Retraction Watch

Tracking retractions as a window into the scientific process

Cancer researcher has dodged accusations for decades (and has a new correction)

with 33 comments

Today isn’t a great day for Carlo Croce, chair of the department of cancer biology and genetics at The Ohio State University (OSU).

The New York Times has a lengthy article detailing the misconduct accusations that have swirled around Croce for years. We’ve covered many, but The Gray Lady obtained documents that show there have been many more.

The story mentions a 2013 letter from Ohio State University to pseudonymous whistleblower Clare Francis (which we reported on in 2014), acknowledging Francis’s allegations against Croce. However, in the letter, an administrator said OSU saw no reason to investigate Croce.

The story didn’t stop there, as the Times reports:

Since 2014, another critic, David A. Sanders, a virologist who teaches at Purdue University, has made claims of falsified data and plagiarism directly to scientific journals where more than 20 of Dr. Croce’s papers have been published.

(For more on Sanders from our archive, click here and here.) These problems are starting to be reflected in the scientific record, the paper notes:

As a result of complaints by Dr. Sanders and others, journals have been posting notices of problems with Dr. Croce’s papers at a quickening pace. From just a handful of notices before 2013 — known as corrections, retractions and editors’ notices — the number has ballooned to at least 20, with at least three more on the way, according to journal editors.

We’ve tracked many of these notices: In January, after one of his papers was retracted for an image “mismatch,” Croce accused his co-authors of fraud. Last year, he retracted a News & Views from a Nature journal after accusing them of censoring his writing, and he lost another paper last year in the Journal of Biological Chemistry for image duplication (Croce denied to us having any role in the paper).

By our count, Croce has logged five retractions (1, 2, 3, 4, 5) and has multiple papers questioned on PubPeer, many of which he’s coauthored with Alfredo Fusco, a cancer researcher in Italy who has nine retractions and is in the midst of an investigation for scientific misconduct.

As a result of queries from the Times, OSU now says it’s reopening its investigation:

At Ohio State, officials said they were unaware of Dr. Sanders’s charges against Dr. Croce until asked about them for this article. Now, in the wake of those and other questions from The Times, the university has decided to take a new look to determine whether it handled those cases properly. “The university is instituting an independent external review,” a spokesman, Christopher Davey, said in a statement, adding that the review “is not an indication that we have discovered any evidence of scientific misconduct or other issues raised in your inquiry.”

Croce’s problems didn’t begin in 2013. In fact, he has faced accusations for decades, as the Times reports:

[In the 1990s], Dr. Croce and a colleague faced federal allegations that they had submitted false claims for payment of grant money for science that was never carried out and that was to be overseen by a scientist who had, in fact, left the United States and gone to Italy. After the case was combined with a second fraud investigation, a civil settlement forced Thomas Jefferson University in Philadelphia, where Dr. Croce worked at the time, to pay $2.6 million to the government. Neither Dr. Croce nor the other parties to the settlement admitted any wrongdoing in the case.

In 2007, the U.S. National Institutes of Health withdrew a grant proposal by Croce, saying it was too similar to one submitted only months earlier by a more junior colleague at OSU. Then, a tipster accused an official in Croce’s lab of using his grant money for personal trips, and a former colleague accused Croce of scientific misconduct. But, as the Times notes:

Dr. Croce was cleared in both of those cases.

Meanwhile, we’ve got another recent notice to report — we’ve found a rather long correction for a 2014 PNAS paper due to “overlap” in the text with content from five other papers, which were unattributed and “reproduced without quotation marks.”

It’s familiar language that was used in a 2015 correction we covered that Croce received for a 2013 paper in PLOS ONE. Then and now, we asked the same question: When does “overlap” become “plagiarism?” (Croce has denied plagiarism allegations in the past.)

PNAS handled the “overlap” in the same way that PLOS ONE did. The journal opted for a milder and less common course of action—a correction that includes references and quotation marks around the overlapping text, instead of a retraction. Here’s the correction notice for “Protective role of miR-155 in breast cancer through RAD51 targeting impairs homologous recombination after irradiation,” which has been cited 41 times since it was published in 2014:

The editors wish to note that, after this article was published, a reader noticed that some fragments of text and some sentences in the abstract, significance statement, introduction, and discussion overlap with text from other articles and were reproduced without quotation marks. No concerns have been raised about the originality of the research or about the results and conclusions. Some text in the abstract and the significance statement overlaps with text from ref. 19.

19-Wang Y, Huang JW, Calses P, Kemp CJ, Taniguchi T (2012) MiR-96 downregulates REV1 and RAD51 to promote cellular sensitivity to cisplatin and PARP inhibition. Cancer Res 72(16):4037–4046.

