THE JOURNAL OF UROLOGYâ
e500
patients (pts) is needed. A systematic evaluation of biomarker expression has been commenced at our tertiary cancer center and here we report on the early results. METHODS: Samples from primary tumor and/or metastases were evaluated for expression of a panel of biomarkers (BMKs) by immunohistochemistry (IHC) including: ERCC1, EGFR, HER2/neu, VEGFR, PDGFR, p53, p63, cKIT, PTEN. Two cohorts were selected: pts with locally advanced (T2-4N+M0) UC receiving peri-operative cisplatin-based chemotherapy (CT) (cohort 1) and metastatic pts receiving first-line platinum-based CT (cohort 2). IHC results were assessed according to standard protocols and dichotomized as positive (1+) or negative for all markers. Tumor was deparaffinized and specific antigen retrieval determined for individual antibodies. Fisher exact test was used to evaluate the association with response for pts with measurable disease. Cox regression model analyzed staining results with PFS and OS in uni/multivariable analysis (UVA/MVA), adjusted for prognostic variables (lymph-node status [cohort 1], Bajorin score [cohort 2]). RESULTS: From 03/2000 to 03/2013, 86 cases were retrieved (N¼30 in cohort 1 and N¼56 in cohort 2). Rates of staining positivity were: 37/63 (59%) ERCC1, 34/50 (68%) EGFR, 41/53 (77%) HER2/ neu, 45/62 (72%) VEGFR, 11/56 (18%) PDGFR, 26/48 (54%) p53, 41/ 48 (85%) p63, 9/47 (19%) cKIT, and 11/38 (29%) PTEN. BMKs were uniformly distributed (p always >0.05) and no association between staining and response was found in assessable pts. Median follow-up was 29.5 mos (IQR: 12-51). There were no significances in outcomes in cohort 2, while in cohort 1 PDGFR (adjusted HR: 30.41, 95% CI, 2.8>100) and p63 (adjusted HR: 0.12, 95% CI, 0.02-0.87) were associated with PFS while only p63 retained significance for OS in UVA (p¼0.005, HR hardly estimable due to small numbers). BMKs were independent each other and of clinical variables. CONCLUSIONS: A significant proportion of UC pts harbor potentially druggable targets, although it is unclear if targeting them translates to improved outcomes. New signals were obtained in relation to prognosis of UC, partly discordant with available literature. A greater sample size and a validation cohort will be required to confirm the prognostic significance of PDGFR and p63 in patients undergoing perioperative treatment. Source of Funding: None
MP50-17 LATE RECURRENCE FOLLOWING RADICAL CYSTECTOMY: PATTERNS, RISK FACTORS, AND OUTCOMES Brian Linder*, Stephen Boorjian, Tvrtko Hudolin, John Cheville, Prabin Thapa, Robert Tarrell, Igor Frank, Rochester, MN INTRODUCTION AND OBJECTIVES: Radical cystectomy (RC) with pelvic lymph node dissection remains the gold standard treatment for muscle-invasive and high risk non-muscle invasive tumors. Nevertheless, up to 50% of patients will experience disease recurrence after RC. The majority of recurrences occur in the first 2-3 years following surgery, and the outcomes of such patients have been well characterized. However, limited data exist regarding late (i.e. 5 years postoperatively) relapse of disease after RC. Thus, we sought to evaluate the patterns, risk factors and outcome of patients with late recurrence (LR) of urothelial carcinoma following radical cystectomy (RC). METHODS: We identified 2091 patients who underwent RC at Mayo Clinic between 1980-2006. Survival was estimated using Kaplan-Meier method and compared based on timing (< versus 5 years) and location (urothelial versus non-urothelial) of recurrence with the log-rank test. Cox proportional hazard regression models were used to evaluate variables associated with LR and death from bladder cancer. RESULTS: Late recurrence was identified in 82 (3.9%) patients. Median postoperative follow-up was 16.6 years. On multivariate analysis, younger age (p¼0.0008), non-muscle invasive disease (p¼0.01), and prostatic urethral involvement (p<0.0001) were
