Progression-Free Survival 2 as a Potential Surrogate for Overall Survival in Glioblastoma: Insights From a Spanish Multicentric Cohort.
In glioblastoma, first-line treatment can influence subsequent treatment lines. Therefore, progression-free survival 2 (PFS2), the time from initial randomization to progression on subsequent treatment or death, may better predict overall survival (OS) compared to PFS. However, this association needs validation in glioblastoma, particularly with novel targeted agents in early clinical trials.
Medical records of glioblastoma patients from a multicentric Spanish cohort with genomic profiling treated between 2018 to 2023 were analyzed. Correlations between OS and PFS or PFS2 were calculated using an iterative multiple imputation approach and time-to-event endpoints with Kaplan-Meier methods.
We analyzed 405 patients. Median OS, PFS, and PFS2 were 25.4, 8.7 and 16.4 months, respectively. Correlation between OS-PFS was 0.69 (95% CI, 0.62-0.75), whilst for OS-PFS2 it was 0.83 (95% CI, 0.78-0.87). In patients who received targeted therapy, OS-PFS correlation was 0.51 (95% CI, 0.02-0.81), whilst for OS-PFS2, it was 0.72 (95% CI, 0.30-0.91). For patients with ESCAT tier I-II molecular targets, OS-PFS correlation was 0.69 (95% CI, 0.26-0.89), weaker than OS-PFS2 (0.83 [95% CI, 0.29-0.97]). PFS2 also performed better than PFS in ESCAT III-IV patients (0.80 [95% CI, 0.75-0.85] vs. 0.69 [95% CI, 0.61-0.75]) and in those without targetable alterations (0.86 [95% CI, 0.74-0.92] vs. 0.68 [95% CI, 0.51-0.80]).
Progression-free survival 2 (PFS2) is a more robust surrogate for OS than PFS in glioblastoma, including patients receiving targeted therapies or without actionable targets. These findings support adoption of PFS2 as a candidate surrogate endpoint in glioblastoma, offering a balance between trial feasibility and meaningful survival outcomes.
Medical records of glioblastoma patients from a multicentric Spanish cohort with genomic profiling treated between 2018 to 2023 were analyzed. Correlations between OS and PFS or PFS2 were calculated using an iterative multiple imputation approach and time-to-event endpoints with Kaplan-Meier methods.
We analyzed 405 patients. Median OS, PFS, and PFS2 were 25.4, 8.7 and 16.4 months, respectively. Correlation between OS-PFS was 0.69 (95% CI, 0.62-0.75), whilst for OS-PFS2 it was 0.83 (95% CI, 0.78-0.87). In patients who received targeted therapy, OS-PFS correlation was 0.51 (95% CI, 0.02-0.81), whilst for OS-PFS2, it was 0.72 (95% CI, 0.30-0.91). For patients with ESCAT tier I-II molecular targets, OS-PFS correlation was 0.69 (95% CI, 0.26-0.89), weaker than OS-PFS2 (0.83 [95% CI, 0.29-0.97]). PFS2 also performed better than PFS in ESCAT III-IV patients (0.80 [95% CI, 0.75-0.85] vs. 0.69 [95% CI, 0.61-0.75]) and in those without targetable alterations (0.86 [95% CI, 0.74-0.92] vs. 0.68 [95% CI, 0.51-0.80]).
Progression-free survival 2 (PFS2) is a more robust surrogate for OS than PFS in glioblastoma, including patients receiving targeted therapies or without actionable targets. These findings support adoption of PFS2 as a candidate surrogate endpoint in glioblastoma, offering a balance between trial feasibility and meaningful survival outcomes.
Authors
Yaringaño Yaringaño, Mirallas Mirallas, Navarro Navarro, GĂłmez-Puerto GĂłmez-Puerto, Velilla Velilla, Vaz Vaz, Gorria Gorria, MartĂnez-Monino MartĂnez-Monino, HernĂĄndez HernĂĄndez, Aguado Aguado, Di Muzio Di Muzio, Valbuena Valbuena, Ruiz-Pace Ruiz-Pace, GonzĂĄlez GonzĂĄlez, MartĂnez-GarcĂa MartĂnez-GarcĂa, Villacampa Villacampa, Balaña Balaña, Carles Carles, Dienstmann Dienstmann, Pineda Pineda, SepĂșlveda SepĂșlveda, Vieito Vieito
View on Pubmed