Archived

This content is available here for research, reference, and/or recordkeeping.

Author ORCID Identifier

https://orcid.org/0009-0003-4943-5662

Date Available

4-30-2026

Year of Publication

2026

Document Type

Doctoral Dissertation

Degree Name

Doctor of Philosophy (PhD)

College

Medicine

Department/School/Program

Radiation Medicine

Faculty

Wei Luo

Faculty

Ellis Lee Johnson

Abstract

The standard of care for locally advanced cervical cancer is chemoradiation, consisting of conventional external beam therapy (EBT) followed by a boost of high dose rate (HDR) brachytherapy. These two radiation treatments will illicit different degrees of biological response depending on the tissue in question, and quantifying the overall effect the combined treatments will have require some form of biological modeling. The linear quadratic (LQ) model of cell survival is the most widely used radiobiological model in clinics across the nation.

It is quantified by the parameters a and ß, which model how radiation generates lethal damage in a population of cells. The ratio of the two parameters a/ß, quantifies the relative sensitivity of tissues to the effects of radiation fractionation (i.e. delivering radiation in fewer larger fractions or vice versa), and is given in units of Gray (Gy).

This ratio can help inform clinicians on the relative efficacy of different radiation schedules. Early cell experiments established the common clinical practice of assigning a lower a/ß of 3 Gy to normal tissue and a high value of 10 Gy to most cancers, establishing the concept of a "therapeutic window". Earlier research into prostate and breast cancer put this assumption into question, as both cancers demonstrated much lower a/ß ratios on the level of normal tissue.

Cervical cancer still uses the commonly assumed 10 Gy, and prior to this study, no significant efforts have been made to reevaluate that assumption. In this thesis we reevaluated the a/ß of cervical cancer, both in vitro and from clinical data.

A large-scale meta-analysis was performed, analyzing 98 published experimental values across 31 studies. In addition, clinical and dosimetric data from 180 cervical cancer patients treated at our own institution was incorporated with locoregional tumor control rates to obtain clinical estimates of a/ß. In vitro results showed a wide spanning distribution of values, though the distribution was log-normal and strongly peaked at an a/ß of 4.25 Gy. For clinical estimation three previously published methods were applied to our patient data: the Poisson tumor control probability model based on the LQ model, logistic regression, and Cox Proportional Hazards regression. All three methods gave very similar estimates, averaging to about 3.14 Gy. This study is significant, as not only does the current treatment of cervical cancer rely on an estimation of a/ß, but these new estimates open the possibility to exploring new radiation schemes.

Digital Object Identifier (DOI)

https://doi.org/10.13023/etd.2026.199

Archival?

Archival

Funding Information

University of Kentucky Department of Radiation Oncology (2023-2026)

Download

Share

COinS