Comments on “Three-Dimensional Imaging for Breast Augmentation: Is this Technology Providing Accurate Simulations?”

I read with interest the article by Roostaeian and Adams titled “Three-Dimensional Imaging for Breast Augmentation: Is This Technology Providing Accurate Simulations?”1 The authors state that 3-dimensional (3D) imaging is more than 90% accurate in predicting postoperative breast volume. They also state that there is greater than 98% accuracy in the differential for surface contour. My experience with the same system has been different. Readers looking through the article will see that the simulations are far from 90% accurate. Perhaps the numbers measured by the system are “accurate,” but I have found that showing patients actual postoperative photographs of similar patients to be more useful in managing expectations. 3D simulation does not always accurately show the type of result normally achieved. (A) Preoperative 3D image of a 36-year-old woman. (B) Simulated image using 400g implant. (C) Actual result at 2 months postoperative using a 400g implant. This 47-year-old woman had 365 g of smooth Inspira (Allergan, Markham, Ontario, Canada) silicone gel-filled breast implants placed in a subglandular pocket. Actual 2D conventional photographs of preoperative (A, C) and just over 2-month postoperative views (B, D) of the patient. Simulated images are significantly different from the actual result, and patients often find the simulations less than satisfactory during the consultation. These images show simulations using different sized breast implants (270 g and 365 g), and the barely discernable difference shows that the software is not good at helping patients determine their preferred size. The three-dimensional frontal views: (A) shown at consultation, (B) the simulated view with 270 g implants, (C) the simulated view with 365 g implants, and (D) the postoperative view with 365 g implants. The three-dimensional oblique views: (E) shown at consultation, (F) the simulated view with 270 g implants, (G) the simulated view with 365 g implants, and (H) the postoperative view with 365 g implants. Continued. The authors evaluated 20 patients for primary subpectoral dual plane augmentations. The images were taken by a patient coordinator unaware of the study, and the results were evaluated by a blinded independent researcher familiar with the 3D imaging system. The authors state that “the time spent performing 3D imaging never exceeded a few minutes and actually was less than the time required for conventional photography.”1 This also is different from my experience. Figure 1 shows that the system is not particularly accurate when the simulated image is compared to the actual postoperative image. This adds additional time during the consultation to reassure patients that the simulated image is not accurate. Additional time is also needed to manually adjust the image to achieve a more realistic outcome. To compare results, I have continued to perform conventional photography. It would have been helpful to readers if the before photographs had also been included. Better simulations can be achieved when the default landmarks are adjusted—and this takes time. However, the simulations in the study were generated using the software's default simulation, which was not manually adjusted. In the original article, figure 4, for example, looks very different from the simulated to the actual result. The size is visibly smaller and the cleavage is much wider in the simulated result. The actual image also shows that both the upper and lower breast borders have been expanded significantly beyond the simulated image. Figure 5 shows similar differences. Figure 6 shows similar problems, but only if the viewer realizes that the images have been mistakenly reversed in the article (the scars are visible on the “simulated” photo). My experience with the system makes me suspect that Figure 7 is also reversed. The authors state that Figure 10 is the most accurate in their series, and it might be close to 90% accurate from a visual standpoint (which is the only one that patients see), but the lateral view shows enough of a difference to require some explanation during the consultation. The authors state that Figure 11 shows the least accurate simulation. Although the postoperative image shows a high-riding implant that is still not centered behind the nipple, this perhaps suggests that 3 months is not a long enough follow-up at 3 months from when the implant is placed under the muscle. I suspect that the result may actually look more like the simulated image with a longer follow-up visit. At one point in the paper, the authors caution that: When greater fill volumes are chosen, the extra volume becomes apparent in the upper pole of the breast; however, the default simulation software is set up to demonstrate an optimally filled breast and is not able to demonstrate this increase in upper-pole fullness without manual manipulation. It is important to discuss this limitation preoperatively with patients who desire a higher-than-optimal fill volume and are relying on the simulations to choose a particular size.1 “Optimal” for the authors is not extending beyond the base diameter of the breast horizontally, but I would argue that optimal for many patients is an implant that expands the base diameter in any direction needed and not just vertically as advocated by so-called dimensional planning. I believe that the patient shown in Figure 8 illustrates that implants larger than 270 g and 205 g would have given her a better result, allowing the lateral aspect of the breast to at least reach the anterior axillary line. A software program that does not actually follow the base diameter of the simulated implant but forces implants of all sizes into the existing breast base diameter falls short when educating patients. The authors conclude by saying, “It is important to understand how closely simulations resemble actual postoperative results and to communicate this to patients considering breast surgery. In this study, the simulations generated by the Vectra M3 Imaging System provided a high degree of accuracy for breast volume (90%) and contour (98%).”1 I believe that the authors’ conclusions are misleading and contradict their statement of caution. The 3D imaging software has potential and may at some point help patients determine size; but at this stage, I believe the simulations are far from 90% accurate. Figure 2 and 3 are of the same patient in my practice with Figure 2 being her before and after images using a 2D camera and Figure 3 being before, simulated and after images of the same patient using a 3D camera. It can be seen that the two simulated images using 100 cc volume difference between the implants does not help with size choices because they both look the same. It is also shows that the simulated shape is quite different from the actual result achieved. Dr Hall-Findlay receives royalties from QMP, Elsevier, and Lippincott.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.