Management of inferior vena cava tumor thrombus in locally advanced renal cell carcinoma

Abstract

The diagnosis of renal cell carcinoma is accompanied by intravascular tumor thrombus in up to 10% of cases, of which nearly one-third of patients also have concurrent metastatic disease. Surgical resection in the form of radical nephrectomy and caval thrombectomy represents the only option to obtain local control of the disease and is associated with durable oncologic control in approximately half of these patients. The objective of this clinical review is to outline the preoperative evaluation for, and operative management of patients with locally advanced renal cell carcinoma with venous tumor thrombi involving the inferior vena cava. Cornerstones of the management of these complex patients include obtaining high-quality imaging to characterize the renal mass and tumor thrombus preoperatively, with further intraoperative real-time evaluation using transesophageal echocardiography, careful surgical planning, and a multidisciplinary approach. Operative management of patients with high-level caval thrombi should be undertaken in high-volume centers by surgical teams with capacity for bypass and invasive intraoperative monitoring. In patients with metastatic disease at presentation, cytoreductive nephrectomy and tumor thrombectomy may be safely performed with simultaneous metastasectomy if possible. In the absence of level one evidence, neoadjuvant targeted therapy should continue to be viewed as experimental and should be employed under the auspices of a clinical trial. However, in patients with significant risk factors for postoperative complications and mortality, and especially in those with metastatic disease, consultation with medical oncology and frontline targeted therapy may be considered.

Introduction

In 2014, there were approximately 63,920 new cases of cancer of the kidney and renal pelvis in the United States, of which the majority are renal cell carcinoma (RCC) [American Cancer Society, 2014]. An unusual hallmark of RCC is the biological predisposition for vascular invasion, with extension of tumor thrombus into the inferior vena cava (IVC) occurring in 10–25% of cases [Marshall, 1989; Slaton et al. 1997]. The natural history of untreated RCC with venous tumor thrombus (VTT) is poor, as demonstrated by a recent population-based study of patients with VTT managed expectantly demonstrated a median survival of 5 months with a 1-year disease-specific survival of only 29% [Reese et al. 2013]. Conversely, previous reports have demonstrated durable cancer-free survival following radical nephrectomy with tumor thrombectomy [Blute et al. 2004; Pouliot et al. 2010; Hirono et al. 2013; Whitson et al. 2013; Haddad et al. 2014]. Given the poor prognosis if these patients are untreated, and with associated risks such as distal embolism and the sequelae of venous congestion, the diagnosis of RCC with VTT necessitates expedient evaluation, preoperative optimization and rapid coordination of a multidisciplinary team to prepare appropriately selected patients for surgical extirpation.

The objective of this clinical review is to outline the preoperative evaluation for, and operative management of patients with locally advanced RCC with VTT involving the IVC.

Diagnostic imaging and preoperative evaluation

The initial presentation of locally advanced RCC with VTT may be incidentally detected; however, the majority of patients with VTT are symptomatic. These symptoms may be related to localized tumor growth such as flank pain or hematuria; constitutional symptoms such as paraneoplastic syndromes, fatigue, or weight loss; or related to venous occlusion secondary to the VTT, including lower extremity edema, acute varicocele, ascites, Caput medusa, Budd Chiari syndrome, or pulmonary embolism which may be present in up to 5% of patients at diagnosis [Blute et al. 2004; Parekh et al. 2005; Abel et al. 2014a]. Importantly, preoperative pulmonary embolism does not necessitate deferral of surgical management of the VTT. As demonstrated by retrospective data from Abel and colleagues, preoperative pulmonary embolism is not associated with increased 90-day postoperative mortality, nor is it associated with RCC-recurrence or cancer-specific mortality [Abel et al. 2014a].

Once RCC with VTT is identified, appropriate preoperative staging studies to assess for the presence of synchronous metastatic disease include cross-sectional imaging via computed tomography (CT) of the chest, abdomen, and pelvis, with and without intravenous contrast as permitted by renal function [Woodruff et al. 2013; Motzer et al. 2015]. The purpose of abdominal cross-sectional imaging includes characterization of the tumor thrombus extent, as well as local extension of the renal tumor into the perirenal adipose tissue, contiguous or metastatic involvement of the ipsilateral adrenal gland, retroperitoneal or intraabdominal adenopathy or metastases, delineation of renal vascular anatomy, and assessment of enlarged collateral vessels.

Preoperative laboratory studies include a complete blood count, comprehensive metabolic panel including calcium, liver function tests, and a urinalysis. Bone scan and brain imaging [preferentially by magnetic resonance imaging (MRI)] should be obtained if clinically indicated by local symptoms. If urothelial carcinoma is in the differential diagnosis after initial imaging, or in the presence of gross or microscopic hematuria, urine cytology and cystoscopy should be considered. The metastatic evaluation should be performed within 30 days of definitive surgery. Additionally, if clinical symptoms are consistent with deep venous thrombosis of the lower extremities, a Doppler ultrasound to assess venous patency of the lower extremities is indicated.

