myometrium, leiomyomas were generally hypointense (n = 34). Thirty-five of 45 leiomyomas were included in a single section of the dynamic study: 14 of these were well perfused (equivalently or better) compared with the myometrium, whereas 21 were poorly perfused. A variety of locations within the myometrium were also noted: 34 were myometrial; seven, submucous; three, subserosal; and one, cervical in origin.
SI Changes
Immediately following embolization, an increase in SI within leiomyomas, compared with the myometrium, was seen on T1-weighted images regardless of their size, location, initial SI, or perfusion characteristics. This was relatively homogenous. Mean leiomyoma-to–adjacent myometrium SI ratios had increased from 1.1 ± 0.1 before embolization to 1.4 ± 0.1 immediately after embolization. This hyperintensity on T1-weighted images persisted at 1 and 4 months, though it gradually became more heterogeneous and less marked (Fig 1). With T2 weighting, there were no observable changes in SI on the immediate postembolization images. However, after 1 and 4 months, there was a reduction in SI on T2-weighted images in 43 of 45 leiomyomas (Fig 2). Two leiomyomas remained isointense with the myometrium.
fig.ommitted
Figure 1. Midsagittal T1-weighted spin-echo (420/20) MR images obtained through the pelvis before (A), immediately after (B), 1 month (C), and 4 months (D) after bilateral UAE show multiple uterine leiomyomas, which are hypo- (wide arrows) and isointense (thin arrows) with myometrium. Immediately after embolization, there is a relative increase in SI in the leiomyomas (arrows) and to a lesser extent in the myometrium. In C, this appearance becomes more heterogenous and in D, it is seen to fade (arrows).
fig.ommitted
Figure 2. Midline sagittal T2-weighted spin-echo (2,500/80) MR images show uterus with leiomyomas before (A) and 4 months after (B) UAE. The large hypo- and hyperintense leiomyomas in A (large and small arrow, respectively) have reduced in size and SI in B (arrows).
Perfusion Changes
Data for ME90 and percentage change in ME90 were normally distributed. Before embolization, the myometrium was better perfused than leiomyomas, with a significantly higher ME90 (P < .001). Immediately following embolization, there was a dramatic reduction in the overall uterine perfusion. However, while perfusion of the myometrium reduced considerably from an ME90 of 110% before embolization to an ME90 of 26% immediately following embolization (Fig 3), perfusion of the leiomyoma had virtually stopped (Table 1, Fig 4). After 1 and 4 months, perfusion of the myometrium recovered to normal levels, while leiomyoma perfusion remained extremely poor (Table 1, Figs 5, 6). Thus, there was a significant difference between ME90 of myometrium before and that immediately after embolization (P < .001) but not between ME90 of myometrium before and that at 1 or 4 months after embolization. Leiomyomas, on the other hand, showed a significant difference between ME90 before embolization and that immediately after embolization, as well as at 1 and 4 months (P < .001).
fig.ommitted
Figure 3. Graph illustrates the changes in myometrial perfusion profiles (mean ± SD for all subjects) following bilateral UAE. There is brisk myometrial enhancement before embolization , which is markedly reduced immediately after embolization but returns to normal levels at 1 and 4 months.
fig.ommitted
TABLE 1. Enhancement of Myometrium and 11 Dominant Leiomyomas after Bilateral UAE
fig.ommitted
Figure 4. Graph illustrates the changes in leiomyoma perfusion profiles (mean ± SD for all subjects) following bilateral UAE. There is relatively brisk leiomyoma enhancement before embolization , which is markedly reduced immediately after embolization and remains suppressed at 1 ({diamond}) and 4 months.
fig.ommitted
Figure 5. Graph illustrates a comparison of perfusion profiles of myometrium and dominant leiomyoma (mean values of all subjects). The differences between preembolization perfusion of myometrium and leiomyoma are appreciated. Immediately after embolization, the preservation of a small amount of myometrial perfusion , compared with complete suppression of leiomyoma perfusion, is seen. At 1 and 4 months, myometrial perfusion has returned to normal, respectively), while leiomyoma perfusion remains suppressed , respectively).
fig.ommitted
Figure 6. Dynamic contrast material-enhanced MR images in a patient with multiple leiomyomas. Sagittal single-section gradient-echo (31/8, 90° flip angle) images were obtained through the middle of the uterus at 0, 30, 60, and 90 seconds following the injection of 0.1 mmol/kg of gadopentetate dimeglumine. Top row: Four images acquired before embolization show that the myometrium (arrow) is better perfused than leiomyomas (arrowhead). Immediately following embolization (row 2), there is an immediate reduction in myometrial and leiomyoma perfusion. At 1 month (row 3) and 4 months (bottom row) after embolization, there is recovery of myometrial but not leiomyoma perfusion.
Immediately following embolization, there was a 6.0% ± 8.1 reduction in volume of the dominant leiomyoma. After 1 and 4 months, dominant leiomyoma volume was further reduced from the preprocedural values by 22.3% ± 17.5 and 36.7% ± 26.5, respectively (Fig 7). The reduction in volume of the whole uterus immediately after embolization and at 1 and 4 months mirrored these values (9.2% ± 9.9, 16.5% ± 9.6, and 30.0% ± 12.2, respectively).
fig.ommitted
Figure 7a. Graphs illustrate the changes in leiomyoma volume following bilateral UAE. (a) Dominant leiomyoma volumes and (b) percentage reduction in dominant leiomyoma volume with time after embolization show a more rapid reduction in the 1st month after the procedure, followed by a slower but continued volume reduction between 1 and 4 months. In b, the percentage reduction in total leiomyoma volume ({lozenge}) and in myometrial volume alone at 4 months are shown for comparison.
fig.ommitted
Figure 7b. Graphs illustrate the changes in leiomyoma volume following bilateral UAE. (a) Dominant leiomyoma volumes and (b) percentage reduction in dominant leiomyoma volume with time after embolization show a more rapid reduction in the 1st month after the procedure, followed by a slower but continued volume reduction between 1 and 4 months. In b, the percentage reduction in total leiomyoma volume ({lozenge}) and in myometrial volume alone at 4 months are shown for comparison.
