Thought I'd come back and post the solution I used in the end for the benefit of anyone else searching.

I used Drew Noakes' Metadata Extractor library: https://github.com/drewnoakes/metadata-extractor-dotnet

Here's some example code to extract IPTC data. In a nutshell, you pass a physical path for your file to ImageMetadataReader.ReadMetadata to get a handle on the set of directories that hold each type of metadata, then use MetadataExtractor.Formats to pull the specific directory you need (e.g. EXIF, IPTC, DICOM). Then instance MetadataExtractor.Formats.xxxx.xxxxDescriptor , where the xxxx refers to your chosen metadata type, and you can then extract the specific values for metadata tags. The exception is Adobe XMP data, which is stored a bit differently, but our man Drew Noakes has you covered: https://github.com/drewnoakes/xmp-core-dotnet

@using MetadataExtractor    
var image = Model.Value<IPublishedContent>("Image");
<!-- get physical path of image -->
var physicalPath = Server.MapPath(image.Url);
<!-- get IPTC metadata -->
var directories = ImageMetadataReader.ReadMetadata(physicalPath);
var IptcDir = directories.OfType<MetadataExtractor.Formats.Iptc.IptcDirectory>().FirstOrDefault();
if (IptcDir != null) 
    {
         var descriptor = new MetadataExtractor.Formats.Iptc.IptcDescriptor(IptcDir);
         // get IPTC values
         String IptcCopyright= descriptor.GetCopyrightNoticeDescription();
         String IptcSource= descriptor.GetSourceDescription();
    }