Upload binary file using XMLHttpRequest and FileReader whyle setting custom content-type and boundary

I'm trying to upload a (DICOM) binary file to the server using XMLHttpRequest and FileReader.

According to DICOM-standard as Content-Type muss be multipart/related;type=application/dicom defined and in the Request Payload must be the Content-Type:application/dicom again, I have managed to build that structure somehow with this code:

    let boundary = Math.random().toString().substr(2);
    let reader = new FileReader();
    reader.readAsBinaryString(fileList[0]);

    reader.onload = function(e) {

        var xmlHttpRequest = new XMLHttpRequest();
        xmlHttpRequest.open("POST", "myurl", true);
        var dashes = '--';
        var crlf = "\r\n";
        if ( fileList[0].type == '' ){
            filetype = 'application/dicom';
        } else {
            filetype = fileList[0].type;
        }
        let content = e.target["result"];
        var data = dashes + boundary + crlf + "Content-Type: " + filetype + crlf + crlf + content + crlf + dashes + boundary + dashes;
        xmlHttpRequest.setRequestHeader("Content-Type", "multipart/related;type=application/dicom;boundary=" + boundary+";");
        xmlHttpRequest.setRequestHeader("Accept", "application/dicom+json");
        xmlHttpRequest.send(data);
    }; xmlHttpRequest.send(data);
        };

The problem with that approach is that it seems that XMLHttpRequest makes an UTF-8 encoding and that's corrupting the binary data (see this post).

• The questions are how can use reader.readAsArrayBuffer() and set the body content-type and boundary? or
• How can I prevent XMLHttpRequest to make UTF-8 encoding?

My second question is how to handle the big files (about 1 TB) with this approach?