Faculty Bibliography

We present a case of a 41-year-old man with symptomatic pulmonary cement embolism following percutaneous vertebral augmentation, which was successfully retrieved via a percutaneous endovascular approach, a novel technique with only two prior cases reported. Cement leakage, including venous embolization of cement into the cardiopulmonary circulation, is a known potential complication following percutaneous kyphoplasty and vertebroplasty. While many patients with pulmonary cement embolism are asymptomatic and likely go undiagnosed, others experience respiratory distress and hemodynamic compromise requiring surgical and medical intervention. The optimal management for pulmonary cement embolism must be tailored to fit each individual patient, dependent upon the acuity of the clinical presentation, coexisting patient comorbidities, and the risks of systemic anticoagulation. In our patient, cement migration was visualized in real-time during vertebral augmentation. Endovascular retrieval by our Interventional Radiology section obviated the need for anticoagulation therapy or more invasive open surgical procedures.

BACKGROUND: Arteriovenous malformation is a short circuit between an organ's arterial and venous circulation. Arteriovenous malformations are classified as congenital and acquired. In the uterus, they may appear after curettage, cesarean delivery, and myomectomy among others. Their clinical feature is usually vaginal bleeding, which may be severe, if curettage is performed in unrecognized cases. Sonographically on 2-dimensional grayscale ultrasound scanning, the pathologic evidence appears as irregular, anechoic, tortuous, tubular structures that show evidence of increased vascularity when color Doppler is applied. Most of the time they resolve spontaneously; however, if left untreated, they may require involved treatments such as uterine artery embolization or hysterectomy. In the past, uterine artery angiography was the gold standard for the diagnosis; however, ultrasound scanning has diagnosed successfully and helped in the clinical management. Recently, arteriovenous malformations have been referred to as enhanced myometrial vascularities. OBJECTIVES: The purpose of this study was to evaluate the role of transvaginal ultrasound scanning in the diagnosis and treatment of acquired enhanced myometrial vascularity/arteriovenous malformations to outline the natural history of conservatively followed vs treated lesions. METHODS: This was a retrospective study to assess the presentation, treatment, and clinical pictures of patients with uterine Enhanced myometrial vascularity/arteriovenous malformations that were diagnosed with transvaginal ultrasound scanning. We reviewed both (1) ultrasound data (images, measured dimensions, and Doppler blood flow that were defined by its peak systolic velocity and (2) clinical data (age, reproductive status, clinical presentation, inciting event or procedure, surgical history, clinical course, time intervals that included detection to resolution or detection to treatment, and treatment rendered). The diagnostic criteria were "subjective" with a rich vascular network in the myometrium with the use of color Doppler images and "objective" with a high peak systolic velocity of >/=20 cm/sec in the vascular web. Statistical analysis was performed and coded with statistical software where necessary. RESULTS: Twenty-seven patients met the diagnostic criteria of uterine enhanced myometrial vascularity/arteriovenous malformation. Mean age was 31.8 years (range, 18-42 years). Clinical diagnoses of the patients included 10 incomplete abortions, 6 missed abortions, 5 spontaneous complete abortions, 5 cesarean scar pregnancies, and 1 molar pregnancy. Eighty-nine percent of patients had bleeding (n = 24/27), although 1 patient was febrile, and 2 patients were asymptomatic. Recent surgical procedures were performed in 55.5% patients (15/27) that included curettage (n = 10), cesarean deliveries (n = 5), or both (n = 1); 4 patients had a remote history of uterine surgery that included myomectomy. Treatment was varied and included expectant treatment alone in 48% of the patients with serial ultrasound scans and serum human chorionic gonadotropin until resolution (n = 13/27 patients), uterine artery embolization (29.6%; 8/27 patients), methotrexate administration (22.2%; 6/27 patients), hysterectomy (7.4%; 2/27 patients), and curettage (3.7%; 1/27 patients). Three patients required a blood transfusion. Of the 9 patients whose condition required embolization, the conditions of 7 patients resolved after the procedure although 1 patient's condition required operative hysteroscopy and 1 patient's condition required hysterectomy for intractable bleeding. Average peak systolic velocity after embolization in the 9 patients was 85.2 cm/sec (range, 35-170 cm/sec); the average peak systolic velocity of the 16 patients with spontaneous resolution was 58.5 cm/sec (range, 23-90 cm/sec). CONCLUSIONS: Acquired enhanced myometrial vascularity/arteriovenous malformations occurred after unsuccessful pregnancies or treatment procedures that included uterine curettage, cesarean delivery, or cesarean scar pregnancy. Triage of patients for expectant treatment vs intervention with uterine artery embolization based on their clinical status, which was supplemented by objective measurements of blood velocity measurement in the arteriovenous malformation, appears to be a good predictor of outcome. Ultrasound evaluation of patients with early pregnancy failure and persistent bleeding should be considered for evaluation of a possible enhanced myometrial vascularity/arteriovenous malformation.

