Hypothesis/Commentary

Diet, Cancer, and the Lipidome

Philippe Bougnoux, Bruno Giraudeau, and Charles Couet

INSERM E 0211; University Hospital, Tours, France

Abstract

The potential for dietary fat to interfere with the develop-ment of breast cancer by delaying its occurrence makes theidentification of defined molecules a mandatory step incancer prevention. In order to circumvent the limitationsand/or bias of dietary exposure assessment tools, bio-markers of past lipid intake such as the fatty acidcomposition of white adipose tissue have been used. Whenconsidered separately, candidate fatty acids identified asfavorable on the basis of their association with breastcancer risk have usually led to inconsistent results inanimal intervention studies. This inconsistency indicatesthat any approach based on a single fatty acid should beabandoned for an integrated view over the complex lipidinteractions which finally determines the lipidome, the

lipid profile that is found in individuals. This articlepresents a reappraisal of the role of the lipid profilethrough a comprehensive reanalysis of adipose tissue fattyacid composition obtained in patients with benign ormalignant breast tumors as well as in experimental animalsduring dietary interventions. Rather than a single fatty acid,a composite indicator combining elevated monounsaturatesand low W6/W3 fatty acid ratio was associated with breastcancer protection. This lipidome may become the templatefor identifying breast cancer risk related to diet, and fordesigning proper dietary modifications to delay the occur-rence of breast cancer, although the universality of thefindings cannot be assessed from a single study. (CancerEpidemiol Biomarkers Prev 2006;15(3):416–21)

Numerous epidemiologic studies have established strikinggeographic differences in the rate of incidence of breastcancer, and those conducted in migrant populations stronglysupport a role for environmental exposure including diet inthe variation of breast cancer rates across countries (1).Indeed, it has been suggested that 20% to 60% of cancers,depending on the anatomic localization of the tumor, may beavoidable by altering the diet (2). Among several dietarycomponents that could modulate breast cancer risk, fat hasbeen extensively examined with little evidence for apromoting effect of total fat intake independent of the fatcontribution to total energy intake (3), which has beenrecently shown to be positively associated with breast cancerrisk in Chinese women (4). Ecological, cohort, and case-control studies aimed at evaluating the association betweenusual or past diet was assessed by either dietary recall,dietary record, and dietary interview, whereas food frequen-cy questionnaire on the one hand and breast cancer risk onthe other have led to discordant conclusions (5). Thediscrepancies between studies are mainly due to thelimitations and/or bias of dietary exposure assessment tools,including subjectivity, unintentional inaccuracy, underreport-ing, and dieting behavior which alter the quality of dietarydata collected in free-living populations (6). Beyond theuncertainties that are common to any study of nutritionalepidemiology, there are specific limitations in evaluating theintake of some dietary fatty acids which are partly or totallyexcluded from food composition tables currently available.

This is usually the case for the fatty acids quantitativelyconsidered as minor, but which are biologically important,including long-chain polyunsaturated fatty acids (PUFA) ofthe N3 and N6 series, trans-fatty acids, and conjugated linoleicacid, for instance.

Lessons Learned from the Use of Biological Markersof Dietary Exposure

In order to circumvent these limitations, biological markers ofdietary exposure such as the fatty acid composition of serumphospholipids, membrane phospholipids and/or white adiposetissue (WAT) triglycerides have been extensively used inpopulation-based studies with special emphasis on N3 and N6PUFA. Briefly, studies dealing with the N3/N6 PUFA compo-sition of either serum or erythrocyte membrane phospholipidsand breast cancer risk were inconclusive (7). It should bestressed that plasma and membrane phospholipid compositionvary according to recent dietary intakes which might be alteredby the occurrence of cancer. Conversely, the fatty acidcomposition of WAT triglycerides best reflects past dietaryintake for the essential fatty acids (linoleic acid, a-linolenic acid,and long-chain N3 PUFA; refs. 8, 9), and seems to be a morerelevant marker of lasting exposure with respect to the breastcancer risk and fat intake relationships. This marker has beenused in several studies aiming to investigate the relationbetween exposure to N3 and N6 PUFA and breast cancer. Inour case-control study published in 2002 (10), and conducted in241 patients with invasive breast carcinoma and 88 controlpatients with benign breast disease, we found a significantinverse association between individual levels of a-linolenic acidand docosahexaenoic acid (DHA) in breast adipose tissue andthe risk of breast cancer, with the strongest inverse associationfound for the ratio of long-chain N3 PUFA to N6 PUFA. Theresults of other studies have been reviewed recently with theoverall conclusion that the protective effect of N3 PUFA onbreast cancer risk depends on background levels ofN6 PUFA (7).