Some text in the introduction overlaps with text from the following references:

4-Nogueira A, Catarino R, Medeiros R (2011) DNA damage repair and cancer: The role of RAD51 protein and its genetic variants, DNA Repair and Human Health, ed Vengrova S (InTech, Rijeka, Croatia).

19-Wang Y, Huang JW, Calses P, Kemp CJ, Taniguchi T (2012) MiR-96 downregulates REV1 and RAD51 to promote cellular sensitivity to cisplatin and PARP inhibition. Cancer Res 72(16): 4037–4046.

21-Le Calvez-Kelm F, et al. (2012) RAD51 and breast cancer susceptibility: No evidence for rare variant association in the Breast Cancer Family Registry study. PLoS ONE 7(12): e52374.

25-Bartel DP (2009) MicroRNAs: Target recognition and regulatory functions. Cell 136(2): 215–233.

In the discussion, paragraph 5 overlaps substantially with text from ref. 44 and is now shown with quotation marks below.

“The DNA damage response process is frequently impaired in aggressive breast cancers, as a consequence of either mutation or deregulation of critical components, such as BRCA1, ataxia telangiectasia mutated (ATM), and p53. Whereas p53 mutations have high frequency, mutations in genes coding for ATM or BRCA1 represent rare events in sporadic breast cancers. Functional impairment of BRCA (‘BRCAness’ phenotype) has been frequently observed in sporadic breast cancers [40, 41]. Existence of alternative mechanisms limiting the expression and functions of either BRCA1 or its regulators, such as ATM, characterize this phenomenon. Among these mechanisms, aberrant activity of miRNAs plays a critical role, as reported for miR-146 and miR-182, which directly target BRCA1 [42, 43] or miR-181a/b targeting ATM [44].” (44)

44-Bisso A, et al. (2013) Oncogenic miR-181a/b affect the DNA damage response in aggressive breast cancer. Cell Cycle 12(11):1679–1687.

We asked PNAS why it didn’t consider this to be a case of plagiarism. The journal declined to comment further on the specific case, but the executive editor Diane Sullenberger explained that the journal subscribes to the Office of Research Integrity (ORI) definition of plagiarism, “particularly regarding how much text is in question and whether it has been misappropriated or materially misleads the reader about the authors’ contributions.” The editor pointed us to this section of the ORI definition:

As a general working definition, ORI considers plagiarism to include both the theft or misappropriation of intellectual property and the substantial unattributed textual copying of another’s work…

Substantial unattributed textual copying of another’s work means the unattributed verbatim or nearly verbatim copying of sentences and paragraphs which materially mislead the ordinary reader regarding the contributions of the author. ORI generally does not pursue the limited use of identical or nearly-identical phrases which describe a commonly-used methodology or previous research because ORI does not consider such use as substantially misleading to the reader or of great significance.

The extent of the overlap in this case is not entirely clear.

Croce did not respond to questions about the new correction, but he has previously denied plagiarism allegations, explaining that OSU had also cleared him of a plagiarism charge from an anonymous whistleblower (a claim we couldn’t verify).

He also denied the recent misconduct allegations to the Times:

During an interview in October, and in a later statement, Dr. Croce, 72, denied any wrongdoing, said he had been singled out in some of the accusations simply because he was a prominent figure, and largely placed the blame for any problems with figures or text on junior researchers or collaborators at other labs.

It seems like Croce will have more opportunities to defend his record. The Times asked four experts to review images from a 2005 PNAS paper on which Croce is last author about WWOX, a gene that may function as a tumor suppressor; all said they believed they were duplicates:

After being contacted by The Times, the journal’s editor, Ms. Sullenberger, consulted with Dr. O’Brien at Berkeley and other analysts. The expert who in 2014 thought the duplication was unlikely, she said, now “accepts the new analyses because of their sophisticated nature.”

The journal, Ms. Sullenberger said, is now planning to issue a notice to readers about concerns regarding the WWOX paper.

This isn’t the first time OSU has had to wade through misconduct allegations. Pharmacy professor Terry Elton was sanctioned in 2012 for manipulating dozens of figures — but only after OSU reopened its initial investigation, which had concluded the problems were due to “disorganization,” not misconduct.

The institution also recently launched an investigation into one of the co-authors of a now-retracted (and heavily criticized) paper that suggested first-person shooter games can train people to become better marksmen.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

  • Morty March 8, 2017 at 2:31 pm

    Croce has over 1000 published articles (listed at PubMed). I just randomly opened one of the more recent top publications from Croce lab (Sun HL et al., Cancer Cell. 2016 Nov 14;30(5):723-736. doi: 10.1016/j.ccell.2016.10.001) and here is what I found after 2 min:

  • LV March 8, 2017 at 3:20 pm

    What a sad story about money and influence and the role of universities. I feel especially for the trainees whose (hopefully) legitimate work some may see as tainted by association. It’s interesting what Croce has to say about his colleagues (‘“Scientists are like cows,” he said. “They like to follow something which has already been established.”’) considering that his view of massive effects of miRNAs is in some ways the “established doctrine” in the field.

