In medicine, academic performance is evaluated quantitatively, by the sheer number of papers. Promotions are granted according to the publication output, often counted in hundreds. Doctors love to throw around sentences like “I have more than 300 papers”, or 400, or 500, which is meant to put their clinician colleagues in their place. Such high-throughput publishing culture heavily relies first of all on the system of “honorary” authorships, i.e. those utterly unrelated to the actual research become co-authors solely by the virtue of their higher hierarchy status or their being friends or even family. Other questionable tactics are salami-publishing (where even a tiniest dataset or analysis is stretched and re-used again and again for several consecutive publications) and good old self-plagiarism, or text re-use. To avoid being busted for double-publishing, clever doctors combine both methods to achieve some variation between their overlapping publications. At the end of the day, where others would publish only one measly paper, these tricksters get two, three and much more. Guess whose publication list will look more impressive, and who will climb the academic career ladder then. Another danger of self-plagiarism: it can lead to “proper” plagiarism and poor quality research. When untreated in one scientist, it also becomes contagious.
The US-based Croatian radiologist Hedvig Hricak is a master of this strategy (see my earlier report on self-plagiarism), in fact she is so good that her junior colleagues seem to have been learning it from her. Hricak leads the radiology department at the prestigious Memorial Sloan Kettering Cancer Center (MSKCC) in New York. Despite her previous retraction for text and data re-use, as well as her solid record of self-plagiarized papers, MSKCC decided to fully support Hricak: their research integrity officer declared in January 2016 that he had “not found any issues that have not already been satisfactorily addressed”. Not only this, MSKCC even moved to employ certain clinicians from under Hricak’s patronage, despite or maybe even because of their own unethical publishing and retractions.
In fact, Hricak seems to feel that safe that she apparently started to plagiarize texts written by others as well. The paper by Kircher, Hricak and Larson, “Molecular imaging for personalized cancer care”, Mol Oncol. 2012, contains at least five sections which are very similar to the contents of the 2010 book chapter “National Cancer Clinical Trials System for the 21st Century: Reinvigorating the NCI Cooperative Group Program”, by Nass, Moses and Mendelsohn. A certain whistleblower alerted the journal to “plagiarism and copyright violation”, this detailed analysis I make available here. The Editor-in-Chief (EiC) of Molecular Oncology, Julio Celis, informed the whistleblower some days ago that he gave the authors 30 days to reply.
While plagiarism is usually a quite straightforward reason for retraction, self-plagiarism is much trickier. The “Text Recycling Guidelines” on the website of the Committee on Publication Ethics (COPE) are muddy at best. Basically, COPE seems to leave it up to the individual editors to decide whether any given self-plagiarized paper warrants a correction or even a retraction. Some editors are honest, some less so, and some bend the rules for those with power and connections. This is why one self-plagiarized Hricak paper was retracted and others were saved, occasionally with COPE’s explicit approval. Hricak is after all a very influential and well-connected clinician; even those “just” working for her enjoy an impressive degree of editorial well-wishing.
Presented below are some cases of Hricak’s and her colleagues’ self-plagiarism and salami-publishing (as forwarded to me by the whistleblower), and how the journal editors chose to condone them.
- Lakhman, Katz, Goldman, Yakar, Vargas, Sosa, Miccò, Soslow, Hricak, Abu-Rustum and Sala, “Diagnostic Performance of Computed Tomography for Preoperative Staging of Patients with Non-endometrioid Carcinomas of the Uterine Corpus”, Ann Surg Oncol. 2016
Lakhman, Yakar, Goldman, Katz, Vargas, Miccò, Zheng, Moskowitz, Soslow, Hricak, Abu-Rustum and Sala, “Preoperative CT-based nomogram for predicting overall survival in women with non-endometrioid carcinomas of the uterine corpus”, Abdom Imaging (now titled Abdominal Radiology), 2015
For these two papers, the whistleblower informed both journals that
“the patient cohort is identical, the imaging readers are identical, the imaging features are identical, the tables on interobserver agreement, baseline characteristics and CT features are all the same”.
The whistle-blower’s detailed critique is available here. Deb Whippen, senior managing editor at Annals of Surgical Oncology, replied then as follows:
“Both journals carefully reviewed the articles and the claims you brought forth according to the COPE guidelines. The Editors investigated the matter and concluded that while the data studies share the same patient population, the questions are sufficiently different to warrant two publications. Each journal serves a different reader base and there is a valid need to educate both audiences. We consider this matter to be officially closed”.
The two journals issued on January 29th 2016 a similar-sounding editorial note, here and here, to inform their readers that self-plagiarism and data reuse is an ethical and actually useful practice, as long as a bit of “different rationales and criteria” is thrown in.
- Vargas, Burger, Donati, Andikyan, Lakhman, Goldman, Schöder, Chi, Sala, Hricak. “Magnetic resonance imaging/positron emission tomography provides a roadmap for surgical planning and serves as a predictive biomarker in patients with recurrent gynecological cancers undergoing pelvic exenteration”. Int J Gynecol Cancer. 2013
Burger, Vargas, Donati, Andikyan, Sala, Gonen, Goldman, Chi, Schöder, Hricak. The value of 18F-FDG PET/CT in recurrent gynecologic malignancies prior to pelvic exenteration. Gynecol Oncol. 2013
Donati, Lakhman, Sala, Burger, Vargas, Goldman, Andikyan, Park, Chi, Hricak. Role of preoperative MR imaging in the evaluation of patients with persistent or recurrent gynaecological malignancies before pelvic exenteration. Eur Radiol. 2013
The whistleblower alerted all three journals as well as COPE in January 2014, the criticisms were similar as above: identical patient populations, hypotheses, results “either identically published or slightly modified”. Detailed criticisms and communications are available here.