416

Cancer Epidemiol Biomarkers Prev 2006;15(3). March 2006

Received 7/21/05; revised 12/26/05; accepted 1/18/06.

The costs of publication of this article were defrayed in part by the payment of page charges.This article must therefore be hereby marked advertisement in accordance with 18 U.S.C.Section 1734 solely to indicate this fact.

Requests for reprints: Philippe Bougnoux, INSERM E 0211, University Hospital of Tours, 2Boulevard Tonnelle, 37044 Tours Cedex, France. Phone: 33-2-4736-6179; Fax: 33-2-4736-6226.E-mail: bougnoux@med.univ-tours.fr

Copyright D 2006 American Association for Cancer Research.

doi:10.1158/1055-9965.EPI-05-0546

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from

Intervention Studies Targeted on Single Fatty Acids: aRepeated Failure

To fully interpret the information that the biomarkers (i.e.,the WAT content of a-linolenic acid, DHA and N6 to N3PUFA ratio) can provide with respect to cancer risk andpossible prevention in humans, dietary interventions tar-geted at either a-linolenic acid or DHA were set up inN-nitrosomethylurea and 7,12-dimethylbenz(a)anthracene-induced mammary tumor models in rats with inconsistentresults (7). Whereas some studies seemed to support theepidemiologic evidence concerning the importance of theratio of N6 to N3 PUFA in mammary tumor growth (11-13),others failed to find any significance (14-16), perhapsbecause of the inhibitory effect of N3 PUFA on mammarytumor growth, which depends on the background levels ofN6 PUFA as well as on antioxidant levels (17). Anexperimental study on rats showed that the addition ofvitamin E to a 15% linseed oil diet rich in a-linolenic acidled to an increase in tumor growth compared with controlswithout vitamin E, whereas the addition of a prooxidantcompound (sodium ascorbate/2-methyl-1,4-naphtoquinone)led to a decrease in tumor growth (18). Thus, despitethe identification through epidemiologic studies of a fewWAT fatty acids as a biomarker of breast cancer risk inhumans, a more complex picture is emerging from theexperimental dietary intervention studies focusing on thosefatty acids.

Breast Cancer Risk Varies Depending on the WATFatty Acid Level Considered

To get a more integrated view of the complex lipid interactionswhich, in turn, give rise to the individual lipid profile in WATtriglycerides, we retrospectively reconsidered the statisticalanalysis of hundreds of WAT fatty acid compositionspreviously obtained from French women with either breastcancer or benign breast tumor (10), using principal componentanalysis. Secondly, we reexamined the conclusions drawnfrom our previous dietary intervention studies in rodents(16, 19) in light of the new and more sophisticated analysis ofhuman data.

Figure 1 shows the adjusted odds ratios (OR) of breastcancer for the fatty acids of WAT sampled from 329 womenand presented by lipid class. A decreased risk of breast cancerwas associated with higher content of N3 PUFA, either a-linolenate or the long chain N3 fatty acids, eicosapentaenoicacid or DHA. In contrast, a high content of N6 PUFA wasassociated with either a trend (linoleic acid, 18:2N6) or anincreased risk (20:2N6). Cis-monounsaturates were all protec-tive, whereas trans-monounsaturates were not (t-16:1N7), orwere strongly associated with an increased risk (elaidic acid, t-18:1N9). No association with breast cancer risk was detectedfor saturates.

WAT Fatty Acid Levels are not Independent Variables

The significance of such associations is unclear when oneconsiders the common metabolic pathways shared by numer-ous fatty acids which link the content of individual fatty acid toeach other. Figure 2 presents the correlation coefficients foundbetween two fatty acids of each PUFA family and the otherfatty acids of WAT triglycerides. For instance, linoleic acidshows a strong commensurate and positive correlation with its

417

Figure 1. Estimated associations of breast cancer and high level ofWAT fatty acid content. Adipose tissue obtained at surgery from241 patients with invasive, nonmetastatic breast cancer (cases), andfrom 88 patients with benign, nonproliferative tumors (controls)were analyzed for fatty acid composition. Prior to analysis,standardization was done for each fatty acid, thus, allowingcomparisons between fatty acids. For this purpose, each individualfatty acid value was subtracted from the mean value of the groupand divided by the SD. Then, ORs were estimated for each fattyacid adjusted for BMI, height, age, and menopausal status in theframework of logistic regression models (23). Total saturatesinclude 14:0, 15:0, 16:0, 17:0, 18:0, and 20:0; total cis-monounsaturates include 14:1N5, 16:1N7, 17:1, 18:1N7, 18:1N9,and 20:1; total N6 PUFA include 18:2, 20:2, 20:3, 20:4, and 22:4;and total N3 PUFA include 18:3, 20:5, 22:5, and 22:6.

