Stress fractures are overuse injuries of the bones resulting from repetitive stresses. Since symptoms are non-specific, an imaging modality is mandatory for accurate diagnosis. Bone scintigraphy used to be the gold standard in evaluating stress fracture, but the advent of MRI led to its underutilization. This study aims to determine the role of three-phase bone scintigraphy in the assessment of stress fracture, and to determine the sensitivity, specificity, positive and negative predictive values and the accuracy rate of bone scan. In a tertiary hospital in Metro Manila, 15 patients referred for three-phase bone scintigraphy to assess for stress fracture from 2004 to 2006 were included in the study. Of these 15 patients, 12 also underwent MRI. Sensitivity, specificity, positive and negative predictive values, and the accuracy rate were computed with MRI as the gold standard. The sensitivity of bone scans approach 100 percent. However, its specificity is low due to other conditions that can produce a positive scan. Because of the limitations inherent to scintigraphy, MRI could be a valid first-line imaging technique in diagnosing stress fracture. MRI provides a greater anatomic detail of the area in question. It may secure an accurate diagnosis if the fracture line is demonstrated. However, extensive marrow edema precludes the visualization of the fracture line in some cases. Bone scintigraphy together with an accurate history is still a very useful tool in diagnosing stress fracture. MRI should be reserved for cases where the radiographic and scintigraphic findings are indeterminate.