Option B is the correct answer. Implementing dynamic data masking in Azure SQL Database allows you to mask the sensitive data automatically based on the user's role or IP address. This provides a flexible and efficient way to protect sensitive data without requiring changes to the application code. Option A is incorrect because partial indexing is used for query performance optimization, not for data masking. Option C is incorrect because Azure Information Protection is used for classifying and protecting documents and emails, not for data stored in databases. Option D is incorrect because implementing a custom stored procedure for data masking can be complex and may not provide a consistent security model.