Figure 2. Coefficients of correlation obtained in adipose tissuebetween main fatty acids and either linoleic acid (A), arachidonic acid(B), a-linolenic acid (C), DHA (D).

Cancer Epidemiology, Biomarkers & Prevention

Cancer Epidemiol Biomarkers Prev 2006;15(3). March 2006

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from

long-chain derivatives according to the number of metabolicsteps involved in their biosynthesis, and a weak positivecorrelation with N3 PUFA (Fig. 2A). In contrast, linoleic acidinversely correlates with saturates and monounsaturates, withthe exception of trans-monounsaturates. Positive and inverseassociations are observed for arachidonic acid, a-linolenic acid,DHA (Fig. 2B, C , and D), and for many others (data notshown). Consequently, a single fatty acid cannot be consideredas an independent biomarker of breast cancer risk, and there isa need to simplify this complex system of correlations into asmaller number of dimensions.

Simplifying through Principal Component Analysis

For this purpose, we did a principal component analysis usingour whole database of 329 patients (cases and controls). Theprincipal component analysis was based on 23 fatty acidswhich belong to the four principal fatty acid classes as follows:saturates (14:0, 15:0, 16:0, 17:0, 18:0, and 20:0), monounsatu-rates (14:1, 16:1c, 16:1t, 17:1, 18:1N7, c-18:1N9, t-18:1N9, and20:1), N6-polyunsaturates (18:2, 20:2, 20:3, 20:4, and 22:4), andN3-polyunsaturates (18:3, 20:5, 22:5, and 22:6), on the basis oftheir level in adipose tissue or their carbon chain length (fattyacids with <14 carbons or with a level <0.2% of total fatty acidswere not included in the analysis). Because age and bodymass index (BMI) are closely associated with the risk of breast

cancer, at least in postmenopausal women, and because theystrongly correlate with long-chain PUFA levels in the WAT,fatty acids were considered through their residuals fromthe regression on age and BMI, thus preventing from artifacteffects due to age and/or BMI. Principal component analysis isaimed at transforming a set of intercorrelated variables (the 23fatty acids) into a set of uncorrelated variables, or principalcomponent (http://www.statsoft.com/textbook/stfacan.html;ref. 20). The first principal component accounts for as much aspossible of the variability between patients, and each succeed-ing component accounts for as much as possible of theremaining variability. Thus, each principal componentexplains a fraction of the variance, which is the fraction ofinformation explained by the principal component. Theinterpretation of the principal components—i.e., the meaningof these new variables—is made in view of their correlationwith the initial variables.

The results of this analysis are shown in Fig. 3. The twoprincipal components accounted for almost half (42.4%) of theinformation (interindividual variability) borne by all 23 fattyacids. The first principal component (X axis) accounted for24.4% and the second principal component (Y axis) accountedfor 18.0%. Fatty acids were not randomly located. Saturatedfatty acid location was clustered into the left part of the scatterplot, monounsaturates (with the notable exception of t-18:1N9)were in the upper part, long-chain polyunsaturates (both N6and N3 PUFA) were on the right part, whereas 18:2N6 was on

418

Figure 3. Principal component analysis of adipose tissue fatty acids. In the scatter plot of the second principal component against the firstprincipal component (in which the X axis represents the first principal component and the Y axis represents the second principal component), thecoordinates of each fatty acid equals the coefficients of correlation between the fatty acid and the principal components. The unity correlationcircle drawn defines the limits in which the fatty acids locate: the closer a fatty acid to this unity circle, the higher its contribution to thedefinition of the principal components. The N6/N3 ratio is located as illustrative variables—i.e., it does not contribute to the definition of theprincipal components, but it is positioned in the scatter plot according to its correlation with the two principal components. Red arrow, increasedrisk of breast cancer, taking into account the OR associated with both the X and Y axis, adjusted for BMI, age, menopausal status, and height.The position of this arrow is almost superposed on the Y axis because the OR associated with the first component is close to 1, whereas the ORassociated with the second principal component is 1.28 (95% CI, 1.11-1.49; P = 0.001).