This is how the journals replied. The response of Gynecologic Oncology was understandable: they declared that theirs was “the original paper in the group of three” and that they are discussing the issue “with the Editors-in-Chief of European Radiology and the International Journal of Gynecological Cancer”. Uziel Beller, EiC of the latter journal, explained to the whistleblower:
“There is a grain of truth in your detailed observations however, in our opinion, we do not believe it warrants or justifies further action. We are certain that Dr. Vargas [Hricak’s subordinate at MSKCC, -LS] and colleagues appreciate and understand the significance of ethical publishing and will keep the rules and regulations in mind when producing original work in the future”.
The EiCs of European Radiology Beth Y. Karlan and Adrian Dixon drew the whistleblowers attention to Beller’s letter and stated on behalf of their journal:
“Following internal investigation (which looked at the detailed submission letters explaining the background of these papers), discussion between the relevant editors and with the lead author who is ultimately responsible for their governance, it has been concluded that these separate publications were justified on the grounds that all three papers investigated a different technique or combination of techniques and tested different hypotheses.
While nobody condones excessive publications on any one group of patients, it is inevitable that large and growing data sets will be analysed in numerous different ways and that such analyses will become more common in the future. It is of course essential that the authors and editors make full reference to other papers published on related topics and similar data sets, especially so that the data do not get counted as separate studies in any subsequent meta-analysis”.
The common author to all above mentioned papers is the Albanian radiologist Evis Sala, at that time lecturer and honorary consultant radiologist at the Addenbrooke’s Hospital in Cambridge, UK. Sala had to retract one of her papers just 3 months after it was published in the journal The Obstetrician & Gynaecologist in January 2012. The retraction note for Moyle et al, “Magnetic resonance imaging of uterine abnormalities” gave only unspecified “factual errors and out-of-date terminology” as rationale.
- Blažić, Maksimović, Gajić and Šaranović. “Apparent diffusion coefficient measurement covering complete tumor area better predicts rectal cancer response to neoadjuvant chemoradiotherapy”. Croat Med J. 2015
Blažić , Lilic and Gajić. “Quantitative Assessment of Rectal Cancer Response to Neoadjuvant Combined Chemotherapy and Radiation Therapy: Comparison of Three Methods of Positioning Region of Interest for ADC Measurements at Diffusion-weighted MR Imaging”. Radiology, 2016
Again the whistleblower laments a “major overlap in patient population”, and that the tables and 2 out of total 3 figures are the same (detailed criticism here). Neither paper references the other, yet both claim that “to our knowledge, this is the first study” to apply the described methodology in order to predict tumor response to chemoradiotherapy.
Interestingly, the academic EiC of the journal Radiology, Herbert Kressel, was originally appointed as Editor-in-Chief by none other than Hricak herself. As I reported before, Kressel took great pains to cover up the issues of Hricak’s self-plagiarism and data re-use in his journal. He also disregarded peer criticism of yet another questionable Hricak publication in Radiology, namely Zhang et al. 2009. Korean radiology professor Byung Kwan Park of Samsung Medical Center in Seoul Korea was a staunch critic of that Hricak paper and demanded its retraction due to “too many critical flaws”. Kressel published Park’s Letter to Editor in the same year, followed by a Response from the first and corresponding author Jingbo Zhang from MSKCC (link to both here).
In 2013, a second correction followed. Zhang, the original “guarantor of integrity of entire study”, disappeared as author completely (together with another co-author, Liang Wang). Instead, a totally new first and corresponding author from MSKCC’s Department of Radiology was surprisingly pulled out of the hat. Correcting a paper by a new corresponding author, with the original one banished, four years after the paper was first published and criticized, is certainly an unconventional event in academic publishing. Not only this, Kressel even recruited his own statistics expert to save the Hricak and Co paper. However, that expert apparently overlooked certain mathematical irregularities which the whistleblower alerted the journal to (detailed analysis here). Most strikingly, Hricak’s team proclaimed that the doubling time of patients’ tumours “ranged from −248 to 72 days (mean, 474 days; median, 811 days)”. Thus, both the mean and median values were beyond the range of the doubling time, a mathematically somewhat improbable situation. But surely not to Kressel and the journal’s publisher, the Radiological Society of North America (RSNA), which Hricak used to preside over.
To me, Park wrote in regard to Hricak’s paper:
“There are a lot of critical flaws in it. None of the authors can answer correctly the questions that readers pose. Their doubts will not be quenched even though any erratum will be published”.
It seems therefore not all of Hricak’s is that awe-inspiringly excellent that the medical research community should be thankful on their knees for her evangelical self-plagiarism. But with Sala, Blažić and others, this practice already went into next generation, all due to Hricak’s protective tutelage.
Welcome to the wonderful world of clinical publishing.