Initial Utilization of An Outcomes Database to Determine the Effects of Planning Technique Change On Patient Outcomes
A Du*, B Morris, P Lewiz, A Beggs, S Hedrick, Provision Center for Proton Therapy, Knoxville, TN
Presentations
PO-GePV-M-27 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: Changes in department standardization of treatment plans must be driven by evidence of improved outcomes. We developed an outcomes database for head and neck patients which combines pre- and post- treatment outcomes and on treatment information with a comprehensive list of treatment plan parameters. Using this database, we ran a study to determine whether a departmental change involving the deletion of spots from the patient’s skin prior to plan finalization (treatment parameter) correlated to reaction dermatitis score (outcome) in head and neck patients.
Methods: Patient outcome data was mined from the oncology information system and a cloud-based patient tumor registry (Vision Tree, San Diego, CA). Outcome data was associated with sets of treatment parameters based on the time frame in which the outcome assessment dates occurred relative to the treatment date. Next, we split this aggregate database based on whether a patient had received spot deletion. No patients fell in both categories. We then reduced each patient’s list of radiation dermatitis ratings (non-existent=0, most severe=3) throughout their assessment history to a max score.
Results: To test the null hypothesis that the max radiation scores were equal for patients that received spot deletion (N=9, M=1.333,SD=1.0), and did not receive it (N=20,M=1.8,SD=0.52) an unpaired t-test was used. A p-value of 0.10795793 was obtained. We note that while our p-value is slightly greater than α=0.1 by 0.007, the difference is insignificant enough to reject the null hypothesis. We can conclude that patients who did not receive the spot deletion had more severe max radiation dermatitis.
Conclusion: Our example study showed the utility of a database which aggregates outcomes assessment with treatment plan information to find correlations between multiple streams of patient data. Future directions include constructing a similar database for other common treatment sites and implementing more sophisticated data analysis techniques.
Funding Support, Disclosures, and Conflict of Interest: Funded by Provision Healthcare Knoxville, TN
Keywords
Treatment Planning, Source Strength, Treatment Techniques
Taxonomy
TH- Response Assessment: Radiomics/texture/feature-based response assessment
Contact Email