Diet, Cancer, and the Lipidome

Cancer Epidemiol Biomarkers Prev 2006;15(3). March 2006

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from

the lower part of the scatter plot. Therefore, the X axis opposedsaturates to long-chain polyunsaturates, and the Y axisopposed 18:2N6 to monounsaturated fatty acids. In addition,we located the N6/N3 ratio as an illustrative variable, i.e., itwas not used to establish the principal components. This ratiostrongly correlated with the Y axis (Fig. 3).

The two axes were then considered as independentcovariates in the framework of a logistic regression modelaimed at assessing the risk of breast cancer. The associationwas not significant for the X axis [i.e., OR, associated with adecrease of one unit on the X axis: OR, 1.02; 95% confidenceintervals (CI), 0.92-1.14; P = 0.661], and highly significant forthe Y axis (OR associated with a decrease of one unit on theY axis: OR, 1.25; 95% CI, 1.10-1.42, P < 0.001). The secondprincipal component remained highly significantly associatedwith the risk of breast cancer after adjustment (OR, 1.28;95% CI, 1.11-1.49; P = 0.001). Considering the four quartilesassociated with the second principal component, this associ-ation may be reexpressed as follows: ORQuartile 1/Quartile 4 =3.23; 95% CI, 1.37-8.07; ORQuartile 2/Quartile 4 = 1.58; 95% CI,0.73-3.47; and ORQuartile 3/Quartile 4 = 1.02; 95% CI, 0.48-2.15.Thus, as shown in Fig. 3, a location in the left lower quadrantof the scatter plot is associated with an increased risk ofbreast cancer. Therefore, a lipid profile of the WAT whichcomprises low linoleic acid or a low N6/N3 ratio along with

elevated cis-monounsaturates is protective against the risk ofbreast cancer, independently of age and BMI.

The Lipidome as a New Insight into the Link BetweenDiet and Breast Cancer

Principal component analysis does not provide any indicationof the interindividual differences in the WAT content of eachfatty acid. Figure 4 presents a lipid profile array of cases andcontrols. There are several differences in the pattern of colorsbetween cases and controls. In controls, there is a spot ofelevated values involving monounsaturates. In cases, moreelevated values of N6 PUFA are observed in the lower rightcorner compared with controls. The ratio of N6 to N3 fatty acids(bottom) appears as a main distinctive feature between cases(left) and controls (right). Thus, similar to the profile array oftranscripted genes which allowed the individualization of thecombinations of gene alterations associated with the risk ofdeath in breast cancer (21), the lipid profile array provides anindication of the combinations of WAT fatty acid levelsassociated with the risk of breast cancer. By analogy with theproteome or genome, the word lipidome has been coined tocharacterize this lipid profile which may be altered through adietary intervention.

419

Figure 4. Fatty acid level array in patients with benign (controls) or malignant (cases) breast tumors. Each lane represents a patient, sortedaccording to its position on the second principal component as shown in Fig. 3. Each line represents one fatty acid, according to its correlationwith the second principal component. Fatty acid values are represented as different colors for each quartile, from green (low) to red (elevated).Bottom, the N6/N3 ratio of PUFAs.

Cancer Epidemiology, Biomarkers & Prevention

Cancer Epidemiol Biomarkers Prev 2006;15(3). March 2006

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from

Failure of Dietary Interventions Targeted on a SingleFatty Acid in Favorably Modifying the Lipidome inRats

On the basis of this new composite biomarker of a low risk forbreast cancer—i.e., low N6/N3 ratio and high cis-monounsa-turates—it was tempting to look back to the changes in fattyacid composition of WAT that we induced in our dietaryinterventions targeted at either a-linolenic acid or DHA inN-nitrosomethylurea-induced mammary tumor in rats. Table 1summarizes the fatty acid composition of WAT triacylglycerolswith or without dietary intervention over 26 weeks (16, 19). Inboth diets, a significant increase in the specific N3 fatty acidadded to the control diet is obtained with a decrease in the N6/N3 ratio as expected. However, despite this change in the N6/N3 ratio, which seems to fit the appropriate direction withrespect to breast cancer risk, a concomitant and significantdrop in monounsaturates was induced. This drop is exactly inthe opposite direction compared with the indications providedby the analysis of the lipidome. Thus, favorable dietary fattyacid exposure was not achieved in these animal studies,stressing the need for a composite dietary modification ratherthan for single nutrient intervention.

