Joyce Shoffner would never have predicted that Duke University, an institution she revered and at one time worked for, would put her in a breast cancer clinical trial testing a fraudulent technology.
“They advertised publicly that this science offered an 80 percent cure rate,” Shoffner said. “To have the type of cancer I had, I was just going to do that, there was nobody that was going to stop me, because this was what I was told and this was what I believed was going to happen.”
In July 2008, Shoffner became patient No. 1 in the trial that promised to choose the best therapy for the unique characteristics of her disease. Alas, the groundbreaking genomic predictors pioneered by Anil Potti and his mentor Joseph Nevins, which the trials were testing, would turn out to be fraudulent.
Now, Shoffner—a 68-year-old Raleigh resident, and formerly a biomedical photographer—has the sad honor of being one of only two living plaintiffs in the patients’ lawsuit against Duke.
Shoffner’s voice, which combines the monotone of pain with southern lilt, will not be heard from the stand at the Durham County Superior Court.
The case was settled—she is prohibited from disclosing the terms, but now otherwise she is free to talk.
“It’s always been about genomics and cell lines and datasets, but not about people. All of it was based on people, but the people were kept in the dark,” Shoffner said. “Why did we not at least have honesty? Why was there not some integrity with this when it came to the human life? Why could decent human beings do this to other human beings?”
Shoffner said she did not learn about Potti’s fraudulent data until November 2010—over two years after she first joined the trial. Altogether, 117 patients were treated in Duke’s three phase II clinical trials. Shoffner was one of the 56 patients in the neoadjuvant breast cancer trial.
Some of these patients were seduced by the televised promise Potti made in a Duke commercial.
“The goal is to be able to tell patients with cancer that I’m not just a cancer doctor, I’m here to treat your particular cancer,” Potti said in the ad. “The way to get to that is to do prospective clinical trials. So what I would say to patients is to inquire about prospective clinical trials that use genomic testing to try to determine whether they’re getting the right chemotherapy option or not.”
Potti and Duke’s promises filled Shoffner with hope that her tumor would “melt away.”
“The trials and the exciting hope of the groundbreaking science were all presented to me verbally first,” Shoffner said. “A researcher, doctor, had come up with a way of treating personal cancer—my own cancer, not general cancer, but mine, my tumor.
“They were going to look at the DNA of my tumor, and from that Duke and Dr. Potti would be able to tell which chemo was correct for my type of cancer. They were saying that with this treatment, these tumors—they used the words—‘would just melt away.’
“It wasn’t until I had already agreed to be part of this groundbreaking trial that they then brought someone in to have me sign a consent form.
“By that point my mind was already made up.”
The promise that she could randomly be placed in the genomically guided arm of the trial was the key reason Shoffner joined the trial, she said.
To get in, Shoffner needed a second biopsy, one that would provide tissue for Potti’s genomic analysis.
“These were painful, they went through under my arm all the way up into my neck doing this biopsy,” Shoffner said. “But it was going to be worth it.
“They implied that people had been having such wonderful results that they would have to put titanium clips around my tumor, because these tumors were melting away rapidly. And by the time I finished my chemo, there may not be any tumor left, and they would have to go by those nine titanium clips that they inserted into my body to remove the residue.
“I thought, ‘Well, this is absolutely remarkable.’”
Before Shoffner was enrolled, signs of trouble were already emanating from Potti’s lab. Bradford Perez, a medical student under Potti’s tutelage, had blown the whistle about irregularities in handling of the data, and MD Anderson biostatisticians Keith Baggerly and Kevin Coombes had published a letter in Nature Medicine, critiquing Potti’s genomic predictor model data.
Unbeknownst to Shoffner, Duke had suspended the trials twice, silenced the whistleblower, and overlooked data from other institutions—while she was in touch with her Duke oncologist, Paul Kelly Marcom.
