New Tool May Help Prostate Cancer Patients Peer Into the Future

From the WebMD Archives

Sept. 28, 2000 -- For some cancers -- prostate cancer being one of them -- choosing the best course of treatment can be an agonizing decision. Now, a newly developed tool may help doctors and patients select the best possible treatment for early-stage prostate cancer. The researchers tell WebMD that their simple, point-based system is superior to all other existing methods for predicting whether cancer will return.

"The purpose of the study was to develop the most accurate prediction tool that we could using currently available data," says lead author Michael W. Kattan, PhD, assistant attending outcomes research scientist at Memorial Sloan-Kettering Cancer Center in New York. The results, he tells WebMD, "are extremely practical. It is useful today. It's not a new marker or treatment coming down the pipeline ... it's exactly what the patient needs to know right now." The paper appears in the October issue of the Journal of Clinical Oncology.

Although there were several methods available to predict the outcome for patients, they all worked by separating the patients into groups based on risk factors, Kattan tells WebMD. He explains that they felt the patients in a risk group were not truly all the same. For example, even if two men are assigned to the same risk group based on having the same PSA test score -- the higher the score the more likely there is prostate cancer present -- they may not have the same outcome for a host of other varying factors between them.


So, he tells WebMD, his team developed "a tool that looks like a ruler on paper, where each variable gives a certain number of points." The doctors enter the patient's PSA number, cancer stage, whether hormone therapy is going to be used, and the planned radiation dose, along with other variables, and "the tool then predicts the probability that the cancer will recur within five years."

To see how well the tool predicted real-life outcomes, the team reviewed their hospital's records for more than 1,000 men treated with radiation for early-stage prostate cancer -- where the tumor is confined to the prostate itself and has not spread to other organs.


When they tested their new method against the eight other tools, it was significantly more accurate at predicting outcomes after five years. They also tested it on what Kattan describes as a yardstick for telling the level of accuracy of such a tool. "On a scale ranging from zero, which is the same 50/50 accuracy of a coin toss, to one, which is like having a crystal ball, our tool was right in the middle," he tells WebMD.


They then repeated the experiment and got the same results, using data from a similar group of patients treated at The Cleveland Clinic in Ohio.

Of all the variables, "PSA drives the boat, [but] it's not the sole predictor, and other factors can change the prediction." That's precisely why a tool like this, which takes several variables into account, is so important, Kattan tells WebMD.

Although this study looked at radiation therapy, says Kattan, "we've got similar tools that are optimized for predicting outcomes of prostate surgery and brachytherapy," which is a procedure where radioactive "seeds" are planted in the prostate. They all work the same way, he says, and the predictive value has been similar for all of treatment types.

According to Kattan, the early-stage prostate cancer patient is "a good example of someone who needs the most accurate prediction we can give. There are lots of cancers where you don't need that," he says, "because there's really only one default treatment, and it doesn't really matter too much how much it works because it's your only shot. Here, it's the opposite situation."


To make the right decision, patients need "to know how effective a treatment is going to be." This new tool "is essentially a formula, represented in a friendly format, that predicts early-stage prostate cancer treatment outcome more accurately than any of the others."

WebMD Health News


© 2000 WebMD, Inc. All rights reserved.