{"paper_id":"e48dcdb5-0e8f-4c68-97b7-fed42b23eb51","body_text":"Mesenchymal tumours of the female genital tract are well described pathologically but\nthe imaging findings are less well documented. We present a case of an unusual\nmesenchymal tumour of the female genital tract, highlight the key sonographic and\nMRI findings and summarise the current diagnostic and management approach.\n\nA 50-year-old female presented with a history of prolonged menstrual bleeding and\nright iliac fossa discomfort. Her past medical history of note included\nendometriosis, a partially septate uterus and two previous lower segment cesarean\nsections. She had no allergies and denied any relevant drug history such as\ntamoxifen or hormonal therapy use.\n\nInitial pelvic ultrasound demonstrated normal uterus and ovaries, but detected a 13\nmm vascular soft tissue nodule in a 28 × 20 × 25 mm cystic lesion in\nthe left posterior vaginal fornix ( Figure 1 ).\nThe differential diagnosis at this time included a cervical or high vaginal polyp or\nan endometrioma, although the solid vascular component was unusual for the\nlatter.\n(a and b) Transvaginal B-mode ultrasound demonstrates a 25 mm cystic lesion\n(arrowheads) in the left lateral vaginal fornix with low-level internal\nechoes and a 13 mm solid nodule (white arrow), which is hypervascular on\nDoppler ultrasound. The uterus and ovaries were sonographically normal.\nThe patient proceeded to have a non-contrast enhanced MR study. This confirmed a\nwell-defined, predominantly cystic, 24 × 31 × 24 mm structure in the\nleft vaginal fornix of mildly hyperintense signal on T2 weighted (T2W) images,\nintermediate to high signal on T1 weighted (T1W) images with a mild increase in\nsignal intensity on T1W fat-suppressed images. The internal solid component was of\nintermediate signal with a hyperintense rim on T2W images and low signal on T1W\nimages. There was no signal suppression on the Short T1 Inversion Recovery (STIR)\nimages to suggest a fatty component ( Figure 2 ).\nThe lesion was not thought to be suspicious and, as the patient was asymptomatic, no\nintervention was undertaken at the time. Two years later, the patient returned with\npersistent vaginal bleeding and repeat imaging showed the cystic component had\nincreased in size and now measured 38 × 44 × 38 mm, but there was no\nchange in the solid component.\n(a–c) Pelvic MRI study shows a well-defined cystic structure\n(arrowhead) in the left vaginal fornix, which demonstrates mildly\nhyperintense internal signal on T1W images (a) and high signal on T2W images\n(b). The internal solid component (white arrow) had a T2 hyperintense rim\nand was hypointense on T1W images. There was no signal suppression on the\nSTIR images (c) to suggest a fatty component. STIR, Short T1 Inversion\nRecovery; T1W, T1 weighted; T2W, T2 weighted.\n\nThe patient underwent total laparoscopic hysterectomy and bilateral\nsalpingo-oophorectomy. The mixed solid cystic vaginal lesion was excised at the same\ntime.\nGross examination revealed a nodular well-circumscribed lesion measuring 16 ×\n12 × 14 mm and a small tissue fragment measuring 5 × 6 × 2 mm.\nBoth specimens were submitted to the laboratory labeled as vaginal cyst.\nMicroscopically, the small tissue fragment was a strip of tissue lined by stratified\nsquamous epithelium with little underlying stroma. The nodule was a\nwell-circumscribed lesion comprised of predominantly bland spindle cells with pale\neosinophilic cytoplasm and ill-defined cell borders, set in loose focally oedematous\nstroma containing abundant thin-walled vascular channels ( Figure 3 ). There were no mitoses and no evidence of necrosis,\nbut the tumour extended to the excision margin.\nHistological features of superficial angiomyofibroblastoma: The lesion is\ncomposed of bland spindle cells with pale eosinophilic cytoplasm set in\nstroma rich in thin-walled vascular channels (× 200).\nImmunohistochemical analysis showed the tumour cells were positive for ER, PgR, CD34,\nDesmin, CD99, BCL-2 and negative for H-caldesmon, SMA, c-Kit, S100 protein, MNF116,\nHMB45 and melan A. Less than 2% of cells were positive for Ki67.\nThe uterus and fallopian tubes were normal and both ovaries displayed cystic changes,\nbut there was no evidence of malignancy. The accompanying peritoneal washings were\nalso negative.\nA histological diagnosis of superficial myofibroblastoma of the lower female genital\ntract was made.\n\nSonographic surveillance over a period of 8 years has not demonstrated local\nrecurrence.\n\nSuperficial myofibroblastomas of the female genital tract are rare tumours but are\nwell documented histologically. They preferentially occur in the lower female\ngenital tract, with majority of cases in the vagina, cervix or vulva. 1 ,  2  The tumours have been described in patients with a wide age range 1  and there is at least one reported case of the tumour presenting in pregnancy. 3\nThe classical histological description is of a tumour with bland spindle cells that\nare weakly eosinophilic, set in loose stroma with multiple vascular channels,\noedematous foci and scattered inflammatory cells including lymphocytes and macrophages. 