Shoffner said that Marcom, who was the principal investigator of her trial, never disclosed this information in over two years. When Marcom did talk with Shoffner in November 2010, he said only that there were “problems with the data,” and that the trials were terminated, Shoffner said.
Duke officials and Marcom did not respond to questions from The Cancer Letter.
“I Don’t Approve of Lying to Patients”
Even under normal circumstances, communication between doctors and patients is complex—nuance can be lost; anger can flash.
At Duke, Potti’s fraud killed every vestige of trust Shoffner had in the institution, leaving her feeling betrayed.
“It’s a tragic, tragic event that slows down research,” said Jimmie Holland, the Wayne E. Chapman Chair in Psychiatric Oncology at Memorial Sloan Kettering Cancer Center, and founding president of the American Psychosocial Oncology Society.
“Informed consent is based on trust between the patient and the institution and the doctor. It’s trust in general; it’s a principle that is much wider than medicine,” Holland said to The Cancer Letter. “I don’t think most oncologists would willfully lie or withhold information—well, obviously that’s bad.
“The examples like this that become widely known and discussed have a tendency to reduce the trust in cancer research by protocol, which is very sad, because we need to have research to go forward and improve the care of cancer patients.”
Lying, or withholding information from patients is “a very bad idea,” said Barrie Cassileth, an integrative medicine specialist and the Laurance S. Rockefeller Chair in Integrative Medicine at MSKCC.
“I don’t approve of lying to patients. It has detrimental effects,” Cassileth said to The Cancer Letter. “It diminishes what the doctor-patient relationship is about, and there will never be any good to come out of misleading patients.”
The research enterprise is the real victim in Shoffner’s story, said Jennifer Griggs, a professor of medicine and medical oncology in the Division of Hematology/Oncology at the University of Michigan.
“[Shoffner] refers to being a lab rat or a guinea pig, and what this does is it erodes trust in the biomedical research enterprise and the system. And this makes it harder for all of us to do research,” said Griggs, who reviewed the Duke protocols and other publicly available data for The Cancer Letter.
“The Potti group has done harm to research progress worldwide.”
Griggs is also a professor in the University of Michigan’s School of Public Health in the Department of Health Management & Policy.
“The genomic predictor was clearly bad (everyone agrees), but I don’t see that this patient was physically hurt because of this trial per se,” said Matthew Goetz, a professor of oncology and pharmacology at Mayo Clinic, who also reviewed Shoffner’s case for The Cancer Letter. “It would be understandable, however, that the emotional anxiety of knowing that she was enrolled onto a clinical trial that was testing a bad classifier caused her anxiety and depression—this sort of claim could be justified.”
After Shoffner completed her chemotherapy regimen in the trial, she learned that she had been assigned to receive Adriamycin-Cytoxan (AC)—standard chemotherapy for breast cancer—based on Potti’s predictor model.
Shoffner’s cancer has not recurred, but she has suffered from common adverse consequences of the AC regimen. These include blood clots and diabetes.
After learning about Potti’s fraud, Shoffner has been re-examining her treatment at Duke—and wondering whether she was duped into participating in the neoadjuvant trial, whether she received the right chemo, and whether she should have received additional treatment post-surgery.
Shoffner said she is being treated for post-traumatic stress disorder, which she attributes to her experience at Duke.
“For me and for the surviving patients, it’s not over,” Shoffner said. “Whether the healing process will actually heal, I don’t know. It’s not resolved for me, it will be a part of my life till the day I die.”
Shoffner Questions Treatment Choices
Shoffner remembers the shock of learning that something had gone awry at Duke.
Her oncologist, Marcom, called on the day before Thanksgiving in 2010. He was informing her that the trial, in which Shoffner participated and which he ran, had been terminated.
“The reason that the trials had been canceled, he said there seemed to be some problems with the data, that’s all he said,” Shoffner recalled.
Unbeknownst to Shoffner, the trials had been suspended in July 2010, after The Cancer Letter reported irregularities in Potti’s CV, triggering a broader investigation. In fact, Marcom was making his calls to patients days before the Institute of Medicine’s Omics Committee held its first meeting to investigate the Duke scandal.
