Ductal carcinoma in situ (DCIS) is a condition we don’t understand. We don’t know its significance, how to describe it, and how to treat it. Worse, we may have created it. Its incidence in the US in 1975 was 1.87 per 100 000; now it’s 32.5. During that time there has been no drop in invasive cancers. The increase is because of mammography. How does a doctor communicate with a woman about such a confusing and uncertain condition?

This question was tackled at a conference in Edinburgh earlier this week, and I was gripped by the consultations between three doctors and two actors playing a woman and her husband. If DCIS creates difficulties for women, it’s much worse for the affected woman. The woman is likely to receive conflicting information, not be sure whether she has cancer or not, and be offered various options for treatment ranging from watchful waiting to mastectomy and radiotherapy.

Uncertainty is “psychonoxious,” said Lesley Fallowfield, professor of psycho-oncology in Sussex, and makes it difficult for the woman to adapt to her condition. In some ways it’s worse than having a diagnosis of cancer.

The audience of around 100, mostly breast cancer surgeons and medical oncologists, were given a chance to vote on how they might describe DCIS to a patient: around half voted for it “might become breast cancer,” a quarter for it is “breast cancer at an early stage,” and another quarter for it “is breast cancer and can’t be left.”

Asked what specialist nurses in their teams might call the condition, two thirds answered “no idea.” A similar proportion had “no idea” what cancer charity websites said. Unsurprisingly a Google search produces a mass of conflicting information.

In a focus group of 13 women who regularly attended for mammography not one of them had heard of DCIS. They went for mammography to be “rather safe than sorry.”

The affected woman played by the actress was 50, married with two children, and had had low grade DCIS of 5 cm picked up on mammography. (We later learnt that there is big variation in how often different units in the UK diagnose “low grade” DCIS, some don’t see it at all, some see it commonly. The pathologists don’t agree.) She had been old that the lesion “could be precancerous cells.”

The audience was asked to vote on how she should be treated: 60% opted for lumpectomy, wide excision, and radiotherapy, 20% for mastectomy, and 20% weren’t sure. (NICE guidelines for what they are worth recommend mastectomy).

The first doctor, a British surgeon, explained DCIS by describing a duct and saying how the cancerous cells were entirely inside the duct and “hadn’t gone anywhere.”

“Was it cancer?” the patient asked.
“Yes, but not invasive.”
“Might it become invasive?”
“Yes.”
“How often?”
“In perhaps 50% of cases.”

This is a disputable answer. It’s based on old, retrospective data. Should the doctor share the uncertainty around the data. As an “aggressive anti-paternalist” I think yes. I think that not sharing uncertainty is in some senses “lying” and a kind of abuse, particularly if the information leads a woman to opt for a treatment that she would not have chosen if fully informed. Most practising clinicians disagree with me.

The consultation then moved on to a discussion of treatment options. The surgeon said that the choice was between lumpectomy and mastectomy, but that as the lesion was 5 cm and there would have to be a margin of excision of at least 1 cm all around—so it would mean a large defect and possibly a poor cosmetic result. It was the woman’s choice, but he leant towards mastectomy.

The surgeon did a good job of explaining the condition but didn’t share the full uncertainty. Even so, I felt that the woman and her husband would get home and find themselves confused. They would understandably search for information on the internet and probably become more confused.

In the role play the couple then visited their daughter in Canada and were able to have an informal discussion with a Canadian oncologist. She talked about how DCIS “can become a cancer” and is “a marker for developing a cancer.” She thought that lumpectomy was a real option but explained that it might be necessary to operate a second time if pathological examination showed that there were “cancer cells” at the margins of the excised tissue, something unsurprisingly that women hate.

Next the couple—back in Britain—had the chance to enter a trial that is being planned of surgical treatment as recommended by the woman’s surgeon against no surgery but annual mammography. One of the problems of gaining fully informed consent is to answer the question of what proportion of low grade DCIS might become invasive. The first surgeon said 50% but some of the doctors thought it very much lower—and the best answer is that we don’t know, hence one reason for the trial. The patient/actress said that it seemed “bit Russian roulettish” but that she probably would enter the trial. On a vote about two thirds of the audience said that their unit would enter a trial, but some observed that trials of surgery against monitoring were difficult because so many women preferred the reassurance of the surgery and found that their anxiety mounted with simply monitoring.

In the discussion one of the medical oncologists called himself an “expert in overtreatment” and described his discomfort in treating patients with chemotherapy for a 2-3% improvement in survival. (He wasn’t talking here about DCIS but more generally.) Another medical oncologist said: “All I ever see around me is more and more treatment—a toxic deluge of overtreatment.”

These were Brits, and, in contrast, some Americans talked about using MRI scans and hormone therapy, potentially opening up many new options. With uncertainty the cultural differences can flourish.
I found myself wondering whether we would be better or worse off if we had never invented mammography and hadn’t opened the Pandora’s box of DCIS.

Richard Smith was the editor of the BMJ until 2004 and is director of the United Health Group’s chronic disease initiative.