Application of deep learning algorithm in the recognition of cryptococcosis and talaromycosis skin lesions

Abstract

Fungal pathogens and infections are a growing global public health concern, with those most at risk being those with underlying health conditions or weakened immune systems. In recent years, the incidence and mortality of invasive fungal infections such as cryptococcus and talaromycosis have increased with the increase in HIV infection, the use of malignant tumours, organ transplantation and immunosuppressive age. 1,2 Cryptococcosis, an invasive fungus caused by Cryptococcus, is the third most common invasive fungal infection. 3,4 Approximately one million people are infected with Cryptococcus, and 630,000 people die each year worldwide. [5][6][7][8] Talaromycosis, also with a high case fatality rate, is an invasive fungal infection caused by Talaromyces marneffei infection, which causes

Generalized cutaneous talaromycosis (Penicilliosis) in an immunocompetent individual

“… 3 However, the presentation can vary in non-HIV patients with different morphologies. 11 According to a study by He et al, nodules and subcutaneous abscesses comprise a higher proportion of skin lesions in non-HIV patients. 9 Another study by Chan et al reported that umbilicated lesions rarely occur in non-HIV patients.…”