© Oxford University Press 2006.
IN THIS ISSUE
Gene Expression Profiling for Neuroblastoma, Breast CancerThe clinical behavior of metastatic neuroblastomas with a nonamplified MYCN gene generally correlates with the patient's age at diagnosis. However, age at diagnosis is not always associated with patient survival. Asgharzadeh et al. (p. 1193) developed a prognostic classifier of disease progression for children diagnosed with metastatic neuroblastomas lacking MYCN gene amplification based on the expression of 55 genes. The classifier identified patients with low and high progression-free survival rates among those older than 12 months at diagnosis with high-risk disease and those older than 18 months at diagnosis. The authors conclude that gene expression profiles of tumors obtained at diagnosis from patients with high-risk, metastatic neuroblastomas can identify subgroups with different outcomes.
Before a gene expression classifier can be used in the clinic, it must be shown to predict outcomes in independent patient populations. A gene signature should also have prognostic value independent of standard clinical and pathologic risk criteria. Buyse et al. (p. 1183) report that a 70-gene signature previously shown to have prognostic value in women with node-negative breast cancer also predicted outcomes in an independent group of patients. The 70-gene signature performed better than standard clinicopathologic risk classifications in predicting all outcomes analyzed: risk of distant metastases, risk of recurrence, and risk of death from any cause. It also yielded prognostic data beyond the clinicopathologic assessment.
In an editorial, Simon (p. 1169) notes that Asgharzadeh et al. evaluated the accuracy of their classifier appropriately. He applauds Buyse et al. for a proper validation study. However, both studies have limitations—for example, neither was performed under real-world conditions. He also writes that outcome predictors identified in gene expression profiling studies should not be confused with surrogate markers of disease, which are much harder to identify.
Breast Cancer Risk Models Including Breast Density
The Gail model's ability to predict the absolute risk of invasive breast in white women was improved through a relative risk model that included mammographic density. Chen et al. (p. 1215) now present the corresponding model for absolute risk. They report that the new model predicts higher attributable risks than the Gail model for women with a high percentage of dense breast area. Average risk projections in various age groups were similar with both models, indicating that the new model was well calibrated. The new model promises modest improvements in discriminatory power but needs to be validated with independent data.
In another new model, Barlow et al. (p. 1204) used prospective risk information from 1 million women to predict a breast cancer diagnosis in women undergoing screening mammography. For premenopausal women, the significant factors for breast cancer diagnosis included age, breast density, family history, and a prior breast procedure. Postmenopausal women had the same factors as well as race, ethnicity, body mass index, natural menopause, hormone therapy, and a prior false-positive mammogram. The authors conclude that breast density is a strong risk factor for breast cancer and that their model may identify women at high risk.
In an editorial, Bondy and Newman (p. 1172) note that, although an anatomic- or tissue-based biomarker would be a more rational tool, breast density measures may be the next-best option. Several practical considerations make obtaining breast density measurements for the general population problematic. Including breast density in risk prediction tools is an exciting development, the authors note, but caution is required because even a perfectly predictive model will be of minimal value if its component factors are unavailable, misunderstood, or inappropriately assigned.
Use of Lay Health Advisors in Mammography Intervention
Rates of screening mammography in the United States have increased in recent years but continue to be low among poor women, minority women, and women who live in rural areas. Paskett et al. (p. 1226) report on a randomized trial using lay health advisors—community women trained to deliver an individualized health education intervention—to improve mammography screening rates in underserved women in rural North Carolina. The advisors attempted to address practical and perceived barriers to mammography screening, to increase women's awareness of screening's benefits, and to provide basic knowledge about breast health. Women who received the intervention were more likely to have a mammogram in the year after it began than the control group women.
Cyclin D1 Overexpression in Breast Cancer and Bortezomib
Cyclin D1 overexpression in breast cancer has been unexpectedly associated with improved outcomes. To determine the mechanism of this association, Ishii et al. (p. 1238) measured the expression of both cyclin D1 and STAT 3 (the antiapoptotic signal transducer and activator of transcription 3) in human breast cancer primary tumors and in cyclin D1–overexpressing cell lines. They also analyzed the effect of the proteasome inhibitor bortezomib on cyclin D1 stability and its subsequent effect on STAT3 and compared the growth of xenograft tumors derived from cyclin D1–overexpressing cells in untreated and bortezomib-treated mice. Expression of cyclin D1 was inversely related to that of STAT3 in the primary tumors and cell lines. Bortezomib treatment increased cyclin D1 levels in the cell lines, which inhibited STAT3 expression and slowed xenograft tumor growth. The authors conclude that cyclin D1 inhibits STAT3 activity in breast cancer cells and that cyclin D1 expression may predict response to bortezomib.
Radiation-Induced Pneumonitis and CD95
Pneumonitis—inflammatory swelling of the lungs associated with respiratory failure—is a dose-dependent side effect of radiotherapy. The CD95/CD95-ligand system, which is involved in apoptosis induction and possibly proinflammatory responses, may be involved in the development of radiation-induced pneumonitis. To find out, Heinzelmann et al. (p. 1248) irradiated CD95/CD95-ligand-deficient mice and wild-type mice and compared breathing frequency, pulmonary resistance, and pulmonary histopathologic changes. After irradiation, wild-type mice developed pneumonitis symptoms like increased breathing frequency and pulmonary resistance; the CD95/CD95-ligand-deficient mice did not.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||