Converting HTML with img tags with base64_encoded data in Lexical

I have HTML that I need converting to Lexical JS. The html contains img tags with base64 encoded values.

But when I run through this code the data for the img tag is none existent.

  useEffect(() => {
    const parser = new DOMParser();
    const dom = parser.parseFromString(content, 'text/html');
    const structure = analyzeNode(dom.body);
    setNodeStructure(structure);

    const editor = createEditor({
      namespace: 'HeadlessEditor',
      onError: (error) => console.error(error),
      nodes: [
        HeadingNode,
        QuoteNode,
        ListNode,
        ListItemNode,
        TableNode,
        TableCellNode,
        TableRowNode,
        LinkNode
      ]
    });

    // Convert HTML to Lexical nodes
    editor.update(() => {
      const root = $getRoot();
      root.clear();

      const nodes = $generateNodesFromDOM(editor, dom);
      console.log('Generated nodes:', nodes);
      nodes.forEach(node => {
        console.log('Appending node:', node);
        root.append(node);
      });
    }, {
      onUpdate: () => {
        // Get editor state after update is complete
        const editorState = editor.getEditorState();
        console.log('Editor State:', editorState);

        const jsonContent = editorState.toJSON();
        console.log('Editor JSON Content:', jsonContent);
        setEditorContent(JSON.stringify(jsonContent, null, 2));
      }
    });

  }, [content]);

The HTML DOM contains something like this:

{
      "type": "element",
      "tag": "p",
      "children": [
        {
          "type": "element",
          "tag": "img",
          "attributes": {
            "src": "data:image/png;base64,iV...

After conversion, it it just this:

      {
        "children": [],
        "direction": null,
        "format": "",
        "indent": 0,
        "type": "paragraph",
        "version": 1,
        "textFormat": 0,
        "textStyle": ""
      }

How do I make it so that Lexical JS has my image tags?

Note that when I use the Editor I can see the image – it is base64_encoded and displaying it.

Thank you!