The mean volume reduction of all 45 leiomyomas was 38.2% ± 25.2 at 4 months, whereas the volume reduction in the myometrium alone (total uterine volume-total leiomyoma volume) was 20.1% ± 22.6 (Fig 7).
Relationship of Initial SI and Perfusion to Late Volume Changes in Leiomyomas
In all groups, the percentage change in the volume data was normally distributed. Leiomyomas that were high in SI on T2-weighted images prior to embolization (n = 11; mean volume, 117.5 cm3 ± 154.0) showed a significantly greater reduction in volume at 4 months, compared with the 34 leiomyomas (mean volume, 81.0 cm3 ± 93.7) that were low in SI (55.8% ± 21.2 vs 32.5% ± 23.9, respectively; P = .006; Fig 8).
fig.ommitted
Figure 8. Graph illustrates a comparison of volume reduction in initially high SI versus low SI leiomyomas on T2-weighted images. Box plots show median (middle line of box), quartiles (top and bottom lines of box), maximum (upper whisker), and minimum (lower whisker) values for the percentage reduction in leiomyoma volume at 4 months. A significant difference is seen between leiomyomas that were high in SI compared with those that were low in SI on initial T2-weighted images.
Perfusion relative to the myometrium was assessed in 35 of 45 leiomyomas, because the other 10 leiomyomas were not seen on single central-section dynamic images. Leiomyomas that were initially well perfused (n = 14; mean volume, 103.2 cm3 ± 114.7) did not show a significantly greater reduction in volume at 4 months compared with the 21 leiomyomas (mean volume, 76.6 cm3 ± 86.8) that were initially poorly perfused (42.6% ± 21.3 vs 36.8% ± 27.2, respectively; P = .50; Fig 9). Also, leiomyomas that were well perfused and high in SI on initial T2-weighted images (n = 7) did not appear to show a significantly greater reduction in volume at 4 months than did those that were poorly perfused and high in SI (n = 2), although the number of leiomyomas in the latter group is too small for a meaningful comparison. Similarly, leiomyomas that were well perfused and low in SI on initial T2-weighted images (n = 7) did not show a significantly greater reduction in volume at 4 months than did those that were poorly perfused and low in SI (n = 19). In addition, leiomyoma size (">="
50 cm3) or location (submucosal vs myometrial) did not have any effect on volume reduction at 4 months. The reduction in ME90 of the dominant leiomyomas immediately after embolization did not correlate with their volume reduction at 4 months.
fig.ommitted
Figure 9. Graph illustrates a comparison of volume reduction in initially well-perfused versus poorly perfused leiomyomas. Box plot shows median (middle line of box), quartiles (top and bottom lines of box), maximum (upper whisker), and minimum (lower whisker) values for the percentage reduction in leiomyoma volume at 4 months. No significant difference is seen between leiomyomas that were well perfused compared with those that were poorly perfused on initial dynamic contrast-enhanced images.
Relationship between Immediate Changes in Perfusion, 4-month Reduction in Leiomyoma Volume, and Clinical Response
A significant correlation was found between the reduction in ME90 of dominant leiomyomas immediately after embolization and the clinical response (Table 2, Spearman {rho} = 0.64, P = .03). However, there was no correlation between the reduction in ME90 of the myometrium immediately after embolization and the clinical response (Table 2, Spearman {rho} = -0.3). Also, no correlation was demonstrated between the reduction in total leiomyoma volume at 4 months and the clinical response (Table 2, Spearman {rho} = 0.03).
fig.ommitted
TABLE 2. Immediate Reduction in Enhancement of Myometrium and Leiomyoma, Late Reduction in Leiomyoma Volume, and Clinical Response
FSH Data
Serum FSH levels remained relatively stable in all patients. Preembolization levels of 4.9 IU/L ± 2.2 did not change significantly from those at 4 months (5.1 IU/L ± 1.6).
DISCUSSION
Findings from this study clearly demonstrate the differential perfusion responses between the myometrium and the leiomyoma after bilateral UAE. Perfusion of the myometrium was retained (ME90, 26%) immediately following embolization, despite having to inject embolization material into both uterine arteries until the flow stopped. This is likely to be the result of collateral circulation and the size of vessels supplying the myometrium. Myometrial vessels tend to be larger: Farrer-Brown et al (6) reported greater average vascular diameter in myometrium than in leiomyomas. The small number of patients remains a limitation of this study, although differences in myometrial versus leiomyoma perfusion were consistent and significant in all cases.
The intrinsic vascular pattern of uterine leiomyomas appears to represent a localized expansion of the myometrial vasculature, with the vessels within these tumors oriented in the direction of the muscle cell bundles (6). Early structural studies in which methods of pigment or radiopaque dye injection were used demonstrated that small leiomyomas were generally less vascular than myometrium, whereas larger leiomyomas were more vascular (7). Findings from blood flow studies (8) in which color Doppler ultrasonography was used confirmed that vascularization of leiomyomas was largely dependent on tumor size. Huang et al (9) demonstrated a negative correlation between leiomyoma size and/or volume and pulsatility index.
In contrast, immunohistochemical data measuring proportional area stained as an indicator of vascular density showed that myometrium had a significantly greater microvascular density than did small leiomyomas and both inner and outer regions of large leiomyomas (10). In addition, microvascular density measured on the basis of discrete microvessel count was significantly higher in myometrium than in all uterine leiomyoma groups (10). In our study, of the 35 leiomyomas included on the dynamic images, the well-perfused and the poorly perfused leiomyomas showed no significant difference in their mean volumes. It is likely that differing regions exist within these tumors in terms of blood flow and tissue perfusion, which account for the variability in reported results (8,11).
The differences in microvascular density in the myometrium and leiomyomas represent a difference in angiogenesis and vascular remodeling in these vascular beds and may be explained in two ways. They may result from a complex differential angiogenic promoter or inhibitor signals on leiomyomas and myometrium. It may also be that the presence of leiomyomas induces vascular changes within the myometrium. It has been postulated that this increase in vascular density is responsible for menorrhagia, which is observed in women with leiomyomas. However, we did not observe a significant difference in the rate or peak enhancement of the myometrium between preembolization and 4-month values.
