Osteoarthritis, the wearing away of cartilage in a joint capsule, nearly always involves some kind of misalignment or compression of the joint. It's often assumed that cartilage can't regenerate, but science is beginning to tell us otherwise.
Before I continue, please note that an arthritic area may not be painful, even when it turns into bone-on-bone. In other words, a physician may find arthritis on an x-ray where you have no pain, or they may find arthritis in the same area where your pain is, and it may not be the cause. This can often be confusing for both patient and physician. More on this below. It should also be noted that, in some cases of osteoarthritis, chronic inflammation is a culprit in tissue degeneration, and the interaction between inflammation and biomechanics is just beginning to be understood (study below).
But in the following picture, you can see the results of a study in which an internal decompression device was installed in the knees of 24 people with end-stage osteoarthritis (average age of 49; all participants qualified for a total knee replacement). According to the study, effective, working cartilage grew back, and was still back two years later (when the study was written). This was an invasive procedure that all participants agreed was preferable to a total knee replacement.
In the next two pictures, you see the result of about a year of Egoscue exercises.
I hope you can see the striking similarity between the images above and below. The big difference, of course, is that using the Egoscue exercises the client didn't have to be cut, and instead was resetting muscle compensations and loading patterns as he went, restoring honest alignment up the chain.
Remember when I said above that arthritis often doesn't correlate with pain? In the knee decompression study, the authors noted that "a clear correlation between [decreased pain] and structural.... tissue repair could not be demonstrated in this study. This indicates that pain sensation is not obviously related to structural changes." In other words, people didn't consistently report feeling better after they had more cartilage. Pain is a complex subject, and there could be many reasons for this. For instance, it may be that the important compensatory relationships weren't sufficiently addressed. (From a P-DTR perspective, some of these relationships can be quite surprising.) It is also possible that the surgery created new compensatory patterns. There is a lot we still don't know about pain, and a lot more I could write about this. But in the case of the Egoscue client, he was feeling quite a bit better and able to get back to what he loved.
Egoscue, NKT, P-DTR and many other modalities work off some similar principles about compensation and loading. If you can figure out why the knee wants to compress, it can speed things up quite a bit. You may not need to get these visible results to start to feel better.
For a really great summary of how restored biomechanics can improve the pain in an arthritic joint, see Austin's The Art of Fitness and NKT's own Jesse James: http://tao-fit.com/dont-blame-arthritis-for-your-pain/
Here's the study of knee decompression: http://ard.bmj.com/content/early/2011/05/12/ard.2010.142364.long
Here's a description of the Egoscue case by Austin's esteemed Egoscue clinic: https://egoscueaustin.wordpress.com/2010/01/27/check-this-out-you-are-going-to-love-this/
For more on the interaction of biomechanics and inflammation in osteoarthritis, see https://www.unispital-basel.ch/fileadmin/unispitalbaselch/Bereiche/Chirurgie/Abteilungen/Orthop%C3%A4die/Arthroseforschung/Biomechanik_Arthrose.pdf.