William D. Dupont

Multifactor-Dimensionality Reduction Reveals High-Order Interactions among Estrogen-Metabolism Genes in Sporadic Breast Cancer

One of the greatest challenges facing human geneticists is the identification and characterization of susceptibility genes for common complex multifactorial human diseases. This challenge is partly due to the limitations of parametric-statistical methods for detection of gene effects that are dependent solely or partially on interactions with other genes and with environmental exposures. We introduce multifactor-dimensionality reduction (MDR) as a method for reducing the dimensionality of multilocus information, to improve the identification of polymorphism combinations associated with disease risk. The MDR method is nonparametric (i.e., no hypothesis about the value of a statistical parameter is made), is model-free (i.e., it assumes no particular inheritance model), and is directly applicable to case-control and discordant-sib-pair studies. Using simulated case-control data, we demonstrate that MDR has reasonable power to identify interactions among two or more loci in relatively small samples. When it was applied to a sporadic breast cancer case-control data set, in the absence of any statistically significant independent main effects, MDR identified a statistically significant high-order interaction among four polymorphisms from three different estrogen-metabolism genes. To our knowledge, this is the first report of a four-locus interaction associated with a common complex multifactorial disease.

Risk Factors for Breast Cancer in Women with Proliferative Breast Disease

To assess the importance of various risk factors for breast cancer in women with benign proliferative breast lesions, we reevaluated 10,366 consecutive breast biopsies performed in women who had presented at three Nashville hospitals. The median duration of follow-up was 17 years for 3303 women, 1925 of whom had proliferative disease. This sample contained 84.4 per cent of the patients originally selected for follow-up. Women having proliferative disease without atypical hyperplasia had a risk of cancer that was 1.9 times the risk in women with nonproliferative lesions (95 per cent confidence interval, 1.2 to 2.9). The risk in women with atypical hyperplasia (atypia) was 5.3 times that in women with nonproliferative lesions (95 per cent confidence interval, 3.1 to 8.8). A family history of breast cancer had little effect on the risk in women with nonproliferative lesions. However, the risk in women with atypia and a family history of breast cancer was 11 times that in women who had nonproliferative lesions without a family history (95 per cent confidence interval, 5.5 to 24). Calcification elevated the cancer risk in patients with proliferative disease. Although cysts alone did not substantially elevate the risk, women with both cysts and a family history of breast cancer had a risk 2.7 times higher than that for women without either of these risk factors (95 per cent confidence interval, 1.5 to 4.6). This study demonstrates that the majority of women (70 per cent) who undergo breast biopsy for benign disease are not at increased risk of cancer. However, patients with a clinically meaningful elevation in cancer risk can be identified on the basis of atypical hyperplasia and a family history of breast cancer.

Power and sample size calculations: A review and computer program☆

Methods of sample size and power calculations are reviewed for the most common study designs. The sample size and power equations for these designs are shown to be special cases of two generic formulae for sample size and power calculations. A computer program is available that can be used for studies with dichotomous, continuous, or survival response measures. The alternative hypotheses of interest may be specified either in terms of differing response rates, means, or survival times, or in terms of relative risks or odds ratios. Studies with dichotomous or continuous outcomes may involve either a matched or independent study design. The program can determine the sample size needed to detect a specified alternative hypothesis with the required power, the power with which a specific alternative hypothesis can be detected with a given sample size, or the specific alternative hypotheses that can be detected with a given power and sample size. The program can generate help messages on request that facilitate the use of this software. It writes a log file of all calculated estimates and can produce an output file for plotting power curves. It is written in FORTRAN-77 and is in the public domain.

Atypical hyperplastic lesions of the female breast. A long-term follow-up study.

