Correspondence

Benign Breast Disease and Breast Cancer

N Engl J Med 2005; 353:1856-1858October 27, 2005

Article

To the Editor:

The article on the risk of cancer in patients with benign breast disease by Hartmann et al. (July 21 issue)1 contains some questionable assumptions. “Atypical ductal hyperplasia” and “lobular hyperplasia,” shown in Figure 1E and Figure 1F of the article, do not belong in the category of benign breast disease. Experienced pathologists know that the differentiation of the so-called atypical ductal hyperplasia from an intraductal carcinoma of the breast is extremely difficult in most cases and impossible in some.2 Lobular hyperplasia, shown in Figure 1F, is closely related to lobular carcinoma in situ, recognized for generations as a precursor lesion of breast cancer.3 Thus, these two types of lesions should be classified as precursors of mammary carcinoma, as demonstrated by a much higher rate of subsequent breast cancer in patients with these lesions than in patients with truly benign disorders. One would have expected an even higher rate of cancer in patients who were apparently not treated, except for excision of the lesion. The issue of lesion classification and risk is even more important if this information is shared with patients, as suggested by Elmore and Gigerenzer, in the accompanying editorial.4

Leopold G. Koss, M.D.
Susan Fineberg, M.D.
Montefiore Medical Center, Bronx, NY 10467
lkoss@montefiore.org

4 References

1
Hartmann LC, Sellers TA, Frost MH, et al. Benign breast disease and the risk of breast cancer. N Engl J Med 2005;353:229-237
Full Text | Web of Science | Medline

2
Rosai J. Borderline epithelial lesions of the breast. Am J Surg Pathol 1991;15:209-221
CrossRef | Web of Science | Medline

3
Page DL, Kidd TE Jr, Dupont WD, Simpson JF, Rogers LW. Lobular neoplasia of the breast: higher risk for subsequent invasive cancer predicted by more extensive disease. Hum Pathol 1991;22:1232-1239
CrossRef | Web of Science | Medline

4
Elmore JG, Gigerenzer G. Benign breast disease -- the risks of communicating risk. N Engl J Med 2005;353:297-299
Full Text | Web of Science | Medline

To the Editor:

In their editorial, Elmore and Gigerenzer offer insightful comments on the risk of risk communication. They state that risk can be communicated in terms of absolute risks, “which foster insight” and are “clearer.” They recommend the use of the number needed to treat as a “clear” form of communication. We agree that decisions usually should be based on absolute risks, but to our knowledge, there is no evidence that absolute risk reduction or its reciprocal number needed to treat is clear or fosters insight. Rather, there is ample evidence that laypersons and professionals misunderstand number needed to treat.1-4

Elmore and Gigerenzer claim that “the use of relative risks suggests greater effects than truly exist, whereas . . . absolute risks . . . prevents this misunderstanding.” Both relative risk and absolute risk, however, are arithmetic reformulations of the same numbers and are therefore equally correct. In order to establish that one of them is more misinterpreted than the other would require a gold standard for correct interpretation. To our knowledge, there is little evidence that relative risks are more misunderstood than absolute ones.

Ivar S. Kristiansen, M.D., Ph.D.
Palle M. Christensen, M.D., Ph.D.
Dorte Gyrd-Hansen, Ph.D.
University of Southern Denmark, DK-5000 Odense, Denmark
ivarsk@c2i.net

4 References

1
Sheridan SL, Pignone MP, Lewis CL. A randomized comparison of patients' understanding of number needed to treat and other common risk reduction formats. J Gen Intern Med 2003;18:884-892
CrossRef | Web of Science | Medline

2
Sheridan SL, Pignone M. Numeracy and the medical student's ability to interpret data. Eff Clin Pract 2002;5:35-40
Medline

3
Halvorsen PA, Kristiansen IS. Decisions on drug therapies by numbers needed to treat: a randomized trial. Arch Intern Med 2005;165:1140-1146
CrossRef | Web of Science | Medline

4
Nexoe J, Kristiansen IS, Gyrd-Hansen D, Nielsen JB. Influence of number needed to treat, costs and outcome on preferences for a preventive drug. Fam Pract 2005;22:126-131
CrossRef | Web of Science | Medline

