Humanitarian Facial Recognition for Rare Craniofacial Malformations

Dear Sir, We read the recent publication titled “Humanitarian Facial Recognition for Rare Craniofacial Malformations” by Hennocq et al1 with a lot of interest. The authors describe a humanitarian use case for artificial intelligence (AI) that includes the identification of birth defects from pictures of conflict areas and natural disasters.1 When medical care is limited, finding these children quickly is important. In one instance, the authors were able to diagnose Apert syndrome using their facial recognition software. Although the findings are preliminary, the potential future applications are promising because it could help increase access to the millions of people born with craniofacial malformations. Below, we explore the potential for broader application of this technology and identify the challenges that must be overcome to facilitate such large-scale implementation. After achieving optimal diagnostic accuracy for craniofacial malformations through algorithm optimization, the subsequent area of focus should be the deployment of this software on mobile smartphone device platforms. This strategic shift would enhance the software’s usability by frontline humanitarian workers due to the inherent portability of these devices. Prior successful transitions from desktop applications to mobile platforms exist, such as Google Lens. The healthcare domain also offers numerous examples, including the utilization of mobile phones for diagnosing surgical site infections in Rwanda.2 Furthermore, facial recognition software could leverage recent advancements in hardware components, particularly graphics and neural processing units, that would allow faster and more accurate diagnosis. Beyond its application in humanitarian contexts, such software deployment on mobile phones has the potential to alleviate access barriers to craniofacial care within low- and middle-income countries, where an estimated 10.8 million individuals have craniofacial malformations.3 Craniofacial care professionals would see their workload reduced, and this benefit cannot be overstated. In Africa for example, the distribution of plastic surgeons is 0.0073 plastic surgeons per 100,000 people compared with 2.18 plastic surgeons per 100,000 people in the United States.4 Within this context, nongovernmental organizations specializing in craniofacial interventions, such as Operation Smile, Smile Train, and Mercy Ships, could experience significant advantages. The proposed software’s ability to expedite screening processes would be particularly valuable during their short-term surgical missions. These programs are often challenged by time constraints, and incomplete phenotypic evaluation during initial screening can unfortunately lead to surgical cancelations. The implementation of this AI-based screening tool has the potential to streamline the process, freeing up valuable time for a more comprehensive evaluation of complex cases. This, in turn, could lead to a significant increase in the number of patients receiving definitive surgical care. AI-based algorithms assessing facial dysmorphia can be an integral part of the preoperative screening, thus reducing and effectively enhancing screening days and contributing to proper logistic preparation before surgical camps. More importantly, the use of AI-based algorithms can facilitate training between local and nongovernmental organization teams before surgery. AI-based algorithms can also play a role in preoperative planning by analyzing patient-specific data and providing personalized surgical recommendations, ultimately leading to more effective and efficient surgery. Furthermore, AI-based algorithms can also assist in postoperative assessments by analyzing facial dysmorphia and providing detailed information on the success of the surgery, which can help surgeons make the necessary adjustments and improve future surgical procedures. In conclusion, AI-based innovations should be introduced in low-middle-income countries to test their effectiveness in addressing unusual dysmorphia and to train them. Moreover, their use should be extended to international organizations. DISCLOSURE The authors have no financial interest to disclose in relation to the content of this article.