Daniel L. Clarke-Pearson

Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach.

Methods of evaluating and comparing the performance of diagnostic tests are of increasing importance as new tests are developed and marketed. When a test is based on an observed variable that lies on a continuous or graded scale, an assessment of the overall value of the test can be made through the use of a receiver operating characteristic (ROC) curve. The curve is constructed by varying the cutpoint used to determine which values of the observed variable will be considered abnormal and then plotting the resulting sensitivities against the corresponding false positive rates. When two or more empirical curves are constructed based on tests performed on the same individuals, statistical analysis on differences between curves must take into account the correlated nature of the data. This paper presents a nonparametric approach to the analysis of areas under correlated ROC curves, by using the theory on generalized U-statistics to generate an estimated covariance matrix.

Cyclophosphamide and Cisplatin Compared with Paclitaxel and Cisplatin in Patients with Stage III and Stage IV Ovarian Cancer

BACKGROUND Chemotherapy combinations that include an alkylating agent and a platinum coordination complex have high response rates in women with advanced ovarian cancer. Such combinations provide long-term control of disease in few patients, however. We compared two combinations, cisplatin and cyclophosphamide and cisplatin and paclitaxel, in women with ovarian cancer. METHODS We randomly assigned 410 women with advanced ovarian cancer and residual masses larger than 1 cm after initial surgery to receive cisplatin (75 mg per square meter of body-surface area) with either cyclophosphamide (750 mg per square meter) or paclitaxel (135 mg per square meter over 24 hours). RESULTS Three hundred eighty-six women met all the eligibility criteria. Known prognostic factors were similar in the two treatment groups. Alopecia, neutropenia, fever, and allergic reactions were reported more frequently in the cisplatin-paclitaxel group. Among 216 women with measurable disease, 73 percent in the cisplatin-paclitaxel group responded to therapy, as compared with 60 percent in the cisplatin-cyclophosphamide group (P = 0.01). The frequency of surgically verified complete response was similar in the two groups. Progression-free survival was significantly longer (P < 0.001) in the cisplatin-paclitaxel group than in the cisplatin-cyclophosphamide group (median, 18 vs. 13 months). Survival was also significantly longer (P < 0.001) in the cisplatin-paclitaxel group (median, 38 vs. 24 months). CONCLUSIONS Incorporating paclitaxel into first-line therapy improves the duration of progression-free survival and of overall survival in women with incompletely resected stage III and stage IV ovarian cancer.

Concurrent Cisplatin-Based Radiotherapy and Chemotherapy for Locally Advanced Cervical Cancer

BACKGROUND AND METHODS On behalf of the Gynecologic Oncology Group, we performed a randomized trial of radiotherapy in combination with three concurrent chemotherapy regimens -- cisplatin alone; cisplatin, fluorouracil, and hydroxyurea; and hydroxyurea alone -- in patients with locally advanced cervical cancer. Women with primary untreated invasive squamous-cell carcinoma, adenosquamous carcinoma, or adenocarcinoma of the cervix of stage IIB, III, or IVA, without involvement of the para-aortic lymph nodes, were enrolled. The patients had to have a leukocyte count of at least 3000 per cubic millimeter, a platelet count of at least 100,000 per cubic millimeter, a serum creatinine level no higher than 2 mg per deciliter (177 micromol per liter), and adequate hepatic function. All patients received external-beam radiotherapy according to a strict protocol. Patients were randomly assigned to receive one of three chemotherapy regimens: 40 mg of cisplatin per square meter of body-surface area per week for six weeks (group 1); 50 mg of cisplatin per square meter on days 1 and 29, followed by 4 g of fluorouracil per square meter given as a 96-hour infusion on days 1 and 29, and 2 g of oral hydroxyurea per square meter twice weekly for six weeks (group 2); or 3 g of oral hydroxyurea per square meter twice weekly for six weeks (group 3). RESULTS The analysis included 526 women. The median duration of follow-up was 35 months. Both groups that received cisplatin had a higher rate of progression-free survival than the group that received hydroxyurea alone (P<0.001 for both comparisons). The relative risks of progression of disease or death were 0.57 (95 percent confidence interval, 0.42 to 0.78) in group 1 and 0.55 (95 percent confidence interval, 0.40 to 0.75) in group 2, as compared with group 3. The overall survival rate was significantly higher in groups 1 and 2 than in group 3, with relative risks of death of 0.61 (95 percent confidence interval, 0.44 to 0.85) and 0.58 (95 percent confidence interval, 0.41 to 0.81), respectively. CONCLUSIONS Regimens of radiotherapy and chemotherapy that contain cisplatin improve the rates of survival and progression-free survival among women with locally advanced cervical cancer.

