A Retrospective Photometric Study of 82 Published Reports of Mastopexy and Breast Reduction

Sir:Figure“A Retrospective Photometric Study of 82 Published Reports of Mastopexy and Breast Reduction”1 is a seriously flawed article deserving critical analysis. The main premise appears to be an effort to disprove claims that “Numerous techniques have been reported that claim to preserve or improve breast projection and upper pole fullness ….” Dr. Swanson describes a photometric technique which, when applied to photographs published in medical journals, appears to disprove these assertions. The author begins with combining disparate articles from various authors using roughly similar but often significantly different techniques. For example, 25 articles by 25 different surgeons are reviewed and lumped together as examples of vertical scar mastopexy/reduction surgery. In the vertical scar breast reduction group, many of these techniques are markedly different. Grouping these different techniques on the one similarity they share—the final orientation of a vertical scar—is an oversimplification that fails to recognize the significant differences among these techniques. Dr. Swanson's hypothesis is not even agreed on by the authors of the articles he uses to support his research. The fact is that some of these authors do not believe that their own techniques increase upper pole fullness, whereas others believe that these operations merely prevent a loss of upper pole fullness.2 Dr. Swanson attempts to disprove a hypothesis that is not even held by all the surgeons whose articles he reviewed. It is difficult to criticize a technique for failing to accomplish a goal it was never designed to achieve. The correct way to prove or disprove a hypothesis is to review a series and analyze the results. Rather than reviewing a series, the author attempts to study the photographs from the published articles of other surgeons. The author admits that some of the published photographs do not fit his measurement tool and were therefore excluded from the study. Clearly, this is an abrogation of the primary rule of medical research that authors must not arbitrarily decide which data they choose to include or exclude. Selection bias is a serious flaw in this experimental design. The author should have devised a measurement tool that would have allowed him to use all the data. Instead, he disregarded data that did not conveniently fit his photometric measurement system. This selection bias represents one of the critical flaws of this article. Another serious shortcoming of this article is the inability to correctly derive measurements from the photographs studied. Measurements can only be accurately derived from a photograph if there exists within the photograph a standard rule. An example of such a standard rule would be the patient holding a ruler or the placement of a premeasured mark on the patient's body. Recognizing that a measured rule was not available, the author took an arbitrary measurement—the average upper arm length for an adult woman—as the standard rule from which to measure breast dimensions. The application of mean upper arm length in different patients renders all the data of questionable value. More significantly, however, the reference length was presumably taken from the before-and-after photographs by observations of surface anatomy such as a point on the shoulder or elbow flexion crease. However, it is well known that patients experience weight loss after breast reduction surgery; this has been reported as high as 11 pounds at long-term follow-up.3 As weight changes, the locations of these landmarks change, thus making measurements dependent on these observations useless. The author states that there were obvious magnification differences between preoperative and postoperative photographs and that the computer program used was then able to correct these magnification differences by matching “for orientation and size.” In Figure 1 of Dr. Swanson's article, he shows an example taken from an article originally authored by Dr. Nahai. In the accompanying legend, Dr. Swanson claims that “the preoperative images are … size- and orientation-matched to the postoperative images ….” If this is the case, the computer program on which this article is based failed to achieve what was required of it. The before-and-after frontal views clearly demonstrate marked differences in magnification, with an interacromial measurement on the preoperative photograph of 45 mm versus 40 mm in the postoperative photograph.Fig. 1: The difference between the relative sizes of two objects in the same photograph taken with lenses of different focal lengths is demonstrated. The relative sizes of the coffee mugs change as the lens focal length changes despite the fact that the scene was not changed. Any attempt to correct the magnification differences between the photographs will not correct the perspective distortion. It is obviously not possible to use Dr. Swanson's technique to measure the preoperative and postoperative relative lengths of these items without standardization of lens focal length. In fact, it appears that a significant change has occurred in the relative sizes of the objects because of alteration of lens focal length alone. Photographs were obtained using (above) 20-mm focal length, (center) 55-mm focal length, and (below) 100-mm focal length.The most serious flaw in this article, however, is the lack of appreciation of the phenomenon common to photography known as “perspective distortion.” Perspective distortion is the alteration in appearance that occurs in subjects of photographs taken with lenses of different focal lengths or with the camera at a different distance from the subject (Fig. 1). Unless a photograph is taken from exactly the same distance from the subject or using a lens of exactly the same focal length, the subject within preoperative and postoperative photographs will have a different appearance, resulting in distortion of any measurements taken directly from these photographs for comparison. Dr. Swanson recognizes that differences in magnification exist in the preoperative and postoperative photographs he studied, and he attempted—albeit unsuccessfully as outlined above—to correct these magnification differences. However, correcting magnification differences does not correct perspective distortion, which will result in significant proportional differences in the measurements between areas of the photograph that are in the center of the photographic field (such as the breast) and areas of the photograph closer to the edges (such as the upper arm). The vast majority of digital cameras sold have non–fixed focal length lenses.4 It is likely that many, if not most, surgeons use non–fixed focal length cameras to capture before-and-after photographs in their practices. The typical photographer/surgeon stands in front of his or her subject and then uses the zoom to frame the photograph. This makes it highly unlikely that the before-and-after photographs were taken at exactly the same distance and using the same focal length lens setting. Obviously, no standardization of distance from the subject or lens focal length exists in Dr. Swanson's study. The author's inability to control these factors makes all the photographs susceptible to perspective distortion and thus renders the measurement data obtained from these photographs useless. With useless data, the conclusions reached by the author are groundless. This study analyzes poor data with measurement tools of no value to arrive at unsubstantiated conclusions. Attempting to apply scientific measures to the field of aesthetic surgery is a laudable goal. However, in the absence of good science, the only thing worse is bad science. Frank Lista, M.D. Jamil Ahmad, M.D. The Plastic Surgery Clinic, Mississauga, Ontario, Canada DISCLOSURE The authors have no financial interest in the content of this communication.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.