  • Mike Briggs March 8, 2017 at 3:50 pm

    NYT says “My sense was, Carlo Croce’s too big to make findings of misconduct on,” Dr. Dahlberg said in an interview. “It just wasn’t going to happen.” Why should any NIH-funded researchers be too big to make findings of misconduct on? Are there other researchers who are too big to make findings of misconduct?

  • Steven Clarke March 8, 2017 at 4:38 pm

    It is of interest to see PNAS referred to a December 1994 document from the Office of Research Integrity that stated “ORI generally does not pursue the limited use of identical or nearly-identical phrases which describe a commonly-used methodology or previous research because ORI does not consider such use as substantially misleading to the reader or of great significance.” I wonder whether the ORI still feels so accepting of plagiarism some 22 years later?

  • Aimless Drifter March 8, 2017 at 6:17 pm

    All language consists of commonly used phrases used without attribution, as they are generally deemed to be of public domain. You led your comment with the phrase, “It is of interest”. This is is a copied, unattributed phrase. However, it does not constitute plagiarism as no reasonable person would give you credit for inventing such a statement. Highly technical language is often reused over and over, as there are only so many ways to describe certain constructs. This is especially true in math/stat, as there are only so many ways to describe concepts, in a mathematical manner, such as a Gaussian distribution, and the precision of the mathematics precludes rewriting things “in your own words”.

    I have not looked at the papers in questions, but the general principle the ORI is advancing is sound. Were it not, virtually all scientist would be guilty of plagiarism.

    • Muster Mark March 9, 2017 at 10:44 am

      You are right. This is the reason why the ORI< with limited resources, chose to not waste resources on this particular kind of lapse. The passage quoted above is entirely factual. There are only so many ways to state these observations. One could of course argue about whether the choice of these particular facts, with prior work described in this articular order, to the exclusion of other related results, constitutes plagiarism.

    • Steven Clarke March 9, 2017 at 1:30 pm

      Plagiarism is obviously not commonly used phrases. And it is not that hard to avoid – put things in your own words and/or provide a reference – even in a methods section! If you don’t, the mistaken assumption may be made that you thought up the method yourself. The ORI may have distanced themselves from making a federal case of out “minor” plagiarism, and journal editors may use this as a crutch, but at best this is sloppy science.

      • Aimless Drifter March 9, 2017 at 4:37 pm

        Carefull now. Do you cite Carl Friedrich Gauss every time you use OLS regression? Do you cite Fisher every time you use ANOVA? Do you cite Robert Recorde every time you use the plus or equal sign?

        Many methods are so commonly used, that no reasonable person would attribute them to you. Lack of attribution in these cases does not constitute plagiarism. This is what the ORI statement is all about.

        • Steven Clarke March 9, 2017 at 5:58 pm

          This is my problem with the ORI statement – it may give students the idea that it may be OK to plagiarize just a little bit, or to not cite relevant papers. If you take it to the extremes that you mention of plus and minus signs, does that mean that nothing really has to be cited because we don’t cite plus and minus signs? Seems like then we should just throw in the towel about research integrity!

          • Aimless Drifter March 9, 2017 at 7:02 pm

            This is an excellent point. But it is not about the ORI statement itself, but rather addresses a concern that students may misinterpret or twist the ORI statement to their advantage. Nevertheless, the ORI policy/statement is necessary, for if you maintain, as some do, that any unattributed copying of methods or text from others constitutes plagiarism, then virtually anyone can be found guilty of this offense. The ORI should not set an impossibly high standard, that almost no-one can attain, and then selectively enforce it when the find someone they don’t like. The ORI statement on plagiarism helps protect against this. In summary, plagiarism is more challenging to determine than one might initially expect.

  • fernandopessoa March 9, 2017 at 5:06 am

    After the case was combined with a second fraud investigation, a civil settlement forced “Thomas Jefferson University in Philadelphia, where Dr. Croce worked at the time, to pay $2.6 million to the government. Neither Dr. Croce nor the other parties to the settlement admitted any wrongdoing in the case.”

    Why did they pay $2.6 million to the government?

    • herr doktor bimler March 9, 2017 at 1:13 pm

      Presumably the government was the victim of the fraud.

  • J March 9, 2017 at 5:29 am

    From the article:
    “In fact, Ohio State decided not to convene a formal investigation committee, in part because “there is very little data for the committee to review” — a reference to the failure of the scientists to locate almost any of their original data.”

