A digital exam for hematologists

Sanford J. Shattil

Medical specialists frequently “own” or “disown” certain procedures. For example, it is my impression that at our institution many physicians do not routinely perform digital rectal examinations, leaving that task to consultant gastroenterologists, urologists, gynecologists, or surgeons. This trend to defer such a traditional part of the admission work-up appears to extend to the medical subspecialties, including hematology-oncology, where even a clearly indicated examination of the peripheral blood smear is all-too-frequently “turfed” to pathologists.1 In an ironic twist, the “digital examination” is now being reintroduced to hematologists, or at least to the readers of Blood. Until now, the journal has relied on the critical eyes of its reviewers and editors to pick up instances in which digital images have been inappropriately manipulated by authors. The ease with which such manipulation can be accomplished by today's computer programs makes Blood's reliance on others to detect and rectify these infractions inadequate. Now all manuscripts accepted for publication in Blood will be routinely inspected for inappropriate manipulation of digital images. Further explanation of this new policy is warranted.

Undoctored microscopic images of stained blood or bone marrow and high-resolution fluorescence microscopic images to detect subcellular structures, even in living cells, are arguably as visually stunning as a beautiful painting. However, these and other images obtained and analyzed by digital methods are frequently modified by authors in an attempt to increase clarity or to put experimental results in the best possible light. Efforts to do so have been facilitated by the universal availability of computer programs suitable for this purpose. While there is nothing wrong with the appropriate processing of digital images, there is something very wrong with their inappropriate manipulation, whether it be an honest but misguided attempt to clarify or a devious attempt to mislead. The latter behavior has received much attention recently due to high-profile cases of digital misconduct in the stem cell field.2,3 But the more frequent cause of the problem appears to be a lack of author education. In a recent pilot study carried out by Blood, approximately 20% of accepted manuscripts contained one or more figures with digital images that had been manipulated inappropriately, as deemed by rules and criteria spelled out in our Author Guide.4 None of these instances had been detected by reviewers or editors before acceptance of the manuscript. In each case, the authors were then contacted and primary data were examined, if necessary, by Blood staff and the editors, to understand the nature of the problem. Final publication of problematic manuscripts was frequently delayed until the issues were resolved. In almost every case, errors could be chalked up to a simple lack of understanding on the part of the authors as to the line between what is appropriate and inappropriate manipulation of digital images. In order to better meet our obligation to the scientific and medical communities, it is Blood's intention to now draw this line.

Figure 1

Bookends, circa 1943. The original photograph by the late Arthur Shattil was scanned at 200 dpi by a Hewlett-Packard scanjet 4670 scanner, and the resulting TIFF file was saved in Adobe Photoshop CS2, without further processing.

The Journal of Cell Biology and its editors, including Managing Editor Mike Rossner, have pioneered efforts to eliminate publication of inappropriately manipulated digital images. Taking its cue from them, Blood will now routinely screen all accepted manuscripts for deviations from good digital-imaging practices. To aid authors, important dos and don'ts when preparing such images are explicitly stated in our Author Guide. Many authors might benefit from viewing additional material on this subject, including explicit examples of inappropriate manipulation of images.5 Cases of deliberate misrepresentation of digital images will result in rejection of the manuscript or reversal of our previous decision to publish a paper. In such instances, hopefully rare, the authors' institution or funding agency will be notified of the infraction. Our major goals are to educate our authors and readership about acceptable practices in preparing digital images for publication and to eliminate improperly processed images from our pages. While these are but two of many important goals of Blood's editorial and publishing staffs, we aim to make the digital images that we publish as pure and as eloquent as blood itself.6