With respect to formal evaluation of the renal mass and characterization of the degree of IVC occlusion and extent of VTT, in addition to assessing for the presence of bland thrombus inferior to the tumor thrombus, most authors recommend obtaining an abdominal MRI within 1–2 weeks of surgery given the propensity for VTT to progress rapidly [Boorjian et al. 2007; Wotkowicz et al. 2008; Boorjian and Blute, 2009; Woodruff et al. 2013]. Gadolinium administration is based upon glomerular filtration rate (GFR) such that a full dose is administered if GFR is at least 60 ml/min/1.73 m²; a decreased dose is given if GFR is 30–60 ml/min/1.73 m²; and gadolinium is contraindicated if GFR is less than 30 ml/min/1.73 m², given the risk of inciting nephrogenic systemic fibrosis. With recent advancements in CT imaging, specifically the use of multidetector CT scanners with the capacity for three-dimensional reformatting, high-resolution imaging is possible, such that CT scans may be used for thrombus characterization.

Comparison of imaging-derived thrombus extent on multidetector CT to final pathological specimens in 41 patients yielded a concordance rate of 84% with 96% accuracy in predicting thrombus extent [Guzzo et al. 2009]. Lawrentschuk and colleagues reported equivalent accuracy in a head-to-head comparison of CT and MRI compared with surgical pathology in eight patients [Lawrentschuk et al. 2005]. The accuracy of CT versus MRI has yet to be directly compared regarding the determination of extent of bland thrombus inferior to the VTT. However, multidetector CT may be utilized as an alternative imaging method in patients in whom MRI is contraindicated due to non-MRI-compatible implants (e.g. pacemakers) or in patients unable to tolerate MRI due to claustrophobia.

Characterization of the tumor thrombus includes assessment of the tumor thrombus level (Table 1). Additionally, various features of the tumor thrombus and the IVC have important utility in preoperative surgical planning. In a cohort of 18 patients, Gohji and colleagues observed that IVC diameter greater than 40 mm on preoperative abdominal CT was prognostic of extensive invasion into the IVC [Gohji et al. 1994]. Work by Zini and colleagues reported that Anterior-Posterior (AP) diameters on preoperative abdominal MRI of the IVC and renal vein at the level of the renal vein ostium (RVo) greater than 18 and 14 mm, respectively, were 90% sensitive for predictive IVC wall invasion [Zini et al. 2008]. Aside from raw vessel diameter, Aslam Sohaib and colleagues proposed that the most reliable sign of vessel wall invasion was a tumor signal on both the intraluminal and extraluminal sides of the IVC wall, and reported that MRI was 92% accurate in predicted vein wall invasion in a series of 12 patients [Aslam Sohaib et al. 2002]. More recently, in a contemporary series of 172 patients with IVC venous tumor thrombi, the ability to predict the need for extensive vascular resection necessitating complex vascular reconstruction (e.g. complete IVC resection, venous patch graft, or tube-interposition graft) at the time of tumor thrombectomy from preoperative imaging was assessed [Psutka et al. 2014]. Optimal radiographic thresholds that univariately predicted the need for extensive vascular resection included a renal vein diameter at the RVo of 15.5 mm, maximal diameter of the IVC of 34 mm, and AP and coronal diameters of the IVC in the anterior-posterior dimension at the level of the RVO of 24 and 19 mm, respectively. On multivariable analysis, the presence of right-sided tumor [odds ratio (OR) 3.3, p = 0.017], AP diameter of the IVC at the RVo at least 24 mm (OR 4.4, p = 0.017), and radiographic evidence of complete occlusion of the IVC at the RVo (OR 4.9, p < 0.001) were associated with a significantly increased risk of need for extensive vascular resection. Furthermore, if a patient had none of these features, the predicted probability of requiring extensive vascular resection was 2% whereas in comparison to 66% of patients with all three risk factors.

Table 1.

Classification of tumor thrombus level.

Tumor thrombus level	Definition
0	Tumor thrombus is limited to the renal vein, detected clinically or during assessment of the pathological specimen
I	Tumor thrombus extends into the IVC, <2 cm above the renal vein
II	Tumor thrombus extends into the IVC, >2 cm above the renal vein but below the hepatic veins
III	Tumor thrombus extends above the hepatic veins but below the diaphragm
IV	Tumor thrombus extends above the diaphragm, including atrial thrombus

IVC, inferior vena cava.

