Photodynamic therapy (PDT) that employs the photochemical interaction of light, photosensitizer and oxygen is an established modality for the treatment of cancer. However, dosimetry for PDT is becoming increasingly complex due to the heterogeneous photosensitizer uptake by the tumor, and complicated relationship between the tissue oxygenation ([3O2]), interstitial light distribution, photosensitizer photobleaching and PDT effect. As a result, experts argue that the failure to realize PDT’s true potential is, at least partly due to the complexity of the dosimetry problem. In this study, we examine the efficacy of singlet oxygen explicit dosimetry (SOED) based on the measurements of the interstitial light fluence rate distribution, changes of [3O2] and photosensitizer concentration during Photofrin-mediated PDT to predict long-term control rates of radiation-induced fibrosarcoma tumors. We further show how variation in tissue [3O2] between animals induces variation in the treatment response for the same PDT protocol. PDT was performed with 5 mg kg−1 Photofrin (a drug-light interval of 24 h), in-air fluence rates (ϕair) of 50 and 75 mW cm−2 and in-air fluences from 225 to 540 J cm−2. The tumor regrowth was tracked for 90 d after the treatment and Kaplan–Meier analyses for local control rate were performed based on a tumor volume ⩽100 mm3 for the two dosimetry quantities of PDT dose and SOED. Based on the results, SOED allowed for reduced subject variation and improved treatment evaluation as compared to the PDT dose.