Phototherapy is a non-invasive therapeutic modality that uses specific wavelengths of light to modulate biological processes and treat disease. Driven by advances in photobiology, regenerative medicine, and clinical technology, phototherapy has evolved from empirical heliotherapy into a mechanism-based intervention with expanding medical applications. This review places the biological effects, therapeutic indications, and safety profiles of ultraviolet, blue, and green light in comparative context, while focusing primarily on the mechanisms, dosimetry, and clinical translation of red and near-infrared photobiomodulation (PBM). Although ultraviolet and blue light therapies have established roles in dermatology and neonatal jaundice, their clinical use is limited by concerns such as photocarcinogenesis, retinal phototoxicity, and circadian disruption. In contrast, red and near-infrared light offer deeper tissue penetration and, when appropriately dosed, a comparatively low risk of direct photochemical injury. Through mitochondrial, redox, inflammatory, and other context-dependent signaling pathways, red/NIR PBM has been investigated in wound repair, chronic pain, osteoarthritis, musculoskeletal disorders, neurological conditions, and dermatological applications. As a representative quantitative example, a 2024 meta-analysis of 10 placebo-controlled randomized trials involving 542 participants with knee osteoarthritis found a moderate reduction in pain at rest with PBM compared with placebo (standardized mean difference, - 0.70; 95% confidence interval, - 1.10 to - 0.20), although the certainty of evidence was very low. However, broader clinical translation remains constrained by heterogeneous dosimetry, insufficiently standardized treatment protocols, and uneven evidence maturity across indications. To address these limitations, we propose a precision-oriented framework that links clinical target, tissue-optical context, wavelength selection, patient- and site-specific modifiers, dosimetric implications, measurable endpoints, and evidence maturity. This framework is intended to support treatment planning, parameter reporting, and prospective validation rather than to prescribe currently validated individualized dosing algorithms.