The immediate reduction in maximal leiomyoma enhancement in the dominant leiomyoma correlated with the clinical response at 1 year. However, initially well-perfused leiomyomas compared with poorly perfused leiomyomas did not show a significant difference in volume reduction at 4 months. This is consistent with the finding that the overall volume reduction of leiomyomas (only 38% at 4 months) did not correlate with improvement in clinical scores at 1 year. Thus, the embolization process, which results in an immediate perfusion deficit in the leiomyomas, probably has a significant functional component that contributes to a favorable clinical outcome rather than the structural presence of the leiomyomas.
Sampson (12) originally suggested that an abnormal vasculature in the leiomyomatous uterus was responsible for menorrhagia. Study findings demonstrate that leiomyomas produce prostaglandins (13) and a number of growth factors, including fibroblast growth factor, vascular endothelial growth factor, heparin-binding epidermal growth factor, platelet-derived growth factor, transforming growth factor, and prolactin (14) that are regulators of angiogenesis. In patients with leiomyomas, dysregulation of growth factors and their receptors is likely to be responsible for menorrhagia (15). Switching off the production of these growth factors and prostaglandins from leiomyomas after embolization may well account for the long-term improvement in symptoms reported by these patients.
The initial SI on T2-weighted images had a significant effect on volume reduction at 4 months. Nondegenerated uterine leiomyomas have a typical appearance at MR imaging: well-circumscribed masses of homogenously decreased SI compared with the myometrium on T2-weighted images. At histologic examination, nondegenerated leiomyomas are composed of whorls of uniform smooth muscle cells with various amounts of intervening collagen (16). Cellular leiomyomas, which are composed of compact smooth muscle cells, have relatively higher SI on T2-weighted images and may demonstrate enhancement on contrast-enhanced images (17). Seven leiomyomas in our cohort were in this category. It is likely that these cellular leiomyomas undergo more necrosis as a result of the embolization process, which leads to a greater volume reduction. Degenerated leiomyomas have a variable appearance on T2-weighted images: Leiomyomas with hyaline or calcific degeneration are generally low in SI, whereas those with cystic degeneration are generally very high in SI, but the cystic areas do not enhance (18).
High SI on T1-weighted images of leiomyomas is normally associated with red degeneration due to T1-shortening effects of methemoglobin (16). In our study patients, an increase in SI on T1-weighted images was seen immediately after embolization, which is too early for the production of methemoglobin and was likely due to reduced blood volume in the leiomyomas or T1 effects of the accumulation of iodine-based contrast medium (T1 of iohexol measured in our laboratory, ~
850 msec at 0.5 T). However, in some patients, patchy areas of increased SI on T1-weighted images at 1 month may be accounted for by the presence of methemoglobin, as was previously suggested (19). Overall, a satisfactory clinical response was achieved in our cohort of patients (only one patient had no improvement in symptom scores) without adverse side effects, including any unwanted increase in serum FSH levels.
In conclusion, MR imaging demonstrates differential changes in perfusion between the myometrium and leiomyomas after bilateral UAE. At 1 month, there is recovery of myometrial perfusion, but perfusion of the leiomyoma remains depressed. The immediate reduction in leiomyoma perfusion correlates with the improvement in clinical score at 12 months. Leiomyoma volume at 4 months, on the other hand, remains at approximately 60% of its original value and does not correlate with clinical response. Leiomyomas high in SI on the initial T2-weighted images show a significantly greater volume reduction than those that are low in SI. Thus, dynamic MR imaging may be used to predict clinical response, while SI on T2-weighted images predicts volume reduction.
ACKNOWLEDGMENTS
We thank Roberto Dina, MD, for his help with understanding the pathology of leiomyomas and Mary Crisp for her secretarial assistance.
REFERENCES
Ravina JH, Herbreteau D, Ciraru-Vigneron N, et al. Arterial embolization to treat uterine myomata. Lancet 1995; 346:671-672.
Worthington-Kirsch RL, Popky GL, Hutchins FL, Jr. Uterine arterial embolization for the management of leiomyomas: quality of life assessment and clinical response. Radiology 1998; 208:625-629.
Goodwin SC, McLucas B, Lee M, et al. Uterine artery embolization for the treatment of uterine leiomyomata midterm results. J Vasc Interv Radiol 1999; 10:1159-1165.
Bradley EA, Reidy JF, Forman RG, Jarosz J, Braude PR. Transcatheter uterine artery embolization to treat large uterine fibroids. Br J Obstet Gynaecol 1998; 105:235-240.
Pelage JP, Le Dref O, Soyer P, et al. Fibroid-related menorrhagia: treatment with superselective embolization of the uterine arteries and midterm follow-up. Radiology 2000; 215:428-431.
Farrer-Brown G, Beilby JO, Rowles PM. Microvasculature of the uterus: an injection method of study. Obstet Gynecol 1970; 35:21-30.
Farrer-Brown G, Beilby JO, Tarbit MH. The vascular patterns in myomatous uteri. J Obstet Gynecol Br Commonw 1970; 77:967-975.
Kurjak A, Kupesic-Urek S, Miric D. The assessment of benign uterine tumor vascularization by transvaginal color Doppler. Ultrasound Med Biol 1992; 18:645-649.
Huang SC, Yu CH, Huang RT, Hsu KF, Tsai YC, Chou CY. Intratumoral blood flow in uterine myoma correlated with a lower tumour size and volume but not correlated with cell proliferation or angiogenesis. Obstet Gynecol 1996; 87:1019-1024.
Casey R, Rogers PA, Vollenhoven BJ. An immunohistochemical analysis of fibroid vasculature. Hum Reprod 2000; 15:1469-1475.
Forssman L. Distribution of blood flow in myomatous uteri as measured by locally injected 133Xe. Acta Obstet Gynecol Scand 1976; 55:101-104.
Sampson JA. The blood supply of uterine myomata. Surg Gynecol Obstet 1912; XIV:15-234.
Rees MC, Turnbull AC. Leiomyomas release prostaglandins. Prostaglandins Leukot Med 1985; 18:65-68.
Dixon D, He H, Haseman JK. Immunohistochemical localization of growth factors and their receptors in uterine leiomyomas and matched myometrium. Environ Health Perspect 2000; 108(suppl 5):795-802.
