© 2001 by Oxford University Press
Journal of the National Cancer Institute, Vol. 93, No. 7, 485,
April 4, 2001
© 2001 Oxford University Press
IN THIS ISSUE |
Retinoids are one of the most promising groups of chemopreventive agents, and they have shown evidence in clinical trials of their benefit in preventing recurrence of breast cancer and first occurrence of ovarian cancer. Because of a need to identify retinoids that are effective at clinically achievable doses, Guruswamy et al. (p. 516) studied the effects of a variety of retinoids and related compounds, including some novel compounds, on ovarian cancer cells in culture. They found that all 11 compounds tested had the ability to reverse the cells’ cancerous phenotype and decrease cell growth. Apoptosis was induced by all compounds that were not dependent on binding to the retinoid receptor and by two that were. Induction of expression of the mucin MUC-1, a marker of glandular differentiation, appeared to be involved in the retinoids’ mechanism of action.
"This study demonstrated that clinically achievable concentrations of retinoids decreased the growth fraction of and induced glandular differentiation and apoptosis in ovarian cancer cells and tissue."
—Guruswamy et al.
Modeling Bone Metastasis
Metastasis to bone is a relatively common event in cancer. However, it is difficult to describe the clinical course of different kinds of bone metastases and to compare treatments based on clinical research. Cook and Major (p. 534) contrast several statistical methods for the analysis of data on treatment of bone metastases to identify appropriate analysis strategies. They conclude that a random-effects Poisson model, which takes into account individual differences in numbers of recurrent adverse events, is most robust and suitable for describing data of this kind. They note that the analysis that is typically performed, events per person-years, can substantially underestimate interindividual variability and therefore increase the probability of a falsely positive clinical trial.
In an editorial, Glynn and Buring (p. 488) point to some of the important lessons learned from analyses such as those done by Cook and Major, such as not ignoring available data on recurrent events, examining assumptions of independence of events, and taking into account varying lengths of time between events.
"The importance of appropriate statistical models for drawing reliable conclusions in trials of chronic diseases with recurrent events has been actively debated."
—Cook and Major
Gender and Smoking-Related Bladder Cancer Risk
Evidence suggests that, when smoking habits are comparable, women may be at higher risk for lung cancer than men. Because cigarette smoke is one of the established risk factors for bladder cancer, it is unclear whether women who smoke may also be at higher risk for bladder cancer. In a population-based, case–control study, Castelao et al. (p. 538) examined possible gender differences in the smoking-related risk of bladder cancer. The authors found that, when smoking habits were comparable, women had a higher risk of smoking-related bladder cancer than men. The risk of bladder cancer for either sex was not modified by use of filtered versus nonfiltered cigarettes, low-tar versus higher tar cigarettes, or the pattern of inhalation. The authors conclude that, when smoking habits are comparable, women may be at higher risk for bladder cancer than men.
Patient Treatment and Outcome by Provider Characteristics
While the management and prognosis of colorectal cancer are largely dependent on features such as tumor stage, there is considerable variation in treatment and outcome not explained by traditional prognostic factors. Hodgson et al. (p. 501) reviewed studies of variation in treatment and outcome by patient and provider characteristics. They observed that the surgeon expertise and case volume are associated with improved tumor control but not necessarily with perioperative mortality or long-term survival. They also did not find any differences in general in the management of patients by health maintenance organizations or fee-for-service providers. They noticed that older patients are less likely to receive adjuvant therapy after surgery, and they found that black patients often receive less aggressive therapy and are more likely to die of this disease than white patients. The authors conclude that more studies are needed to understand mechanisms underlying associations between patient and provider characteristics and the treatment outcomes.
Diet and Colorectal Cancer Risk
Several recent studies have failed to show a beneficial association between consumption of fruit, vegetables, and dietary fiber and the risk of colorectal cancer. Terry et al. (p. 525) hypothesized that this association might be more apparent in a population that normally consumes relatively low amounts of fruit and vegetables and high amounts of dietary fiber. The authors therefore determined the dietary habits of a large cohort of Swedish women and found that women who consumed the lowest amounts of fruit and vegetables had the highest risk for developing colorectal cancer, while women who consumed even the highest amounts of cereal fiber did not have a lower colorectal cancer risk. The authors suggest that recommendations to increase fruit and vegetable consumption to decrease colorectal cancer risk may be most beneficial for those individuals who consume less than two servings per day of fruit and vegetables.
"Nevertheless, our data suggest that even moderate increases in fruit and vegetable consumption among persons with very low intake may confer benefits to this group."
—Terry et al.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||