The presence of complete occlusion of the IVC as denoted by failure of contrast to pass around the tumor thrombus has also been independently associated with increased risk of tumor thrombus invasion into the IVC wall [Zini et al. 2008; Psutka et al. 2014]. Taken together, these radiographic characteristics may be useful in predicting the extent of surgical resection of the IVC that will be necessary to excise the VTT, which has implications both for patient counseling as well as informing a surgeon’s decision to reserve operative resources such as cardiopulmonary bass or to involve a vascular sur-geon in the care of the a patient preoperatively. Intraoperative imaging adjuncts that further confirm and provide important guidance with respect to the upper extent, consistency, and mobility of the tumor thrombus include duplex ultrasound or transesophageal echocardiography.

In addition to assessing the upper extent of the tumor thrombus, the presence of bland thrombus inferior to the tumor thrombus has implications for surgical planning. Blute and colleagues characterized different patterns of bland thrombus involvement in cases with VTT (Table 2), proposing primary cavorrhaphy in cases without venous occlusion or distal or bland thrombus, and permanent IVC interruption in cases with associated partially or fully occlusive distal or bland thrombus, with or without segmental IVC resection [Blute et al. 2007].

Table 2.

Preoperative classification of bland thrombus and degree of venous occlusion, and proposed surgical management [Blute et al. 2007].

Preoperative classification	Definition	Surgical approach
A	IVC without venous occlusion, no associated distal or bland thrombus	Primary cavorrhaphy
B	IVC partially occluded with distal bland thrombus in the pelvis only	Intraoperative deployment of Greenfield (IVC) filter
C	Partial IVC occlusion by tumor thrombus with total occlusion distally by bland thrombus	Staple/ligate the IVC inferior to the level of the IVC tumor thrombus with permanent interruption of the IVC
D	Total occlusion of the IVC by tumor thrombus/bland thrombus	Segmental resection of the IVC with permanent interruption of the IVC

IVC, inferior vena cava.

Coordination of the surgical team and preoperative optimization

In preparation for surgery, medical clearance is necessary. Additional consultation from anesthesia and cardiothoracic surgery are recommended for patients over 50 years of age as well as those with level III and IV tumor thrombi who are expected to require either veno-venous bypass or cardiopulmonary bypass with or without circulatory arrest. If the primary surgeon does not have expertise in vascular reconstruction or liver mobilization, these surgical specialists should be involved in planning and the conduct of the operation in patients with imaging findings of extensive intravascular tumor and higher level tumor thrombus [Calero and Armstrong, 2013]. Additionally, in patients with concurrent metastatic disease in whom simultaneous metastasectomy is planned, the relevant surgical subspecialties should be involved (e.g. thoracic surgery, hepatobiliary, colorectal surgery).

In patients with completely occlusive tumor thrombus, near-completely occlusive thrombus, recent deep venous thrombosis or embolic event, or bland thrombus in addition to tumor thrombus, preoperative therapeutic anticoagulation should be considered. We advocate the use of low-molecular weight heparin (LMWH) started in the outpatient setting in such cases [Woodruff et al. 2013]. LMWH is preferentially utilized given level one evidence demonstrating a decreased risk of recurrent venous thromboembolic events without increased risk of major bleeding episodes, and increased overall survival in comparison to warfarin among patients with malignancy [Lee, 2009; Lyman et al. 2013]. Anticoagulation should be held 24 h prior to surgery. For patients with a contraindication to LMWH or who are unable to obtain long-term LMWH, warfarin may be started with appropriate bridging therapy, such as an unfractionated heparin drip as an inpatient or short-term LMWH at therapeutic dose until the international normalized ratio reaches a target of 2–3.

Preoperative placement of IVC filters is generally avoided in surgical candidates as these devices increase the surgical complexity of thrombectomy and can become incorporated in the tumor thrombus [Woodruff et al. 2013]. However, suprarenal IVC filters may be employed in patients with continued pulmonary embolism (PE) despite anticoagulation or in patients in whom anticoagulation is contraindicated.

Preoperative renal angioinfarction has been recommended prior to radical nephrectomy with tumor thrombectomy [Wotkowicz et al. 2008]. This practice has been proposed to facilitate nephrectomy through reduction of intraoperative blood loss and operative time [Singsaas et al. 1979; Bakal et al. 1993], increasing the ease of dissection though local tissue edema from hypoxia and tissue necrosis [Wotkowicz et al. 2008]. In the setting of preoperative interruption of arterial supply to the kidney, the anterior renal vein may then be ligated intraoperatively prior to identification of the renal artery without increasing the risk of significant hemorrhage from venous collaterals or large volume blood volume sequestration within the kidney. Preoperative renal angioinfarction may also prevent of tumor cell spillage following vascular ligation, and has been proposed to have an immunomodulatory role, augmenting natural killer cell activity, triggered by tumor necrosis factor release [Lopatkin et al. 1979; Bakke et al. 1982; Nakano et al. 1983; Kaisary et al. 1984]. It is important to monitor patients for postinfarction syndrome caused by the immune response to the infarcted kidney, characterized by fevers, chills, flank pain, malaise, hematuria, transient hypertension, and hyponatremia. These sequelae have been reported in 74–89% of patients following renal artery embolization, and are generally self limited [Schwartz et al. 2007; May et al. 2009]. Additionally, one must keep in mind that postinfarction syndrome as well as potential postinfarction renal insufficiency and ileus have been reported to delay definitive surgery by up to 14 days [Zielinski et al. 2000].