Shoffner said Marcom did not mention Potti’s fraud.
Another aspect of her conversation with Marcom ended up worrying Shoffner even more.
“He said, ‘I do wish I had given you Taxol,’” Shoffner recalled. “Being devastated that this holy grail of medicine that was going to cure my type of cancer, which is a bad one—that’s all I could think about at that time. I thought nothing about him saying that about Taxol.
“A couple of days later, we spoke again, and during that conversation, once again, he made a comment that he wished he’d given me Taxol.
“But shortly thereafter, I started thinking about it when I was coming down off my horror, and I thought, ‘Well, what is Taxol? And why is he upset that he didn’t give it to me?’
“So I looked it up. And I believed from what I read and understood that I should’ve had Taxol for this type of cancer, and I didn’t get it.
“Then I found out that [the fraud] was all over the news.
“I had not heard it, or didn’t relate it to me, of course, but it had already been in The New York Times, locally in the news, TV and newspapers,” Shoffner said. “I began to feel that I was the last to know.”
She found past coverage of the Duke story, and ultimately called Keith Baggerly, a biostatistician at MD Anderson Cancer Center, who had spent hundreds of hours examining the Nevins and Potti data.
Shoffner asked Baggerly whether she was harmed. Baggerly said he wasn’t qualified to answer this question.
“[Duke said] that there was no harm done to the patients, that we all got the standard of care anyway,” she said. “But I did realize at that time that everybody knew [about the fraud] before I did.”
Suddenly, Duke, the place she trusted with her life, no longer seemed safe.
“I was terrified,” Shoffner said. “Where was I going to go? What was I going to do?”
University of Michigan breast cancer expert Griggs said Duke should have informed Shoffner and other trial participants about the fraud much earlier.
“It seems that the institution knew things—and The Cancer Letter is intimately involved in this—it looks like there were things that were known well before the institution made it public,” Griggs said.
“That information would have been better off being disclosed to the patients earlier than it was.”
The Question of Taxanes
Shoffner’s disease was HER2 negative, and estrogen receptor (ER) and progesterone receptor (PR) positive.
In an interview with The Cancer Letter, Shoffner said that Marcom was able to determine that she was receiving neoadjuvant Adriamycin-Cytoxan chemotherapy based on the genomic predictor model.
This would mean that Marcom was able to unblind her predictor results.
The protocol reads:
“Both the patient, treating physicians, and research nurses/staff will be blinded to randomization assignment until after the primary treatment endpoint is met. (i.e. progression on therapy; stable disease/minimal response after completing chemotherapy; or proceeding to definitive surgery).”
The protocol is posted here.
“I learned for the very first time in the lawsuit that Dr. Marcom had unblinded my predictor results after my fourth cycle of chemo, but before a decision was made about what further treatment I may receive,” Shoffner said. “He saw that I had been placed in the guided AC arm of the trial and that the predictor said I was resistant to taxanes. I also learned that he used the predictor results to go into his judgment not to give me a taxane, either before surgery or after.
“If I hadn’t been given the fourth cycle of AC, my treatment couldn’t be considered as a data-point in the clinical trial results.
“Dr. Marcom never discussed giving me a taxane before or after he sent me for surgery. I never was given the opportunity to consider whether to take it or not because he never discussed it with me. I would have chosen to take it from what I now know about TAC as the standard of care that I learned in the lawsuit.
“This is why I was surprised when Dr. Marcom even mentioned taxanes in his first telephone call to me in November 2010 nearly two years later, when he said he should have given me a taxane.”
If this is correct, Marcom continued to use the faulty predictor model to determine Shoffner’s treatment choices after her participation in the trial ended.
According to the Duke trial protocol, she was not precluded from receiving a taxane, or other additional therapy.