Phase III Trial of Carboplatin and Paclitaxel Compared With Cisplatin and Paclitaxel in Patients With Optimally Resected Stage III Ovarian Cancer: A Gynecologic Oncology Group Study

PURPOSE In randomized trials the combination of cisplatin and paclitaxel was superior to cisplatin and cyclophosphamide in advanced-stage epithelial ovarian cancer. Although in nonrandomized trials, carboplatin and paclitaxel was a less toxic and highly active combination regimen, there remained concern regarding its efficacy in patients with small-volume, resected, stage III disease. Thus, we conducted a noninferiority trial of cisplatin and paclitaxel versus carboplatin and paclitaxel in this population. PATIENTS AND METHODS Patients with advanced ovarian cancer and no residual mass greater than 1.0 cm after surgery were randomly assigned to receive cisplatin 75 mg/m2 plus a 24-hour infusion of paclitaxel 135 mg/m2 (arm I), or carboplatin area under the curve 7.5 intravenously plus paclitaxel 175 mg/m2 over 3 hours (arm II). RESULTS Seven hundred ninety-two eligible patients were enrolled onto the study. Prognostic factors were similar in the two treatment groups. Gastrointestinal, renal, and metabolic toxicity, as well as grade 4 leukopenia, were significantly more frequent in arm I. Grade 2 or greater thrombocytopenia was more common in arm II. Neurologic toxicity was similar in both regimens. Median progression-free survival and overall survival were 19.4 and 48.7 months, respectively, for arm I compared with 20.7 and 57.4 months, respectively, for arm II. The relative risk (RR) of progression for the carboplatin plus paclitaxel group was 0.88 (95% confidence interval [CI], 0.75 to 1.03) and the RR of death was 0.84 (95% CI, 0.70 to 1.02). CONCLUSION In patients with advanced ovarian cancer, a chemotherapy regimen consisting of carboplatin plus paclitaxel results in less toxicity, is easier to administer, and is not inferior, when compared with cisplatin plus paclitaxel.

Randomized Comparison of Fluorouracil Plus Cisplatin Versus Hydroxyurea as an Adjunct to Radiation Therapy in Stage IIB-IVA Carcinoma of the Cervix With Negative Para-Aortic Lymph Nodes: A Gynecologic Oncology Group and Southwest Oncology Group Study

PURPOSE In 1986, a protocol comparing primary radiation therapy (RT) plus hydroxyurea (HU) to irradiation plus fluorouracil (5-FU) and cisplatin (CF) was activated by the Gynecologic Oncology Group (GOG) for the treatment of patients with locally advanced cervical carcinoma. The goals were to determine the superior chemoradiation regimen and to quantitate the relative toxicities. METHODS All patients had biopsy-proven invasive squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma of the uterine cervix. Patients underwent standard clinical staging studies and their tumors were found to be International Federation of Gynaecology and Obstetrics stages IIB, III, or IVA. Negative cytologic washings and para-aortic lymph nodes were required for entry. Patients were randomized to receive either standard whole pelvic RT with concurrent 5-FU infusion and bolus CF or the same RT plus oral HU. RESULTS Of 388 randomized patients, 368 were eligible; 177 were randomized to CF and 191 to HU. Adverse effects were predominantly hematologic or gastrointestinal in both regimens. Severe or life-threatening leukopenia was more common in the HU group (24%) than in the CF group (4%). The difference in progression-free survival (PFS) was statistically significant in favor of the CF group (P = .033). The sites of progression in the two treatment groups were not substantially different. Survival was significantly better for the patients randomized to CF (P = .018). CONCLUSION This study demonstrates that for patients with locally advanced carcinoma of the cervix, the combination of 5-FU and CF with RT offers patients better PFS and overall survival than HU, and with manageable toxicity.

American Society of Clinical Oncology Guideline: Recommendations for Venous Thromboembolism Prophylaxis and Treatment in Patients With Cancer

PURPOSE To develop guideline recommendations for the use of anticoagulation in the prevention and treatment of venous thromboembolism (VTE) in patients with cancer. METHODS A comprehensive systematic review of the medical literature on the prevention and treatment of VTE in cancer patients was conducted and reviewed by a panel of content and methodology experts. Following discussion of the results, the panel drafted recommendations for the use of anticoagulation in patients with malignant disease. RESULTS The results of randomized controlled trials of primary and secondary VTE medical prophylaxis, surgical prophylaxis, VTE treatment, and the impact of anticoagulation on survival of patients with cancer were reviewed. Recommendations were developed on the prevention of VTE in hospitalized, ambulatory, and surgical cancer patients as well as patients with established VTE, and for use of anticoagulants in cancer patients without VTE to improve survival. CONCLUSION Recommendations of the American Society of Clinical Oncology VTE Guideline Panel include (1) all hospitalized cancer patients should be considered for VTE prophylaxis with anticoagulants in the absence of bleeding or other contraindications; (2) routine prophylaxis of ambulatory cancer patients with anticoagulation is not recommended, with the exception of patients receiving thalidomide or lenalidomide; (3) patients undergoing major surgery for malignant disease should be considered for pharmacologic thromboprophylaxis; (4) low molecular weight heparin represents the preferred agent for both the initial and continuing treatment of cancer patients with established VTE; and (5) the impact of anticoagulants on cancer patient survival requires additional study and cannot be recommended at present.