Aside from postinfarction syndrome, there are conflicting data regarding the utility of preoperative renal artery embolization. In a matched retrospective cohort series from Zielinski and colleagues, the authors reported that 5-year overall survival following renal artery embolization prior to radical nephrectomy in 90 patients with pT3 disease was superior to those treated with radical nephrectomy alone (56% versus 43%, p < 0.01) [Zielinski et al. 2000]. Conversely, May and colleagues compared 227 patients treated with preoperative renal angioinfarction with 607 patients treated with radical nephrectomy alone. Propensity score matching yielded two matched cohorts of 189 patients, between whom there were no significant differences in postoperative complications other than decreased intraoperative blood loss among those who had undergone prior embolization, cancer-specific or overall survival [May et al. 2009]. Similarly, the International Renal Cell Carcinoma-Venous Thrombus Consortium evaluated 1042 patients, of whom 228 (23%) had undergone prior renal artery embolization, and found no difference in perioperative mortality, or 5-year cancer-specific or overall survival [Martinez-Salamanca et al. 2014]. At this time, given the lack of apparent operative or oncologic benefit and the increased risk and discomfort associated with the postinfarction cytokine storm as well as the potential to delay surgery, it is not our institutional practice to perform preoperative renal artery angioembolization.

The paradigm of using neoadjuvant systemic therapy to downstage locally advanced tumors in an effort to facilitate operative excision and improve survival is used in a variety of malignancies. The rapid development and approval of antiangiogenic targeted therapies for metastatic RCC since 2006 [Bukowski et al. 2007; Motzer et al. 2007, 2010, 2013; Hutson et al. 2010; Escudier and Gore, 2011] has garnered interest in the potential utility of these agents to debulk and downstage the tumor thrombus level, which may permit less invasive thrombectomy approaches, potentially avoiding bypass procedures or entry into the chest. For example, in a study by Cost and colleagues, the median decrease in tumor thrombus height was 1.5 cm, and one patient was successfully downgraded from level IV to III. However, the results of this approach are variable, with reports of reduction in thrombus height in 5–44% of cases, stable disease in 28–91%, and tumor thrombus progression in 5–28% [Cost et al. 2011; Bex et al. 2012; Bigot et al. 2014; Peters et al. 2014; Sassa et al. 2014]. As of yet, there is no level one evidence to support the use of targeted agents prior to tumor thrombectomy. Thus, at this time, the use of neoadjuvant-targeted therapy in the preoperative setting should continue to be considered experimental, and as such, should be employed in the context of a clinical trial.

Perioperative monitoring and the role for vascular bypass

Following appropriate preoperative clearance, laboratory and staging studies, the patient is brought to the operative suite. Central venous access and arterial line placement provide necessary real-time monitoring of hemodynamic and respiratory status, while also permitting rapid volume and blood product resuscitation should it become necessary. In patients with level III–IV tumor thrombi in whom either cardiopulmonary or veno-venous bypass is planned, placement of a Swan-Ganz catheter permits simultaneous monitoring of pressures in the right atrium and ventricle, pulmonary artery, and the filling pressure or ‘wedge’ pressure of the left atrium.

For level II–IV tumor thrombi, intraoperative trans-esophageal echocardiography (TEE) may be performed to delineate the level of the thrombus prior to incision, characterization of the nature of the thrombus, assessment for thrombus embolization during IVC manipulation, as well as assessment of cardiac function [Koide et al. 1998; Oikawa et al. 2004; Gonzalez et al. 2013]. In cases where TEE is not employed, sudden hemodynamic instability should prompt emergent TEE to assess for embolization of thrombus. The incidence of intraoperative embolization is approximately 1.5% overall, with increasing risk among higher level thrombi, and is associated with a 75% risk of mortality [Shuch et al. 2009]. Technical points that we employ to minimize the risk of embolization include gentle vascular mobilization permitting early proximal control of the upper extent of the tumor thrombus. Furthermore, once the vascular tourniquets are positioned, they are not removed until the tumor thrombus is completely cleared and the endothelium has been irrigated and inspected for residual thrombus. Additionally, the thrombus is removed en bloc with the nephrectomy specimen. Transection of the vein with separate extraction of the thrombus may cause embolization of thrombus fragments and is to be avoided. In the case of a large intraoperative pulmonary embolism, surgical extraction via median sternotomy by cardiothoracic surgery may be life saving.

