David Li, Bradley B. Pua and David C. Madoff. Semin Intervent Radiol. 2014 March; 31(1): 70-81.
Renal cell carcinoma (RCC) is the most common malignant renal tumor, comprising an estimated 2 to 3% of all malignancies with an estimated 64,770 new cases diagnosed in the United States in 2012. Common symptoms associated with RCC include hematuria, flank pain, and a palpable mass. Patients presenting with symptoms usually have advanced stage disease. The natural history of RCC has changed, with the majority (∼70%) now diagnosed incidentally on routine abdominal imaging. Despite the associated stage migration due to increase in incidental detection, there remain a significant proportion of patients presenting with advanced stage disease. Approximately 30 to 40% of patients with renal tumors will either present with or later develop metastatic disease, and death rate from RCC remains high, with 13,570 disease-related deaths in 2012.
Prognosis significantly changes with stage, ranging from a 5-year overall survival (OS) of 90% for localized disease to 10% for advanced stage disease. For localized RCC, surgical resection is the treatment of choice. Partial nephrectomy (nephron-sparing surgery) to preserve renal function is preferred over radical nephrectomy if technically feasible and has been shown to not compromise survival outcomes. For patients unable to undergo surgery, local ablative therapies remain an option.
Renal angiomyolipoma (AML) is the most common mesenchymal renal tumor, with a reported incidence of 0.3 to 3% in the general population. Renal AMLs can occur either in a sporadic form or in association with tuberous sclerosis (TS). AMLs are generally considered to be benign; though in rare cases, it can be locally aggressive with either extension into the renal vein and inferior vena cava or dissemination into adjacent lymph nodes. AMLs are typically asymptomatic and found incidentally on imaging. However, a minority present with hematuria, flank pain, or spontaneous hemorrhage. Thus, treatment of symptomatic AMLs or those at risk of hemorrhage is recommended.
Almgård first popularized renal artery embolization (RAE) in the 1970s and initial interest in the technique resulted in multiple published series in the 1980s and 1990s. Indications for RAE for the treatment of renal masses include as an adjunctive preoperative treatment before nephrectomy for primary kidney masses; palliation of symptoms related to advanced stage RCC; and primary treatment of AML.
Anatomic and technical considerations
Multiple previous review articles have detailed the anatomic and technical considerations for RAE. Classically, single renal arteries arise from the abdominal aorta at the level of the L1–L2 interspace. The main renal arteries branch into anterior and posterior divisions, followed by segmental, lobar, interlobar, and arcuate arteries. Variant anatomy can occur in greater than 30% of the general population, with either early division of the main renal arteries or extrarenal arteries further subdivided into accessory (hilar) or aberrant (polar) entry into the kidney.
Arterial vascular access is typically gained via the right common femoral artery and a vascular sheath (5 French) placed. Through the vascular sheath, aortography can be performed to locate the origin of the renal arteries with a flush catheter placed slightly superior to the expected origin of the renal arteries. Selection of the renal arteries is performed with shaped catheters such as a SOS-shaped, Cobra, or Simmons-shaped catheter. Superselective catheterization of the renal arteries can be performed with a microcatheter and microwire. Embolization materials can be introduced into the desired target vessels. If ethanol is chosen as the embolic material, occlusion balloon catheter delivery systems are typically used.
There is wide variation in embolization agents used for RAE as highlighted in Table 1. No specific agent has been demonstrated to be more efficacious than others in the setting of RAE. Several principles specific to RAE should be noted in choosing an embolic agent. First, vessels in the kidney are considered “end arteries” without significant intrarenal collaterals. Second, RCCs are hypervascular tumors that recruit extrarenal collateral vascular supply. As such, utilization of an agent that results in permanent small vessel/capillary bed occlusion/sclerosis such as NBCA glue, ethanol, PVA, or Embospheres (Merit Medical, South Jordan, UT) in addition to large vessel occlusion with coils is desired.
Table 1. Case series published in the English literature, in order of year of publication, agents used for therapeutic embolization, number of patients, and indication for embolization
Current indications for RAE include palliation for advanced stage RCC; preoperative embolization before nephrectomy; treatment for angiomyolipoma; and as an adjunctive therapy to ablation for RCC.