A total of 10,542 breast biopsy specimens obtained between 1950 and 1968 were studied. Examples of atypical “ductal” (ADH) and atypical lobular hyperplasia (ALH), defined as having only some features of carcinoma in situ (CIS), were diagnosed in 3.6% of these specimens. In the same series, CIS was diagnosed in 1.7% of biopsy specimens excluding those with invasive cancer. The subsequent risk of invasive breast carcinoma after ALH or ADH was 4–5 times that of the general population. Follow‐up was 90% successful and extended 17 years after biopsy. History of breast cancer in a mother, sister, or daughter doubled the risk of subsequent invasive carcinoma development (to 8 times for ALH and 10 times for ADH). The authors conclude that among the epithelial hyperplastic lesions of the human breast, a minority may be recognized by their resemblance to CIS which have a clinically significant elevation of subsequent breast cancer risk. This risk is one‐half that of CIS.

The effects of ibuprofen on the physiology and survival of patients with sepsis

BACKGROUND In patients with sepsis the production of arachidonic acid metabolites by cyclooxygenase increases, but the pathophysiologic role of these prostaglandins is unclear. In animal models, inhibition of cyclooxygenase by treatment with ibuprofen before the onset of sepsis reduces physiologic abnormalities and improves survival. In pilot studies of patients with sepsis, treatment with ibuprofen led to improvements in gas exchange and airway mechanics. METHODS From October 1989 to March 1995, we conducted a randomized, double-blind, placebo-controlled trial of intravenous ibuprofen (10 mg per kilogram of body weight [maximal dose, 800 mg], given every six hours for eight doses) in 455 patients who had sepsis, defined as fever, tachycardia, tachypnea, and acute failure of at least one organ system. RESULTS In the ibuprofen group, but not the placebo group, there were significant declines in urinary levels of prostacyclin and thromboxane, temperature, heart rate, oxygen consumption, and lactic acidosis. With ibuprofen therapy there was no increased incidence of renal dysfunction, gastrointestinal bleeding, or other adverse events. However, treatment with ibuprofen did not reduce the incidence or duration of shock or the acute respiratory distress syndrome and did not significantly improve the rate of survival at 30 days (mortality, 37 percent with ibuprofen vs 40 percent with placebo). CONCLUSIONS In patients with sepsis, treatment with ibuprofen reduces levels of prostacyclin and thromboxane and decreases fever, tachycardia, oxygen consumption, and lactic acidosis, but it does not prevent the development of shock or the acute respiratory distress syndrome and does not improve survival.

Power and sample size calculations for studies involving linear regression

This article presents methods for sample size and power calculations for studies involving linear regression. These approaches are applicable to clinical trials designed to detect a regression slope of a given magnitude or to studies that test whether the slopes or intercepts of two independent regression lines differ by a given amount. The investigator may either specify the values of the independent (x) variable(s) of the regression line(s) or determine them observationally when the study is performed. In the latter case, the investigator must estimate the standard deviation(s) of the independent variable(s). This study gives examples using this method for both experimental and observational study designs. Cohens method of power calculations for multiple linear regression models is also discussed and contrasted with the methods of this study. We have posted a computer program to perform these and other sample size calculations on the Internet (see http://www.mc.vanderbilt.edu/prevmed/psintro+ ++.htm). This program can determine the sample size needed to detect a specified alternative hypothesis with the required power, the power with which a specific alternative hypothesis can be detected with a given sample size, or the specific alternative hypotheses that can be detected with a given power and sample size. Context-specific help messages available on request make the use of this software largely self-explanatory.

Intraductal carcinoma of the breast: Follow‐up after biopsy only

Twenty‐eight women with ductal carcinoma in situ (DCIS) of the breast treated by biopsy only were identified in a histologic review of 11,760 biopsies performed between 1950 and 1968. Seven of the 25 women followed for more than three years developed invasive breast carcinoma, all in the same breast with a previously detected DCIS. Average follow‐up interval for the 18 women not developing invasive carcinoma was 16 years. The invasive carcinomas presented clinically from three to ten years (average, 6.1) after the biopsies demonstrating DCIS. Four women with invasive carcinoma developed distant metastases following mastectomy. This study suggests that 28% of women treated with biopsy only for DCIS presenting as an incidental histologic finding will develop invasive carcinoma in a follow‐up period of approximately 15 years.