Vascular bypass is utilized to facilitate safe and complete resection, and is classically indicated in level III and IV tumor thrombi, bulky intraarterial thrombus, or when the patient is unable to tolerate the reduction in cardiac output secondary to cross clamping of the IVC [Blute et al. 2004; Granberg et al. 2008; Pouliot et al. 2010]. Cardiopulmonary bypass permits maintenance of hemodynamic stability, while optimizing visibility during thrombus extraction. The use of deep hypothermia, in which the patient is cooled to less than 30°C, may be employed concurrently with cardiopulmonary bypass to prevent ischemic sequelae associated with circulatory arrest and temporary interruption of the bypass circuit while the thrombus is extracted [Chowdhury et al. 2007]. Although hypothermic circulatory arrest has been associated with perioperative mortality rates of 3–16% among patients undergoing tumor thrombectomy [Ciancio et al. 2010a], a recent multi-institutional series of 162 patients with level III and IV tumor thrombus found no association between cardiopulmonary bypass or deep hypothermic circulatory arrest and postsurgical complications or mortality following tumor thrombectomy on multivariate analysis [Abel et al. 2013b].

In patients with levels II–III thrombi in which bypass is indicated due to the need for complex vascular resection or reconstruction to permit thrombus extraction, veno-venous bypass may be used instead of cardiopulmonary bypass. After adequate mobilization of the IVC, a 20 Fr venous cannula is placed through a purse-string suture in the IVC inferior to the caudal-most extend of the thrombus. An 8–14 Fr venous cannula is then introduced through the right atrium or right brachial vein to permit venous return, which is then connected to the bypass circuit using heparinized shunt tubing [Boorjian et al. 2007]. The advantages of this technique include avoidance of circulatory arrest, as well as hypothermia and systemic anticoagulation, which may result in both intraoperative and postoperative coagulopathy. When compared with cardiopulmonary bypass with circulatory arrest, veno-venous bypass has been associated with decreased median operative, anesthesia, and bypass times, as well as trends towards decreased intraoperative blood loss, requirement for transfusion, and length of hospitalization [Granberg et al. 2008]. Recently, Patil and colleagues proposed an alternative to bypass to obtain vascular control in a series of 87 patients with level III–IV caval thrombi through obtaining control of the intrapericardial IVC via a right thoracoabdominal incision [Patil et al. 2014]. The authors reported a 30-day mortality rate of 9.2% with a 19.5% incidence of major complications, with increased risk noted among patients with supradiaphragmatic disease and poor performance status.

Surgical approach

The objective of surgical management of RCC with IVC tumor thrombus is complete resection of all tumor burden. In addition to radical nephrectomy and tumor thrombectomy, this may also involve extensive resection of the cava with or without complex vascular reconstruction, retroperitoneal lymphadenectomy, or metastasectomy. In patients with a solitary kidney with RCC and VTT, nephron-sparing surgery has been reported in limited numbers for patients with solitary kidney [Sengupta et al. 2005; Kolla et al. 2010; Woldu et al. 2010; Kim et al. 2012; Abaza and Angell, 2013].

For locally advanced RCC with VTT, multiple surgical approaches are feasible, depending on the surgeon’s experience, preference, patient-specific anatomy, and tumor thrombus extent. Although the open approach is generally favored, minimally invasive approaches have been successfully applied using either hand-assisted laparoscopic or robot-assisted laparoscopic techniques by surgeons with expertise [Varkarakis et al. 2004; Abaza, 2011; Sun et al. 2012; Bansal et al. 2014]. These approaches are most applicable to patients of RCC with renal vein thrombus or level I tumor thrombus in which the thrombus can be milked back into the renal vein with gentle traction, thus permitting proximal renal vein division with a negative vascular margin utilizing either a vascular clip or endovascular stapler. As reviewed by Sun and colleagues, robotic surgery for selected level I and II caval thrombi is also feasible [Sun et al. 2014]. Of paramount importance, however, is the need for surgeons attempting minimally invasive IVC tumor thrombectomy for level II thrombi to be able to obtain complete caval isolation, with circumferential control of the infrarenal and suprarenal IVC, the contralateral renal vein, and lumbar veins. For level III caval thrombi, surgeons may need to employ additional robotic maneuvers, including the control of the short hepatic veins, control of the porta hepatis, and possibly trans-thoracic thoracoscopic control of the suprahepatic IVC [Sun et al. 2014]. It is important to note that the available reports of minimally invasive nephrectomy and IVC tumor thrombectomy come from high-volume centers of excellence with significant expertise in minimally invasive techniques.