To date, no prospective randomized controlled clinical trials have been performed to evaluate the role of RAE in these settings. In 1999, Kalman and Varenhorst published a survey of the pertinent literature on RAE, which was later adapted by Madoff et al in 2006.
Table 1 highlights some of the limitations of the previous evidence accumulated in the 20th century. These studies were limited by design, many being observational with significant variation in both patient selection criteria and embolic agents used, making it difficult to draw firm conclusions.
Given the wide variation in results and lack of strong clinical evidence, preoperative RAE has largely fallen out of favor. Scattered case series have been published since 2000 as summarized in Table 2, and have not altered the role of RAE.
Table 2. Case series published in the English literature since 2000 for renal arterial embolization performed in the setting of renal cell carcinoma, in order of year of publication, number of patients, indication for embolization, conclusions, and study limitations
Benefits of RAE in the preoperative setting include a decrease in perioperative blood loss, creation of a tissue plane of edema facilitating dissection, and reduction in tumor bulk including extent of vascular thrombus, when present. Wide variation in reporting markers such as reduction in intraoperative blood loss, transfusion requirements, surgical procedure time, surgical complications, and survival outcomes has limited its use to local practice patterns (Fig. 1).
Figure 1. A 53-year-old woman with 11.7 cm left renal cell carcinoma that underwent preoperative left renal arterial embolization followed by uneventful left nephrectomy with a reported total estimated blood loss of 300 mL. (A) Axial contrast-enhanced computed tomography image of the enhancing left renalmass (arrow). (B) Selective left renal arteriogramdemonstrating a large hypervascular renal cell carcinoma (arrow) before embolization. (C) Microcatheter-mediated embolization of the hypervascular mass using tris-acryl gelatin microspheres to stasis. Note the decreased regions of tumor vascularity (arrow). (D) Postembolization selective left renal arteriogram demonstrating near-complete embolization of the left renal artery to stasis.
In an observational study by Schwartz et al performed at the authors’ institution, there was a perceived benefit of both decreased intraoperative transfusions and operative time. The study population included those with advanced stage disease with mean tumor size of 11.2 cm. Almost half of all patients had vascular invasion, with reported blood loss of only 1,048 mL (median 725 mL) and average patient transfusion requirements over their entire hospital course of 3.9 units. In the subset of patients who did not have vascular invasion, mean blood loss was lower (mean 647 mL, median 425 mL). Described complications were minor (none requiring hospitalization) and surgical times and survival outcomes were not analyzed.
Recent studies by May et al and Zielinski et al include case-controlled matched historical cohorts with survival outcomes that are contradictory. May et al concluded no survival benefit from RAE, while Zielinski et al concluded significant survival benefit. This difference may be a result of different selection processes of the control groups. In the study by Zielinski et al, controls were chosen from the same historical cohort and matched in age, sex, tumor size, grade, and stage. In the report by May et al, controls were chosen from the timeframe of 1992 to 2006, with baseline oncology characteristics matched based on propensity scores to the RAE cohort performed within the timeframe of 1992 to 1997. In both series, surgical complications were similar in both control and RAE groups; intraoperative blood loss and intraoperative times were not directly evaluated. May et al did demonstrate increased postoperative transfusion requirements in the RAE cohort.
Subramanian et al provided comparisons of those undergoing preoperative RAE to controls in regard to intraoperative times, blood loss, and immediate perioperative complication and mortality rates. In this study, preoperative RAE was associated with increased median operative times (390 vs. 313 minutes), median intraoperative transfusion requirements of 8 versus 4 units of packed red blood cells, and perioperative mortality of 13% versus 3%. However, the control population was not equal with the preoperative RAE group, with the latter being composed of patients of higher tumor stage, American Society of Anesthesiology (ASA) scores, and need for the use of intraoperative cardiopulmonary bypass.
