Vascular lesions of the uterus are rare and the vast majority reported in the literature are arteriovenous malformations (AVMs). Uterine AVMs can be congenital or acquired. Acquired lesions are believed to result from pelvic surgery, trauma, curettage, trophoblastic disease, diethylstilbestrol (DES) exposure, neoplasm or infection. Hemangiomas of the uterus are more localized than are AVMs and the vessels within them are uniform, unlike those forming an AVM1. There are insufficient data in the recent literature regarding uterine hemangiomas, probably due to their silent clinical courses.
In recent years, there has been an increasing number of reports in the literature regarding vascular lesions of the uterus. It can readily be seen from these reports that there is confusion concerning the terminology of uterine vascular lesions. For example, in some reports the terms ‘uterine arteriovenous malformation’2 and ‘uterine vascular malformation’3 have been used to describe uterine lesions showing a hypervascular appearance with turbulent flow on color Doppler ultrasound, although many of them resolved spontaneously within a few months. The term ‘malformation’, however, is generally used to describe defects in the structure of an organ or region of the body resulting from an intrinsically abnormal process of development4. Therefore, spontaneous resolution of a malformation in a short period of time is unlikely. Mulliken and Glowacki investigated the histopathological differences between hemangiomas and vascular malformations based on endothelial characteristics5. They found that AVMs are errors of morphogenesis with stable cellularity and do not show spontaneous regression, unlike hemangiomas, which are benign neoplasms with increased mitotic activity and which go through a proliferating and involuting phase. Therefore, the term either ‘vascular abnormality’ or ‘vascular lesion’ seems to best describe hypervascular areas within the uterus on color Doppler ultrasound unless they are proved to be an AVM by angiography or pathological examination.
The diagnosis of a uterine AVM is difficult because the condition is rare and the diagnostic methods used in routine practice such as pelvic examination, curettage, gray-scale sonography or hysteroscopy usually fail to provide findings which strongly suggest this malformation. Gray-scale sonography has a limited role in the diagnosis of uterine AVMs and often shows an ill-defined uterine mass consisting of mildly echogenic tissue that is interspersed with multiple small hypoechoic spaces of varying sizes3, 6, 7. On magnetic resonance (MR) imaging, uterine AVMs are manifested as a bulky uterus, a focal uterine mass, disruption of the junctional zones, serpiginous flow-related signal voids, and prominent parametrial vessels7. At the present time, color Doppler sonography is the preferred method of diagnosing uterine AVMs2, 3, 6, 7. On color Doppler ultrasound, a uterine AVM appears as a hypervascular lesion with turbulent flow within the myometrium2, 3, 6, 7. Spectral analysis of the vessels within the lesion shows low-impedance, high-velocity flow, with resistance index values ranging from 0.25 to 0.55 and peak velocity values ranging from 40 cm/s to 96 cm/s in previously reported cases3, 6–8. Doppler and MR imaging features of uterine AVMs may overlap with other causes of arteriovenous shunting, including trophoblastic disease, retained products of conception and abnormal placentation6, 7, 9. The AVMs can be distinguished from these pregnancy-related conditions using serum human chorionic gonadotropin (hCG) measurement6, 7. Angiography, which is the gold standard for definitive diagnosis of AVMs, has the advantage of demonstrating the feeding arteries and draining veins of the AVM. On angiography a uterine AVM appears as a hypervascular mass with typical early venous filling. Angiography, however, is an invasive procedure, and should be reserved for cases in which surgical intervention or therapeutic embolization of the lesion is planned due to persistent and severe uterine hemorrhage3, 6, 10, 11.
