Faculty Bibliography

BACKGROUND/PURPOSE: Vascular endothelial growth factor (VEGF) has been shown previously to correlate with tumor growth and metastasis in an experimental model of anaplastic Wilms' tumor. The authors hypothesized that treatment with anti-VEGF antibodies would suppress both primary tumor growth and metastasis in this model. METHODS: Tumors were induced in the right kidneys of nude mice by the injection of cultured Wilms' tumor cells. After 1 week, anti-VEGF treatment was begun with injection of either vehicle or an anti-VEGF antibody intraperitoneally. Mice were killed after 4.5 weeks of treatment and tumor weights and the incidence of metastases evaluated. RESULTS: Anti-VEGF treatment resulted in a greater than 95% reduction in tumor weight (P < .0001). Anti-VEGF treatment also abolished the establishment of lung metastases (40% in control animals, P < .003). Cessation of treatment resulted in rebound tumor growth. CONCLUSION: Anti-VEGF therapy can suppress both primary tumor growth and the establishment of metastases in experimental anaplastic Wilms' tumor.

This phase I dose-escalation study was performed to determine the tolerability of three-drug combination high-dose BCNU (B) (450 mg/m2), escalating-dose thiotepa (500-800 mg/m2) and etoposide (1200 mg/m2) in divided doses over four days in 22 adults with malignant primary brain tumors. Patients received G-CSF and hematopoeitic support with peripheral blood progenitor cells (PBPC) (n = 18) or both PBPC and marrow (n = 4). The maximum tolerated dose of thiotepa with acceptable toxicity was determined as 800 mg/m2. The 100-day mortality rate was 9% (2/22). Grade III/IV toxicities included mucositis (71%), diarrhea (29%), nausea/vomiting (19%), and hepatic toxicity (14%). Neurological toxicities occurred in 24% and included seizures (two patients) and encephalopathy (three patients). Encephalopathy was transient in two patients and progressive in one patient. All patients had neutropenic fever. Median time to engraftment with absolute neutrophil count (ANC) >0.5 x 10(9)/l was 10 days (range 8-30 days). Platelet engraftment >20 x 10(9)/l occurred after 11 days (range 9-65 days). In the eighteen patients supported solely with PBPC, there was a significant inverse correlation between CD34+ dose and days to ANC (rho = -0.78, p = 0.001) and platelet engraftment (rho = -0.76, p = 0.002). Overall, 11% of evaluable patients (2/18) had a complete response to BTE. Median time to tumor progression (TTP) was 9 months, with an overall median survival of 17 months. BCNU (450 mg/m2), thiotepa (800 mg/m2) and etoposide (1200 mg/m2) in divided doses over four days is a tolerable combination HDC regimen, the efficacy of which warrants further investigation in adults with optimally resected chemoresponsive brain tumors.

Sequential high-dose chemotherapy may increase the threshold dose of CD34+ cells necessary for rapid and successful hematologic recovery. There are limited data regarding the pharmacodynamics and threshold CD34+ cell dose required for engraftment following high-dose paclitaxel. To determine the dose of CD34+ PBPC sufficient for rapid engraftment, 65 women with metastatic breast cancer undergoing a sequential high-dose paclitaxel, melphalan, and cyclophosphamide, thiotepa, and carboplatin (CTCb) chemotherapy regimen were evaluated. The intertreatment interval was a median of 27 days. Paclitaxel was escalated from 400 to 825 mg/m2, infused continuously (CI) over 24 h on day -4 with PBPC reinfusion on day 0. Following marrow recovery, 90 mg/m2/day of melphalan was given over 30 min on days -2 and -1, with PBPC reinfusion on day 0. On recovery, patients received CTCb on days -7 to -3, with PBPC reinfusion on day 0. G-CSF was administered after each cycle until WBCC recovery. For paclitaxel, an ANC >0.5 x 10(9)/L occurred at a median of 6 days (range 0-7 days) after PBPC reinfusion. The median nadir platelet count was 63 x 10(9)/L (range 6 x 10(9)/L-176 x 10(9)/L). Eight patients (12%) had platelet nadir <20 x 10(9)/L, and all recovered their counts to >20 x 10(9)/L on day 7. There was no clinical difference in days to engraftment between women receiving <2 or > or =2 x 10(6) CD34+ PBPC/kg following paclitaxel. All patients recovered neutrophil and platelet counts within 7 days after reinfusion of > or =1 x 10(6) CD34+ cells/kg and G-CSF. The data suggest that a paclitaxel dose of 825 mg/m2 is not myeloablative. For melphalan, median days to ANC >0.5 x 10(9)/L was 10 days (range 9-15), and platelet recovery to >20 x 10(9)/L was 13 days (range 0-28) after PBPC reinfusion. Median time to engraftment was more rapid in patients receiving > or =2 x 10(6) CD34+/kg versus <2 x 10(6)CD34+/kg, for both neutrophils (11 days versus 10 days, p = 0.05) and platelets (14 days versus 12 days, p < 0.01). Ninety-eight percent of patients infused with > or =2 x 10(6) CD34+/kg engrafted within 21 days. Following CTCb in this sequential regimen, a dose of > or =2 x 10(6) CD34+ cells/kg provided for significantly more rapid neutrophil engraftment than <2 x 10(6) CD34+ cells/kg (9 days versus 10 days,p = 0.01), but a dose > or =3 X 10(6) CD34+ cells/kg is necessary for reliable, rapid, and sustained neutrophil and platelet engraftment by day 21.

