Phil B.FontanarosaMD, Deputy EditorIndividualAuthorStephen J.LurieMD, PhD, Fishbein FellowIndividualAuthor
To the Editor: We agree with Dr Barton and
colleagues1 that there are few data to evaluate
the clinical breast examination (CBE) as a clinical test. However, we believe
the likelihood ratio (LR) they estimate for the presence of a mass in the
breast is at least an order of magnitude too low.
The Health Insurance Plan of New York (HIP)2
data, which the authors cite as the reference for Table 6 in their article,
is a report of results of 449 biopsies of screen-detected breast masses.3 In the HIP study, 20% of masses were malignant
and, as Mushlin4 pointed out, none of the
clinical characteristics of a mass, once a mass was identified, were sufficient
to lower the probability of malignancy to below 10%. Since posterior probability=prior
probability × LR, then it is also true that LR=posterior probability/prior
probability; in this case, LR=20/0.35=57 for the presence of a mass.
Goodson III WH, Moore II DH. Clinical Breast Examination for Detecting Breast Cancer. JAMA. 2000;283(13a):1687-1689. doi:10.1001/jama.283.13.1685