Three papers regarding uterine AVMs, by Timmerman et al.12, Kelly et al.13 and Ichikawa et al.14, are published in this issue of the Journal. I think the paper by Timmerman et al.12, together with at least two recently published studies2, 3, will lead to a discussion: have we recently over-diagnosed uterine AVMs? In these reports2, 3, 12 which are three of the largest reported case series, nearly all patients were diagnosed with uterine vascular malformation following a term delivery, a molar pregnancy, or an induced or spontaneous abortion. In all patients except for true AVM and molar pregnancy cases, a spontaneous resolution occurred within a short period of time. The average period of spontaneous resolution was 69.5 days (range, 40–112 days) in the study by Wiebe and Switzer2, 5.3 weeks (range, 1 week to 6 months) in the study by Timmerman et al.12 and 6 weeks (range, 2–8 weeks) in the other study by Timmerman et al.3. In these studies2, 3, 12, a uterine AVM was diagnosed based only on color Doppler findings including hypervascularity and turbulence without angiographic or pathological confirmation in most of the cases. However, we know that these color Doppler findings are also observed in a variety of conditions including normal pregnancy, gestational trophoblastic disease and retained products of conception4, 9. In normal pregnancy, due to vasodilatation and decreasing resistance in the spiral arteries, a hypervascular appearance with turbulent flow which is also referred to as ‘peritrophoblastic flow’ is observed especially in the myometrium where implantation has occurred4. This appearance in the myometrium can also be observed in very early intrauterine gestation even before a gestational sac visible on ultrasound appears, or following a recent spontaneous or induced abortion with an empty uterus4, 15. The hypervascularity and turbulence within the myometrium during gestation does not immediately disappear following spontaneous or induced abortion15. The hypervascularity gradually regresses and then disappears within the following days. This regression period is further prolonged in the presence of retained products of conception or molar gestation4, 9, 15. Dillon et al.15 investigated the post-abortion appearance of the uterus on color Doppler sonography in 19 women who underwent elective first-trimester abortions. In four of eight patients, typical peritrophoblastic flow was observed on the second and third days after the abortions were performed. After the third day, flow was observed in two of 11 patients. In the three recent studies2, 3, 12 mentioned above, however, the diagnosis of uterine vascular malformation was made from as early as the second day after a delivery or an abortion2, 12. The average period between pregnancy and diagnosis of an AVM was 7.6 weeks (range, 2 days to 25 weeks) in the study by Timmerman et al.12, 21.7 days (range, 2–37 days) in the study by Wiebe and Switzer2 and 6 weeks in the other study by Timmerman et al.3. Therefore, the hypervascular areas with turbulent flow in the myometrium in at least some of the cases in these reports2, 3, 12 may have represented ‘peritrophoblastic flow’ in spiral arteries rather than a vascular malformation. On the other hand, in the study by Timmerman et al.12, which is the largest case series of uterine AVM in the literature, 12 of 30 patients had retained products of pregnancy at the time of diagnosis, which was pathologically confirmed to be trophoblastic tissue in six cases and molar tissue in the other six patients. In this study serum hCG monitoring was not performed12. In the study by Wiebe and Switzer2, a uterine AVM was diagnosed based on gray-scale and color Doppler findings in seven patients who had prolonged bleeding after an induced abortion and who were managed expectantly with the monitoring of weekly serum hCG levels. Interestingly serum hCG was elevated in all cases2, which obviously demonstrated the presence of retained trophoblastic tissue within the uterus. When hCG levels returned to normal limits, all uterine lesions disappeared spontaneously. Kido et al.9 reported a patient who presented with sudden heavy vaginal bleeding 6 weeks after an induced abortion. Gray-scale sonography showed several serpiginous and tubular anechoic spaces within the myometrium and color Doppler imaging revealed hypervascularity, marked turbulence, and low-impedance, high-velocity flow within the lesion, suggesting a uterine AVM9. Findings on MR imaging also suggested an AVM and the serum hCG level was slightly elevated. Because the presumptive diagnosis was a uterine AVM and severe hemorrhage persisted, hysterectomy was performed in this patient and pathological examination showed retained placental products that were necrotic and encroached into the myometrium9. In the study by Timmerman et al.3 there was an interesting case (Case 8) who presented with heavy uterine bleeding following