Stewart EA, Nowak RA. Leiomyoma-related bleeding: a classic hypothesis updated for the molecular era. Hum Reprod Update 1996; 2:295-306.
Murase E, Siegelman ES, Outwater EK, Perez-Jaffe LA, Tureck RW. Uterine leiomyomas: histopathologic features, MR imaging findings, differential diagnosis, and treatment. RadioGraphics 1999; 19:1179-1197.
Yamashita Y, Torashima M, Takahashi M, et al. Hyperintense uterine leiomyoma at T2-weighted MR imaging: differentiation with dynamic enhanced MR imaging and clinical implications. Radiology 1993; 189:721-725.
Okizuka H, Sugimura K, Takemori M, Obayashi C, Kitao M, Ishida T. MR detection of degenerating uterine leiomyomas. J Comput Assist Tomogr 1993; 17:760-766.
Burn PR, McCall JM, Chinn RJ, Vashisht A, Smith JR, Healy JC. Uterine fibroleiomyoma: MR imaging appearances before and after embolization of uterine arteries. Radiology 2000; 214:729-734., 百拇医药(Nandita M. deSouza MD FRCR and Andreanna D. Williams DCR MSc)
ABSTRACT
PURPOSE: To monitor changes in perfusion and volume of uterus and leiomyomas after bilateral uterine artery embolization (UAE) and to correlate immediate perfusion changes with subsequent reduction in leiomyoma volume and clinical outcome.
MATERIALS AND METHODS: Eleven consecutive women underwent magnetic resonance (MR) imaging before UAE, immediately after, and at 1 and 4 months. Reduction in maximal enhancement above baseline at 90 seconds (ME90) after injection of the dominant leiomyoma immediately after embolization was correlated with its volume reduction at 4 months and with clinical response at 12 months.
RESULTS: Forty-five leiomyomas were noted (mean, four per patient). Myometrium enhanced briskly (ME90 of 110%), with a reduction in ME90 to 26% immediately after embolization. Initial leiomyoma ME90 was lower (P < .001), but it suppressed to baseline levels immediately after embolization. At 1 and 4 months, myometrial perfusion returned to normal, but leiomyoma perfusion remained suppressed (P < .001). Immediate reduction in leiomyoma ME90 correlated with clinical response (Spearman {rho} = 0.64). Leiomyomas initially high in SI on T2-weighted images showed significantly greater volume reduction than those low in SI (P = .006). Well-perfused leiomyomas did not show greater volume reduction than those that were poorly perfused. Volume reduction did not correlate with improvement in clinical symptom score.
CONCLUSION: Immediate reduction in leiomyoma perfusion after bilateral UAE correlates with clinical response, whereas leiomyomas initially high in SI on T2-weighted images indicate a likely greater volume reduction.
Index terms: Arteries, therapeutic embolization, 969.1264 • Arteries, uterine, 969.12 • Leiomyoma, 854.318 • Uterine neoplasms, MR, 854.121411, 854.121412, 854.12143 • Uterine neoplasms, therapy, 854.1264
INTRODUCTION
Uterine arterial embolization (UAE) is gaining popularity as a safe and cost-effective way of treating leiomyomas that produce symptoms of menorrhagia, sensation of pressure, or a mass (1–4). Although the procedure obliterates flow in both uterine arteries, findings from clinical follow-up suggest that perfusion of the myometrium recovers, with reports of successful full-term pregnancies (5). However, documentation of immediate or late changes in perfusion of the leiomyoma compared with the surrounding uterine muscle is poor, and the mechanism of selective infarction of leiomyomas is poorly understood. The purpose of this study was to use magnetic resonance (MR) imaging to monitor changes in uterine and leiomyoma perfusion and volume immediately and 1 and 4 months after embolization of both uterine arteries and to correlate immediate perfusion changes with subsequent reduction in leiomyoma volume and with clinical outcome.
MATERIALS AND METHODS
Eleven consecutive women (age range, 29–48 years; mean age, 40 years ± 6.6 [SD]) who were to undergo bilateral UAE of leiomyomas, which were causing symptoms of menorrhagia, were recruited. The research ethics committee of the hospital granted approval for MR imaging, and informed consent was obtained after the MR imaging procedure was explained. All patients had menorrhagia, whereas abdominal distension (feeling of a mass) was present in eight cases. In one patient, images were obtained with a 1.5-T system (Eclipse; Marconi Medical Systems, Cleveland, Ohio) and a body coil and in the other 10 patients with a 0.5-T system (Apollo; Marconi Medical Systems) and a pelvic phased-array coil.
Imaging was performed before and within 30 minutes of the completion of embolization and 1 and 4 months after the procedure. MR imaging consisted of coronal and sagittal T1-weighted spin-echo sequences (496/20 repetition time msec/echo time msec) obtained before and after contrast material administration, sagittal T2-weighted fast spin-echo sequences (4,000/84 [effective], echo train length of 16), and single-section dynamic gadolinium-enhanced (Omniscan [0.1 mmol per kilogram of body weight of gadopentetate dimeglumine]; Nycomed Amersham, Oslo, Norway) sequences (31/13; 90° flip angle) obtained with a temporal resolution of 4 seconds. In the patient who underwent MR imaging with the 1.5-T system, an echo time of 8 msec was used for dynamic contrast-enhanced sequences (192 x 256 matrix, two signal averages, 35-cm field of view, and 8-mm section thickness). At least 5 hours had elapsed between the administration of the preembolization dose of gadolinium-enhanced contrast material and the immediate acquisition of postembolization images. In eight cases, preembolization imaging was performed the day prior to the procedure.
Embolization was performed with a femoral approach by using a 5-F cobra (Cordis, Miami, Fla) catheter. Both uterine arteries were selectively catheterized in all cases. Embolization was achieved by injecting 355–500-µm polyvinyl alcohol particles (Contour Emboli, St Albans, England) into each uterine artery until the flow had ceased. Pain was managed with intravenously administered opiates, either as a bolus or with a patient-controlled pump. In two cases, no intravenous postprocedural analgesia was required.
