Site Editor
Gregory J. Riely, MD, PhD
Gregory J. Riely, MD, PhD
Posted: Monday, December 9, 2024
E.M.W. van de Garde, PhD, of St Antonius Hospital, Utrecht, the Netherlands, and colleagues developed and evaluated the portability of a text-mining algorithm for prospectively capturing disease progression in electronic health data from patients with metastatic non–small cell lung cancer (NSCLC) who were treated with immunochemotherapy. Their findings were published in JCO Clinical Cancer Informatics.
“Manual [chart] review is not suitable for remote real-time detection of [disease] progression events in large cohorts under follow-up, as it would require repeated electronic health record reviews per patient over time, which is time-consuming,” the investigators commented. “[Our] algorithm was portable across different hospitals, and the performance over time was good, making it an interesting approach for prospective follow-up of multicenter cohorts.”
Using electronic health data from patients with stage IV disease who received first-line pembrolizumab plus chemotherapy in four Dutch hospitals, the investigators developed and optimized a text-mining algorithm for capturing disease progression in two hospitals (n = 75 and 91 patients) and transferred it to the other two (n = 50 and 50 patients) for external evaluation. To simulate real-time application, incremental data were provided at weekly intervals, and the algorithm was evaluated at each.
According to the investigators, the algorithm performed well in all hospitals, with all performance scores exceeding 90%. Real-time simulation revealed performance scores above 90% from week 15 in all four hospitals. The exact dates of disease progression were found to vary in 46 of the 157 patients with progressive disease (range: –116 to 384 days); however, the numbers of those labeled with such advancement too early (n = 24) and too late (n = 22) were well balanced. The progression-free survival curves constructed from text-mining analyses and manual chart review appeared to be similar across hospitals.
Disclosure: For full disclosures of the study authors, visit ascopubs.org.
JCO Clinical Cancer Informatics
JNCCN Spotlights
