Portal de Periódicos da CAPES

Investigating Semi-Quantitative In Vivo Spatial and Dosimetric Analytical Algorithms to Report GBM Patterns of Failures Following Chemoradiation in the IMRT Era

2018; Elsevier BV; Volume: 102; Issue: 3 Linguagem: Inglês

10.1016/j.ijrobp.2018.07.807

1879-355X

H. Elhalawani, T Lin, Karine A. Al Feghali, S. Radwan, Neil Chevli, Abdallah Mohamed, Clifton D. Fuller, C. Chung,

Medical Imaging Techniques and Applications

Prior studies have generally classified glioblastoma multiforme (GBM) failures following concurrent chemoradiation (CRT) and temozolomide (TMZ) as either marginal or in-field. This study aims to investigate GBM patterns of failure in a semi-quantitative manner through the use of three-dimensional (3D) spatial dose distribution and deformable registration processes in GBM patients (pts) exclusively treated with intensity-modulated RT (IMRT). Eligible patients had a radiologically documented recurrence following treated with post-operative definitive TMZ and RT with IMRT for biopsy-proven GBM at a single institution between 2009 and 2013 were scanned. Planning CT (pCT) scans along with 3D dose grid, contours and RT plan were extracted in DICOM format. Gross recurrent tumor volume (rGTV) was segmented as the enhancing lesion on post-contrast T1 MRI documenting the earliest radiographic evidence of recurrence (rT1C), and the centroid of the rGTV was placed. Validated deformable image registration software was used to co-register the rT1C to pCT. Deformably mapped rGTVs were compared dosimetrically to the initial 3D planned dose and spatially to the target volumes (TV). Failures types were classified based on: (1) location of centroid relative to the initial TV, where central = centroid within TV and peripheral = centroid outside TV and (2) dosimetry based on dose to 95 % failure volume (fD95%), where high-dose = fD95% is >95 % dose prescribed to initial TV) and low-dose = fD95% is <95% dose prescribed to initial TV. Recurrences were classified into 5 types: A (central high-dose), B (peripheral high-dose), C (central low-dose), D (peripheral low-dose), and E (extraneous). Twenty-five eligible pts were analyzed. All pts were prescribed 60 Gy in 30 fractions. Median time to recurrence was 14.2 months (IQR: 6 – 61). Recurrences had a median volume of 6.35 cc (IQR: 3.8-25.6) and were all ipsilateral to the original tumor. The following recurrence patterns were seen: 15 (62.5%) type A, 6 type B (25%) and 3 type E (12.5%). No types C or D recurrences were seen. The recurrence nidus originated within the initial gross tumor volume (GTV) in 70.8%, within clinical target volume (CTV) but outside GTV in 16.7%, and outside PTV in 12.5%. Type B failures showed a non-significant trend for higher median volume (24.6 cc) than types A (4.5 cc) or E (5.4 cc). Median fD95% for types A, B & E were 61, 55.9 and 25.4 Gy, respectively. Using this approach that incorporates serial 3D tumor volumetric and dosimetric data to evaluate patterns of recurrence, the majority of recurrences following 60Gy IMRT for newly diagnosed GBM were within the high-dose volume. Further investigation using deformable registration of serial images to map patterns of failure in relation to dosimetry may help identify dose and spatial features associated with tumor failure.