A postal questionnaire on clinical symptoms was completed by each patient before embolization and at 12 months after embolization. All recorded data were related to the symptoms of menses: namely length, blood loss, and associated pain. Each symptom was scored from 1 to 3 to produce a total final clinical response score of 9. Menses length of less than 3 days was scored 1; 4–5 days, 2; and 6 or more days, 3. Light menstrual loss (spotting) was scored 1; loss requiring one to two changes of sanitary napkins per 24 hours, 2; and heavy loss with or without clots and requiring more than two changes of sanitary napkins per 24 hours, 3. Occasional twinges of pain were scored 1; cramps, 2; and cramps requiring analgesia, 3. The change in the clinical score at 12 months from that at preembolization levels was calculated as a percentage with the following formula: difference in score/preembolization score x 100.
Serum follicle-stimulating hormone (FSH) levels were measured by using a microparticle enzyme immunoassay before embolization and at the same stage of the menstrual cycle 4 months after the procedure.
Image Analyses
In each patient, the number of leiomyomas and the individual signal intensities (SIs) relative to the myometrium were recorded on initial, immediate, and 1- and 4-month postembolization T2-weighted images as hypointense, isointense, or hyperintense. All leiomyoma-to–adjacent myometrium SI ratios were also recorded on T1-weighted images obtained before embolization, immediately after, and at 1 and 4 months. Leiomyoma locations (submucous, myometrial, and subserosal) were also recorded.
In every patient, perfusion data were obtained from the dynamic enhanced series before embolization, immediately after, and at 1 and 4 months. Regions of interest were drawn over an area of maximal enhancement within the dominant leiomyoma and an adjacent area of myometrium in the same z plane (head to foot) to avoid SI variations due to the pelvic phased-array coil. Regions of interest were placed by one author (N.M.d.S.) and were between 0.3 and 1.5 cm2 in area (median, 0.5 cm2). SIs were recorded with an imaging perfusion profile software (Marconi Medical Systems, Highland Heights, Ohio). In addition, for every leiomyoma depicted on each single section of the dynamic series, the enhancement above baseline was noted. Leiomyomas were well perfused when the plateau enhancement (after 90 sec) was equivalent to or greater than that of the adjacent myometrium, or poorly perfused, when the plateau enhancement was less than that of adjacent myometrium.
Volume measurements of the dominant leiomyoma and the entire uterus (including all leiomyomas) were made by one author (N.M.d.S.) by drawing around the area with an electronic caliper and by using standard scanner software to compute the area per section. The sum of the areas was multiplied by the section thickness. Measurements were made before embolization, immediately after, and at 1 and 4 months. Individual volume measurements of all leiomyomas were made similarly on preembolization and 4-month postembolization images. Measurements obtained before and after embolization were made while viewing corresponding images simultaneously so that individual leiomyomas could be clearly identified and compared.
Statistical Analyses
Statistical analyses were done by using software (Unistat Ltd, London, England). Data were tested for normality by using normality profiles and a Shapiro-Wilk normality test to determine subsequent use of appropriate tests for statistical comparison. For each time point of the dynamic series for both myometrium and dominant leiomyoma, the percentage enhancement above baseline was calculated by using the equation (SIpost - SIpre)/SIpre, where post is after enhancement and pre is before enhancement. The calculated change in SI from that at baseline would eliminate any T2* effects in the absolute values of SI. The mean and SD of the pooled patient data were plotted before embolization, immediately after, and at 1 and 4 months. The maximal enhancement above baseline at 90 seconds (ME90) of myometrium and that of leiomyoma before embolization was compared by using a paired t test. The ME90 of myometrium and leiomyoma before embolization was similarly compared with the values obtained immediately after embolization and at 1 and 4 months.
All distributions tested normal. The percentage change in volume at 4 months between areas of high SI and those of low SI on T2-weighted images of leiomyomas or between well- and poorly perfused leiomyomas was compared by using a nonpaired two-sample t test with unequal variance. Changes in individual leiomyomas were considered independently of each other. High SI well- versus poorly perfused leiomyomas were compared as were low SI well- versus poorly perfused leiomyomas, large (">="
50 cm3) versus small (<50 cm3) volume leiomyomas, and submucosal versus myometrial leiomyomas.
In dominant leiomyomas, the immediate reduction in ME90 (reduction in a percentage enhancement between preembolization and immediate postembolization images) was correlated with their reduction in volume at 4 months by using Spearman rank correlation coefficient. Similar correlation coefficients were obtained between immediate reduction in ME90 of myometrium and leiomyoma and clinical response (percentage change in symptom scores at 12 months) and between total volume reduction of all leiomyomas in each patient at 4 months and clinical response at 12 months.
The significance of the difference in serum FSH levels at 4 months from those obtained before embolization was assessed by using a paired t test.
RESULTS
Patients had multiple leiomyomas (range, one to eight leiomyomas; mean, four; median, four). Forty-five leiomyomas were noted, and it was possible to obtain volume measurements from all of them. Leiomyoma volumes ranged from 0.6 to 434 cm3 (mean, 89.9 cm3 ± 110.5; median, 41.2 cm3). Although 11 leiomyomas were isointense or hyperintense on T2-weighted images, compared with the myometrium, leiomyomas were generally hypointense (n = 34). Thirty-five of 45 leiomyomas were included in a single section of the dynamic study: 14 of these were well perfused (equivalently or better) compared with the myometrium, whereas 21 were poorly perfused. A variety of locations within the myometrium were also noted: 34 were myometrial; seven, submucous; three, subserosal; and one, cervical in origin.
SI Changes
Immediately following embolization, an increase in SI within leiomyomas, compared with the myometrium, was seen on T1-weighted images regardless of their size, location, initial SI, or perfusion characteristics. This was relatively homogenous. Mean leiomyoma-to–adjacent myometrium SI ratios had increased from 1.1 ± 0.1 before embolization to 1.4 ± 0.1 immediately after embolization. This hyperintensity on T1-weighted images persisted at 1 and 4 months, though it gradually became more heterogeneous and less marked (Fig 1). With T2 weighting, there were no observable changes in SI on the immediate postembolization images. However, after 1 and 4 months, there was a reduction in SI on T2-weighted images in 43 of 45 leiomyomas (Fig 2). Two leiomyomas remained isointense with the myometrium.
fig.ommitted
Figure 1. Midsagittal T1-weighted spin-echo (420/20) MR images obtained through the pelvis before (A), immediately after (B), 1 month (C), and 4 months (D) after bilateral UAE show multiple uterine leiomyomas, which are hypo- (wide arrows) and isointense (thin arrows) with myometrium. Immediately after embolization, there is a relative increase in SI in the leiomyomas (arrows) and to a lesser extent in the myometrium. In C, this appearance becomes more heterogenous and in D, it is seen to fade (arrows).
fig.ommitted
Figure 2. Midline sagittal T2-weighted spin-echo (2,500/80) MR images show uterus with leiomyomas before (A) and 4 months after (B) UAE. The large hypo- and hyperintense leiomyomas in A (large and small arrow, respectively) have reduced in size and SI in B (arrows).