The open approach for radical nephrectomy may be performed via a midline, chevron or subcostal incision, each of which provide excellent exposure to bilateral renal hilums and may be utilized for levels I–III tumor thrombi [Agochukwu and Shuch, 2014]. For supradiaphragmatic thrombi, the incision is extended cephalad via median sternotomy, thus providing exposure for institution of cardiopulmonary bypass, and access to the suprahepatic and retrohepatic IVC as necessary. Alternatively, a right-sided thoracoabdominal approach may be utilized in level III and IV cases to provide excellent exposure to the suprahepatic and retrohepatic IVF, while providing intrathoracic access to the heart for initiation of bypass. However, this incision generally necessitates placement of a chest tube and is associated with significant postoperative pain. As such, we favor a midline or anterior subcostal incision if bypass is unlikely. However, if cardiopulmonary bypass is anticipated preoperatively, we favor an extended midline incision with sternotomy.

Following entry into the peritoneum, the affected kidney and the anterior surface of the IVC and aorta are exposed. Exposure is secured using a self-retaining retractor. The next step is early ligation of the renal artery, which reduces blood flow through venous collaterals and potentially limits blood loss later in the case [Ciancio et al. 2003]. Additionally, early ligation of renal circulation may also permit retraction of the VTT. For right-sided tumors, approaching artery in the inter-aortocaval space decreases early manipulation of the IVC and right renal vein [Blute et al. 2004].

Next, the vena caval thrombectomy is performed. The tumor thrombus level guides the approach to tumor thrombectomy. In the case of a level 0 or I tumor thrombus with minimal extension into the IVC, the thrombus may be able to be milked back into the renal vein, then isolated with a vascular clamp which is placed around the RVo in such a manner as to avoid occluding the IVC completely. The ostium of the renal vein is then sharply circumscribed, permitting removal of the tumor thrombus en bloc with the nephrectomy specimen and attached renal vein. The cavotomy is then closed primarily, with a continuous 4–0 polypropylene suture in a running fashion.

Level II tumor thrombi necessitate mobilization of the IVC and the contralateral renal vein to allow proximal and distal vascular control above and below the tumor thrombus. Once the IVC is circumferentially mobilized via ligation and division of the lumbar veins, Rummel tourniquets or vascular clamps are placed sequentially on the suprarenal IVC proximal to the cephalad extent of the thrombus, then on the contralateral renal vein, and lastly on the infrarenal IVC. A test clamp should be performed, as the IVC is cross clamped initially to ensure the patient is able to remain hemodynamically stable during this procedure. In most cases when clamping below the hepatic venous confluence, bypass is not necessary due to collateral venous return via the lumbar system and portal venous system for level II tumor thrombi. Ligation of the accessory hepatic veins from the caudate lobe to the IVC may also be helpful at this point to obtain proximal control beyond the most cephalad extent of the tumor thrombus. Once vascular control is achieved, an ‘L’-shaped cavotomy is performed longitudinally along the isolated IVC and extending over the RVo [Blute et al. 2004]. The thrombus is then dissected from the IVC endothelium and the ostium of the renal vein is circumferentially excised, again permitting the thrombus to be removed in continuity with the kidney. The lumen of the IVC is then inspected for residual tumor thrombus, is flushed, and closed in a running manner, as per level I, ensuring that all thrombus (tumor or bland) and air is aspirated from the IVC lumen prior to completion of the cavorrhaphy. This is performed by placing the operative bed in the Trendelenberg position, then releasing the infrarenal clamp to allow back-bleeding prior to completing the cavorraphy. The contralateral renal vein and suprarenal IVC clamps are then sequentially released.

Level III tumor thrombi represent a relatively heterogeneous entity and necessitate precise characterization of the tumor level on preoperative imaging and intraoperative TEE. Although it is possible to obtain control of the suprahepatic IVC, permitting complete surgical excision of a level III tumor thrombus intraabdominally as described for level II thrombi, mobilization of the liver using liver transplant maneuvers may be necessary [Ciancio et al. 2011; Zhang et al. 2013]. Although urologists may be familiar with such techniques, involvement of a hepatobiliary or transplant surgeon may be helpful for this portion of the procedure. The liver is thus mobilized via dissection and division of the ligamentum teres followed by cautery division of the falciform ligament. This dissection is continued to the right side with division of the right superior coronary ligament and then to the left side, with division of the left triangular ligament. The visceral peritoneum overlying the right hepatic hilum and the infrahepatic IVC are then incised together with the right inferior coronary and hepato-renal ligaments, permitting rolling of the liver to the left side of the peritoneal cavity. The lesser omentum is opened, permitting isolation and control of the porta hepatis with either the surgeon’s hand or a Rummel tourniquet to permit a Pringle maneuver. In doing so, temporary occlusion of the portal venous and arterial inflow of the liver prevents hepatic venous congestion and risk of subsequent hepatic capsule fracture and hemorrhage while the hepatic veins are clamped [Boorjian et al. 2007; Ciancio et al. 2011]. If the tumor thrombus can be milked down below the hepatic venous confluence, however, a tourniquet may be applied to the IVC below the hepatic venous outflow, thereby avoiding liver congestion [Parekh et al. 2005]. With hepatic mobilization, the retrohepatic and suprahepatic IVC is exposed, thus permitting clamping cephalad to the most proximal extent of the tumor thrombus. Not infrequently, however, the combination of IVC cross clamp and the Pringle maneuver result in hemodynamic instability due to insufficient venous return or substantial blood loss from venous collaterals. Thus, as noted for level II thrombi, a test clamp is recommended prior to proceeding with cavotomy. In the case of significant hypotension, vascular bypass via either cardiopulmonary bypass with hypothermic cardiac arrest or veno-venous bypass is indicated.

