Clinical predictors of pseudoprogression in glioblastoma: a retrospective cohort analysis

Distinguishing pseudoprogression (PsP) from true progression (TP) in glioblastoma (GBM) remains a diagnostic challenge, yet is essential for guiding treatment and counseling prognosis. This study retrospectively assessed the incidence, clinical predictors, and survival impact of PsP compared to TP.

Patients with surgically treated GBM and postoperative (chemo)radiotherapy in two Dutch hospitals (2006-2021) were included. Reports of magnetic resonance imaging (MRI) scans performed 4 months post-radiotherapy and at 3-month intervals, as well as reports of MRI scans prompted by neurological decline, were evaluated for PsP, TP, or mixed response (MR). Associations with clinical, tumor, and treatment characteristics and overall survival (OS) were analyzed.

Of 424 GBM patients, 175 were eligible for PsP analysis. The incidence of PsP was 29.1%, and PsP was associated with longer OS (median 16.6 months, 95% CI 12.0-21.2) compared to MR (14.1 months, 95% CI 11.1-17.2) and TP (11.6 months, 95% CI 10.0-13.2; p = 0.010). However, PsP occurring < 4 months after chemoradiotherapy was linked to shorter OS (11.3 months) than PsP > 4 months (17.4 months; p = 0.027). Male sex was significantly associated with outcome in univariate analysis, showing a trend toward significance in multivariate analysis. Treatment completion remained significant only in the multivariate model.

PsP is associated with improved survival compared to TP, though early-onset PsP portends poorer outcomes. None of the evaluated factors were a significant predictor of PsP in both univariate and multivariate analyses. Future research should focus on validating molecular markers, and refining PsP definitions using standardized criteria.