The literature review on predictive models and factors for keratoconus detection highlights the potential of machine learning techniques to improve efficiency in diagnosing and monitoring this corneal disease. Machine learning models, such as random forests and convolutional neural networks, have proven to be valuable tools with remarkable combined sensitivity and specificity in keratoconus detection. However, many studies do not follow established reporting standards, underscoring the need to improve the clinical translation of these models. This review also addresses the prevalence and risk factors associated with keratoconus, including genetic, environmental, and behavioral factors. Familial aggregation of keratoconus and its prevalence in different populations are discussed, as well as genetic contributions and associated comorbidities. Research suggests that integration of demographic and risk factor data, as well as combining data from multiple imaging devices, could improve machine learning models for keratoconus detection. Furthermore, the need for global collaboration and improvement in the quality of research is emphasized.