A prognostic nomogram to predict overall survival in women with recurrent ovarian cancer treated with bevacizumab and chemotherapy.

OBJECTIVE: To develop a nomogram to predict overall survival (OS) in women with recurrent ovarian cancer treated with bevacizumab and chemotherapy. METHODS: A multicenter retrospective study was conducted. Potential prognostic variables included age; stage; grade; histology; performance status; residual disease; presence of ascites and/or pleural effusions; number of chemotherapy regimens, treatment-free interval (TFI) prior to bevacizumab administration, and platinum sensitivity. Multivariate analysis was performed using Cox proportional hazards regression. The predictive model was developed into a nomogram to predict five-year OS. RESULTS: 312 women with recurrent ovarian cancer treated with bevacizumab and chemotherapy were identified; median age was 59 (range: 19-85); 86% women had advanced stage (III-IV) disease. The majority had serous histology (74%), high grade cancers (93.5%), and optimal cytoreductions (69.5%). Fifty-one percent of women received greater than two prior chemotherapeutic regimens. TFI (AHR=0.98, 95% CI 0.97-1.00, p=0.022) was the only statistically significant predictor in a multivariate progression-free survival (PFS) analysis. In a multivariate OS analysis, prior number of chemotherapy regimens, TFI, platinum sensitivity, and presence of ascites were significant. A nomogram to predict five-year OS was constructed and internally validated (bootstrap-corrected concordance index=0.737). CONCLUSION: Our multivariate model identified prior number of chemotherapy regimens, TFI, platinum sensitivity, and the presence of ascites as prognostic variables for OS in women with recurrent ovarian cancer treated with bevacizumab combined with chemotherapy. Our nomogram to predict five-year OS may be used to identify women who may benefit from bevacizumab and chemotherapy, but further validation is needed.

Outcomes of obese versus non-obese subjects undergoing robotic-assisted hysterectomy: a multi-institutional study.

The goal of our study was to determine whether there was a difference in operative outcomes in obese versus non-obese subjects undergoing robotic-assisted hysterectomies of varying levels of difficulty. Secondarily, we sought to analyze the published outcomes between robotic-assisted hysterectomy and total laparoscopic hysterectomy in obese women at each of these levels of difficulty. This was a multi-institutional retrospective cohort study of all patients undergoing robotic-assisted hysterectomy by five gynecologic oncologists at four geographically separate locations from April 2003 to March 2008. The cohort was stratified into obese vs. non-obese groups, and defined surgical outcomes compared between groups, then further divided into three subgroups based on case difficulty level. Univariate analysis and regression analysis using SAS 9.1 was performed. We then conducted a literature search of total laparoscopic hysterectomy outcomes in obese women, dividing the resulting studies into three comparative subgroups based on surgical difficulty levels for comparison with our robotic-assisted hysterectomy results. Our cohort had 228 obese and 323 non-obese subjects. Overall, the obese group had higher blood loss and longer operative time. When further stratified by level of difficulty, obese subjects also had a higher average blood loss and longer operative time in the hysterectomy-alone subgroup. No clinically significant differences in operative outcomes exist between obese and non-obese women when utilizing the da Vinci robotic system to perform a hysterectomy, independent of case difficulty level. More prospective, controlled studies which compare the two surgical approaches of robotic-assisted and laparoscopic hysterectomy approaches are needed.

Panniculectomy with simultaneous gynecologic oncology surgery.

OBJECTIVE: The objective was to report the utility and morbidity of panniculectomy in obese gynecologic oncology patients undergoing exploratory laparotomy. METHODS: A retrospective chart review of 41 consecutive women who had a panniculectomy as part of an abdominal gynecologic oncology procedure between July 1996 and May 2000 was performed. Obese patients possessing a large pannus, the majority with a BMI > or = 30 kg/m(2), were included. Demographic, preoperative, operative, and postoperative data were obtained. Statistical analyses were performed using Statistical Analysis System (SAS) Version 6.13. RESULTS: Panniculectomy was performed on 41 patients with a mean age of 55, weight of 126 kg, and BMI of 48 kg/m(2). The most common comorbidities in this population were hypertension, diabetes, and osteoarthritis. Wound infection occurred in 4 (9.8%) patients; 88% of the patients received a hysterectomy. The average EBL was 358 cc. Operative time and length of hospital stay were on average 203 min and 5.5 days, respectively. A prior history of diabetes increased the risk of early complications (P = 0.03). Late complications were more likely to occur in older women (P = 0.05). Wound complications were increased in patients with larger BMI's (P = 0.05). CONCLUSIONS: This study supports the safety of the panniculectomy procedure in this high-risk group of morbidly obese patients for whom a technical advantage may be achieved by improved operative exposure.

