Understanding Your Cancer Prognosis

How much time do I have? This is a question that most people with advanced cancer — and many people with earlier stage cancers — consider asking their doctor. It’s also a question that is asked less often than you might think. Most oncologists dread the subject — partly because it is impossible to provide an accurate answer, but also since doctors prefer to offer hope and encouragement.

It has been my experience that sometimes a doctor does not give this information, whether or not she has been asked, until things are truly dire. At that point, hearing the prediction may be shocking to the patient, who possibly hoped for better projections or perhaps needed more time to take care of business. I have never heard of a doctor saying something like, It’s time to go home and get your affairs in order, but this message needs to be more gently delivered — while there is plenty of time to do so.

Would we make different decisions about our lives if we knew that time was limited?

There are some people who never want to hear that there is little or no useful treatment still available; there also are those who don’t ever want to have this discussion about time. However, as upsetting as it is, most of us would prefer to have an honest understanding of our situations. We may need to provide plans for our children or other family members. We may never have finished an estate plan or even written a will. We may own a small business that needs a succession plan. We may just need some time to grieve and be with the people whom we most love.

There are many difficult decisions that need to be made while someone lives with advanced cancer. Does it make sense to try another chemotherapy option, which may have little chance of success? Is there a clinical trial that seems to be a good option? Is it time to involve palliative care or even transition to hospice care? Some studies have found that people live longer and have a better quality of life under hospice care than they do while continuing chemotherapy. This is never a simple choice, but it is very important for cancer patients to understand the realities of their condition and have the opportunity to make the best decision for themselves as they approach the end of life.

When pressed with questions about the future, some doctors may offer a disappointing answer: I don’t have a crystal ball. We already know that. But, if we bring up the subject, we are asking because we need their best-informed guess. I have therefore suggested to many patients that these questions have a better chance of producing a helpful answer: Given my situation, what do you think is the best possible outcome? And what do you think is the worst? For any one individual, the outcome is likely to fall somewhere between those two answers, and most doctors are much more comfortable responding to these questions than to a single prognostic estimate.

I have been recently thinking about cancer outcomes because a new predictive model has been developed by researchers at City of Hope National Medical Center in California. Led by Dr. Finly Zachariah, the research team used data from almost 30,000 patients with metastatic cancer to create a model for predicting three-month mortality. Their research plan was fairly simple: oncologists were asked, “Would you be surprised if this patient died within the next three months?” Doctors responded “not surprised” about 13% of the time, and the mortality rate within that group was higher than in the other. However, within three months, almost 10% of the patients in the “surprised” group had also died.

Since this predictive model can be computer-driven, the research team developed a program in which a computer would also miss 70% of the expected deaths. However, the researchers noted that the program could be modified to produce significantly more accurate results. Artificial intelligence is growing at a rapid pace, and it seems probable that programs like this will become more common. Once again, computers can more accurately predict prognosis than doctors.

This leaves me with the same question: Do you really want to know? Clearly nothing — not a computer, not an experienced oncologist — can 100% accurately predict an individual’s prognosis. However, this new computer model does give us an opportunity, if we want it, to have a better estimate. Would we make different decisions about our lives if we knew that time was limited?

There should not be a single answer to these musings, since we are all unique. However, it seems worthwhile for us to consider that finding a more accurate prognosis is possible; we should also reflect on whether having this information would enhance quality of life over our final months.

In addition to suggesting that my patients ask the pair of questions above, I now also mention this new artificial intelligence possibility for making projections. But the most important advice that I can provide is: Don’t ask unless you are sure that you want the answer.