The Burning Questions
While predicting how the best plastic surgery treatments will evolve, we will try to glean into the future while standing on the firm ground of what has been achieved already in the immediate past. The questions addressed here will be:
- What are some common artificial intelligence technologies currently available?
- How can they benefit individuals who want plastic surgery?
- How does artificial intelligence for plastic surgery work today?
- Will AI become more important in the future?
Plastic surgery is an ever-evolving field in medical sciences that uses different technologies to help many patients achieve the wonderful results they desire. As a specialty, it has come a long way from its early days, and so have the surgical techniques used to make people look good. Many of the aesthetic procedures we use today are a far cry from those practiced just a decade ago. However, old meets new when it comes to artificial intelligence in plastic surgery of the future.
“ What all of us have to do is to make sure we are using AI in a way that is for the benefit of humanity, not to the detriment of humanity.”
To answer the burning questions, we will start by analyzing two of the vital techniques that are now being performed as a part of the available technologies in artificial intelligence: natural language processing, facial recognition, and deep learning.
Natural Language Processing
The utilization of NLP (Natural Language Processing) has increased in recent years. Different applications check spelling and grammar on different types of word processors, autocorrect text messages, and predict future texts. The fundamental changes started mainstream about a decade ago when some data researchers developed a system for extracting data from medical records that anchored their text to the medical records.
Recently, applying NLP in reconstructive and plastic surgery has shown that public opinion about plastic surgery is increasing in this area. A large amount of data related to plastic surgery is always found within Twitter's microblogging site. While analyzing Twitter, it was seen that some of the top aesthetic plastic surgeons in the past had increasingly used different types of social media in their practice. These media platforms were used to market their plastic surgery practice and also provide some amount of patient education at the same time.
In one study, some data researchers attempted to quantify how most of the general population perceived plastic surgery. They employed a text analysis technique called ‘hedonometry’ to study tweets related to plastic surgery that occurred for 5 continuous years. Hedonometry is an awesome machine learning technique that uses an algorithm to analyze words in the context of their surrounding words. After analyzing over a million relevant tweets, the analyzed data showed that “plastic” was the most popular term. But, of all the terms, the word 'plastic' had the least favorable impression in the minds of the people. The other related terms which are not so popular in common usage like 'aesthetic', 'cosmetic', and 'reconstruction' were considered by common people to be more positive. This was even more evident when compared to the other words used to describe plastic surgery such as 'plastic surgery' and 'nip and tuck'.
In another study by a separate group of researchers, their results corroborated the earlier facts. In light of these results, it was suggested that these results could be used to make informed decisions about the title that a plastic surgeon would use - would it be an aesthetic or would it be a cosmetic surgeon. These results highlighted the future potential for such intelligent applications to shape the nature of digital marketing strategies. They also emphasized focusing on the public perception about different aspects of plastic and reconstructive surgery.
Let us look at another different case scenario. We all are aware of the benefits of having a smartphone while answering FAQs about any topic. Some data scientists developed an application that would give answers to questions about patient-related queries. This application had been rigorously trained to answer questions related to 10 selected topics that were frequently asked by the patients before their surgery. Then they conducted a study by interviewing the participants to see if they understood how their technology worked. The results showed that the participants were able to accurately determine the applications of each one of the selected topics with about a 92% accuracy rate. In addition, in about 83% of the cases, they were able to find an adequate response. The results have fuelled a hope that this amazing technology could be integrated into clinical practice. The hope is to provide an improved patient support system that would enable the surgeons to have more free time to devote to more meaningful work.
Facial Recognition
The commercialization of a brand-new technology with facial recognition is increasing day by day, as more smartphones are getting enabled with high-quality camera lenses and applications, and increasingly more smartphone users are taking advantage of this technology. The process begins by creating a pattern that is used to measure the features of a face. The measurements are then analyzed and compared with an existing database of biometric facial measurements to identify a person.
Utilizing this technology, a model was used to improve the facial care of patients by enabling it with certain characteristics. Post-operative patients were classified into groups that were beneficial in assessing patient satisfaction rates as well as setting proper realistic expectations before taking them up for surgery. This greatly enhanced doctor-patient communication, leading to more happy and satisfied patients.
Additional applications with this technology show great promise in diagnosing developmental disorders with their typical characteristics in facial proportions and shapes. Also, it would help in forming a predictive analysis of the success of complicated craniofacial surgeries, and cosmetic plastic surgery treatments.
Deep Learning
Over the last two decades, research in deep learning for plastic surgery has progressed by leaps and bounds as unstructured data is abundantly available after they are collected through widely used informational technologies. In particular, aesthetic plastic surgeons regularly collect images of the patients before as well as after their plastic surgery procedures. These are subsequently made available to a large national database, thus creating an enormous amount of primary data sources.
A study was conducted on facial plastic surgery procedures. The researchers found that an ANN (artificial neural network) was able to accurately classify the status of nose correction surgeries, also called rhinoplasty, in 85% of the photos that were tested. The results had a good level of both specificity and sensitivity. The accuracy was nearly comparable to that of the residents and treating doctors of that hospital.
In another Deep Learning application developed for the recognition of Melanoma, ANN had been utilized to identify melanomas in the photographs taken from the biopsied lesions. The images were obtained through smartphones attached with a dermoscopic lens. The ANN could identify the characteristics of the lesions diagnosed with melanoma, and it was able to learn straight from all the data it was receiving.
This same application was again tested by providing it with high-quality images published in dermatology journals. The probability of this application in diagnosing melanoma through the images was evaluated. When compared to the clinical accuracy of the treating clinicians, the application was found to have an equivalent level of accuracy. This study showed the Deep Learning application’s potential as a diagnostic and decision-making tool for physicians, which could help treat skin cancer in the early stages and prevent the dreaded metastasis of malignant melanoma and other skin cancers.
Conclusion
Artificial intelligence is still early in its developmental cycle. Still, this technology has been shown to help the best surgeon-scientists accurately predict the type of treatment a new patient would require. In the future, it is widely expected that AI will be able to assess the appearance of a face and body with unparalleled precision and accuracy.