Commentary on: The Use of a Novel Artificial Intelligence Platform for the Management of Rhytids

See the Original Article here.

We have read with great interest the recent article by Yoelin et al, titled “The Use of a Novel Artificial Intelligence Platform for the Management of Rhytids.” ¹ The authors used a mobile phone application, Appiell (Los Altos, CA), to evaluate the severity of glabellar frown lines. Appiell, which is based on artificial intelligence (AI) technology, has demonstrated its ability to evaluate wrinkles reliably and accurately.

Facial wrinkles, the first and most visible morphological change of aging, are among the features most targeted for rejuvenation treatments. The evaluation of facial wrinkles is therefore of great importance in assessing the efficacy of these treatments. Currently, a variety of wrinkle evaluation systems—both subjective and objective—exist.

Among the subjective systems, the 5-level Wrinkle Severity Rating Scale² and the 4-level Fitzpatrick Wrinkle Scale³ are the most widely used to assess deep facial wrinkles, such as nasolabial folds. Other examples include the Lemperle Scale,⁴ which is used to evaluate the effect of nasolabial fold correction, and the 5-level Validated Assessment Scales,⁵ which are used to evaluate the severity of crows’ feet, frontal lines, and glabellar frown lines. All subjective evaluations are based on observers’ experience, and interrater agreement is sometimes not guaranteed. Subjective methods are usually more applicable for evaluating deep facial wrinkles or folds but are less suited for assessing fine lines.

Objective systems typically depend on specific imaging tools or lighting conditions. For example, manual wrinkle selection on digital images followed by saturation value measurement and comparison with unwrinkled areas can be used to evaluate perioral wrinkles.⁶ The 3-dimensional fringe projection method based on the FotoFinder Mediscope (FotoFinder Systems, Inc., Columbia, MD) can also be used to analyze fine facial lines.⁷ The SWIRL method analyzes the severity of wrinkles from photographs taken under obliquely angled light.⁸ All these objective evaluations are based on algorithms and usually demonstrate high interinstitutional agreement. However, current objective evaluation systems require either large equipment or special lighting conditions. Moreover, follow-up patient visits can only occur at clinics, which restricts the popularity of these systems.

With the various wrinkle evaluation methods used by different researchers, evaluating the efficacy of different treatments between studies is becoming increasingly difficult. There is a need for a single standardized method for measuring the severity of facial wrinkles and for comparing the efficacy of different treatments. A reliable, capable, and portable platform is a potentially perfect evaluation system for this purpose.

The Appiell AI platform was designed as an objective automatic evaluation tool for facial wrinkles. This platform represents a great advance in medical assistance technology because it allows patients to take follow-up photographs at home and assesses wrinkle severity without manual intervention. Consequently, it helps to reduce patient loss to follow-up and makes detailed follow-ups possible, which helps doctors better understand the treatment time course. Despite these advantages, the system still needs to be perfected in the following aspects:

• Accuracy: The accuracy of an evaluation is the most important feature of a tool like this. Although Appiell showed good agreement with assessments by experienced raters, its accuracy and sensitivity still need improvement with additional training, as stated by the authors. Training should include patients with different Fitzpatrick skin types, deep wrinkles, and fine lines in different regions of the face.

• Standardization: Appiell is installed on a mobile phone, which makes it convenient for patients to take follow-up photographs at the required time. However, standardization of the photographs is a concern. Patients may take photographs at different distances, from different angles, and under different lighting conditions. In addition, patients may not be able to perform maximal facial expressions, as required. These variations could potentially influence the evaluation results. Therefore, if AI is to be used as a popular evaluation tool, the problem of standardization must be solved.

• Interface: An evaluation tool designed for patients should help them adhere to the follow-up schedule. However, it may be impossible to train each patient to operate Appiell in the manner of a professional doctor. Therefore, a detailed and user-friendly interface is crucial for the platform. The authors have not described this in detail in the manuscript. In our opinion, more assistive features, such as detailed user instructions, guidance on photography, lighting condition detection, and follow-up reminders, should be added. These features can also help standardize follow-up photographs.

We believe that advances in AI will have a profound effect on medicine. These technologies are helping to acquire and process medical data, improve patients’ experiences, and provide medical professionals with new insights.

Disclosures

The authors declared no conflicts of interest with respect to the research, authorship, and publication of this article.

Funding

The authors received no financial support for the research, authorship, and publication of this article.

REFERENCES