STUDY DESIGN:Comparative reliability and prospective validity. OBJECTIVE:First, to evaluate the reliability of four methods of assessing magnetic resonance imaging (MRI) bone marrow oedema (BMO) of the posterior vertebral arch of the lumbar vertebrae of elite junior fast bowlers. Second, to evaluate the validity of the most reliable method for the early detection of lumbar bone stress injury. SUMMARY OF BACKGROUND DATA:MRI has demonstrated utility in identifying bone marrow oedema (BMO) in lumbar vertebrae. Methods to grade the severity of BMO may provide valuable insight to inform clinical management, particularly in elite athletes where detection of early-stage bone stress may prevent progression to more severe and costly bone stress injury. METHODS:65 male elite junior cricket fast bowlers had repeat MRI scans during a cricket season. A subset of 19 bowlers' images were re-assessed by experienced musculoskeletal radiologists to determine intra- and inter-rater reliability. All images were aligned with independent medical records of lower back symptoms and diagnosed bone stress injuries to establish the relationship of BMO and lumbar bone stress injury. RESULTS:Clinical detection of abnormal BMO, whether the pars region of the vertebra was considered in its entirety or subdivided into regions, had fair-to-moderate inter-rater reliability, and fair-to-almost perfect intra-rater reliability. Measurement of BMO signal intensity using an imaging software tool had excellent intra-rater and inter-rater reliability (ICC = 0.848, 0.837). BMO signal intensity was positively associated with subsequent LBSI (p < 0.001), and differentiated between asymptomatic and symptomatic bowlers (p < 0.001). CONCLUSIONS:Measurement of BMO signal intensity using an imaging software tool proved a reliable and valid measure of the severity of lumbar bone stress injury in elite junior fast bowlers. LEVEL OF EVIDENCE:2.