Continued local recurrence of carcinoma 15–25 years after a diagnosis of low grade ductal carcinoma in situ of the breast treated only by biopsy

Background. The stratification of ductal carcinoma in situ (DCIS) of the human breast into prognostically relevant categories by size and histologic pattern is a current concern. Few studies have been able to follow women after the identification of any type of DCIS when they have had biopsy only.

Lobular neoplasia of the breast: Higher risk for subsequent invasive cancer predicted by more extensive disease

We have stratified the cancer risk implications of lobular pattern in situ neoplasias of the breast by separating marked examples of this histologic spectrum (lobular carcinoma in situ [LCIS]) from lesser examples (atypical lobular hyperplasia). The lesser-developed examples have been shown previously to have a lower relative risk (RR) of later invasive carcinoma of the breast (IBC). Forty-eight examples of LCIS were found in 10,542 otherwise benign breast biopsies, representing an incidence of 0.5%. Nine patients were excluded from follow-up because of bilateral mastectomy within 6 months of entry biopsy, IBC within 6 months of entry biopsy, or prior IBC. Follow-up of the remaining 39 patients was complete, averaged 18 years, and revealed an RR of subsequent IBC of 6.9 (P less than .00001). Average overall follow-up for LCIS patients was 19 years; it was 25 years for those alive and free of IBC at the time of their follow-up interview. Neither family history of IBC nor postmenopausal estrogen therapy further affected risk. The absolute risk of IBC after LCIS was 17% at 15 years (adjusted for withdrawals), and the RR was 8.0 in the first 15 years of follow-up compared with the general population. An analysis based on a time-dependent hazards model found that during the first 15 years following biopsy women with LCIS had 10.8 times the risk of breast cancer compared with biopsied women of comparable age who lacked proliferative disease. Some previously published articles reporting lobular neoplasia (LN) suggest that those series with the greatest incidences of LN (whether termed LN or LCIS) have the lowest RR of subsequent breast cancer. Those series with higher incidences of LN include less well-developed histologic patterns of LN (atypical lobular hyperplasia). We conclude that our study of LN and studies performed by others support the higher risk of IBC after histologically flagrant examples (LCIS, about nine times higher) and a relatively lower but definable risk after more histologically subtle examples (atypical lobular hyperplasia, four to five times lower). This relative cancer risk is probably not constant over more than 15 years; thus, cancer risk 15 to 25 years after initial diagnosis of LCIS is uncertain.

Atypical lobular hyperplasia as a unilateral predictor of breast cancer risk: a retrospective cohort study

BACKGROUND Clinical decisions about atypical lobular hyperplasia are based on the belief that later invasive breast-cancer risk is equal in both breasts. We aimed to show laterality and subsequent risk implications of invasive breast cancer in women with atypical lobular hyperplasia. METHODS We did a retrospective cohort study of 252 women who had undergone 261 benign surgical biopsies that showed atypical lobular hyperplasia from 1950 to 1985, as part of the Nashville Breast Studies. Primary outcomes were development of invasive breast cancer and laterality of cancer compared with side of the biopsied breast. FINDINGS 50 (20%) of 252 women treated by biopsy only developed invasive breast cancer. Relative risk of breast cancer in women with atypical lobular hyperplasia was 3.1 (95% CI 2.3-4.3, p<0.0001). Of these 50 women, the breast with invasive cancer was the same breast diagnosed with atypical lobular hyperplasia (ipsilateral) in 34 (68%) and the contralateral breast in 12 (24%). The ratio of ipsilateral/ contralateral cancers for atypical lobular hyperplasia without other atypical lesions was 17/5. For six women with atypical lobular hyperplasia plus atypical ductal hyperplasia, the ratio was 1/1. INTERPRETATION Invasive carcinoma after atypical lobular hyperplasia is about three times more likely to arise in the breast diagnosed with atypical lobular hyperplasia than in the opposite breast without these initial findings. Our findings suggest a model of premalignancy for atypical lobular hyperplasia intermediate between a local precursor and a generalised risk for both breasts. See Commentary page 96