Keloids are a major complication related to surgical wound healing and very challenging condition to treat. Many treatment options are available, but the efficacy of the treatment is poor in most of cases and some keloids do not respond to the treatment at all. We compared the efficacy of intralesional 5‐fluorouracil (5‐FU) and triamcinolone (TAC) injections in a double‐blind randomized controlled trial (RCT). Forty‐three patients with 50 keloid scars were treated with either intralesional TAC or 5‐FU‐injections over 6 months. We wanted to find out whether biological features (cell density, cell proliferation rate, vascular density, myofibroblast numbers, steroid hormone receptor expression) in keloids could be used to predict the response to therapy and define the biological changes that take place in patients receiving a response. As there was no statistically significant difference in the remission rate between TAC and 5‐FU treatments, all patients were combined and analyzed as responders and nonresponders. Although responders have slightly more myofibroblasts than the nonresponders in their keloids in the pretreatment biopsy samples, we could not identify a single predictive factor that could identify those patients that respond to drug injections. The good clinical response to therapy is associated with the simultaneous reduction of myofibroblasts in the keloid. This study demonstrates that myofibroblasts are reduced in number in those keloids that were responsive to therapy, and that both 5‐FU and TAC injections are useful for keloid treatment.