Back to Journal

SM Otolaryngology

Artificial Intelligence in Facial Plastic Surgery

[ ISSN : 2574-2418 ]

Abstract Introduction Skin Cancer Facial Plastics Conclusion References
Details

Received: 23-Apr-2024

Accepted: 29-Apr-2024

Published: 02-May-2024

Abdullah Aldaihani*

Dalhousie University, Department of Otolaryngology, Canada

Corresponding Author:

Abdullah Aldaihani, Dalhousie University, Department of Otolaryngology, Canada

Keywords

Facial Plastics; Otolaryngology; Dermatology; Artificial Intelligence; Skin Cancer

Abstract

Using Artificial Intelligence Facial Plastic surgery is a new important tool that has been implemented recently into helping surgeons, we focused in two aspect of AI uses in facial plastics which is in the use of early detection of skin lesions and the use of it in the aspect of cosmetic facial plastic surgery by developing an algorithm that determines attractive facial features most closely related to post-operative target variables.

Introduction

Understanding the combination of technology and Facial Plastic Surgery is curranty and will be very important to position Otolaryngologist and Facial Plastic surgeons at the forefront of surgical innovation, which is a hot topic to talk about nowadays. This account of the future applications of Artificial intelligence in facial plastic surgery introduce us to a subset of issues amenable to support this technology. It equips facial plastic surgeons with the knowledge to navigate technical conversations with peers, trainees, patients, and technical partners for collaboration and to usher in a new era of technology in facial plastic surgery. From the mathematical basis of Artificial Intelligence to its commercially viable applications, topics introduced herein constitute a framework for design and execution of quantitative studies that will better outcomes and benefit patients. Finally, adherence to the principles of quality data collection will leverage and amplify plastic surgeons’ creativity and undoubtedly drive the field forward [1].

As the technology community romanticizes Artificial Intelligence and packages it into commercial products for resale to the healthcare industry, we on the front lines cannot help but wonder how to stay ahead of the curve. How might a research team grappling with study design and difficult data collection best position itself to leverage open-source AI resources in the name of efficiency and better patient outcomes? How might the individual surgeon use AI tools to assist in optimizing surgical plans or improving the patient experience? This account should provide facial plastic surgeons with a practical approach to understanding the uses, limitations, and potential of artificial intelligence in research and clinical practice. Data collection in facial plastics remains a challenge as they are innately visual specialty thus are uniquely suited to embrace artificial intelligence. Before and after photos are widely publicized and are a testament to the success or failure of a plastic surgery procedure, intimately linking visual features to outcomes. Progress made in computer vision and facial recognition outside of healthcare is being leveraged to segment facial anatomy, quantify visual appeal, model procedure outcomes, and predict aging [2].

Although there is largely increasing appetite in the facial plastic surgery community for expert peer-reviewed literature on the applications of artificial intelligence in surgery, it is important to note that the democratization of access to AI resources currently makes it possible for facial plastic surgery innovation to come from within our community. Surgical skills and clinical intuition are necessary but insufficient conditions to implementing such powerful technological advances as AI in the field of plastic surgery. I believe that facial plastic surgeons will be able to navigate technical conversations with peers, trainees, patients, and technical partners for collaboration and to usher in a new era of technology in plastic surgery.

Skin Cancer

Dermatologists and facial plastic surgeons works in combination in managing treat skin lesions and cancers medically or surgically. When dermatologists detect a skin lesion or cancer on the face, they will either proceed with Mohs surgery or consult a facial plastic surgeon to perform resection the lesion or to reconstruct the defect that has been left behind after performing Mohs’ surgery. Early detection of skin cancer is a very, if not the most, important aspect to successful treatment and cure. Therefore using Artificial Intelligence to improve automatic detection of skin cancers from photos or videos is incredibly useful.

Recently, a team of researchers in Stanford University in California, USA were able to train an Artificial Intelligence software to detect and classify skin cancer with greater accuracy than dermatologists that use dermoscopes. There is another study done at Department of Dermatology at McGill University in Montreal Canada that shows the promise of Artificial Intelligence in screening for melanoma. A melanoscoan, which is Sophisticated whole-body photographic scanners have been developed that could use Artificial Intelligence to help quantify and detect dangerous skin cancers [3]. It’s this kind of Artificial intelligence that holds the bright future for skin cancer early detection and prompt treatment.

