- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Faculty Research and Publications /
- Machine Learning and Computed Tomography Radiomics...
Open Collections
UBC Faculty Research and Publications
Machine Learning and Computed Tomography Radiomics to Predict Disease Progression to Upfront Pembrolizumab Monotherapy in Advanced Non-Small-Cell Lung Cancer: A Pilot Study Janzen, Ian; Ho, Cheryl; Melosky, Barbara; Ye, Qian; Li, Jessica; Wang, Gang; Lam, Stephen; MacAulay, Calum; Yuan, Ren
Abstract
Background/Objectives: Pembrolizumab monotherapy is approved in Canada for first-line treatment of advanced NSCLC with PD-L1 ≥ 50% and no EGFR/ALK aberrations. However, approximately 55% of these patients do not respond to pembrolizumab, underscoring the need for the early intervention of non-responders to optimize treatment strategies. Distinguishing the 55% sub-cohort prior to treatment is a real-world dilemma. Methods: In this retrospective study, we analyzed two patient cohorts treated with pembrolizumab monotherapy (training set: n = 97; test set: n = 17). The treatment response was assessed using baseline and follow-up CT scans via RECIST 1.1 criteria. Results: A logistic regression model, incorporating pre-treatment CT radiomic features of lung tumors and clinical variables, achieved high predictive accuracy (AUC: 0.85 in training; 0.81 in testing, 95% CI: 0.63–0.99). Notably, radiomic features from the peritumoral region were found to be independent predictors, complementing the standard CT evaluations and other clinical characteristics. Conclusions: This pragmatic model offers a valuable tool to guide first-line treatment decisions in NSCLC patients with high PD-L1 expression and has the potential to advance personalized oncology and improve timely disease management.
Item Metadata
| Title |
Machine Learning and Computed Tomography Radiomics to Predict Disease Progression to Upfront Pembrolizumab Monotherapy in Advanced Non-Small-Cell Lung Cancer: A Pilot Study
|
| Creator | |
| Contributor | |
| Publisher |
Multidisciplinary Digital Publishing Institute
|
| Date Issued |
2024-12-28
|
| Description |
Background/Objectives: Pembrolizumab monotherapy is approved in Canada for first-line treatment of advanced NSCLC with PD-L1 ≥ 50% and no EGFR/ALK aberrations. However, approximately 55% of these patients do not respond to pembrolizumab, underscoring the need for the early intervention of non-responders to optimize treatment strategies. Distinguishing the 55% sub-cohort prior to treatment is a real-world dilemma. Methods: In this retrospective study, we analyzed two patient cohorts treated with pembrolizumab monotherapy (training set: n = 97; test set: n = 17). The treatment response was assessed using baseline and follow-up CT scans via RECIST 1.1 criteria. Results: A logistic regression model, incorporating pre-treatment CT radiomic features of lung tumors and clinical variables, achieved high predictive accuracy (AUC: 0.85 in training; 0.81 in testing, 95% CI: 0.63–0.99). Notably, radiomic features from the peritumoral region were found to be independent predictors, complementing the standard CT evaluations and other clinical characteristics. Conclusions: This pragmatic model offers a valuable tool to guide first-line treatment decisions in NSCLC patients with high PD-L1 expression and has the potential to advance personalized oncology and improve timely disease management.
|
| Subject | |
| Genre | |
| Type | |
| Language |
eng
|
| Date Available |
2025-01-17
|
| Provider |
Vancouver : University of British Columbia Library
|
| Rights |
CC BY 4.0
|
| DOI |
10.14288/1.0447775
|
| URI | |
| Affiliation | |
| Citation |
Cancers 17 (1): 58 (2025)
|
| Publisher DOI |
10.3390/cancers17010058
|
| Peer Review Status |
Reviewed
|
| Scholarly Level |
Faculty; Researcher
|
| Rights URI | |
| Aggregated Source Repository |
DSpace
|
Item Media
Item Citations and Data
Rights
CC BY 4.0