Sclerosing hemangioma of the lung is a benign neoplasm with a widely debated histogenesis. It has a polymorphic histomorphology characterized by a biphasic cell population of "surface cells" and "round cells" arranged in four general patterns: Papillary, solid, angiomatous, and sclerotic. This variability in histomorphology makes it difficult to diagnose sclerosing hemangioma by fine needle aspiration (FNA). We present a case of sclerosing hemangioma diagnosed on FNA with immunohistochemistry performed on an accompanied cell block. The clinical presentation, cytomorphology, immunohistochemistry, and differential diagnoses are discussed.

Purpose: To identify variables predicting active extravasation and contrast-induced nephropathy (CIN) in angiography performed for acute bleeding. Material and methods: We retrospectively reviewed all conventional angiograms performed for acute bleeding (1/2013 to 6/2015), including 75 angiograms (46 gastrointestinal; 29 other sources of intra-abdominal bleeding cases) in 70 patients (26 women; 44 men) with a mean age of 59 (range 21-92) years. We recorded demographic information, comorbidities, vital signs immediately preceding angiography, post-procedure complications, and change in hematocrit (DELTAHct) and fluids and blood products administered over the 24 hours prior to angiography. Results: Of 75 exams, 20 (27%) showed contrast extravasation. DELTAHct was the only independent predictor of contrast extravasation (p=0.017). Patients with active extravasation had larger DELTAHct than those without extravasation (-17% vs. -1%; p=0.01). CIN occurred after 10 of 66 angiograms (15%), with GFR being the only independent predictor (p=0.03); 8% of patients with normal renal function (GFR>60) developed CIN; 29% of patients with mild-moderate impairment (GFR 30-60) developed CIN; and 67% of patients with severe renal impairment (GFR<30) developed CIN after angiography. Conclusion: Greater decrease in hematocrit over the 24 hours prior to angiography for intra-abdominal bleeding is an independent predictor of active extravasation, whereas pre-existing renal impairment is the independent predictor of CIN. The potential benefit of diagnosing active extravasation in patients with large hematocrit declines can be weighed against the risk of renal impairment

Purpose: Delineate sources of increased radiation during, cost of, and difficult retrieval after IVC filter (IVCF) placement. Material and methods: All 299 IVCFs (8/2013-12/2014) were identified by PACS search, 252 placed in a fluoroscopy suite (FS) and 47 in the operating room (OR), and reviewed for radiation exposure, fluoroscopy time, filter type, and angulation. Filter removals were assessed for the number of retrieval devices needed and fluoroscopy time. Results: Multiple linear regressions revealed that jugular versus femoral access and filter type had no impact on radiation exposure. However, filters placed in the OR entailed more radiation than in the FS (156.3 vs 71.4 mGy; p=0.001), longer fluoroscopy time (6.1 vs 2.8 min; p<0.0001), and resulted in greater filter angulation (4.8 vs 2.6degree; p<0.0001). Filter angulation was primarily dependent on the filter type (p=0.02), with the Venatech and Denali filters associated with decreased angulation (2.2 and 2.4degree, respectively), and the Option, Celect, and Meridian filters associated with greater angulation (4.2, 4.6, 4.7degree, respectively). There was a 32% retrieval rate. Filter angulation, but not filter type or filtration duration, independently predicted cases requiring more >1 retrieval device (p=0.0008) and >30 min fluoroscopy time (p=0.02). Cost savings for IVCF placement in the FS versus OR were estimated at $444.50/case. Conclusion: Increased radiation and cost were associated with placement in the OR, compared to the FS. Filter angulation was the primary determinant of difficulty in removing filters, while angulation was determined by filter type. Performing IVCF placement in the FS using specific filters may reduce radiation and cost, while enabling subsequent ease of retrieval

