有谁用poi读过word文档吗?格式不能保持一致,大家看看!

<%@ page import="java.io.*"%>
<%@ page import="org.textmining.text.extraction.*"%>
<%@ page contentType="application/msword;charset=gb2312"%><!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML>
    <HEAD><TITLE></TITLE>
        <META http-equiv="Content-Type" content="application/msword; charset=gb2312">
    </HEAD>
<body>
     <%
        response.setContentType("application/msword");
        FileInputStream in = new FileInputStream ("E:\\angel\\1刘桃英\\1刘桃英.doc"); 
        WordExtractor extractor = new WordExtractor(); 
        String str = extractor.extractText(in); 
        out.println(str); 
      %>     
</body>
</html>能读出数据来,但是格式不正确,我加了response.setContentType("application/msword");
还是不行,我查了半天,好象没有这种问题的帖子,大家看看,有没有谁解决过这种问题,帮忙指点一下!