Evaluation of Monoclonal Humanized Anti-HER2 Antibody, Trastuzumab, in Patients With Recurrent or Refractory Ovarian or Primary Peritoneal Carcinoma With Overexpression of HER2: A Phase II Trial of the Gynecologic Oncology Group

PURPOSE To evaluate the feasibility, toxicity, and efficacy of single-agent monoclonal antibody therapy targeting the human epidermal growth factor receptor 2 (HER2)/neu receptor in ovarian and primary peritoneal carcinoma. PATIENTS AND METHODS Eligible patients had measurable persistent or recurrent epithelial ovarian or primary peritoneal carcinoma with 2+ or 3+ HER2 overexpression documented by immunohistochemistry. Intravenous trastuzumab was administered initially at a dose of 4 mg/kg, then weekly at 2 mg/kg. Patients without progressive disease or excessive toxicity could continue treatment indefinitely. Those with stable or responding disease at 8 weeks were offered treatment at a higher weekly dose (4 mg/kg) at time of progression. Patient sera were analyzed for the presence of the soluble extracellular domain of HER2, host antibodies against trastuzumab, and trastuzumab pharmacokinetics. RESULTS A total of 837 tumor samples were screened for HER2 expression, and 95 patients (11.4%) exhibited the requisite 2+/3+ expression level. Forty-five patients, all of whom received prior chemotherapy, were entered, and 41 were deemed eligible and assessable. There were only mild expected toxicities and no treatment-related deaths. Although an elevated level of the soluble extracellular domain of HER2 was detected in eight of 24 patients, serum HER2 was not associated with clinical outcome. There was no evidence of host antitrastuzumab antibody formation. Serum concentrations of trastuzumab gradually increased with continued therapy. An overall response rate of 7.3% included one complete and two partial responses. Median treatment duration was 8 weeks (range, 2 to 104 weeks), and median progression-free interval was 2.0 months. CONCLUSION The clinical value of single-agent trastuzumab in recurrent ovarian cancer is limited by the low frequency of HER2 overexpression and low rate of objective response among patients with HER2 overexpression.

Retrospective Analysis of Selective Lymphadenectomy in Apparent Early-Stage Endometrial Cancer

PURPOSE Selective lymphadenectomy is widely accepted in the management of endometrial cancer. Purported benefits are individualization of adjuvant therapy based on extent of disease and resection of occult metastases. Our goal was to assess effects of the extent of selective lymphadenectomy on outcomes in women with apparent stage I endometrial cancer at laparotomy. PATIENTS AND METHODS Patients with endometrial cancer who received primary surgical treatment between 1973 and 2002 were identified through an institutional tumor registry. Inclusion criteria were clinical stage I/IIA disease and procedure including hysterectomy and selective lymphadenectomy (pelvic or pelvic + aortic). Exclusion criteria included presurgical radiation, grossly positive lymph nodes, or extrauterine metastases at laparotomy. Recurrence and survival were analyzed using Kaplan-Meier analysis and Cox proportional hazards model. RESULTS Among 509 patients, the median number of lymph nodes removed was 15 (median pelvic, 11; median aortic, three). Pelvic and aortic node metastases were found in 24 (5%) of 509 patients and 11 (3%) of 373 patients, respectively. Patients with poorly differentiated cancers having more than 11 pelvic nodes removed had improved overall survival (hazard ratio [HR], 0.25; P < .0001) and progression-free survival (HR, 0.26; P < .0001) compared with patients having poorly differentiated cancers with 11 or fewer nodes removed. Number of nodes removed was not predictive of survival among patients with cancers of grade 1 to 2. Performance of aortic selective lymphadenectomy was not associated with survival. Three (27%) of 11 patients with microscopic aortic nodal metastasis are alive without recurrence. CONCLUSION These data add to the literature documenting the possible therapeutic benefit of selective lymphadenectomy in management of patients with apparent early-stage endometrial cancer.

Screening for Ovarian Cancer

A 56-year-old woman presents to her physician, requesting screening for ovarian cancer. She reports the recent death of a friend from ovarian cancer at the age of 65 years. The patient has no family history of ovarian or breast cancer. The physical examination, including pelvic and rectal examination, is normal. Should the physician recommend screening for ovarian cancer?

Overexpression of HER-2/neu in endometrial cancer is associated with advanced stage disease.

Prior studies have shown that overexpression of HER-2/ neu occurs in one third of breast and ovarian cancers and that overexpression is associated with poor prognosis. We used a monoclonal antibody to assess immunohistochemically the level of HER-2/ neu expression in normal and malignant endometrium. In 24 normal endometrial samples light to moderate (1+ to 2+) staining for HER-2/ neu was seen in the glands, and there was no variation in intensity of staining during the menstrual cycle. Among 95 endometrial adenocarcinomas, nine (9%) were found to have heavier staining for HER-2/ neu than was seen in normal endometrium (3+). High expression of HER-2/ neu was found in 27% of patients with metastatic disease compared with 4% of patients with disease confined to the uterus (p neu expression also was associated with absence of estrogen receptor (p neu overexpression in endometrial cancer.