OBJECTIVES: To evaluate the management, clinical courses, and outcomes of cesarean scar pregnancies diagnosed in the first trimester. METHODS: We identified 60 cases of cesarean scar pregnancies diagnosed between 5 and 14 weeks. Group A contained 48 patients with fetal/embryonic cardiac activity; group B comprised 12 patients without cardiac activity; and group C included 11 patients with cardiac activity who chose expectant management. RESULTS: Five of the 48 patients (10.4%) in group A were successfully treated for vaginal bleeding. Thirty-three (68.7%) received methotrexate injections, and all had full resolution. Three (6.3%) required uterine artery embolization for late-developing arteriovenous malformations. Ten of the 12 patients (83.3%) in group B were managed expectantly and had full recovery. Two of the 10 (20.0%) had arteriovenous malformations; 1 had unsuccessful uterine artery embolization followed by a hysterectomy, and the second requested a hysterectomy. Ten of the 11 patients (90.9%) in group C continued the pregnancies. One declined local injection. Four of the 10 (40.0%) delivered live offspring by successive elective cesarean deliveries. Three (30.0%) had hysterectomies for placenta percreta, and 1 did not have a hysterectomy after delivery. Five (50%) had second-trimester complications, all leading to hysterectomies. Of the 60 patients, 20 (33.3%) had serious complications: 5 had arteriovenous malformations; 4 had uterine artery embolization; and 11 had hysterectomies. CONCLUSIONS: A cesarean scar pregnancy is a serious complication for patients who have had cesarean deliveries. Counseling, treatment, and follow-up are challenging for patients and caregivers. However, emerging data from different management approaches confirm that a cesarean scar pregnancy may progress and result in a live neonate at the expense of further fertility. This study confirmed that expectant management of a cesarean scar pregnancy is associated with a high risk of hysterectomy due to morbidly adherent placenta.

Purpose: TIPS creation fails to control ascites in 40% or more of patients, but the variables predicting outcome are unclear, with prior studies highlighting pre-TIPS portosystemic gradient (PSG) (Nair et al 2004; JVIR 15:1431). We studied which variables predict outcome of TIPS for refractory ascites. Materials and Methods: We retrospectively identified patients who underwent TIPS for refractory ascites between 1/12 and 5/14, yielding 40 patients. We excluded 17 patients due to insufficient peri-procedural documentation or technical failures, leaving 23 patients (16 men, 7 women, mean age 60 +/-2 yrs) for assessment of variables influencing osmotic (albumin and sodium levels) and hydrostatic (pre- and post- TIPS PSG and large varices) pressure. Responders were defined as those requiring fewer or no paracenteses; nonresponders had persistent ascites, with similar pre-TIPS frequency of therapeutic paracentesis. Complications within 1 month requiring hospitalization were noted. Multiple logistic regression, Mann-Whitney U tests, and one-tailed chi² tests assessed group differences. Results: Ten patients (43%: responders) had documented improvement in ascites. Multiple logistic regression including pre- and post-TIPS PSG significantly impacted outcome (p=0.04). Post- but not pre-TIPS PSG predicted outcome (p=0.04 vs. p=0.84). Responders had significantly lower post- TIPS gradient (5.8) compared with non-responders (7.6) (p=0.02). In contrast, responders and non-responders did not differ in albumin (2.7 vs. 2.7) or sodium (136 vs. 134) levels, or pre-TIPS gradient (13.9 vs. 14.7 mmHg) (p>0.05). Similar numbers of responders (50%) had large varices compared to non-responders (61%) (p=0.3). Responders (50%) had significantly more complications compared to non-responders (15%) (p=0.04), mostly encephalopathy (85%) requiring hospitalization. Conclusion: Only post-TIPS PSG predicted which patients had significantly reduced ascites, in contrast to prior studies suggesting importance of pre-TIPS gradient. Findings suggest aggressively lowering the gradient below 6 mmHg may be the most reliable technique to improve outcomes, although with expected higher risk of complications

Purpose: To evaluate if there is a significant difference in retrieval difficulty as determined by fluoroscopy time and equipment use when comparing three different filters: Denali, ALN, and Option/Option Elite. Materials and Methods: Retrospective analysis was performed of 33 filter retrievals performed over a 24 month period (12 men, 21 women; mean age 60). There were 8 Denali (Bard Peripheral Vascular), 9 ALN (ALN Implants), and 16 Option/ Option Elite (Argon Medical Devices) filter retrievals. Demographics, filter dwell time, fluoroscopy time during retrieval, and equipment used for retrieval were collected for each case. One Option case was excluded due to unsuccessful retrieval. Kruskal-Wallis H, two sample t-test and Chi-square analysis were used to determine significant differences in fluoroscopy time and equipment use between the three filter groups. Results: There were no significant differences between groups based on demographics or filter dwell time. Mean retrieval times were 4.9 min (SD: 3.6, range 1.2-10.3) for Denali, 9.9 min (SD: 5.5, range 2.9-18.6) for ALN, and 23.9 min (SD: 20.0, range 2.5 - 58) for Option/Option Elite. A Kruskal-Wallis H test showed a significant difference in these times (chi²=7.17, p=0.028), subsequent post-hoc analysis showed a significant difference only between Denali and Option but not between Denali and ALN or ALN and Option. Given normal distribution of fluoroscopy times for Denali and ALN, a separate t-test comparing these two filters did demonstrate a significant difference in fluoroscopy time (p=0.023). None of the Denali retrievals required additional equipment beyond a snare and sheath, 1/9 (11%) of ALN and 10/15 (67%) of Option retrievals required at least a tip deflecting wire or angled catheter (chi²=11.07, p=0.004). Sub-analysis separating Option and Option Elite filters did not change any results significantly. Conclusion: Denali filter retrieval involves significantly less fluoroscopy time than ALN or Option/Option Elite filter retrieval. In this series, Option/Option Elite retrieval also required significantly more equipment to achieve technical success compared to Denali and ALN retrieval. A larger sample size is required to corroborate these early results. (Table Presented)

