2546 - Site-Specific Associations between NCCN Distress Thermometer Screening Data and NLP-Identified Toxicities during Radiotherapy
Presenter(s)
M. Zhao1, R. Benson2, M. V. Elia3, I. Friesner4, M. Ho5, W. T. Li5, C. I. Nnadi6, J. Chew2, N. W. Cho2, H. Vasudevan2, S. Sinha2, M. Garcia7, M. Rabow6, L. Boreta2, S. E. Braunstein2, D. Spiegel8, and J. C. Hong9; 1Department of Radiation Oncology, University of California San Francisco, San Francisco, CA, San Francisco, CA, 2Department of Radiation Oncology, University of California San Francisco, San Francisco, CA, 3Bakar Computational Heath Sciences Institute, University of California San Francisco, CA, San Francisco, CA, 4University of California, San Francisco, Bakar Computational Health Sciences Institute, San Francisco, CA, 5Bakar Computational Heath Sciences Institute, University of California San Francisco, San Francisco, CA, 6University of California San Francisco, San Francisco, CA, 7University of California, San Francisco, San Francisco, CA, 8Department of Radiation Oncology, Beth Israel Deaconess Medical Center, Boston, MA, 9University of California San Francisco, Department of Radiation Oncology, San Francisco, CA
Purpose/Objective(s):
Cancer and its treatment carry psychological burdens alongside physical symptoms. We previously demonstrated that patients with higher distress, measured with the NCCN Distress Thermometer (DT) from 0 to 10, are more likely to have symptoms such as pain, fatigue, and nausea documented during radiotherapy (RT). The objective of this study was to determine whether patients being treated for different cancers experience different associations between distress and symptomatology.Materials/Methods:
We analyzed a single-institution, retrospective cohort of adult patients who received RT between December 2018 and March 2021 and completed the DT by the first week of RT. We identified Common Terminology Criteria for Adverse Events (CTCAE)-defined symptoms during and up to 30 days following RT from clinical notes in the EHR utilizing a previously validated cTAKES-based NLP pipeline. We used Mann-Whitney U tests to compare distress distributions. We applied multiple logistic regression to identify associations between NCCN DT scores and a selection of symptoms within each of the most common cancer types, controlling for clinical covariates.Results:
Of 1,187 RT courses for 1,181 patients, the most documented symptoms were pain, fatigue, nausea, vomiting, and constipation. The diagnoses most represented were metastatic (21.8% of patients), breast (13.2%), head and neck (HN, 12.7%), and prostate (12.7%) cancer. The median NCCN DT score was 2.0 (IQR 0.0–5.0). Patients with metastatic (mean DT 2.86) and HN (2.91) cancers had higher distress than patients with prostate cancer (2.26). Higher distress in patients with metastatic cancer was associated with general pain (OR 1.18, 95% CI [1.01-1.38]), fatigue (1.21, [1.07-1.38]), nausea (1.13, [1.01-1.27]), generalized muscle weakness (1.17, [1.05-1.30]), and back pain (1.12, [1.01-1.24]) during/after RT. Patients with HN cancer with higher distress showed increased fatigue (1.21, [1.03-1.42]) during/after RT, but not site-specific symptoms (ie. dysgeusia, oropharyngeal pain, oral mucositis). Higher distress in patients with breast and prostate cancers was neither associated with general nor site-specific symptoms (i.e. hot flashes, dermatitis radiation, dysuria, urinary incontinence) during/after RT.Conclusion:
This study found that distress-symptom associations vary by cancer diagnosis. Higher distress was associated more with broad, non-localized symptoms in patients with HN cancer than with site-specific, objectively defined symptoms that may be radiation-related. Our study is potentially limited by discrepancies between patient reporting and provider documentation. Future work may adjust for RT target and dosage, as well as disease stage. Further investigation of population-specific patterns in distress responses may provide insight into potential barriers and facilitators to effective symptom management.