Conclusions and Future Prospects

It should be pointed out that the validation of any linkbetween a biomarker and the emergence of any type ofcancer requires parallel experimental animal and humanstudies. An appropriate animal model is usually set up todetermine the role of the marker, whether associative orcausal, in the disease pathway and to establish the dose-effectrelationship. Thus, to fully interpret the information providedby the analysis of the lipidome with respect to breast cancerrisk, appropriate dietary intervention reproducing twoprominent features of the Mediterranean diet, i.e., a lowN6/N3 PUFA ratio and elevated monounsaturate levels, mustbe carried out in animal models and human pilot studies inthe future. If successfully implemented in an appropriatemodel of breast cancer in rat, one may speculate that theMediterranean diet would bring as much comprehensiveknowledge for dietary exposure and cancer relationships asdid the ‘‘cafeteria’’ diet for the diet and obesity relationship.The reappraisal of the role of the environmental exposure to

dietary fatty acids in breast cancer risk must be done beforepublic health applications based on the concept of cancerprevention by delay (22), either primary or secondary, can beconsidered. The lipidome may be the most appropriate meanto address this challenge. The use of the composite biomarker(OR) defined by the analysis of the lipidome provides a wayto individualize women with a high risk of breast cancer dueto dietary habits and to follow the effect of dietaryinterventions in those women. Such a new approach wouldbe even more opportune in women with a genetic predispo-sition to breast cancer. However, the universality of ourfindings cannot be assessed from a single study and aconfirmation is needed from studies in which the lipidome isderived from adipose tissue of women living in countrieswhere the breast cancer incidence and dietary fat intakediffers strongly from that of the French women.

AcknowledgmentsWe are deeply grateful to Arnaud Basdevant, Denis Hemon, andNorman Salem, Jr. for their fruitful comments and suggestions duringthe preparation of the article.

References1. Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study:

exploring genes, lifestyle and cancer risk. Nat Rev Cancer 2004;4:519 – 27.2. Doll R. The lessons of life: keynote address to the nutrition and cancer

conference. Cancer Res 1992;52:2024 – 9s.3. Willett WC. Dietary fat intake and cancer risk: a controversial and

instructive story. Semin Cancer Biol 1998;8:245 – 53.4. Malin A, Matthews CE, Shu XO, et al. Energy balance and breast cancer risk.

Cancer Epidemiol Biomarkers Prev 2005;14:1496 – 501.5. World Cancer Research Fund. Food, nutrition and the prevention of cancer:

a global perspective. American Institute for Cancer Research, WashingtonDC; 1997.

6. Bingham SA, Luben R, Welch A, Wareham N, Khaw KT, Day N. Areimprecise methods obscuring a relation between fat and breast cancer?Lancet 2003;362:212 – 4.

7. Chajes V, Bougnoux P. N-6/N-3 polyunsaturated fatty acid ratio and cancer.World Rev Nutr Diet 2003;92:133 – 51.

8. London SJ, Sacks FM, Caesar J, Stampfer MJ, Siguel E, Willett WC. Fatty acidcomposition of subcutaneous adipose tissue and diet in postmenopausal USwomen. Am J Clin Nutr 1991;54:340 – 5.

9. Kohlmeier L, Kohlmeier M. Adipose tissue as a medium for epidemio-logic exposure assessment. Environ Health Perspect 1995;103 Suppl3:99 – 106.

10. Maillard V, Bougnoux P, Ferrari P, et al. n-3 and n-6 fatty acids in breastadipose tissue and relative risk of breast cancer in case-control study inTours, France. Int J Cancer 2002;98:78 – 83.

420

Table 1. Changes in rat WAT fatty acid composition induced by either A-linolenic acid or DHA supplementation (16, 19)

Fatty acids (mol%) a-Linolenic acid DHA

Control (n = 12) Supplemented (n = 14) Control (n = 7) Supplemented (n = 6)

Saturates16:0 15.2 17.3 22.9 20.818:0 2.7 2.9 2.8 3.4Total* 19.4 21.8 27 28.9

Monounsaturatesc-18:1N9 56.9 48.3

c52.7 43.8

c

Totalb

62.4 54.7c

57.9 48.9c

N6 PUFA18:2N6 13.9 13.4 12.2 11.1Totalx 14.4 13.7 12.7 11.3

N3 PUFA18:3N3 1.3 6.9

c0.6 0.8

22:6N3 0.1 0.1 0.1 8.2c

Totalk 1.4 7.0c

0.7 9.3c

RatioN6/N3 10.3 1.9

c18.1 1.2

c

*Including: 14:0, 15:0, 17:0, 20:0, 21:0, 22:0, 23:0, and 24:0.cWithin diet, significantly different from control values (P < 0.05).bIncluding: 14:1, 15:1, 16:1, 17:1, c -18:1N7, and 20:1.xIncluding: 18:3N6, 20:2N6, 20:3N6, and 20:4N6.kIncluding: 20:3N3, 20:5N3, 22:5N3, and 22:6N3.