Several recent small series (n = 8–25 patients) have focused on RAE in the setting of palliation, demonstrating favorable results for symptom control. Treatment indications commonly included palliation of hematuria and/or flank pain, and less commonly control of paraneoplastic syndromes and palpable masses. Mukund and Gamanagatti demonstrated symptom control in 7/8 (88%) patients, with gross hematuria controlled in 6/7 (86%) patients and flank pain controlled in 1 patient at 6 months. Maxwell et al demonstrated symptom control in 18/19 (95%) patients with hematuria controlled in 13/13 (100%) patients and flank pain in 8/9 (89%) patients. Munro et al also demonstrated a 70% effectiveness in symptom control, with only a 12% rate of recurrent hematuria requiring hospital admission. Onishi et al compared 30 historical controls with 24 patients who underwent RAE who were matched for age, sex, Eastern Cooperative Oncology Group (ECOG) performance status, and stage. Similar to the other studies, RAE was 75% effective in symptomatic control (14 patients with gross hematuria and 5 patients with flank pain). Interestingly, the RAE group showed a statistically significant median survival benefit of 229 versus 116 days for the control group, suggestive of a possible improvement in outcome either through a cyto-reductive effect or via immunomodulation. This result, while encouraging, is still inferior to surgical palliation, with 7-month median survival in the RAE group versus 17.8 months median survival gained in the setting of palliative nephrectomy.
RAE in the setting of palliation is well tolerated, although postembolization syndrome is common after RAE (reported in 14/24 (57%) patients in one study). Postembolization syndrome is well controlled with antipyretics, analgesics, and antiemetics, usually resolving in 3 days. Median hospital stay after RAE is 4 to 5 days.
Treatment for angiomyolipoma
RAE is well established as the first-line therapy for the treatment of AML. Indications for RAE include symptomatic AMLs such as those with prior or active hemorrhage, flank pain, hematuria, or mass effect. AMLs larger than 4 cm have been shown to be associated with symptoms (80–90%) and spontaneous hemorrhage (50–60%). Therefore, prophylactic treatment of AMLs greater than 4 cm serves as another indication for RAE (Fig. 2). Multiple retrospective series have been published since 2000 that provide long-term follow-up confirming safety and long-term efficacy of RAE for AML (Table 3).
Table 3. Case series published in the English literature since 2000 for renal arterial embolization performed in the setting of angiomyolipoma, in order of year of publication, number of patients, mean size, indication for embolization, embolization agents and outcomes
Figure 2. A 45-year-old woman with incidental 9.4 cm right renal angiomyolipoma that underwent renal arterial embolization. (A) Axial contrast-enhanced computed tomography image of angiomyolipoma demonstrating predominantly fatty components, vascular components (arrow), and origin of the mass from the kidney. (B) Selective right renal arteriogram demonstrating a small area of heterogeneity and hypervascularity overlying themidpole region of the right kidney (arrow). (C) Microcatheter-mediated embolization of the hypervascular aspects of the mass using tris-acryl gelatin microspheres to stasis (arrow). (D) Postembolization selective right renal arteriogram demonstrating complete embolization of the right AML.
These studies show overall durable tumor treatments with low surgical salvage rates, with recurrence rates ranging from 0 to 37%. The majority of recurrences are successfully managed with repeat RAE. Complications are rare with no episodes of renal insufficiency attributable to embolization documented in any series. Postembolization syndrome is the most common complication, occurring in up to 64% of patients in one series. Most cases are self-limited and successfully treated with antipyretics and nonsteroidal anti-inflammatories. Major reported complications are limited to abscess formation, coagulative necrosis, and self-limited nontarget embolization to normal renal parenchyma, all successfully treated without long-term adverse sequelae.
The proven efficacy of nephron sparing surgery (NSS) in the setting of RCC has renewed interest in surgical options for AML, and multiple recent series evaluating the use of NSS as an alternative to embolization for AMLs have been published. Proponents of NSS as compared with embolization for AML argue that NSS provides more durable outcomes and also offers pathology in case of malignancy. However, NSS has increased long-term adverse sequelae associated with the procedure, with Boorjian et al reporting a 5% urinary leak rate and Heidenreich et al reporting a 7% urinary fistula rate. Therefore, NSS is currently limited to the setting of failed RAE.