Purpose: Various endpoints of embolization have been advocated by investigators during uterine artery embolization (UAE) for the treatment of symptomatic uterine fibroids. Described endpoints are based on angiographic appearance. Our intent is to objectively define a non-imaging endpoint of embolization in patients undergoing UAE. Materials and Methods: 55 consecutive patients with symptomatic uterine fibroids underwent UAE using an outer 5- and inner 3-Fr. catheters. Prior to and following embolization, systolic, diastolic, and mean intra-arterial blood pressures (BP) were measured within the ipsilateral internal iliac and the uterine (UA) arteries. For uniformity, the fluoroscopic endpoint was defined as the clearing of contrast within the main UA in or after 5 cardiac cycles. The mean+/-standard deviation of the within-subject percentage change in BPs from pre- to post-embolization were measured. p values, statistically significant at <=0.05, were assigned from a paired sampled t and Wilcoxon matched-pairs signed rank tests to assess whether the percentage change in BP was relevant. Results: Mean changes (pre- and post-embolization) in systolic, diastolic, and mean left UA BP measurements were: 17.74+/-17.87, 9.15+/-12.39, and 13.24+/-12.98, respectively. In the right UA, 29.76+/-24.05, 13.44+/-11.33, and 20.88+/-15.66. All assigned p values were <0.0001. Conclusion: In this study, parametric differences in uterine artery blood pressures before and after UAE were shown to correlate to an accepted angiographic endpoint. UA BP measurements can be used as an equivalent parametric non-imaging endpoint to UAE. A non-angiographic endpoint may lead to the reduction of radiation dosage to the patient and the interventional radiologist when performing UAE. (Table Presented)

Learning Objectives: Learn to utilize artificial ascites to safely perform radiofrequency ablation of subdiaphragmatic hepatocellular carcinoma. Background: RFA is an effective treatment of HCC. A relative contraindication is subdiaphragmatic tumor location because the diaphragm is at increased risk of thermal injury. Artificial ascites of 5-10 mm thickness creates a protective space between the tumor and diaphragm. 5% dextrose in water (D5W) is used because it provides greater electric isolation than normal saline. The ascites decreases post-procedural pain and recovery time without sacrificing efficacy because there is no significant associated heat sink effect. Clinical Findings/Procedure Details: Ultrasound guidance is used to advance a 20-gauge needle into the peritoneal cavity. If the tumor is in the right hepatic lobe, a subcostal approach along the anterior axillary line is used. If the tumor is in the left hepatic lobe, a subcostal epigastric approach is used. A small amount of D5W (20-50 mL) is injected to create a space for placement of an 8 French pigtail catheter. Placement of the catheter close to the tumor, between the liver and diaphragm is desirable. The catheter is infused with D5W and creation of artificial ascites is monitored under ultrasound or CT. Artificial ascites improves the sonic window for ultrasound guided placement of the RFA probe. If the tumor remains difficult to visualize, CT guidance is used for probe placement. Following completion of ablation, the catheter is drained to gravity. Blood tinged fluid is common following the procedure. The catheter is left in place 12-24 hours to drain residual ascites and prevent hemoperitoneum. Other potential complications include peritonitis, pleuritic chest pain, and pleural effusion. Conclusion and/or Teaching Points: Artificial ascites can effectively decrease risk of thermal injury to the diaphragm, post-procedural pain, and recovery time without sacrificing efficacy of subdiaphragmatic HCC radiofrequency ablation. The infusion catheter should be placed close to the tumor, between the liver and diaphragm, to create 5-10 mm thickness of D5W artificial ascites. The catheter should be drained for 12-24 hours post-procedurally to remove residual ascites and decrease risk of hemoperitoneum