    The lesson to be learned: No data, no problems!

    • LV March 9, 2017 at 9:28 am

      In light of the concerns, if the data have gone missing, the papers must be retracted. It’s that simple.

      • J March 9, 2017 at 11:54 am

        I completely agree. If there is no data, there is no research and no paper. But unfortunately, journal editors usually don’t view it as that simple.

    • MRY March 9, 2017 at 5:24 pm

      Fabrication of research data needn’t original data! Concealing evidence of data falsification by removing original data – also no data! Are there problems?

  • Steven Clarke March 9, 2017 at 6:44 am

    Interesting how PNAS will defend poor practices among its authors until pushed in a corner!

  • fernandopessoa March 9, 2017 at 7:34 am

    “One other interesting tidbit from the piece: Croce, a native of Italy, told the paper that he travels professionally more often than he’s on campus in part because he disdains “the lack of culture in Columbus.”

    • Len Dobens March 15, 2017 at 7:28 pm

      Wolverine fan

  • Richard Goldstein, Esq. March 9, 2017 at 12:06 pm

    The Times article exposes a major problem in the processes for investigating research misconduct, i.e., too much discretion is given to institutions in these matters. For good or ill, ORI is simply not equipped – except in the most extreme cases – to ensure institutions run their processes fairly. The limited nature of of ORI (or NSF) oversight has two negative consequences. One consequence, as shown by the Times article, is that ORI is struggles to ‘police’ a major university that has a strong interest in protecting its own ‘superstars.’ The other consequence is a situation not covered by the Times article but one that is a very real problem for scientists who are not ‘superstars,’ i.e., when an institution tramples on the rights of an accused researcher, ORI is not in a position to protect the interests of the accused. As a lawyer who often represents scientists accused of misconduct, I have found that appeals to ORI to ‘rein in’ unfair or excessive institutional investigations are rarely heeded. Fortunately, the Department of Justice, which enforces the false claims act and the criminal statutes, has the power to ensure institutions do not allow their ‘superstar’ researchers to produce flawed research. Unfortunately, there is little or no government protection for the junior scientist who is abused by the institution.

    • fernandopessoa March 10, 2017 at 3:59 am

      “As a lawyer who often represents scientists accused of misconduct”. That is self-advertising.

      “I have found that appeals to ORI to ‘rein in’ unfair or excessive institutional investigations are rarely heeded”. Please provide some evidence of that statement.

    • Anonymous March 10, 2017 at 4:53 am

      Exactly. The institution usually has a strong COI and should only be able to contribute to the investigation not to lead it.

    • Dani March 13, 2017 at 3:04 pm

      Mr Goldstein presents the fundamental conundrum. The University is solely interested in maintaining the credibility of “superstars,” who often know the rules that they can bypass to advance in the University environment. Meanwhile, those who decry falsification of credentials and data by these superstars are vilified both locally and Nationally, because they adhere to the fundamental rules that govern academic medical research. In effect, the rights of the accuser are often trampled to afford privileges for those that are (rightfully) accused.

  • fernandopessoa March 10, 2017 at 4:24 am
  • Alan R. Price March 11, 2017 at 11:12 pm

    In response to the debate above, by Stephen with Aimless and Muster, yes, the ORI continued to follow this ORI “what-is-not-plagiarism” federal policy for decades. Making findings of research misconduct, to be published in the Federal Register, NIH Guide, and ORI Website forever, for minor copying of literature background, methods, and work with collaborators seemed inappropriate and much too time-consuming, given huge difficult cases of falsification and fabrication that ORI did and continues to handle.

    I helped guide the creation and implemented the ORI policy on what plagiarism is not for this federal agency, published in the December 1994 ORI Newsletter and kept on the ORI website — which was confirmed by my successor as ORI Associate Director for Investigative Oversight on Page 4 of the September 2007 ORI Newsletter

    However, we did make 20 ORI findings involving plagiarism in my tenure 1989-2006 (8 of the respondents were debarred from federal funding for plagiarism, often accompanied by falsification of the copied data). I described these ORI case in a paper in Plagiary 1:1-11 (2006) – see;view=fulltext

  • fernandopessoa March 22, 2017 at 10:45 am

    “On Thursday, the president of Ohio State University said he stands behind the school’s decision to not discipline Croce.”

    “President Michael Drake says the school asked a team of people “with national reputations” in the field to look into the allegations and review Ohio State’s response to the misconduct claims.

    “And all the things that they found showed that our policies were appropriate, and we followed them appropriately,” Drake said on All Sides With Ann Fisher. “So, that’s the part that we really focused on. I think it turned out like we expected it would.”

    Drake said the team’s report was being finalized and would be publicly available when it was finished.”

  • Post a comment

    Threaded commenting powered by interconnect/it code.