Author/Editor Response

Drs. Koss and Fineberg correctly note that atypical ductal hyperplasia and atypical lobular hyperplasia form a morphologic continuum with ductal carcinoma in situ and lobular carcinoma in situ, respectively. The microscopical distinction can be difficult and is based on a number of criteria including severity of cytologic atypia, degree of alteration in architectural cellular relationships, and extent of lesions. Not all these features can be represented in a single high-power view. The epidemiologic results in our study support the accuracy of our pathological assessment. Namely, the relative risk for atypical hyperplasia in our study is 4.2, entirely consistent with other major reports.1,2 If our atypia group had included significant numbers of cases of ductal carcinoma in situ and lobular carcinoma in situ, we would expect relative risks in the range of 10 or greater.3

We do not understand the comments by Koss and Fineberg regarding precursor lesions. Many experts in the field consider atypia, especially atypical ductal hyperplasia, to be a precursor lesion to breast cancer. Our data show a slightly increased risk of ipsilateral cancer in women with atypia, compatible with the hypothesis that some atypias represent direct precursors.

Daniel W. Visscher, M.D.
Amy C. Degnim, M.D.
Lynn C. Hartmann, M.D.
Mayo Clinic College of Medicine, Rochester, MN 55905

3 References

1
Dupont WD, Page DL. Risk factors for breast cancer in women with proliferative breast disease. N Engl J Med 1985;312:146-151
Full Text | Web of Science | Medline

2
London SJ, Connolly JL, Schnitt SJ, Colditz GA. A prospective study of benign breast disease and the risk of breast cancer. JAMA 1992;267:941-944[Erratum, JAMA 1992;267:1780.]
CrossRef | Web of Science | Medline

3
Allred DC, Mohsin SK, Fuqua SAW. Histological and biological evolution of human premalignant breast disease. Endocr Relat Cancer 2001;8:47-61
CrossRef | Web of Science | Medline

Author/Editor Response

It should be alarming that most physicians and patients do not understand the crucial numbers that medical research provides for them. Collective innumeracy is an impediment to the efficacy of evidence-based medicine and shared decision making. Research on how to frame numbers so that physicians and patients understand them is still scarce.

Dr. Kristiansen and colleagues ask whether there is evidence that relative risks confuse people more than absolute risks or number needed to treat. In fact, several studies have shown that statements referring to relative risks make benefits appear greater to patients as well as to physicians and policymakers, whereas statements referring to absolute risks or number needed to treat make benefits appear smaller. For example, people's willingness to accept medication or screening is substantially greater when the benefits are framed in terms of relative risks.1,2 Physicians' treatment decisions can also be affected by the way in which research results are summarized.3 One notable study even found that health authorities were much more likely to fund a proposal for cardiac interventions and breast-cancer interventions when the benefits were framed in terms of relative risks than they were to fund an identical proposal with benefits that were framed in terms of absolute risks and number needed to treat.4

Kristiansen et al. cite studies by Sheridan et al., who conclude that discussions of absolute risks and number needed to treat may not foster insight. Yet the apparently discrepant results from Sheridan et al. might be due to the researchers' unusual expression of these numbers.5

Physicians rarely quantify risks or benefits when talking with patients. Instead, we tend to use qualitative terms that might be ambiguous, such as “You are at increased risk of breast cancer” or “You are at high risk of breast cancer.” The way patients interpret such qualitative statements can vary widely, with a corresponding effect on the medical care that we provide. Thus we suggest using numbers along with qualitative terms to minimize ambiguity.

At issue is transparency and clarity in communications with patients. All of the foregoing enforces one conclusion: we need to overhaul our educational systems. Evidence-based medicine must declare war on innumeracy. Now.

Joann G. Elmore, M.D., M.P.H.
University of Washington School of Medicine, Seattle, WA 98195-9780

Gerd Gigerenzer, Ph.D.
Max Planck Institute for Human Development, 14195 Berlin, Germany

5 References

1
Malenka DJ, Baron JA, Johansen S, Wahrenberger JW, Ross JM. The framing effect of relative and absolute risk. J Gen Intern Med 1993;8:543-548
CrossRef | Web of Science | Medline

2
Sarfati D, Howden-Chapman P, Woodward A, Salmond C. Does the frame affect the picture? A study into how attitudes to screening for cancer are affected by the way benefits are expressed. J Med Screen 1998;5:137-140
Web of Science | Medline

3
Forrow L, Taylor WC, Arnold RM. Absolutely relative: how research results are summarized can affect treatment decisions. Am J Med 1992;92:121-124
CrossRef | Web of Science | Medline

4
Fahey TJ, Griffiths S, Peters TJ. Evidence based purchasing: understanding results of clinical trials and systematic reviews. BMJ 1995;311:1056-1059
CrossRef | Web of Science | Medline