Facial Plastics

In a recent study, researchers created an automated classifier for facial beauty — trained using extracted facial features from 165 images of “attractive” female faces that were also independently graded by human referees. In this model, a decision tree algorithm assessed a set of  descriptive attributes — which in this particular investigation included different facial ratios — and attempted to determine attractive facial features most closely related to post-operative target variables. When subjected to the testing set of images, the automated classifier was just as good as humans at assessing beauty. A quantitative measurement of aesthetic improvements could not only set expectations but also discourage patients from undergoing procedures that offer marginal improvement [4].

Conclusion

The Implementation of Artificial Intelligence in the field of Facial Plastic Surgery is still in its early phases and a lot of research must be invested in this field as it will help both surgeons and patients into having the optimal care and outcome, especially in early detection of skin cancer and determining attractive facial features most closely related to postoperative target variables in an objective way.

References

1. Kanevsky J, Corban J, Gaster R, Kanevsky A, Lin S, et al . (2016). Big Data and Machine Learning in Plastic Surgery: A New Frontier in Surgical Innovation. Plast Reconstr Surg. 137: 890e-897e.

2. Jarvis T, Thornburg D, Rebecca AM, Teven CM. (2020). Artificial Intelligence in Plastic Surgery: Current Applications, Future Directions, and Ethical Implications. Plast Reconstr Surg Glob Open. 8: e3200.

3. Drugge RJ, Nguyen C, Drugge ED, Gliga L, Broderick PA, et al. (2009). Melanoma screening with serial whole body photographic change detection using Melanoscan technology. Dermatol Online J. 15: 1.

4. (2020). nnovación: Artificial intelligence is changing the face of plastic surgery News selected by América Retail: Haydée Alarcón. 2 noviembre, 20.

Citation

Aldaihani A, et al. (2024) Artificial Intelligence in Facial Plastic Surgery. SM Vasc Med 4: 2.

Other Articles

Article Image 1

Independent Predictors of Malignancy in Patients with Multinodular Goiter

Introduction: Thyroid nodule is a common clinical finding. Ultrasound (US) and Fine-needle Aspiration (FNA) are the main methods used for investigating thyroid nodules, with questionable predictive values in Multinodular Goiter (MNG) compared to Solitary Thyroid Nodule (STN) due to the presence of multinodularity.

Objective: The present study was conducted to detect the independent predictors of malignancy in patients with MNG.

Patients & Methods: Medical records of patients who were admitted for thyroidectomy at Alexandria Main University Hospital and Medical Research Institute Hospital between January 2013 and January 2016 were reviewed. Demographic and clinical data, US reports, FNA reports, and final histopathological results were recorded and analyzed by univariate and multivariate analysis. Patients with a STN, hyper- or hypo-thyroidism, previous history of surgery for thyroid cancer or those with incomplete data were excluded.

Results: Reports of 1014 patients were reviewed, 419 patients with euthyroid MNG were included in the study. Malignancy rate was 19.3% (81/419). Micro-calcification, solid consistency, ill-defined margins, and presence of suspicious cervical lymph nodes by US were statistically significant predictors. Rate of malignancy in Bethesda II was 8.9% (false negative). Malignant rate increased with increasing in Bethesda rating from Bethesda IV (17.4%) to Bethesda V (51.7%), and lastly Bethesda VI (90.9%). Multivariate analysis revealed that micro-calcifications, suspicious cervical lymph nodes and Bethesda VI FNA were the independent predictors of malignancy. Other malignancy predictors that were only significant by univariate analysis were solid component of the nodule, and ill-defined margins.