Level IV or supradiaphragmatic thrombi generally necessitate involvement of cardiothoracic surgery with cardiopulmonary bypass and hypothermic circulatory arrest, however veno-venous bypass may be utilized following mobilization of the liver if the thrombus is free floating and is able to be reduced below the diaphragm. Alternatively, Ciancio and colleagues have described an intraabdominal approach in which the liver is completely mobilized, and then the central diaphragm tendon is dissected, the pericardium is incised, thus exposing the supradiaphragmatic, intrapericardial IVC and permitting the right atrium to be retracted and clamped intra abdominally to permit caval thrombectomy without entering the chest [Ciancio et al. 2010b].

Additional principles of tumor thrombectomy include assessment of the need for more extensive resection of the IVC wall with or without complex reconstruction. Invasion of the IVC wall by tumor thrombus confers additive risk for disease recurrence and poor prognosis [Van Poppel et al. 1997; Lang et al. 2004; Zini et al. 2008; Abel et al. 2013a; Hirono et al. 2013]. Accordingly, in the American Joint Committee on Cancer 7th edition of the Cancer Staging Manual, renal masses with tumor thrombi invading into the caval wall, regardless of level, are classified as pT3c [Edge et al. 2010]. As such, if caval wall invasion by tumor thrombus is suspected, resection of the vascular wall is indicated to completely excise all involved tissue. Frozen sections of the vascular margins may be obtained to ensure complete resection. In a comparison of 47 patients with positive vascular margins and 209 patients with negative vascular margins, Abel and colleagues observed an increased risk of recurrence among patients with positive vascular margins (median recurrence-free survival of 22.1 versus 70.2 months) [Abel et al. 2013a].

Should vascular resection result in narrowing of the IVC lumen by more than 50%, a biological, autologous or synthetic patch graft may be used for reconstruction to restore the IVC diameter [Hyams et al. 2011]. Alternatively, if segmental resection is necessary to obtain negative vascular margins, a tube graft may be employed. Finally, in cases where the IVC is totally occluded by either tumor or bland thrombus, segmental resection of the IVC may be performed and the IVC may be left in discontinuity [Blute et al. 2004, 2007]. One important technical point for cases where the IVC is either ligated or occluded is, if possible, preservation of venous collaterals and lumbar veins not involved by tumor during IVC mobilization to preserve venous return in the setting of the obliterated or resected IVC. For patients who undergo either caval excision or ligation, chronic postoperative anticoagulation with warfarin is generally employed [Boorjian et al. 2007].

Following ligation of the renal artery, excision of the tumor thrombus with vascular reconstruction as indicated and completion of radical nephrectomy, regional retroperitoneal lymphadenectomy may be performed. Although there is no level one evidence assessing the benefit of lymphadenectomy in RCC with VTT, retrospective data from the overall RCC population have demonstrated an association between improved long-term oncologic outcomes and lymphadenectomy at the time of surgery without significantly increasing morbidity or mortality [Whitson et al. 2011; Capitanio et al. 2013a, 2013b, 2014; Barrisford et al. 2014; Bazzi et al. 2014]. In a recent multicenter series of patients treated for level III–IV tumor thrombi, neither regional lymphadenectomy nor extended lymphadenectomy were independently associated with 90-day postoperative mortality or complications [Abel et al. 2014b]. Currently, the approach to lymph node dissection following nephrectomy with tumor thrombectomy is nonstandardized [Abel et al. 2014b]. However, lymph node metastases independently predict inferior prognosis among patients with RCC and VTT [Blute et al. 2004; Ayati et al. 2006; Haddad et al. 2014; Patil et al. 2014]. Therefore, for primary right-sided renal lesions, we proceed with excision of the paraaortic and interaortocaval lymph nodes, and for primary left-sided tumors, the paracaval and interaortocaval lymph nodes are resected. In the case of metastatic disease or evidence of lymph node involvement outside of these boundaries, rendering a patient surgically disease free is associated with improved disease-specific survival [Hirono et al. 2013]. Thus, excision of all resectable disease is attempted so long as the patient remains stable and it is safe to proceed.