After the surgery, Shoffner was found to have 10 positive lymph nodes, which makes her disease Stage IIIC.
Shoffner received an aromatase inhibitor, which is the standard of care for ER-positive patients like her.
Breast cancer experts say that patients like Shoffner would have ordinarily been considered for taxane-based therapy. However, Shoffner’s oncologist could have also decided that her toxicity incurred by AC was so severe that additional chemotherapy wouldn’t be appropriate.
“An oncologist may believe that, based upon a patient’s performance status and complications from AC, that the risks of giving taxanes outweighed the benefit,” Mayo’s Goetz said. “This is clinical judgment and I have these conversations all of the time with my patients.
“In my mind, even though the classifier was bad, it appears to me that trial design was attempting to answer an important clinical question of whether a genomically-driven test could guide treatment between two regimens that were known to result in similar breast cancer outcomes, but with very different toxicity regimens (e.g. AC causes can result in heart failure in 2 to 5 percent, whereas TC does not),” Goetz said.
“For those patients randomized to the AC arm, it would have been important AFTER surgery that they be at least offered a standard taxane-based regimen (either paclitaxel or docetaxel), which was known to reduce the risk of recurrence for patients that had received AC.”
Goetz said that oncologists would usually have a conversation about taxanes with patients like Shoffner, and this would be documented in medical records.
Shoffner said her medical record does not show that Marcom discussed this matter with her. The complaint against Duke states that Marcom “chose not to prescribe” a taxane.
University of Michigan’s Griggs said most oncologists would have given Shoffner a taxane.
“I don’t think she was given what we would consider the standard of care at the time,” Griggs said to The Cancer Letter. “I think pretty much every oncologist would’ve given her AC followed by a taxane. I would have.
“That the oncologist had the opportunity to treat her with a taxane post-surgery and didn’t in somebody with that high a risk of recurrence—10 nodes—most people would say that the benefit is there with a taxane.
“She’s free of disease, so I can’t say her outcome could have been better. She hasn’t had a recurrence. If she had a recurrence, I think there’d be a much bigger concern about the fact that she didn’t get a taxane.
“Trial or not, I would’ve given her both a taxane and an anthracycline, and I think most oncologists—if you polled 100 oncologists, it would be an unusual oncologist who would say, ‘No, I’d stop at AC.’
“I think the biggest issue is the fact that she was enrolled in a trial that’s based on fraudulent data. And that was the problem that The Cancer Letter was so instrumental in revealing: the fraudulent study and the fraudulent investigator.
“This case, and those of other patients who participated in these trials, demonstrates the importance of the integrity of investigators who conduct clinical research. We can make progress against cancer only through conducting high quality research. If we compromise the trust our patients have in us, we compromise that progress.”
Shoffner’s Message to Duke
What would Shoffner like to say to Duke—Potti, his mentor Nevins, and the deans who protected them?
“Why didn’t you level with us and let us have a choice, instead of just using us for human experiment? Where was your mindset when you overlooked the human factor?” Shoffner said. “I didn’t think that human experimentation was possible in this age and time.
“I’d love to say, ‘Do you have any regrets about just the human side of it?’ I’m sure they all regret that the research didn’t come out like they wanted. Oncogenomics didn’t come out like they wanted in the end.
“Now, have you had the time to think about the people and what happened to them? At this point, do you have any regrets or concern, or are you so mad at us for filing a lawsuit that you’d wish we’d all die?
“I have felt for a long time like they were waiting to see whether I would die within the five years. And then if I didn’t die in the five years, ‘Oh, we cured you!’
“If they said, ‘I’m sorry,’ I don’t think I could believe them. I mean, don’t need to come to me and tell me, ‘Sorry,’ because you have shown me—actions speak louder than words—you’ve showed me that you did not care about me as a human being.
“Someone said along the way, ‘Duke sent you a letter of apology, but that it came back. They didn’t have your address.’
“And I said, ‘Tell them to go to their billing department. They know how to get up with me.’”