Conclusion: Based on the data presented, it may be concluded that the independent predictors of malignancy in patients with MNG were US findings of micro-calcifications and suspicious cervical LNs as well as Bethesda VI on FNA

Mahmoud Sakr¹, Heba Jaheen², Essam Gabr¹, and Ahmed Talha²

 


Article Image 1

Numerical Survey of the Different Shapes of Human

The major aim of this article is to report experimental results about the numerical quantities of the different shapes of human are ear concha as well as to classify them into defined shape groups. To the best of the author knowledge this approach has never been applied. The results are based on 2425 pictures of ears. The major results are: (1) The author has succeeded to classify all ears concha into 36 groups demonstrated in Figure 2. On the left is the absolute number of concha of a certain type where on the right is the percentage of it. (2) The most widespread concha, number 29, is that which appeared 393 times and is 16.2% of the total number. The less widespread shape, number 34, appears one time namely 0.04% of the total. In addition to the numerical results the following subjects are elaborated with respect to human’s ear: its anatomy, its different shapes as well as their relation to human’s character

Abraham Tamir1 *


Article Image 1

Diagnosis and Surgical Treatment of Mucosal Contact Point Headache: Mucosal Contact Point Headache May Not be Accurately Diagnosed before Surgery

Objective: We aimed to investigate the differences in incidence of nasal anatomic abnormalities between the patients with and without headache and the outcome of surgical treatment for the headache patients with mucosal contact point.

Background: “Mucosal contact point headache” has been defined by Headache Classification Committee of the International Headache Society. Whether the type of nasal anatomic abnormalities associates with incidence of headache and whether surgical treatment is necessary remain undefined.

Study Design: Observational study.

Methods: We recruited 107 subjects without headache and 78 subjects with refractory headache for more than 2 years. Subjects underwent high-resolution sinus CT scans and the incidence of nasal anatomic abnormalities was calculated in both groups and results were compared. An additional 25 patients underwent surgical treatment.

Results: Mucosal contact points were observed in 85.9% of patients with refractory headache and also in 80.4% of subjects without headache. The most common mucosal contact point among headache group patients was between deviated nasal septum and lateral nasal wall (55.1%) and the incidence of this type of contact point was significantly different between groups (P <0.05). An additional 25 headache patients with mucosal contact point were corrected via endoscopic surgery. Average pain scale scores decreased significantly between pre- and post-operative measurements (P<0.001). Only 44% of patients had recovered from headache 7 days postoperatively.

Conclusion: Some patients with recurrent headache and mucosal contact point may not meet the diagnostic criteria of mucosal contact point headache since pain was not resolved within 7 days after surgical treatment.

Jia Wang MS, Jin-shu Yin MS* and Hong Peng BS


Article Image 1

Analysis on Correlation between SP and NK-1R and Intranasal Mucosal Contact Point Headache

Objectives: To observe difference of Substance P (SP) and NK-1 Receptor (NK-1R) expression in tissues at contact point and non-contact point among the patients with intranasal mucosal contact point headache and speculate the role of SP and NK-1R in mucosal contact point headache.

Methods: SP and NK-1R in tissues of contact point and non-contact point among 40 patients with intranasal mucosal contact point headache were stained histologically by immunohistochemistry, and the mRNA level was detected by RT-PCR.

Results: SP was located in cytoplasm of acini epithelial cells, distributed in nasal mucosa tissues at both contact point and non-contact point. But stain intensity was significantly increased at contact point (Z=-2.554, P<0.05). NK-1R was located in cytoplasm of acinar epithelial cells, inflammatory cells and nerve fibers of contact point; only in cytoplasm of acinar epithelial cells and nerve fibers of non-contact point. Tinctorial rate of NK-1R increased significantly at contact point (χ2 =40.438, P<0.01). mRNA level of SP and NK-1R was up-regulated in nasal mucosa at contact point compared with non-contact point.

Conclusion: Distribution of SP and NK-1R, especially NK-1R, in nasal mucosal tissue at contact point was higher compared with non-contact point. It was suggested that SP and NK-1R were associated with mucosal contact point headache.

