Abstract 4442
Background
Real-world (rw) evidence derived from electronic health records (EHRs) is being explored for use as a comparator to single arm clinical trials in oncology. While RECIST is a standard for assessing tumor response in trials, alternative approaches are needed in rw cohorts. Our objective was to examine whether rwRR derived from abstracted “real-world tumor response” (rwTR) might perform similarly to RECIST-based response rates in a variety of settings to support assessments of drug efficacy.
Methods
We focused on randomized clinical trials supporting FDA approvals in aNSCLC between 2015-2017 (4 trials, 5 arms). We selected rw cohorts of aNSCLC patients from de-identified Flatiron Health EHR-derived database by applying the eligibility criteria of each trial as feasible, following an established internal process; trials were considered infeasible for comparison and not reported when fewer than 30 eligible rw patients were identified at the time of analyses. rwTR was abstracted from unstructured clinical notes and was based on the treating oncologist’s assessment of disease change (e.g., partial response) indicated by imaging at each point of radiographic disease evaluation. We compared the rwRRs observed in rw cohorts to objective response rates (ORRs) from trials.
Results
rw cohorts included 862 patients, and clinical trial cohorts (2 control arms and 3 experimental arms) included 701 patients, see response rates in the table (all except ALEX were confirmed RRs).Table:
1581P
Trial | Trial arm | ORR | rwRR |
---|---|---|---|
CheckMate-017 | E: Nivolumab / 2L+ squamous | 20% (27/135) | 25% (17/68) |
CheckMate-057 | E: Nivolumab / 2L+ nonsquamous | 19% (56/292) | 17% (51/294) |
KEYNOTE-021 | E: Carboplatin, Pembrolizumab, Pemetrexed/ 1L nonsquamous | 55% (33/60) | 42% (37/88) |
KEYNOTE-021 | C: Carboplatin and Pemetrexed/ 1L nonsquamous | 29% (18/63) | 34% (101/294) |
ALEX | C: Crizotinib/ 1L | 75% (114/151) | 68% (80/118) |
Conclusions
Data curated from EHRs of rw patients with aNSCLC can provide estimates of response rates that are comparable to RECIST-based response rates observed in similar patient populations studied in clinical trial settings. These findings demonstrate the potential of leveraging rw cohorts as a comparator to single arm clinical trials in oncology.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Flatiron Health, Inc.
Funding
Flatiron Health Inc., independent subsidiary of the Roche group.
Disclosure
X. Ma: Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group.: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche. N.C. Nussbaum: Shareholder / Stockholder / Stock options, Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group.: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche. K. Magee: Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche. A.B. Bourla: Shareholder / Stockholder / Stock options, Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche. M. Tucker: Shareholder / Stockholder / Stock options, Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group.: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche. L. Bellomo: Shareholder / Stockholder / Stock options, Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche. C. Bennette: Shareholder / Stockholder / Stock options, Full / Part-time employment, Flatiron Health, Inc. is an independent subsidiary of the Roche Group.: Flatiron Health, Inc.; Shareholder / Stockholder / Stock options: Roche.
Resources from the same session
2182 - Evaluating the prevalence of the expression of PD-L1 in NSCLC specimens with short-duration formalin fixation using IHC 22C3 pharmDx
Presenter: Keiichi Ota
Session: Poster Display session 1
Resources:
Abstract
5255 - [18F]-FDG PET/CT in predicting PD-L1 status in nasopharyngeal carcinoma
Presenter: Liang Zhao
Session: Poster Display session 1
Resources:
Abstract
4910 - Expression of PD-L1 in Chinese Patients with Common Cancers
Presenter: Min Zheng
Session: Poster Display session 1
Resources:
Abstract
4227 - The clearance of EGF by tumor-associated macrophages is suppressed by chemotherapeutic agent cisplatin
Presenter: Irina Larionova
Session: Poster Display session 1
Resources:
Abstract
5222 - VHIO-300 and a thousand one nights, a tale of Precision Medicine
Presenter: Ginevra Caratù
Session: Poster Display session 1
Resources:
Abstract
5668 - Matched Whole-Genome Sequencing and Whole-Exome Sequencing with Circulating Tumor DNA (ctDNA) Analysis are complementary modalities in clinical practice
Presenter: Robin Imperial
Session: Poster Display session 1
Resources:
Abstract
5772 - Exploring the role of genes associated with familial cancer syndromes on the development of multiple primary tumors
Presenter: Atanaska Mitkova
Session: Poster Display session 1
Resources:
Abstract
4784 - Doxorubicin resistance: early and advanced tumors can use two different strategies based on initial and profound abnormalities in microRNA expression signature
Presenter: Volodymyr Halytskiy
Session: Poster Display session 1
Resources:
Abstract
3456 - From tumor transcriptomes to underlying cell type proportions to better predict prognosis and response to treatments
Presenter: Yuna Blum
Session: Poster Display session 1
Resources:
Abstract
4976 - Optimization of automated germline DNA extraction from non-tumoral formalin-fixed paraffin embedded (FFPE) tissues
Presenter: Omar Youssef
Session: Poster Display session 1
Resources:
Abstract