Abstract BACKGROUND Evaluation of response to radiotherapy in diffuse glioma remains a diagnostic challenge, as treatment-associated changes (TACs - radionecrosis, pseudoprogression) are radiologically difficult to differentiate from progressive disease (PD). We aimed to validate previously described predictive clinical and magnetic resonance imaging (MRI) criteria in a large cohort with histological confirmation available for all patients, which offers additional certainty over the radiological diagnosis. METHODS We included patients with WHO grade 2-4 diffuse gliomas treated with radiotherapy, who had further resection or biopsy for radiological progression. Grade 2 tumours were included in the presence of radiological and/or histological progression to a higher grade. TACs were defined as absence of proliferative activity, confirmed by histopathology. A neuroradiologist assessed seven previously described predictive conventional, diffusion and perfusion MRI characteristics of PD. Two previously described clinical characteristics were also evaluated (time from radiotherapy to radiological progression, TTP and total radiotherapy dose). Statistics included binomial univariable and multivariable logistic regression analysis with odds ratios (OR) and 95%-confidence intervals (95%-CI). RESULTS We evaluated 170 resections of 155 patients (69% males) with a median age of 55 years (interquartile range, IQR; 46-63). Most patients had a glioblastoma or astrocytoma grade 4 not-otherwise specified (N=106, 62%). Radiotherapy alone was applied in 118 patients (69%). A partial/complete resection was performed in 84% and 16% had a biopsy. We found PD in 141 resections (83%) and TACs in 29 (17%). The median TTP was 11.6 months (IQR, 4.6-33.8). Perfusion MRI was not evaluated, because of a high number of missing values (85%) due to the retrospective nature of the study. In univariable analysis, soap bubble enhancement (OR for PD, 2.58; 95%-CI, 1.12-5.95; p-value, 0.027), iso-intense apparent diffusion coefficient (ADC) compared to hyperintensity (OR, 4.5; 95%-CI, 1.28-15.81; p-value, 0.019) and a longer TTP of more than 6 months compared to less than 3 months (OR, 3.09; 95%-CI, 1.19-8.03; p-value, 0.021) were associated with PD. In multivariable analysis, longer TTP (OR, 3.12; 95%-CI, 1.12-8.68; p-value, 0.030) and iso-intense ADC (OR, 4.09; 95%-CI, 1.05-15.95; p-value, 0.043) predicted PD. CONCLUSION We validate previously described predictive clinicoradiological criteria in a large cohort of patients with histologically-confirmed outcome to differentiate TACs from PD. Soap bubble enhancement, iso-intense ADC and longer TTP were confirmed as predictors for PD. Perfusion MRI criteria was not evaluated herein, however it should be included into future predictive models. Support/Disclosure: The StophersenkankerNU Foundation, T&P Bohnenn Fund, Vrienden UMC Utrecht
Flies et al. (Wed,) studied this question.