Guomin Zhao¹, Jinshu Yin²*, Hong Peng², and Jing Wang³


Article Image 1

Numerical Survey of the Different Shapes of Human Nose

The major aim of this article is to report experimental results about the numerical quantities of the different shapes of human nose as well as to classify them into defined shape groups. To the best of the author knowledge this approach has never been applied. The results are based on 1793 pictures of noses of which 403 are artworks, 498 are photographs taken by the author in Europe - Holland, Belgium and France - while visiting these places, 801 were photographed in Israel and 91 are not clear where exactly photographed. 1081 were photographs of men and 712 of women. The author has succeeded to classify all the noses into 14 groups demonstrated in Figure 2 where for each nose shape it was possible to find also an artistic demonstration. The detailed results for the number of noses versus their shape and sources: artworks and photographing places - Europe and Israel - are reported in Table 1. The most widespread nose is shape 1, the fleshy nose in Figure 2, 24.2%, where the less widespread nose is shape 4, turned-up nose, 0.45%. It should be emphasized also that shape 14 photographed in Israel is extremely unique and the only one existing among the 1793 noses that were considered. In addition to the numerical results the following subjects are elaborated with respect to human’s nose: its anatomy, its different shapes as well as their relation to human’s character and facial attractiveness. And finally it is important to mention the development of electronic/artificial noses as systems for the automated detection and classification of odors, vapors, and gases.

Abraham Tamir*


Article Image 1

Associated Factors of Unilateral Sensorineural Hearing Loss among Sudanese Patients

Background: Hearing loss is a leading cause of disability worldwide and it is a significant public health problem.

Objectives: To determine the associated factors of unilateral sensorineural hearing loss among different age groups in Sudanese patients.

Methods: 115 Sudanese patients presented to Khartoum state ENT hospitals complaining of unilateral hearing loss were included.

Results: In this study 115 patients were included. Male to female ratio 1:1.3, 49 patients (42.6%) were from Khartoum state and the rest were from different areas in Sudan. The most frequent age group was from 1-15 years representing (38.3%) with a mean age of 24.48 ± 1.5. The commonest associated factor in this study was mumps and it was found in 47 of patients (40.9%), tinnitus was the commonest presenting symptoms in unilateral Sensorineural hearing loss 81 patients (70.4%) were assessed audiologically with pure tone audiometry and 34 patients (59.1%) assessed by Auditory Brain Stem Response (ABR).

Conclusion: Unilateral Sensorineural hearing loss commonly presents during school-age. Mumps is the most common associated factor. Pure tone audiometry and auditory brain stem response are suitable method for the assessment and the diagnosis.

Mahmoud Abdelbagi Mahmoud Taha¹, Sharfi Abdelgadir Omer Ahmed²* and Muna Ahmed Abdulrahim³


Article Image 1

Cochlear Implantation in Far Advanced Otosclerosis: A Case Report

T he term far-advanced otosclerosis was introduced by House and Sheehy  to indicate clinical otosclerosis with air conduction thresholds of more than 85 dB HL, Theotosclerotic process commonly involves the otic capsule and may cause quite widespread demineralisation which leads to a progressive and often profound bilateral sensorineural hearing loss.

Rogério Hamerschmidt¹, Gislaine Richter¹, Adriana Kosma Pires de Oliveira², Mohamad Feras Al-Lahham²* and Valéria Kutianski¹


Article Image 1

Subglottic Lobular Capillary Hemangioma: A Case Report

Lobular Capillary Hemangioma (LCH) is a benign proliferation of capillary blood vessels adopting a lobular configuration. A laryngeal origin of LCH is exceedingly rare. Here, we describe a case of an 11-year-old boy presenting with a subglottic lesion, leading to a subglottic stenosis. Histopathologic findings of the lesion implicated an LCH, which was removed successfully by a coblator. This is the first report of a subglottic LCH. Physicians should be aware of this unique lesion and laryngeal LCH should be considered in diagnosing the cause of a subglottic stenosis. Additionally, coblation should be an effective treatment for laryngeal LCH.

Vinh Ly Pham Hoang¹, Nguyen Pham Dinh¹, Hoang Nguyen²*, Joseph P Kitzmiller², Linh Bui Doan Hai¹, and Huy Nguyen The¹


Article Image 1

Mutational Analysis of GJB2 Gene in Non-Syndromic Hearing Loss from Patients at Children

Purpose: Hereditary hearing loss is associated with several mutated genes; among them mutant GJB2 is the main cause. We conduct this study to provide initial information about types, rates, and influences of mutations in Vietnamese patients.

Patients and Methods: Genomic DNA is extracted from peripheral blood samples of 96 childhood patients and 100 healthy control subjects. Primers are designed to amplify GJB2 gene, encoding Cx26 protein, followed by detecting mutations with Sanger sequencing technology.

