Machine learning-based text mining for cutaneous myiasis and potential value of an accidental maggot therapy for complicated skin and soft tissue infection with sepsis.

Background: Cutaneous myiasis, one of the most frequently diagnosed myiasis types, is defined as skin or soft tissue on a living host infested by dipterous larvae (maggots). However, bibliometric analysis of this disease remains sparse. Machine learning techniques and updated publications provide an opportunity for such an investigation. Materials and methods: All the studies were retrieved from PubMed and were processed using R software in the bibliometric analysis and latent Dirichlet allocation (LDA) topic modeling. Furthermore, the clinical management of two diabetes patients with serious soft tissue infection-associated sepsis was analyzed. Results: A total of 211 results were retrieved and 50 topics relevant to cutaneous myiasis were determined by the LDA algorithm. The topics of uncommon fly species, nasal infestation, and physician discussion of cutaneous myiasis were consistently common over the last 20 years. Case report remains one of the key features in myiasis. Four major clusters were identified, i.e., case report related, disease type and development, travel in the tropics, and skin disease. To further delve into clinical practice, the clinical features of two patients with soft tissue infection-related sepsis were demonstrated, and a distinct beneficial role of myiasis was found. The levels of white blood cell, blood glucose, and C-reactive protein in the case with cutaneous myiasis were more stable than the other case without cutaneous myiasis but with sepsis shock. Conclusion: Maggot debridement therapy may be a promising treatment and beneficial for soft tissue infection-related sepsis. The model analysis of maggot therapy and its clinical advantages shows increasing research value and possible application in future clinical practice.