Ovarian cancer metastatic to the brain: what is the optimal management?

PURPOSE: To better define determinants of survival and optimal management strategies for patients with ovarian cancer and brain metastases. METHODS: A review of literature using Medline identified 15 case series of ovarian cancer patients with brain metastases (OBM). Each article was abstracted for survival data, and in all cases, the intervals between ovarian cancer diagnosis and brain metastasis identification, and between brain metastasis identification and last follow-up were recorded. Cases were categorized by patient characteristics and treatment modality for brain metastases. Estimated survival probabilities were plotted using the Kaplan-Meier method with differences between subgroups analyzed by the log-rank test. Cox proportional hazards model was used to identify independent prognostic factors age, number of metastasis, and treatment modality associated with survival. RESULTS: The median interval from ovarian cancer diagnosis to brain metastasis in 104 identified patients was 19.5 months. Brain metastasis was single in 43%, multiple in 41%, and not reported in 16% of cases. About 81.7% of patients were treated for their brain metastases using external radiation therapy (XRT), chemotherapy, and surgery. XRT was utilized in 76% of 104 patients and in 93% of treated patients. The most commonly used modalities were XRT alone (40%) and craniotomy and XRT (17%). The median survival (MS) for all patients regardless of treatment type was 6 months. Patients who received any treatment lived longer than those not receiving surgery/chemotherapy/XRT (MS; 7 months vs. 2 months, P = 0.0001). Patients with single brain metastasis had a longer median survival (21 months vs. 6 months, P = 0.049) when treated with craniotomy plus radiation and/or chemotherapy compared to treatment regimens that excluded craniotomy. In a multivariate analysis, only treatment type was significant in predicting survival. CONCLUSION: OBM portends a poor prognosis, however, long-term survival is possible. Patients appear to benefit from therapy, especially selected groups of OBM patients with single brain metastasis treated with radiation therapy and surgery.

Predictors of complications after inguinal lymphadenectomy.

OBJECTIVE: The aim of this study was to identify predictors of complications in patients undergoing inguinal lymphadenectomy (LND) in the treatment of vulvar carcinoma. METHODS: Clinical information was abstracted from records of patients with invasive vulvar cancer. All patients underwent LND. Closed suction drains were placed in groin incisions and removed when output was less than 30 ml/24 h. Associations between variables were studied by chi(2) and t tests. RESULTS: Sixty-seven patients undergoing 112 LND were evaluated. Eighty-eight percent of patients underwent radical vulvar surgery and LND while 12% underwent LND alone. Patients were treated with either unilateral (22) or bilateral (45) LND. Seventy-three percent received extended postoperative prophylactic antibiotics. The median duration of suction drainage was 15 days for one drain and 14 days for two drains. Early postoperative cellulitis (<30 days after surgery) developed in 35.4%, early wound breakdown in 19.4%, early lymphedema in 4.8%, and early lymphocyst formation in 13.1%. Late cellulitis (>30 days after surgery) developed in 22.2%, late wound breakdown in 3.2%, late lymphedema in 29.5%, and late lymphocysts in 5%. Patients developing early cellulitis were more likely to have early wound breakdown (P = <0.001, RR 14.2) or early lymphocyst formation (P = 0.016, RR 7.6). Type of procedure, antibiotic use, need for adjuvant therapy, and duration of suction drainage did not influence early complications. Early complications and management strategies did not predict late complications. CONCLUSIONS: Chronic lymphedema occurs in nearly 30% of patients after LND. Late complications after LND were not predicted by early complications. New strategies for prevention of chronic lymphedema are needed.