a long period of menorrhagia. Gray-scale sonography demonstrated a tubular anechoic area (5 mm in width) within the endometrium, which was consistent with an AVM on color Doppler imaging. Although hysteroscopy revealed the presence of retained placental tissue which was confirmed by pathological examination, and gray-scale ultrasound and color Doppler findings within the endometrium disappeared within 2 weeks after hysteroscopic removal of placental tissue, this lesion was considered to be an AVM3. All this evidence suggests that retained products of conception can lead to a hypervascular appearance with turbulent flow on color Doppler imaging, not only in the endometrial cavity but also within the myometrium or in the subendometrial region. The presence of trophoblastic tissue in the myometrium or subendometrial region could be explained by the encroachment of placental tissue into these regions through damaged endometrium due to curettage. This hypervascular image due to the peritrophoblastic flow disappears within a varying period of time as the retained placental tissue is expelled from the uterus transcervically or resorbed by the uterus, with a concomitant decline in serum hCG levels2. It is also known that gray-scale ultrasound for the detection of retained products of conception is specific but not very sensitive. Ultrasound cannot reliably rule out retained placental tissue and especially a moderate endometrial thickness of 2–5 mm is not diagnostic16. Consequently, to make a diagnosis of uterine AVM based only on color Doppler findings could result in over-diagnosis; hence the conclusion in some recent reports2, 3, 12 that uterine AVMs are much more common than previously thought should be interpreted with caution. This issue should be investigated by further studies.
It has been recently suggested that curettage should not be performed in a patient who presents with abnormal uterine bleeding after an abortion or a delivery when there is a hypervascular area with turbulent flow within the myometrium2, 3, 12. In recent reported large case series3, 12 of uterine vascular malformation, however, there was no information showing that life-threatening hemorrhage occurred immediately after curettage although a curettage was performed in many patients in these series3, 12. In fact, there is no evidence in the literature to show that curettage resulted in severe uterine bleeding in a patient with a uterine lesion as described above, except for true AVM cases. In routine practice, the cause of uterine bleeding in the vast majority of patients presenting with these clinical and color Doppler sonographic findings is retained placental tissue, and the treatment is curettage. If curettage is not performed for fear of heavy bleeding, many patients will undergo preventable blood loss. Some other patients may also be exposed to unnecessary invasive procedures such as embolization or hysterectomy due to intractable bleeding. Curettage should not be performed only in cases in which medical history, current clinical symptoms and findings on color Doppler or MR imaging strongly suggest the presence of a uterine AVM. What should we do when a hypervascular lesion within the myometrium, showing turbulence and low-impedance, high-velocity flow on Doppler examination is identified in a patient presenting with abnormal vaginal bleeding following an abortion or a delivery? If the patient is clinically stable, conservative management is a reasonable option after retained products of conception have been ruled out. Serum hCG measurement is important in distinguishing between uterine AVM and retained placental tissue or trophoblastic disease. If hemorrhage is persistent and severe, however, angiography is required both to make a definitive diagnosis and to perform a therapeutic embolization3, 11, 12. On angiography, a hypervascular mass with early venous filling is diagnostic for AVMs8, 12. In other words, an intramyometrial hypervascular lesion without early venous filling is not an AVM and represents other hypervascular conditions including non-AVM vascular abnormalities of the uterus3, 12 and peritrophoblastic flow. Therapeutic embolization is effective in controlling uterine bleeding not only in AVMs but also in other causes of hemorrhage including uterine leiomyomas17 and non-AVM uterine vascular abnormalities3, 12. Some cases of embolization failure have also been reported11, 18. If there is no desire for pregnancy or embolization fails, hysterectomy remains the treatment of choice.
To summarize, at present a consensus exists regarding the diagnosis and management of uterine AVMs. However, the histopathological features, diagnostic criteria and management of the other vascular abnormalities of the uterus remain unclear. Future studies should aim to define criteria to distinguish between so-called uterine vascular malformations and retained products of conception or AVMs.