Perioperative outcomes

IVC tumor thrombectomy for locally advanced RCC is a technically challenging procedure that is associated with substantial risk for perioperative morbidity and mortality. Early postoperative complications have been reported to occur in 15–78% of patients following this operation [Boorjian et al. 2007; Toren et al. 2013; Martinez-Salamanca et al. 2014]. Toren and colleagues recently reported on perioperative outcomes from a population-based sample of 816 radical nephrectomies with IVC thrombectomy performed between 1998 and 2007 in which the in-hospital mortality rate was 7%, of which 75% of deaths occurred within a surgeon’s first two tumor thrombectomy cases [Toren et al. 2013]. Perioperative morbidity and mortality in this study were associated with age, comorbidity, and the intraoperative use of cardiac bypass and inversely proportional in this study to individual surgeon experience. At high volume centers, 30-day mortality ranges from 1.5% to 10% [Boorjian et al. 2007; Abel et al. 2014b]. On multivariate analysis, poor performance status, hypoalbuminemia, and elevated aspartate aminotransferase were independently associated with increased risk of postoperative mortality [Abel et al. 2014b].

Not surprisingly, complication rates and intraoperative adverse events have also been linked to tumor thrombus level. Intraoperative blood loss, transfusion rates, and length of stay as well as 30-day postoperative complication rates increase with tumor thrombus level [Boorjian et al. 2007; Abel et al. 2014b]. Among patients with level III and IV tumor thrombi, independent predictors of major 30-day complications include the presence of preoperative systemic symptoms, thrombus level, and elevated aspartate aminotransferase (AST) and alkaline phosphatase above the upper limits of normal [Abel et al. 2014b]. The most common late complications (defined as within 1 year after tumor thrombectomy) primarily include renal insufficiency, as defined by creatinine over 2.0 mg/dl, and proteinuria, which occur in nearly two-thirds of patients, independently of tumor thrombus level [Boorjian et al. 2007]. Given the significant risk of morbidity and mortality associated with these procedures, Abel and colleagues have proposed an algorithm to aid in preoperative decision making in which patients with hypoalbuminemia, poor performance status, systemic symptoms, and supradiaphragmatic tumor thrombi are considered poor risk, and as such, especially in the presence of concurrent metastases, are recommended to undergo evaluation by medical oncology for potential systemic therapy and deferral of definitive treatment to permit preoperative optimization [Abel et al. 2014b].

Oncologic outcomes

If left untreated, RCC with VTT is associated with a median survival of 5 months with increased risk of cancer-specific mortality in the setting of pT3b and pT3c disease, as well as the presence of concurrent metastases [Reese et al. 2013]. In a population-based study among patients treated with nephrectomy and venous tumor thrombectomy, patients with metastatic disease at diagnosis had a 60% 1-year overall survival in comparison to 90% among patients without systemic involvement [Whitson et al. 2013]. At 5 years, disease-specific survival for patients following nephrectomy and tumor thrombectomy ranges from 40% to 60% [Blute et al. 2004; Pouliot et al. 2010; Hirono et al. 2013; Haddad et al. 2014]. Cancer-specific survival is associated with the presence of caval (as opposed to renal vein) thrombi, nodal and systemic metastases, advanced Fuhrman grade, non-clear cell histology, and increasing tumor size [Boorjian et al. 2007; Martinez-Salamanca et al. 2011, 2014; Kaushik et al. 2013; Tilki et al. 2013, 2014; Whitson et al. 2013; Haddad et al. 2014].

Among patients with metastatic disease at presentation, 5-year cancer-specific survival is approximately 50% [Tilki et al. 2014]. Furthermore, despite the technical challenges of undertaking concurrent tumor thrombectomy, cytoreductive nephrectomy, and metastasectomy, overall survival is similar when compared with patients undergoing cytoreductive nephrectomy alone [Westesson et al. 2014].

Summary

The diagnosis of RCC is accompanied by intravascular tumor thrombus in up to 10% of cases, of which nearly one-third of patients also have concurrent metastatic disease. Surgical resection in the form of radical nephrectomy and caval thrombectomy represents the only option to obtain local control of the disease and is associated with durable oncologic control in approximately half of these patients. Cornerstones of the management of these complex patients include obtaining high-quality imaging to characterize the renal mass and tumor thrombus preoperatively, with further intraoperative real-time evaluation using transesophageal echocardiography, careful surgical planning, and a multidisciplinary approach. Operative management of patients with high-level caval thrombi should be undertaken in high-volume centers by surgical teams with capacity for bypass and invasive intraoperative monitoring. In patients with metastatic disease at presentation, cytoreductive nephrectomy and tumor thrombectomy may be safely performed with simultaneous metastasectomy if possible. In the absence of level one evidence, neoadjuvant targeted therapy should continue to be viewed as experimental and should be employed under the auspices of a clinical trial. However, in patients with significant risk factors for postoperative complications and mortality, and especially in those with metastatic disease, consultation with medical oncology and frontline targeted therapy may be considered.