Knee osteoarthritis is the most prevalent joint disease and a frequent cause of pain, functional loss and disability. Conventional treatments have demonstrated only modest clinical benefits whereas cell-based therapies have shown encouraging results, but important details, such as dose needed, long-term evolution or number of applications required are scarcely known. Here we have reanalyzed results from two recent pilot trials with autologous bone marrow-derived mesenchymal stromal cells using the Huskisson plot to enhance quantification of efficacy and comparability. We find that cell doses of 10, 40 and 100 million autologous cells per knee provided quite similar healing results and that much of the effect attained 1 year after cell application remained after 2 and 4 years. These results are encouraging because they indicate that, apart from safety and simplicity: (i) the beneficial effect is both significant and sizeable, (ii) it can be achieved with a single injection of cells, and (iii) the effect is perdurable for years.