Nodal distribution and its significance in FIGO stage IIIc endometrial cancer.

OBJECTIVE: The aim of this study was to describe the distribution of nodal disease in FIGO Stage IIIc endometrial cancer (EC) and to evaluate whether nodal distribution is related to recurrence and survival. METHODS: Charts from EC patients with FIGO Stage IIIc disease from 1989 to 1998 were abstracted for clinicopathologic data, pelvic (PLN) and para-aortic (PALN) nodal involvement, number of positive/removed nodes, and extranodal disease spread. Patterns of nodal distribution were evaluated for site of first recurrence and survival. Associations between variables were tested by chi(2) and Wilcoxon rank sums. Survival analyses were performed by the Kaplan-Meier method. RESULTS: Of 607 EC patients evaluated, 47 were identified with FIGO Stage IIIc disease. All 47 patients underwent hysterectomy and PLN sampling, and 42/47 had PALN sampling. The median number of PLN removed was 16 (range 2-35), and the median number of PALN was 7 (0-18). Stage IIIc disease was defined by positive PLN alone in 43%, positive PLN and PALN in 40%, and positive PALN alone in 17%. Positive peritoneal cytology and/or adnexal metastasis were present in 12 patients. Only 1/12 of these patients had isolated positive PLN whereas 11/12 had positive PALN (P = 0.007). An increasing number of positive PLN was associated with PALN metastasis (P = 0.0001), and of the 10 patients with bilateral PLN involvement, 9/10 also had positive PALN (P = 0.001). Sites of first recurrence were similar regardless of whether PALN were positive. At a median follow-up of 37 months, the 3-year survival estimate was 70% for patients with positive PALN versus 87% for those with isolated PLN disease (P = 0.22). For all patients neither the total number of positive PLN nor the total number of PLN or PALN removed was associated with survival. CONCLUSIONS: PALN involvement is common in patients with FIGO Stage IIIc endometrial cancer, suggesting that PLN sampling alone may result in underdiagnosis of disease. Patients with positive PALN had more extensive disease, but survival and patterns of failure were not significantly different from those with disease confined to PLN, suggesting that lymph node dissection may have a therapeutic role.

A retrospective analysis of radical hysterectomies done for cervical cancer: is there a role for the Pfannenstiel incision?

OBJECTIVE: The goal of this work was to review patients with early-stage cervical cancer undergoing radical hysterectomy, comparing Pfannenstiel and vertical midline incisions for surgical feasibility, complications, and length of stay. METHODS: Patients were identified by searching our institutional database. Data were collected from review of each patient's medical record, including demographics, cancer stage, histology, procedural information, length of stay, and complications. Associations between variables were studied using chi(2) and two-tailed t tests. Multivariate analysis was performed using logistic regression. RESULTS: Between March 1996 and June 2000, 113 patients from the University and Presbyterian Hospitals, Oklahoma City, Oklahoma, underwent radical hysterectomy and pelvic and paraortic lymph node dissection with records available for review. Group 1 consisted of 40 patients who had vertical incisions and group 2 consisted of 73 patients who had Pfannenstiel incisions. There was no difference in race, number of previous abdominal surgeries, distribution of stage, histology, percentage of type III hysterectomies, estimated blood loss, nodal counts, pathologic margin positivity, and postoperative complications among the two groups. Group 2 were younger (41.6 vs 46.5, P = 0.02) and had a lower average QI than group 1 (24.9 vs 28.9, P = 0.001). Group 2 also had a shorter average hospital stay (4.6 days vs 5.8 days, P = 0.04) and shorter operative time (215 min vs 273 min, P = 0.09). Multivariate analysis resulted in Pfannenstiel incisions (P = 0.002), younger age (P = 0.004), and smaller body mass index (P = 0.01) being significant predictors of length of stay. CONCLUSIONS: Pfannenstiel incisions are feasible without increased morbidity and equal nodal retrieval as compared with vertical midline incisions in patients with early-stage cervical cancer. Pfannenstiel incisions may offer an advantage besides cosmesis in the form of shorter operating room time and earlier discharge from the hospital.