Background The study aimed to evaluate the agreement between the radiographic union scale (RUST) and modified RUST (mRUST) in humeral shaft fractures treated with different techniques, and the effect of surgeons' experience and thresholds for determining bone union. Materials and Methods A total of 20 orthopedic surgeons reviewed and scored radiographs of 30 patients with humeral shaft fractures treated by external fixation, intramedullary nailing, and plating using the RUST and mRUST on the 0 day, 6 weeks, 12 weeks and 24 weeks follow-up radiographs. Bone healing, interrater agreement between RUST and mRUST scores, and the threshold for radiographic union were evaluated. Results The intraclass correlation coefficient (ICC) was slightly higher for the mRUST score than the RUST score (0.71 versus [vs.] 0.67). There was substantial agreement between the mRUST and RUST scores for external fixation (0.75 and 0.69, respectively) and intramedullary nailing (0.79 and 0.71); there was moderate agreement between them for plating (0.59 and 0.55). Surgeons with varying experience had a similar agreement for both scores and scores for each humeral cortex. The external fixation and intramedullary nailing group had higher RUST and mRUST scores than the plating group. The ICC for union was substantial (0.64; external fixation: 0.68, intramedullary nailing: 0.64, and plating: 0.61). More than 90% of the reviewers recorded scores of 10/12 for RUST and 13/16 for mRUST at the time of union. Conclusions RUST and mRUST scores can be used reliably for the evaluation of bony union in humeral fractures treated with an external fixator and intramedullary nailing. In cases of humeral plating, a more sensitive tool for evaluation of fracture union is needed.