Parse malformed XML
The HTML Agility Pack will parse html, rather than xhtml, and is quite forgiving. The object model will be familiar if you've used XmlDocument.
You might want to check out the answer to this question.
Basically somewhere between a .NET port of beautifulsoup and the HTML agility pack there is a way.
It's unlikely that you will be able to build an XmlDocument that has this level of malformed structure. XmlDocument (to my knowledge) requires that xml content adhere to proper nesting and closure syntax.
However, you suspect that you could parse this with an XmlReader instead. It may still throw exceptions if certain egregious errors are encountered, but according to the MSDN docs, it can at least disclose the location of the errors.
If you're just dealing with HTML, there is the HTML Agility Pack, which may serve your purposes.