1 ,  2,4  The tumours also have a very characteristic immunohistochemical profile and\nare positive for oestrogen and progesterone, CD34, vimentin and desmin, 1 ,  2  but do not express smooth muscle actin, which distinguishes them from leiomyomas. 5\nAlthough these tumours do not have any malignant potential, 1  they can recur locally with one series demonstrating recurrence after a\nperiod of 9 years, 6  therefore long-term follow up is advised.\nThe aetiology of the tumours remains unclear. Approximately 94% of the\nreported cases (32 of 35 in 2010) occurred in peri-menopausal or menopausal females,\nand 32% of patients (11 of 35) had a history of tamoxifen or HRT use. 1 ,  7  Ganesan et al postulated that tamoxifen may drive growth through the estrogen\nreceptor or alternatively, patients on tamoxifen therapy are likely to be more\nvigilant as they are under surveillance for endometrial changes, so increasing the\ndetection rate. 1\nMesenchymal tumours in other locations have been linked to viral infections but to\ndate, only one study has addressed this possibility in myofibroblastomas of the\nfemale genital tract and revealed no association to human papilloma virus, human\nherpes virus 8 or Epstein Barr virus. 8\nThe imaging features have not been previously described in the medical literature.\nSonographically, the tumour in our case was a hyperechoic solid mass with intense\nintralesional vascularity and demonstrated slow interval growth ( Figure 1 ). The soft tissue nodule was isointense\nto muscle on T1W, intermediate signal intensity with a hyperintense rim on T2W\nimages, and of high signal on the STIR sequences ( Figure 2 ). Interestingly, the tumour in our patient occurred within a\nthin-walled vaginal cyst, which has not been previously described.\nThe finding of intralesional vascularity excludes lesions such as Bartholin cysts,\nrectoceles or urethral diverticula. 7  It also makes the potential diagnosis of a simple endometriotic cyst unlikely\nas these have been shown to be purely cystic with diffuse internal echoes and scanty\nvascularity on ultrasound. 9  The polypoidal, superficial cervical and vaginal subtypes of endometriosis\nhave a more solid soft tissue component, but their immunohistochemical profile\ndiffers from that of the superficial myofibroblastoma. 10\nThere is considerable overlap in the imaging and histological features of many of the\nmesenchymal tumours of the genital tract. The majority of diagnostic strategies are\naimed at distinguishing lesions on the benign end of the spectrum from the more\naggressive angiomyxoma subtype, which is invasive and warrants a different\nmanagement strategy. There is a subset of tumours that are relatively site-specific\nand arise from the superficial tissues of the genital tract. 7 ,  11  These include aggressive angiomyxoma, angiomyofibroma, fibroblastoma and\ncellular angiofibromas, as well as the tumour described in our case, superficial\nmyofibroblastoma. The second subset of tumours that frequently occur in this region\nbut also occur elsewhere includes fibroepithelial stromal polyps, leiomyomas and\nsuperficial angiomyxomas. 7\nAngiomyofibroblastoma subtypes are well-defined, homogeneous vascular lesions of\nmedium echogenicity on ultrasound and are homogeneously low intensity solid masses\non T2W images. 11  Cellular angiofibromas, also known as angiomyofibroblastoma-like tumour,\npresent as well-circumscribed tumours of inhomogeneous echotexture but are\nrelatively isoechoic to surrounding subcutaneous fat. On MRI, the tumours are\nisointense or hypointense to muscle on T1W images, heterogeneous intermediate to\nhigh signal on T2W images with scattered low signal areas depending on the fibrous\ntissue component, and enhance following administration of Gadolinium. 12  The majority of the more aggressive angiomyxoma subtypes are well-defined,\nhypoechoic, cystic or multiseptate vascular masses on ultrasound, of high T2 signal\non MRI and have a swirled or lamellate appearance. 13 ,  14\nUltimately, definitive diagnosis requires surgical excision and careful histological\nexamination. Long-term imaging follow up is recommended, as there has been at least\none case of recurrence after a period of 9 years. 6  Our patient has been free of recurrence for 8 years.\n\nWe have presented a unique case of superficial myofibroblastoma of the lower genital\ntract and discussed the imaging findings and main differential diagnosis.\n\nSuperficial myofibroblastomas are mesenchymal tumours that occur\npreferentially in the female genital tract.\nThe less aggressive spectrum of tumours appear to have a more solid,\nhyperechoic appearance on ultrasound and characteristic MR features when\ncompared to the more aggressive angiomyxoma subtypes which may be\nhypoechoic, cystic or septated; however, definitive diagnosis based on\nimaging alone is not possible and surgical resection is required.\nThese tumours may be locally aggressive and recur locally; therefore,\nfollow-up imaging is advised, although there is no consensus about the\noptimum length of time at present.\nThe authors declare no conflicts of interest and have obtained written\ninformed consent from all parties involved in this publication.","source_license":"CC-BY-4.0","license_restricted":false}