Perfusion Changes
Data for ME90 and percentage change in ME90 were normally distributed. Before embolization, the myometrium was better perfused than leiomyomas, with a significantly higher ME90 (P < .001). Immediately following embolization, there was a dramatic reduction in the overall uterine perfusion. However, while perfusion of the myometrium reduced considerably from an ME90 of 110% before embolization to an ME90 of 26% immediately following embolization (Fig 3), perfusion of the leiomyoma had virtually stopped (Table 1, Fig 4). After 1 and 4 months, perfusion of the myometrium recovered to normal levels, while leiomyoma perfusion remained extremely poor (Table 1, Figs 5, 6). Thus, there was a significant difference between ME90 of myometrium before and that immediately after embolization (P < .001) but not between ME90 of myometrium before and that at 1 or 4 months after embolization. Leiomyomas, on the other hand, showed a significant difference between ME90 before embolization and that immediately after embolization, as well as at 1 and 4 months (P < .001).
fig.ommitted
Figure 3. Graph illustrates the changes in myometrial perfusion profiles (mean ± SD for all subjects) following bilateral UAE. There is brisk myometrial enhancement before embolization , which is markedly reduced immediately after embolization but returns to normal levels at 1 and 4 months.
fig.ommitted
TABLE 1. Enhancement of Myometrium and 11 Dominant Leiomyomas after Bilateral UAE
fig.ommitted
Figure 4. Graph illustrates the changes in leiomyoma perfusion profiles (mean ± SD for all subjects) following bilateral UAE. There is relatively brisk leiomyoma enhancement before embolization , which is markedly reduced immediately after embolization and remains suppressed at 1 ({diamond}) and 4 months.
fig.ommitted
Figure 5. Graph illustrates a comparison of perfusion profiles of myometrium and dominant leiomyoma (mean values of all subjects). The differences between preembolization perfusion of myometrium and leiomyoma are appreciated. Immediately after embolization, the preservation of a small amount of myometrial perfusion , compared with complete suppression of leiomyoma perfusion, is seen. At 1 and 4 months, myometrial perfusion has returned to normal, respectively), while leiomyoma perfusion remains suppressed , respectively).
fig.ommitted
Figure 6. Dynamic contrast material-enhanced MR images in a patient with multiple leiomyomas. Sagittal single-section gradient-echo (31/8, 90° flip angle) images were obtained through the middle of the uterus at 0, 30, 60, and 90 seconds following the injection of 0.1 mmol/kg of gadopentetate dimeglumine. Top row: Four images acquired before embolization show that the myometrium (arrow) is better perfused than leiomyomas (arrowhead). Immediately following embolization (row 2), there is an immediate reduction in myometrial and leiomyoma perfusion. At 1 month (row 3) and 4 months (bottom row) after embolization, there is recovery of myometrial but not leiomyoma perfusion.
Immediately following embolization, there was a 6.0% ± 8.1 reduction in volume of the dominant leiomyoma. After 1 and 4 months, dominant leiomyoma volume was further reduced from the preprocedural values by 22.3% ± 17.5 and 36.7% ± 26.5, respectively (Fig 7). The reduction in volume of the whole uterus immediately after embolization and at 1 and 4 months mirrored these values (9.2% ± 9.9, 16.5% ± 9.6, and 30.0% ± 12.2, respectively).
fig.ommitted
Figure 7a. Graphs illustrate the changes in leiomyoma volume following bilateral UAE. (a) Dominant leiomyoma volumes and (b) percentage reduction in dominant leiomyoma volume with time after embolization show a more rapid reduction in the 1st month after the procedure, followed by a slower but continued volume reduction between 1 and 4 months. In b, the percentage reduction in total leiomyoma volume ({lozenge}) and in myometrial volume alone at 4 months are shown for comparison.
fig.ommitted
Figure 7b. Graphs illustrate the changes in leiomyoma volume following bilateral UAE. (a) Dominant leiomyoma volumes and (b) percentage reduction in dominant leiomyoma volume with time after embolization show a more rapid reduction in the 1st month after the procedure, followed by a slower but continued volume reduction between 1 and 4 months. In b, the percentage reduction in total leiomyoma volume ({lozenge}) and in myometrial volume alone at 4 months are shown for comparison.
The mean volume reduction of all 45 leiomyomas was 38.2% ± 25.2 at 4 months, whereas the volume reduction in the myometrium alone (total uterine volume-total leiomyoma volume) was 20.1% ± 22.6 (Fig 7).
Relationship of Initial SI and Perfusion to Late Volume Changes in Leiomyomas
In all groups, the percentage change in the volume data was normally distributed. Leiomyomas that were high in SI on T2-weighted images prior to embolization (n = 11; mean volume, 117.5 cm3 ± 154.0) showed a significantly greater reduction in volume at 4 months, compared with the 34 leiomyomas (mean volume, 81.0 cm3 ± 93.7) that were low in SI (55.8% ± 21.2 vs 32.5% ± 23.9, respectively; P = .006; Fig 8).
fig.ommitted
Figure 8. Graph illustrates a comparison of volume reduction in initially high SI versus low SI leiomyomas on T2-weighted images. Box plots show median (middle line of box), quartiles (top and bottom lines of box), maximum (upper whisker), and minimum (lower whisker) values for the percentage reduction in leiomyoma volume at 4 months. A significant difference is seen between leiomyomas that were high in SI compared with those that were low in SI on initial T2-weighted images.