The choice of embolization agent for AML is controversial. Common agents include ethanol, PVA, and Embospheres (Merit Medical). The use of Onyx (Covidien, Plymouth, MN) in AML treatment has also been reported. The utility of coils remains unclear with some authors advocating their use, and others suggesting that their use promotes collateral formation around the level of occlusion. Lenton et al reported a high intraprocedural aneurysm rupture rate (30%) as compared with most groups that reported no episodes of rupture. These authors conveyed that this may be attributed to the use of PVA alone rather than PVA and coils. Villalta et al found that smaller microspheres (< 150 μm) were associated with both an increased recurrence rate requiring repeat embolization (odds ratio (OR) 5.88) and risk of pulmonary complications; these complications resulted in the authors advocating for larger sized particles.
Large AMLs (those sized 10 cm or greater) are considered more resistant to RAE. In a series reported by Chan et al, subgroup analysis demonstrated that recurrences necessitating surgery only occurred when treating AMLs > 10 cm in size. Larger tumors are thought to be more difficult to embolize due to their size, multiple feeding vessels, and greater difficultly in isolating them from normal renal parenchyma.
While several series suggest efficacy of treatment of AMLs associated with TS, active surveillance of this high-risk group is prudent. In fact, Kothary et al suggest that AMLs treated with RAE in this group may be associated with higher rate of regrowth.
Significant variation in the literature exists for the optimal time from embolization to nephrectomy. In their review of the literature, Kalman and Varenhorst concluded that the optimal delay from embolization to nephrectomy should be less than 48 hours. Postembolization syndrome (characterized by lumbar pain, nausea, and fever) occurs in a majority of patients 1 to 3 days postprocedure, and can be minimized with nephrectomy performed within this timeframe. Additionally, surgery is technically more difficult 72 hours after embolization, which is thought to be secondary to collateral vessel formation. Controversy surrounds the optimal timing of resection; some authors suggest a delay of 24 to 48 hours after embolization, which allows edema to develop facilitating surgical dissection, while other authors have suggested that there should be as minimal delay as possible to prevent collateral vessel formation.
Reducing the time between RAE and surgery can minimize postembolization syndrome. In a small series by Lin et al, 8 patients underwent concomitant RAE and surgical resection as compared with 14 control patients who underwent conventional staged preoperative embolization. The concomitant group had no instances of postinfarction syndrome as compared with 36% in the staged group, thereby increasing patient comfort, decreasing hospital stay, and as a result reducing health care costs. Another series by Carvajal et al studied seven patients who underwent concomitant RAE and surgical resection. This study was limited to tumors > 13 cm in size (Stage II, IIIA–B), but did demonstrate safety and increased patient comfort.
In treatment of primary hepatic malignancies, there is a growing body of evidence indicating that transarterial embolization therapy in conjunction with localized ablative therapies result in improved patient outcomes as compared with either monotherapy alone. The two therapies are thought to be synergistic, with the decreased blood flow from embolization resulting in decreased heat loss during radiofrequency ablation. RAE may also serve as a useful adjunctive technique to thermal ablative therapies performed in the setting of nonoperable RCC. The synergistic effect of the two modalities may also have a positive impact on immunomodulatory mechanisms. Li et al combined cryoablation and RAE, and showed a decrease in T-regulatory cells thought to be related to a large volume of cell necrosis and release of tumor antigens. The promotion of T-regulatory cells and inflammation is thought to be one method by which tumor cells can subvert normal immune regulation methods. In clinical practice, the role of RAE in combination with local ablation setting is still largely undefined, with only a few case reports and small series performed demonstrating feasibility and safety.
RAE is a safe and effective technique that is well tolerated with few complications, particularly if the time interval from embolization to surgery is reduced to less than 48 hours.
To date, RAE has not been evaluated in a randomized controlled setting, which has contributed to its underutilization. RAE serves as the initial treatment in the management of symptomatic AMLs or those at risk for spontaneous rupture (> 4 cm). In addition, RAE is extremely effective for palliation of symptoms in the setting of nonoperative advanced stage RCC. Potential benefits of this procedure for preoperative RCC treatment include reduction in intraoperative blood loss and operative time by facilitating dissection. Finally, there is suggestion of improved survival benefit after RAE, possibly through immunomodulatory mechanisms. All of these potential benefits warrant the need for prospective studies for further validation.