Artificial Intelligence to Advance Adaptive Radiation Therapy
| dc.contributor.advisor | Smith, Wendy Lani | |
| dc.contributor.advisor | Schinkel, Colleen | |
| dc.contributor.author | Weppler, Sarah Joanne | |
| dc.contributor.committeemember | Davidsen, Jörn | |
| dc.contributor.committeemember | Quon, Harvey C. | |
| dc.date | 2020-11 | |
| dc.date.accessioned | 2020-09-03T18:24:51Z | |
| dc.date.available | 2020-09-03T18:24:51Z | |
| dc.date.issued | 2020-09-03 | |
| dc.description.abstract | The precision of head and neck cancer radiotherapy may be adversely affected by changes in patient and tumor anatomy occurring over the 6-7 weeks of daily treatment. Adaptive radiation therapy (ART) is used to correct precision losses by replanning treatment in response to anatomical changes. However, the resource costs associated with routine ART may be prohibitive. This thesis considers various open questions in head and neck ART including: protocol performance; workflow streamlining and patient selection criteria; and the potential clinical implications of plan adaptation. We first proposed a framework to compare physician dose-monitoring priorities against protocol capabilities and assessed a common ART protocol monitoring changes in patient external contour. We found this protocol’s performance is comparable to randomly selecting patients for ART. Artificial intelligence techniques allowed us to more effectively model interactions between anatomical and dosimetric changes, and propose new patient selection practices. A novel heuristic converted models into simple criteria for clinical use. Various ART correction goals were considered with promising performance on an external validation dataset, including: parotid gland sparing (sensitivity=0.82, specificity=0.70), and pharyngeal constrictor sparing (sensitivity=0.84, specificity=0.68). Selection criteria relied only on pre-treatment patient data, allowing ART consults to be scheduled in advance. Deformable image registration (DIR) is a workflow tool capable of further streamlining ART. However, not all DIR implementations produce equivalent results. We proposed a data clustering method to identify representative examples of DIR differences that most affect ART output. In addition, we provided a general framework to derive workflow-specific DIR performance requirements that ensures ART workflow equivalence. To estimate the potential clinical benefit of ART, we compared patient-reported outcomes with planned and delivered radiotherapy doses. Results indicated that ART may be most beneficial in reducing patient-reported dysphagia, conferring a ≥5% decrease in absolute dysphagia risk in 1.2% of patients with dose increases, with a ≥5% decrease in relative risk in 23.3% of patients. This suggests a novel goal for ART, given that the literature primarily focusses on xerostomia reduction. The results from these studies have prepared our centre to be among the first to lead a randomized clinical trial in ART. | en_US |
| dc.identifier.citation | Weppler, S. J. (2020). Artificial Intelligence to Advance Adaptive Radiation Therapy (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. | en_US |
| dc.identifier.doi | http://dx.doi.org/10.11575/PRISM/38162 | |
| dc.identifier.uri | http://hdl.handle.net/1880/112490 | |
| dc.publisher.faculty | Science | en_US |
| dc.publisher.institution | University of Calgary | en |
| dc.rights | University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. | en_US |
| dc.subject.classification | Oncology | en_US |
| dc.subject.classification | Physics--Radiation | en_US |
| dc.subject.classification | Artificial Intelligence | en_US |
| dc.title | Artificial Intelligence to Advance Adaptive Radiation Therapy | en_US |
| dc.type | doctoral thesis | en_US |
| thesis.degree.discipline | Physics & Astronomy | en_US |
| thesis.degree.grantor | University of Calgary | en_US |
| thesis.degree.name | Doctor of Philosophy (PhD) | en_US |
| ucalgary.item.requestcopy | true | en_US |