A group from Michigan has published a paper in the January 2013 edition of The Journal of Urology describing a strategy for using biopsy and size to help assort patients into treatment (surgery) or “active surveillance” groups. Biopsy was used to group tumors into Favorable, Intermediate, and Unfavorable risk groups. The algorithm below shows how this information was combined with tumor size to make a decision.
Most kidney tumors are discovered when a patient gets a CAT scan, ultrasound, or MRI for an unrelated problem. Because of this, many tumors are small. In a past era, tumors smaller than 3-4 cm were often observed. The risks of the alternative (removing the entire kidney through a flank incision ) was considered worse than the risk of spread during observation. Biopsy was shown to be so poorly predictive of a tumor’s ability to spread that it was not considered to add useful information beyond that of size alone.
Much has changed. First, open partial nephrectomy, and then laparoscopic and robotic partial nephrectomy were developed. These have allowed the surgeons to use small incisions to remove small masses and leave the kidney behind. If a patient is too sick for even these minimally invasive surgeries, a the bulk of a tumor can be destroyed in place with freezing or heat treatments deployed at the end of a needle guided by CAT scan. Now, most people don’t want to hear about watching their kidney tumor; they want it out. This momentum is increased by the fact that these treatments are easier while a tumor is small. But the treatments still carry some risk. With image guided biopsy getting more accurate, is it time to reconsider biopsy and observation?
To be clear, the authors didn’t actually employ their proposed strategy. This study looked back at 151 patients who had a biopsy and then went on to surgery regardless of what was found. Even if they had a biopsy that predicted that the mass was not cancer, or if it added no useful information they still had surgery. All the masses were less than 4 cm. The biopsy often predicted not only cancer, but specific types of kidney cancer, or different levels of aggressiveness. Since everything was taken out, the authors could judge the accuracy of these predictions. The authors then created a “what if” scenario to see how a decision tree for deciding who to operate on or who to watch would work. The decision tree combined biopsy information and size to guide the decision.
So what would have happened? Of 151 biopsies, 4 were read as not cancer (so no treatment). 14 biopsies didn’t reveal any information one way or the other and would likely have been repeated in a real world situation. 36 patients would have been observed in the Active Surveillance group, and 97 patients would have gone directory to surgery. So, great: 40 patients out of 151 avoid surgery (at least at first), and 14 go back for another biopsy. It’s important to remember that, in this scenario, the people who would be observed at this initial point still might end up getting surgery. This could happen later if the mass grew, or they got another biopsy in the future that showed something more dangerous, or both.
We know that all the patients who would have been assigned to surgery really did have more aggressive looking tumors on final pathology. We don’t know what would have happened to the active surveillance group because in real life, everyone got surgery. Of the 36 active surveillance patients, almost a third (11) ended up in reality having more aggressive findings than the biopsy suggested. The term “Active Surveillance” suggests that these people who did not have surgery would have been watched closely, and many of them would ended up having surgery either because they were mis-categorized (as shown in the 11 patients) or because the mass eventually grew. So the 36 is just a starting point for observation, and does not reflect the true number of patients who get treatment over time.
The real important group to think about are the 36 patients in the Active Surveillance group. What would have happened to them over time? Only 25 should be there. Ideally, the other 11 would likely end up getting surgery after subsequent CAT scans show growth, or a decision was made to re-biopsy and new information is gained. It is probably fair to say that even an aggressive tumor is unlikely to spread while it is still less than 2 cm. the authors looked at the 11 and found that they could alter what they considered to be non-aggressive and then not as many patients would be mis-characterized. Some of the 25 will go on to surgery as well, as the masses grow, or anxiety drives them towards treatment.
A further study – the one that would make a real convincing argument about whether this strategy should be used – would be to watch a real life Active Surveillance group. How many of them would end up getting a bunch of CAT scans and/or biopsies and then have surgery in the end? Would any show up with metastatic disease while being watched? If the end result is that 151 people got biopsies and most go on to surgery either immediately or after a delay of some years and many additional test, then the answer is not to use this strategy. If a good portion end up being watched safely, then this has real value.