Diet, Cancer, and the Lipidome

Cancer Epidemiol Biomarkers Prev 2006;15(3). March 2006

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from

11. Cohen LA, Chen-Backlund JY, Sepkovic DW, Sugie S. Effect of varyingproportions of dietary menhaden and corn oil on experimental ratmammary tumor promotion. Lipids 1993;28:449 – 56.

12. Ip C, Ip MM, Sylvester P. Relevance of trans fatty acids and fish oils inanimal tumorigenesis studies. Prog Clin Biol Res 1986;222:283 – 94.

13. Karmali RA, Doshi RU, Adams L, Choi K. Effect of n-3 fatty acids onmammary tumorigenesis. In: Samuelsson B, Paoletti R, Ramwell PW,editors. Advances in prostaglandin, thromboxane, and leukotriene research.New York: Raven Press; 1987. p. 86 – 9.

14. Gonzalez MJ, Schemmel RA, Dugan L, Jr., Gray JI, Welsch CW. Dietary fishoil inhibits human breast carcinoma growth: a function of increased lipidperoxidation. Lipids 1993;28:827 – 32.

15. Sasaki T, Kobayashi Y, Shimizu J, et al. Effects of dietary n -3 to n-6polyunsaturated fatty acid ratio on mammary carcinogenesis in rats. NutrCancer 1998;30:137 – 43.

16. Maillard V, Hoinard C, Steghens JP, et al. Interaction of dietary h-caroteneand a-linolenic acid: effect on promotion of experimental mammarytumours. In: Riboli E, Lambert R, editors. Nutrition and lifestyle:

opportunities for cancer prevention. Lyon: IARC Scientific Publication no.156; 2002. p. 403 – 4.

17. Bougnoux P. n-3 polyunsaturated fatty acids and cancer. Curr Opin ClinNutr Metab Care 1999;2:121 – 6.

18. Cognault S, Jourdan ML, Germain E, et al. The effect of a a-linolenic acidrich diet on rat mammary tumor growth depends on the dietary oxidativestatus. Nutr Cancer 2000;36:33 – 41.

19. Colas S, Paon L, Denis F, et al. Enhanced radiosensitivity of rat autochtonousmammary tumors by dietary docosahexaenoic acid. Int J Cancer 2004;109:449 – 54.

20. Anderson TW. An introduction to multivariate statistical analysis. 2nd ed.New York: John Wiley & Sons, Inc.; 1984.

21. van de Vijver MJ, He YD, van’t Veer LJ, et al. A gene-expression signature asa predictor of survival in breast cancer. N Engl J Med 2002;347:1999 – 2009.

22. Lippman SM, Hong WK. The biology behind cancer prevention by delay.Clin Cancer Res 2002;8:305 – 13.

23. Hosmer DW, Jr., Lemeshow S. Applied logistic regression. New York: JohnWiley & Sons, Inc; 1989.

421Cancer Epidemiology, Biomarkers & Prevention

Cancer Epidemiol Biomarkers Prev 2006;15(3). March 2006

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from

2006;15:416-421. Cancer Epidemiol Biomarkers Prev   Philippe Bougnoux, Bruno Giraudeau and Charles Couet  Diet, Cancer, and the Lipidome

  Updated version

  http://cebp.aacrjournals.org/content/15/3/416

Access the most recent version of this article at:

   

   

  Cited articles

  http://cebp.aacrjournals.org/content/15/3/416.full#ref-list-1

This article cites 18 articles, 3 of which you can access for free at:

  Citing articles

  http://cebp.aacrjournals.org/content/15/3/416.full#related-urls

This article has been cited by 8 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cebp.aacrjournals.org/content/15/3/416To request permission to re-use all or part of this article, use this link

Research. on November 9, 2020. © 2006 American Association for Cancercebp.aacrjournals.org Downloaded from