PURPOSE: This study was conducted to obtain and compare the preferences assigned to cancer states and prevention measures by women who had breast cancer, were at high risk for breast cancer, or had neither condition. PATIENTS AND METHODS: We administered a time trade-off questionnaire to 21 breast cancer patients, 28 women with a personal history of multiple breast biopsies or a family history of breast cancer, and 135 women without these conditions (the reference group). We stratified the reference group into two groups aged 20 to 32 years and 33 to 50 years, respectively. RESULTS: All four groups assigned higher preference to breast cancer than to ovarian cancer. Both reference groups preferred using a tamoxifen-like drug to having mastectomy or oophorectomy for cancer prevention; the high-risk and breast cancer groups did not. None of the four groups had a preference between prophylactic mastectomy and breast cancer. All the groups were willing to subtract more years from their life expectancy to protect offspring from genetic risk than to protect themselves. Reference group members in the 33- to 50-year age range had lower mean ratings than the breast cancer group for almost all the health states, and breast cancer patients were less willing than other respondents to trade time for health. Most of these differences were not statistically significant. The high-risk group was similar to the older reference group in time trade-off ratings. DISCUSSION: The time trade-off-based preferences of healthy women may be used to predict the treatment preferences of women with BRCA1/2 mutations. Obtaining healthy women's ratings of treatment outcomes may help health care policy makers envision the consequences of the difficult choices that high-risk women face.

BACKGROUND: For patients with metastatic prostate cancer, treatment is primarily palliative, relying mainly on the suppression of systemic androgen hormone levels. To help document the achievement of palliation and to characterize positive and negative effects of treatment, we evaluated quality-of-life (QOL) parameters in patients with metastatic prostate cancer who were randomly assigned to two methods of androgen deprivation. METHODS: Patients (n = 739) with stage M1 (bone or soft tissue metastasis) prostate cancer were enrolled in a QOL protocol that was a companion to Southwest Oncology Group INT-0105, a randomized double-blind trial comparing treatment with bilateral orchiectomy (surgical castration) plus either flutamide or placebo. Patients completed a comprehensive battery of QOL questionnaires at random assignment to treatment and at 1, 3, and 6 months later. Data were collected on three treatment-specific symptoms (diarrhea, gas pain, and body image), on physical functioning, and on emotional functioning. All P values are two-sided. RESULTS: Questionnaire return rates for this study never dropped below 80%; only 2% of the patients did not submit baseline QOL assessments. Cross-sectional analyses (corrected for multiple testing) identified statistically significant differences that favored orchiectomy plus placebo for two of the five primary QOL parameters as follows: patients receiving flutamide reported more diarrhea at 3 months (P = .001) and worse emotional functioning at 3 and 6 months (both P<.003). Longitudinal analyses replicated these findings. Other analyzed QOL parameters favored the group receiving placebo but were not statistically significant after adjustment for multiple testing. CONCLUSIONS: We found a consistent pattern of better QOL outcomes at each follow-up assessment during the first 6 months of treatment for orchiectomized patients with metastatic prostate cancer who received placebo versus flutamide. Improvement over time was evident in both treatment groups but more so for patients receiving placebo.

Longitudinal quality of life measurements from an advanced-stage cancer clinical trial are analysed using a variety of methods, and the results compared. The methods used require different assumptions about the mechanism that produces the missing data. They include analyses that require the data to be missing completely at random; fixed-effects models and weighted generalized estimating equations, which require missing at random data; and a fully parametric approach where the outcomes and the missingness mechanism are jointly modelled, allowing non-ignorable missing data. The data show evidence of non-random missingness, but a formal test of non-ignorable missing data is not significant.

We summarize issues that arise when considering quality of life (QOL) data in cancer clinical trials, especially those related to missing data. We describe different types of missing data mechanisms, and discuss ways of assessing and testing missing data mechanisms. A section on presentation of study design and results describes how graphical displays can effectively document the extent of the missing data problem, as well as describe its impact on interpretation of results. Finally, we describe several different statistical methods used to analyse repeated measures, with an emphasis on their properties and their ability to adequately handle different types of missing data mechanisms. We make recommendations as to the most appropriate methods, and suggest important directions for future research.

Accurate estimation of quality of life is critical to cost-effectiveness analysis. Nevertheless, development of sampling algorithms to maximize the accuracy and efficiency of estimated quality of life has received little consideration to date. This paper presents a method to optimize sampling strategies for estimating quality-adjusted life years. In particular, the authors address the questions of when to sample and how many observations to sample at each sampling time, assuming realistically that the sample variance of quality of life is not constant over time. The method is particularly useful for the design problems researchers face when time or research budget constraints limit the number of individuals that can be surveyed to estimate quality of life. The article focuses on cross-sectional sampling. The method proposed requires some knowledge of survival in the population of interest, the approximate variances in utilities at various points along the curve, and the general shape of the quality-adjusted survival curve. Such data are frequently available from disease registries, the literature, or previous studies.

We propose weighted estimating equations for data with nonignorable nonresponse in order to reduce the bias that can occur with a complete case analysis. A survey concerning medical practice guidelines, malpractice litigation, and settlement provides the framework. The survey was sent to recipients in two waves: those who responded on the first or second wave are used to estimate a nonignorable nonresponse model, while the fraction of recipients who never responded is used to allow the percentage of missing data to change with each wave. We use the structure of the GEE of Liang and Zeger (1986, Biometrika 73, 13-22), adding weights equal to the inverse probability of being observed. We present simulations demonstrating the bias that can occur with an unweighted analysis and use the survey data to illustrate the methods.