Perfusion relative to the myometrium was assessed in 35 of 45 leiomyomas, because the other 10 leiomyomas were not seen on single central-section dynamic images. Leiomyomas that were initially well perfused (n = 14; mean volume, 103.2 cm3 ± 114.7) did not show a significantly greater reduction in volume at 4 months compared with the 21 leiomyomas (mean volume, 76.6 cm3 ± 86.8) that were initially poorly perfused (42.6% ± 21.3 vs 36.8% ± 27.2, respectively; P = .50; Fig 9). Also, leiomyomas that were well perfused and high in SI on initial T2-weighted images (n = 7) did not appear to show a significantly greater reduction in volume at 4 months than did those that were poorly perfused and high in SI (n = 2), although the number of leiomyomas in the latter group is too small for a meaningful comparison. Similarly, leiomyomas that were well perfused and low in SI on initial T2-weighted images (n = 7) did not show a significantly greater reduction in volume at 4 months than did those that were poorly perfused and low in SI (n = 19). In addition, leiomyoma size (">="
50 cm3) or location (submucosal vs myometrial) did not have any effect on volume reduction at 4 months. The reduction in ME90 of the dominant leiomyomas immediately after embolization did not correlate with their volume reduction at 4 months.
fig.ommitted
Figure 9. Graph illustrates a comparison of volume reduction in initially well-perfused versus poorly perfused leiomyomas. Box plot shows median (middle line of box), quartiles (top and bottom lines of box), maximum (upper whisker), and minimum (lower whisker) values for the percentage reduction in leiomyoma volume at 4 months. No significant difference is seen between leiomyomas that were well perfused compared with those that were poorly perfused on initial dynamic contrast-enhanced images.
Relationship between Immediate Changes in Perfusion, 4-month Reduction in Leiomyoma Volume, and Clinical Response
A significant correlation was found between the reduction in ME90 of dominant leiomyomas immediately after embolization and the clinical response (Table 2, Spearman {rho} = 0.64, P = .03). However, there was no correlation between the reduction in ME90 of the myometrium immediately after embolization and the clinical response (Table 2, Spearman {rho} = -0.3). Also, no correlation was demonstrated between the reduction in total leiomyoma volume at 4 months and the clinical response (Table 2, Spearman {rho} = 0.03).
fig.ommitted
TABLE 2. Immediate Reduction in Enhancement of Myometrium and Leiomyoma, Late Reduction in Leiomyoma Volume, and Clinical Response
FSH Data
Serum FSH levels remained relatively stable in all patients. Preembolization levels of 4.9 IU/L ± 2.2 did not change significantly from those at 4 months (5.1 IU/L ± 1.6).
DISCUSSION
Findings from this study clearly demonstrate the differential perfusion responses between the myometrium and the leiomyoma after bilateral UAE. Perfusion of the myometrium was retained (ME90, 26%) immediately following embolization, despite having to inject embolization material into both uterine arteries until the flow stopped. This is likely to be the result of collateral circulation and the size of vessels supplying the myometrium. Myometrial vessels tend to be larger: Farrer-Brown et al (6) reported greater average vascular diameter in myometrium than in leiomyomas. The small number of patients remains a limitation of this study, although differences in myometrial versus leiomyoma perfusion were consistent and significant in all cases.
The intrinsic vascular pattern of uterine leiomyomas appears to represent a localized expansion of the myometrial vasculature, with the vessels within these tumors oriented in the direction of the muscle cell bundles (6). Early structural studies in which methods of pigment or radiopaque dye injection were used demonstrated that small leiomyomas were generally less vascular than myometrium, whereas larger leiomyomas were more vascular (7). Findings from blood flow studies (8) in which color Doppler ultrasonography was used confirmed that vascularization of leiomyomas was largely dependent on tumor size. Huang et al (9) demonstrated a negative correlation between leiomyoma size and/or volume and pulsatility index.
In contrast, immunohistochemical data measuring proportional area stained as an indicator of vascular density showed that myometrium had a significantly greater microvascular density than did small leiomyomas and both inner and outer regions of large leiomyomas (10). In addition, microvascular density measured on the basis of discrete microvessel count was significantly higher in myometrium than in all uterine leiomyoma groups (10). In our study, of the 35 leiomyomas included on the dynamic images, the well-perfused and the poorly perfused leiomyomas showed no significant difference in their mean volumes. It is likely that differing regions exist within these tumors in terms of blood flow and tissue perfusion, which account for the variability in reported results (8,11).
The differences in microvascular density in the myometrium and leiomyomas represent a difference in angiogenesis and vascular remodeling in these vascular beds and may be explained in two ways. They may result from a complex differential angiogenic promoter or inhibitor signals on leiomyomas and myometrium. It may also be that the presence of leiomyomas induces vascular changes within the myometrium. It has been postulated that this increase in vascular density is responsible for menorrhagia, which is observed in women with leiomyomas. However, we did not observe a significant difference in the rate or peak enhancement of the myometrium between preembolization and 4-month values.
The immediate reduction in maximal leiomyoma enhancement in the dominant leiomyoma correlated with the clinical response at 1 year. However, initially well-perfused leiomyomas compared with poorly perfused leiomyomas did not show a significant difference in volume reduction at 4 months. This is consistent with the finding that the overall volume reduction of leiomyomas (only 38% at 4 months) did not correlate with improvement in clinical scores at 1 year. Thus, the embolization process, which results in an immediate perfusion deficit in the leiomyomas, probably has a significant functional component that contributes to a favorable clinical outcome rather than the structural presence of the leiomyomas.
Sampson (12) originally suggested that an abnormal vasculature in the leiomyomatous uterus was responsible for menorrhagia. Study findings demonstrate that leiomyomas produce prostaglandins (13) and a number of growth factors, including fibroblast growth factor, vascular endothelial growth factor, heparin-binding epidermal growth factor, platelet-derived growth factor, transforming growth factor, and prolactin (14) that are regulators of angiogenesis. In patients with leiomyomas, dysregulation of growth factors and their receptors is likely to be responsible for menorrhagia (15). Switching off the production of these growth factors and prostaglandins from leiomyomas after embolization may well account for the long-term improvement in symptoms reported by these patients.