Results: Exon 2 of GJB2 gene was sequenced from all 96 samples. We detected 9 variants of GJB2 gene from 56 patients (c.235delC, c.299_300delAT, c.634T>A, c.79G>A, c.608T>C, c.368C>A, c.11G>A, c.341A>G and c.109G>A). Of which, c.634T>A is a novel variant expected to be a causing disease mutation. Variants detected in 56 cases create 14 genotypes, including 2 causing disease genotypes (c.235delC and c.79G>A / c.299_300delAT), 7 genotypes with controversial role in disease (c.109G>A; c.11G>A; c.109G>A / c.341A>G; c.109G>A / c.608T>C; c.109G>A / c.634T>A; c.79G>A / c.109G>A / c.341A>G and c.79G>A / c.341A>G / c.368C>A), and 5 genotypes containing benign single nucleotide polymorphisms (c.79G>A; c.608T>C; c.79G>A / c.341A>G; c.79G>A / c.368C>A and c.79G>A / c.341A>G / c.368C>A).

Conclusion: We have detected the types and rate of mutations appearing on protein coding region of GJB2 gene from 96 Non-Syndromic Hearing Loss children. Most of the mutations are missense heterozygous or compound heterozygous and under controversy. We suspect that in Vietnamese population, hereditary hearing loss might be caused by interaction between disturbance of GJB2 and other genes such as GJB6, SLC26A4 or 12S rRNA on mitochondria. Next generation sequencing should be use to clarify multigenic defects of hereditary hearing loss in Vietnamese children.

Nguyen Dinh¹, Dung Huu Nguyen², Tran Huyen Lam², Kieu Tho Nguyen², Chuong Dinh Nguyen², Hoang Nguyen³*, Joseph P Kitzmiller³, Vu Anh Hoang⁴, Thuong Chi Bui⁵, Toan Thanh Pham⁶


Article Image 1

A De Novo from Mutational Analysis in PAX3 in Hearing Loss Children Caused Waardenburg Syndrome Type 1 at Children Hospital-1 in Ho Chi Minh City, Vietnam

Purpose: To detect a de novo paired box gene 3 (PAX3) mutations, mutational analysis, and associated phenotypes in the Vietnamese Children patients with Waardenburg syndrome type I (WS1). We conduct this study to provide initial information about types, rates, and influences of this particular PAX3 mutations gene in the Vietnamese patients.

Patients and Methods: Three unrelated patients with suspected WS1 were selected from our Genomic extraction DNA, probe and sequence the primers, PCR technique. PAX3 were amplified by polymerase chain reaction (PCR), and then amplicons PCR were analyzed by cycle sequencing. Variations were detected, documented and recognized as “de novo” mutation. We also compared the severity of hearing impairment, phenotypically variations among these individual affected patients.

Results: Number variations of mutations in PAX3 were detected in three patients, respectively: c.955delC (Gln319fsX380), c.667C>% (Arg223Stop). One mutation proved to be de novo as their parents did not carry the mutations, and had never been found in any previous study in Vietnam. All three patients with PAX3 mutations had different iris color, fundi between their two eyes, dystopia canthorum and profound hearing loss. We report the phenotypic expression of WS1 in these three patients and explore the implications for possible genotype phenotype correlations. Sensorineural hearing loss was present in 80% of affected individuals, and spectrum of hearing loss with very high frequency.

Conclusion: De novo mutation in PAX3 has not been reported in Vietnam before. Finally, our finding has detected deletion mutation c.667>T (p.Arg223X) in PAX3 identified in three children individual with WS1. Our analyses indicated that these mutations might constitute a pathogenic-genetic-hearing loss associated with WS1. It is very clinical importance as such patients may be misdiagnosed as congenital hearing loss since it is uncommon in Southern Vietnamese population.

Nguyen Dinh Pham¹, Dung Huu Nguyen², Tran Huyen Lam², Kieu Tho Nguyen², Chuong Dinh Nguyen², Hoang Nguyen³*, Joseph P Kitzmiller³, Vu Anh Hoang⁴, Thuong Chi Bui⁵, and Toan Thanh Pham⁶