The Oncologist’s Perspective on Statistics: Part Three

As you are likely starting to notice, I have so much to say on cancer. I am breaking up big topics into sections to make them more manageable. This is part 3 on statistics.

As I mentioned in part one, statistics can be very helpful to give you reasonable expectations about your dog’s cancer, but as discussed in part two, statistics can be frustrating when your pet suffers an uncommon side effect or doesn’t live up to the statistics that you are quoted.

In this part, I will review a few of the more common statistical terms you may hear. In the Dog Cancer Survival Guide, I discuss many of these terms for 12 common canine cancers.

In canine cancer, a common term you will hear is the median survival time (MST).  Median survival refers to the time it took for half of the dogs to die in a study.  So by definition, half the dogs lived longer than the median survival. For example, if a study had sixty dogs and the median survival time is twelve months, thirty dogs were still alive at twelve months, and thirty dogs were no longer with us.

But why median and not average (typically reported as the mean survival time)? For an average, add up the total survival times for all sixty dogs in the study, and then divide the result by sixty. While this average might seem like a useful statistic, it may not always be so.

If one dog in the study was an extremely long-term survivor, but most other dogs weren’t, the average survival time might give the impression that most of the dogs in the study lived a very long time. The opposite could also be true: if one dog, sadly, died very early after diagnosis, averaging in his short survival time might give the impression that most of the dogs lived a very short time.  The advantage of using MST is that it is less likely to be skewed by these unusual cases. I cannot stress enough: while it is important to learn the MST for your dog’s cancer, no one is “average” and many dogs have much better outcomes than the median.

Other statistical terms you may come across:

Response rates tell us the percentage of dogs that respond to a particular treatment or protocol. This is a common statistic for chemotherapy protocols. For example, the response rate for canine lymphoma to a multi-agent CHOP protocol is 85 to 90%, but the response rate to a COP protocol is lower, approximately 60 to 70%.

Cancer recurrence is defined as the return of cancer after treatment. For example, the overall recurrence rate for a soft tissue sarcoma after surgery is 15% and usually within one year, but for dogs with soft tissue sarcoma removed with incomplete or dirty margins, the recurrence rate jumps to 30%.

Metastasis rates tell us how likely the cancer is to spread to other parts of the body. These rates help oncologist with treatment recommendations. For example, oral melanomas are not only aggressive in the oral cavity, they metastasize to the local lymph nodes, liver, lungs, and kidneys at rates of greater than 60-80%. Preventing the spread with something like melanoma vaccine is is just as important as treating the tumor in the mouth with surgery and/or radiation. On the other hand, for an oral cancer like a low grade fibrosarcoma, metastatic rates are pretty low, so chemotherapy is not typically part of the treatment plan after surgery and/or radiation.

So hopefully, understanding the usefulness and limitations of statistics will help you understand the options and recommendations for your dog. I also remind you that worrying over your dog’s outcome – whether he will live up to or exceed statistics – will not change the outcome. It is wasted energy. Instead, I recommend you focus on caring for your dog through his treatment, giving medications and supplements, and giving your vet feedback on how your dog is doing at home. And most importantly, enjoy each and every day you share together. That is where energy should be focused.