Multi-disciplinary surgical approach to the management of patients with renal cell carcinoma with venous tumor thrombus: 15 year experience and lessons learned

Patient selection

Clinical data from electronic medical records of patients treated by radical nephrectomy (RN) for RCC with renal vein or IVC thrombus at our institution from January 1998 to June 2012 were retrospectively analyzed and placed into a UT Southwestern Medical Center IRB approved database. Research was carried out in compliance with the Helsinki Declaration. We did not obtain informed consent from patients, as this was a retrospective study. Relevant clinical data, pathological features, surgical techniques, hospital stay, perioperative morbidity and mortality, follow up and survival data were collected.

Preoperative evaluation and surgical techniques

The tumors were routinely staged by using abdominal and chest CT scans and chest radiography. MRI was used for better evaluation of VTT level at the discretion of the treating physician. Bone scans were used selectively when clinically indicated. Tumor thrombus extension was classified into 4 levels: level I, extension into the renal vein; level II, extension into the infrahepatic IVC; level III, IVC extension to the level of hepatic veins but below the diaphragm; and level IV, IVC extension above the diaphragm [6].

Preoperative renal artery angioembolization was performed at the discretion of the surgeon to facilitate arterial vascular control in patients with bulky tumor thrombus, hilar adenopathy, or hypervascularity. Surgeries were managed by a multidisciplinary team which included an experienced urologic oncology surgeon, cardiothoracic surgeon and cardiac anesthesiologist, and cardiac scrub team. Trans-esophageal echocardiography was used intraoperatively by the anesthesiologist to verify the cephalad extent of the thrombus, to monitor for tumor emboli, to confirm the complete removal of VTT, and to assess hemodynamic stability. Data regarding the duration of surgery, estimated blood loss and intraoperative complications were recorded.

Pathologic evaluation

Pathologic staging was assigned according to the 2010 TNM staging system [7]. Grading of the tumors was evaluated according to Fuhrman classification [8]. Additionally, pathological tumor size, adrenal involvement, regional lymph node (LN) involvement, tumor necrosis, histopathological cell type and the presence of sarcomatoid differentiation were recorded.

Outcome evaluation and statistical analysis

Perioperative morbidity and mortality within the first 30 and 90 days were recorded and graded according to the Clavien-Dindo grading system [9]. Patients without metastases were routinely followed after surgery every 3 months in the first year, every 6 months in the second year and then annually. Follow up included history, physical examination, metabolic panel, liver function tests, chest x-ray and an abdominal CT scan. Bone scan, chest CT, positron emission tomography or MRI were performed when clinically indicated.

Survival time was calculated from the date of the operation to the date of last follow up or date of death. Disease recurrence was defined as local failure in the RN bed or regional LNs, or distant metastasis. Disease-free survival (DFS) was defined as the time between the date of surgery and the development of local recurrence or distant metastasis. Censored survival values represent patients who were alive without clinical evidence of disease at the last follow up. Cancer-specific survival (CSS) and overall survival (OS) were defined as the time between the date of surgery and death due to cancer (CSS) or due to any cause (OS). The following factors that could potentially affect outcomes were analyzed: age, gender, body mass index and performance status; T stage, VTT level, pathological tumor size, nodal involvement, metastasis at presentation, grade, sarcomatoid differentiation, histological subtype, fat invasion, adrenal involvement and tumor necrosis. Endpoints were CSS and OS. DFS was analyzed only in M0 patients. Finally, independent predictors of disease recurrence and cancer specific mortality were determined using multivariate Cox Regression analyses including only factors significant in univariate analyses. Statistically significant difference was set at p < .05. All statistical tests were performed with SPSS version 19.0.

Clinico-pathological features

Surgical intervention

Surgical parameters and postoperative hospital stay data are included in the Additional file 1: Table S1. Preoperative renal artery angioembolization was performed in 27.4 % of all patients with any venous thrombus, and in 49.3 % of patients with level 2–4 thrombi. Chevron incision was the most common approach and was performed in all patients with level III and IV VTT (with midline strenotomy in cases where cardio-pulmonary bypass was needed). IVC clamping was used in most cases of level II and III VTT. Suprahepatic control of the IVC and control of porta hepatis were gained in level III VTT. More aggressive cardiothoracic procedures were reserved for patients with level IV and 3 patients with level III VTT who were hemodynamically unstable during the initial cross-clamping of the IVC. Right heart venovenous bypass was used to assist in removal of VTT in these 3 patients. Cardio-pulmonary bypass was needed in 5 patients with level IV VTT with mean bypass, aortic cross clamping and circulatory arrest times of 124, 59 and 25 min respectively.

Mean estimated blood loss was 1.5 L and blood loss was greatest in patients with level IV VTT. The mean operative time was around 5 h and it was correlated to level of thrombus (5.5, 6 and 6.5 h in level II, III and IV; respectively). Mean hospital stay was 8.8 days (range, 1–63) and mean ICU stay was 3 days (range, 0–51). Three patients had a complicated postoperative course and required longer care.

Peri-operative morbidity and mortality

Oncological outcomes

Patients were followed up after RN for a median of 16 months (mean 26, range 0–163 months). At the time of the analysis, overall mortality was 44 % with a median survival of 47 ± 4 months (range 38–56 months) and 34 % cancer specific mortality with a median CSS of 62 ± 17 months (range 29–94 months). Kaplan-Meier Survival analysis showed CSS at 2,3 and 5 years to be 70 %, 62 % and 51 %; OS at 2, 3 and 5 years to be 64 %, 57 % and 40 %; respectively (Fig. 1a). There was no significant difference in survival comparing the last 6 years to an earlier period with 3Y- CSS 62 % in both eras.

Prognostic factors

Kaplan-Meier survival analysis (Fig. 1b) demonstrated a significant difference between CSS rates in M0 and M+ patients (5Y-CSS was 68 % and 17 % in M0 and M+ patients; respectively, p < 0.001). Median CSS was 11 ± 4 months for M+ patients, while it was not yet reached for M0 patients. Multivariate Cox regression analyses (data not shown) demonstrated that M+ was the only independent predictor factor for CSS (HR 3.8, CI 1.9-7.6 and p < .001) and OS (HR 2.6, CI 1.5-4.7 and p = .001). LN+ was associated with a trend toward a poor CSS (HR 1.9, CI .97-3.6 and p = .06). In M+ patients, LN+ was the only factor significantly associated with poor oncological outcomes as shown from CSS analysis (HR 2.3, CI 1–5 and P = .03), and in OS analysis (HR 2, CI .96 – 4.3 and P = .06). In M0 patients, high VTT level (III and IV compared to I and II) was among the independent predictors of disease recurrence (HR 4.4, CI 2.1-9.4 and P < .001) and cancer specific mortality (HR 6.5, CI 2–21.2; p = .002) in multivariate Cox regression analysis. Other independent predictors of poor oncological outcomes included larger tumor size (>13 cm) and sarcomatoid differentiation (data not shown).