When you manipulate your images so they’ll be good figures for publication, did you ever think about how much manipulation is used to make them look good and when you might over-step on the manipulation to make them more “significant”?  There are numerous examples in publishing where authors needed to make a deadline, get a promotion, receive more grant funding, or just become known for their findings where images were “enhanced” selectively to prove a result.  For example, authors have copied parts of a gene sequence and pasted them into another one to show results, or they have added more evidence of abnormal cells in a microscopic slide to make their case.  In some cases, these are naive researchers who have demonstrated the results they hoped with the original figure, but want to make sure everyone can see their findings clearly.  In some famous cases, the authors / researchers committed an unethical act that resulted in setbacks to legitimate research.  I have not seen any evidence of this in figures submitted to JDI, but want to try to set some guidelines for authors / researchers.

If your figures come from medical images and you want to enhance the entire image using window / level, magnification, edge enhancement, unsharp masking and other common tools, this is usually OK.  When you select a region of interest and only manipulate that section and you do not describe what you did to the image accurately, then that is usually not OK.  So what I believe should be our rule of thumb for manipulation of medical images for publication is this.  If you can perform the action using a typical PACS workstation, then that action is permissible and probably desirable.  Please do window / level to make the best possible figure.  Do crop the figure so we can focus on the important parts.  If edge enhancement or unsharp masking help make the region of interest more clearly defined, then use it, but apply these filters to the entire image.  I recommend you make the image look as good as possible on the PACS workstation prior to capturing it and resizing it using the tools described in Tony Siebert’s e-tutorials.  See www.siimweb.org/e-tutorials for Tony’s work.