The initial SI on T2-weighted images had a significant effect on volume reduction at 4 months. Nondegenerated uterine leiomyomas have a typical appearance at MR imaging: well-circumscribed masses of homogenously decreased SI compared with the myometrium on T2-weighted images. At histologic examination, nondegenerated leiomyomas are composed of whorls of uniform smooth muscle cells with various amounts of intervening collagen (16). Cellular leiomyomas, which are composed of compact smooth muscle cells, have relatively higher SI on T2-weighted images and may demonstrate enhancement on contrast-enhanced images (17). Seven leiomyomas in our cohort were in this category. It is likely that these cellular leiomyomas undergo more necrosis as a result of the embolization process, which leads to a greater volume reduction. Degenerated leiomyomas have a variable appearance on T2-weighted images: Leiomyomas with hyaline or calcific degeneration are generally low in SI, whereas those with cystic degeneration are generally very high in SI, but the cystic areas do not enhance (18).
High SI on T1-weighted images of leiomyomas is normally associated with red degeneration due to T1-shortening effects of methemoglobin (16). In our study patients, an increase in SI on T1-weighted images was seen immediately after embolization, which is too early for the production of methemoglobin and was likely due to reduced blood volume in the leiomyomas or T1 effects of the accumulation of iodine-based contrast medium (T1 of iohexol measured in our laboratory, ~
850 msec at 0.5 T). However, in some patients, patchy areas of increased SI on T1-weighted images at 1 month may be accounted for by the presence of methemoglobin, as was previously suggested (19). Overall, a satisfactory clinical response was achieved in our cohort of patients (only one patient had no improvement in symptom scores) without adverse side effects, including any unwanted increase in serum FSH levels.
In conclusion, MR imaging demonstrates differential changes in perfusion between the myometrium and leiomyomas after bilateral UAE. At 1 month, there is recovery of myometrial perfusion, but perfusion of the leiomyoma remains depressed. The immediate reduction in leiomyoma perfusion correlates with the improvement in clinical score at 12 months. Leiomyoma volume at 4 months, on the other hand, remains at approximately 60% of its original value and does not correlate with clinical response. Leiomyomas high in SI on the initial T2-weighted images show a significantly greater volume reduction than those that are low in SI. Thus, dynamic MR imaging may be used to predict clinical response, while SI on T2-weighted images predicts volume reduction.
ACKNOWLEDGMENTS
We thank Roberto Dina, MD, for his help with understanding the pathology of leiomyomas and Mary Crisp for her secretarial assistance.
REFERENCES
Ravina JH, Herbreteau D, Ciraru-Vigneron N, et al. Arterial embolization to treat uterine myomata. Lancet 1995; 346:671-672.
Worthington-Kirsch RL, Popky GL, Hutchins FL, Jr. Uterine arterial embolization for the management of leiomyomas: quality of life assessment and clinical response. Radiology 1998; 208:625-629.
Goodwin SC, McLucas B, Lee M, et al. Uterine artery embolization for the treatment of uterine leiomyomata midterm results. J Vasc Interv Radiol 1999; 10:1159-1165.
Bradley EA, Reidy JF, Forman RG, Jarosz J, Braude PR. Transcatheter uterine artery embolization to treat large uterine fibroids. Br J Obstet Gynaecol 1998; 105:235-240.
Pelage JP, Le Dref O, Soyer P, et al. Fibroid-related menorrhagia: treatment with superselective embolization of the uterine arteries and midterm follow-up. Radiology 2000; 215:428-431.
Farrer-Brown G, Beilby JO, Rowles PM. Microvasculature of the uterus: an injection method of study. Obstet Gynecol 1970; 35:21-30.
Farrer-Brown G, Beilby JO, Tarbit MH. The vascular patterns in myomatous uteri. J Obstet Gynecol Br Commonw 1970; 77:967-975.
Kurjak A, Kupesic-Urek S, Miric D. The assessment of benign uterine tumor vascularization by transvaginal color Doppler. Ultrasound Med Biol 1992; 18:645-649.
Huang SC, Yu CH, Huang RT, Hsu KF, Tsai YC, Chou CY. Intratumoral blood flow in uterine myoma correlated with a lower tumour size and volume but not correlated with cell proliferation or angiogenesis. Obstet Gynecol 1996; 87:1019-1024.
Casey R, Rogers PA, Vollenhoven BJ. An immunohistochemical analysis of fibroid vasculature. Hum Reprod 2000; 15:1469-1475.
Forssman L. Distribution of blood flow in myomatous uteri as measured by locally injected 133Xe. Acta Obstet Gynecol Scand 1976; 55:101-104.
Sampson JA. The blood supply of uterine myomata. Surg Gynecol Obstet 1912; XIV:15-234.
Rees MC, Turnbull AC. Leiomyomas release prostaglandins. Prostaglandins Leukot Med 1985; 18:65-68.
Dixon D, He H, Haseman JK. Immunohistochemical localization of growth factors and their receptors in uterine leiomyomas and matched myometrium. Environ Health Perspect 2000; 108(suppl 5):795-802.
Stewart EA, Nowak RA. Leiomyoma-related bleeding: a classic hypothesis updated for the molecular era. Hum Reprod Update 1996; 2:295-306.
Murase E, Siegelman ES, Outwater EK, Perez-Jaffe LA, Tureck RW. Uterine leiomyomas: histopathologic features, MR imaging findings, differential diagnosis, and treatment. RadioGraphics 1999; 19:1179-1197.
Yamashita Y, Torashima M, Takahashi M, et al. Hyperintense uterine leiomyoma at T2-weighted MR imaging: differentiation with dynamic enhanced MR imaging and clinical implications. Radiology 1993; 189:721-725.
Okizuka H, Sugimura K, Takemori M, Obayashi C, Kitao M, Ishida T. MR detection of degenerating uterine leiomyomas. J Comput Assist Tomogr 1993; 17:760-766.
Burn PR, McCall JM, Chinn RJ, Vashisht A, Smith JR, Healy JC. Uterine fibroleiomyoma: MR imaging appearances before and after embolization of uterine arteries. Radiology 2000; 214:729-734., 百拇医药(Nandita M. deSouza MD FRCR and Andreanna D. Williams DCR MSc)