While radiotherapy damages tumor cells, it can also harm the surrounding healthy tissue. This damage can lead to acute and/or late toxicities. For patients with mediastinal tumors like Hodgkin lymphoma and thymic cancer, this could include pneumonitis, esophagitis, secondary cancers, or cardiac toxicity. The purpose of this thesis was to investigate how to best apply modern techniques in radiotherapy to reduce the dose to normal tissue and therefore the risk of toxicity for patients treated for cancer in the mediastinum. The aims of this thesis ranged from applied to more exploratory. The first aim was to calculate the treatment margins for patients with mediastinal lymphoma treated at our institution with the modern techniques of deep inspiration breath hold (DIBH) and daily image guidance. The second aim was to investigate the impact of the two modern techniques of DIBH and proton therapy, relative to treatment in free breathing and photon therapy, for patients with Hodgkin lymphoma and thymic cancer. The third aim was to investigate biological uncertainties in proton therapy due to the linear energy transfer (LET) of the protons at the end of their range for pediatric patients with Hodgkin lymphoma. And finally, the fourth aim was to create photon therapy treatment plansthat minimized the risk of both tumor recurrence and mortality from late toxicity from radiotherapy for patients with Hodgkin lymphoma. The methods for these aims included a retrospective analysis of set-up uncertainties, treatment planning comparisons for different combinations of techniques, Monte Carlo simulations for proton therapy plans, and using an in-house optimizer. For aim 1, we found that a margin of approximately 1 cm was needed to include uncertainties from contouring, setup, and intra-fractional motion for the lymphoma patients treated at our institution with DIBH and daily image guidance. For aim 2, we found that both proton therapy and DIBH reduced the dose to normal tissues, but the technique that had a larger impact depended on the patient group. For aim 3, we did not find a clinically concerning distribution of LET or impact on biologically-weighted dose for pediatric patients with Hodgkin lymphoma. And for aim 4, we created “outcome-optimized” plans, but the optimizer was sensitive to model parameters and not ready for clinical use. The results of this thesis can be used to inform clinical practice and future research for patients with Hodgkin lymphoma and thymic cancer to reduce the radiation dose to normal tissue, and therefore the risk of acute and late toxicity.