Vol. 191, No. 4S, Supplement, Monday, May 19, 2014
significantly associated with an increased risk of LR. Five year post recurrence cancer-specific survival (CSS) was significantly worse following recurrence within 5 years from RC (17%) versus after LR (37%; p¼0.001). Meanwhile, patients with non-urothelial LR demonstrated adverse 5-year CSS compared to urothelial LR (19% versus 67%; p<0.0001). On multivariate analysis, younger patient age (HR 1.01; p¼0.003), the presence of muscle invasive disease (HR 1.31; p<0.0001) and non-urothelial site of recurrence (HR 2.76; p<0.0001), but not the time to recurrence (p¼0.38) were associated with a significantly increased risk of death from bladder cancer following recurrence after RC. CONCLUSIONS: Late recurrence is uncommon following RC. Younger patient age, non-muscle invasive disease, and prostatic urethral involvement were associated with a significantly increased risk of LR. Interestingly, time to recurrence was not associated with patient’s subsequent risk of death. Multivariate analysis of factors associated with death from bladder cancer following recurrence after radical cystectomy HR
95% CI
p-value
Age at surgery
Variable
1.01
1.00, 1.02
0.003
Female gender
0.91
0.81, 1.21
0.91
ECOG perf status
1.36
1.18, 1.57
< 0.0001
Symptoms at recurrence
1.05
0.88, 1.26
0.59
Pathologic tumor stage (pT2-4 vs pT1)
1.31
1.19, 1.46
< 0.0001
pN+
1.03
0.85, 1.24
0.79
Late recurrence (vs early recurrence)
0.87
0.63, 1.19
0.38
Nonurothelial site of recurrence (vs urothelial)
2.76
2.14, 3.56
<0.0001
Source of Funding: None
MP50-18 COMPLICATIONS OF RADICAL CYSTECTOMY IN THE NEOADJUVANT CHEMOTHERAPY ERA Patrick N. Espiritu*, Gautum Agarwal, Julio M. Pow-Sang, Philippe E. Spiess, Wade J. Sexton, Michael A. Poch, Tampa, FL INTRODUCTION AND OBJECTIVES: Radical cystectomy for urothelial carcinoma of the bladder (UC) is associated with significant perioperative morbidity. The purpose of this study was to describe early (within 30 days) complication rates of RC in the neoadjuvant chemotherapy era. METHODS: Using an IRB approved database a retrospective review of patients who received neoadjuvant chemotherapy followed by radical cystectomy between 2001-2013 was evaluated to determine incidence of early complications. Using Martin criteria all complications within 30 days of surgery were identified and graded by the ClavienDindo system. Chi-square and logistic regression analyses were performed to analyze relationships between preoperative covariates including age, gender, status, grade and stage of tumor, smoking status, Charlson Comorbidity Index (CCI), BMI, American Society of Anesthesiologists Score, previous surgery, prior pelvic radiation, chemotherapy regimen, intraoperative variables and incidence of early postoperative complications. RESULTS: 169 patients were included in the study with median age 67 years (IQR 59-74) and 45 patients (27%) were female. Ileal conduit diversion was performed in 120 patients (71%). Mean estimated blood loss was 900 cc (SD 579). Median hospital stay was 7 days (IQR 6-9). Ninety-two patients (54%) experienced at least one complication. Twenty-eight patients (16%) experienced a major complication (Clavien grade IIIa) and 4 patients died (2%). The most common complication was ileus occurring in 22 patients (13%) and anemia requiring transfusion in 13 patients (7%). On multivariable analysis, intraoperative transfusion of > 4 units of packed red blood cells (p < 0.04, CI 1.06, 14.17) was associated with developing any complication. Carboplatin based chemotherapy regimen (p <0.02) and operative time (p <0.02) were associated with developing a major complication on multivariable analysis.