Researcher ORCID Identifier

Graduation Year


Date of Submission


Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts



Reader 1

Dr. Alejandro Bertolet

Reader 2

Dr. Janet Sheung

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

© 2022 Marcus J Lindsey


Trans-arterial radioembolization (TARE) is a protracted modality of radiation therapy where radionuclides labeled with Yttrium-90 (90Y) are inserted inside a patient's hepatic artery to treat hepatocellular carcinoma (HCC). While TARE has been shown to be a clinically effective and safe treatment, there is little understanding of the radiobiological relationship between absorbed dose and tissue response, and thus there is no dosimetric standard for treatment planning. The Biologically Effective Dose (BED) formalism, derived from the Linear-Quadratic model of radiobiology, is used to weigh the absorbed dose by the time pattern of delivery. BED is a virtual dose that can be thought of as a common 'language' with which various forms of radiation therapy can use to 'communicate'. BED allows conclusions to be drawn about the biological response of TARE by putting it in conversation with what is known about the biological response of other treatment modalities, namely external beam radiation therapy (EBRT). A Python program was developed to calculate BED from absorbed dose distributions of six HCC patients treated at Massachusetts General Hospital, and optimized treatment activity levels with respect to biological response were determined. Within the limits of the BED analysis, none of the patients had originally received the optimal dose, with some patients having been overdosed and some having been underdosed. The results show there is a disconnect between the current clinical treatment planning standard for TARE and the tissue's biological response. This study suggests the need for patient-specific TARE dosimetry which considers biological response, such as BED or another adaptive model.