Prospective Study of Dominant Intraprostatic Lesion (DIL) Response Before, During, and Post Radiation Treatment using Multi-Parametric MRI Biomarkers
Dumas M, Elshaikh MA, Lee JK, Feldman AM, Pantelic M, Hearshen D, Movsas B, Chetty IJ, and Wen N. Prospective Study of Dominant Intraprostatic Lesion (DIL) Response Before, During, and Post Radiation Treatment using Multi-Parametric MRI Biomarkers. Int J Radiat Oncol Biol Phys 2019; 105(1):S84.
Int J Radiat Oncol Biol Phys
Purpose/Objective(s): An IRB approved prospective study was developed to identify imaging biomarkers from multi-parametric MRI (mp-MRI) (DWI, DCE) to assess treatment (tx) response for prostate cancer. The study aims to characterize longitudinal changes of imaging features between the dominant intraprostatic lesion (DIL) & normal prostatic tissue (NP) over the radiation tx course & after it. Materials/Methods: MRIs were taken at 4 time points: pre-tx, half way through tx, end of tx, and 2 months post tx. So far 6 patients have been enrolled in the study. Patient 2 did not have a distinguishable DIL. Another was only able to complete two scans. Patient 5 missed the end of tx scan. 78 Gy was prescribed to the whole prostate in 39 fx. MR scan parameters were as follows: T2 weighted imaging with a FRFSE (fast recovery fast spin echo) sequence: TR/TE = 3500/110 ms, 18 cm FOV, 0.42 x 0.42 x 2.5 mm3 voxel size; Diffusion Weighted (DW) echo planar imaging & trace diffusion sensitization for 4 b-values: 0, 500, 1000, and 1500 s/mm2, and following parameters: TR/TE = 5000/77.7 ms, 1.25 x 1.25 x 5.5 mm3 voxel size; DCE & T1 mapping (flip angles 2°, 5°, 10°, 15°, 20°, 25°) using a 3D SPGR sequence with TR/TE = 3.6 ms/1.3 ms/15°, 22 cm FOV, 0.75 × 0.75 × 2 mm voxel size and 9 s temporal resolution. The extended Tofts model was used for pharmacokinetic modeling. Parametric maps were calculated from the DCE data: extracellular, extravascular space fraction (ve) and volume transfer constants Ktrans & Kep. An expert radiologist drew DILs on each MR modality. A similar sized NP contour was also drawn on the contralateral side. To demonstrate tumor response relative to the NP, Mp-MRI features from the DIL were normalized by dividing each DIL value by the corresponding NP mean value. For histogram skewness normalization, NP skewness was subtracted from DIL skewness. Results: Seven imaging features revealed significant dose response in the DIL relative to NP: mean ADC; histogram’s ADC 10th percentile (%ile); histogram 90th %ile of exp. ADC, Ktrans, and Kep; skewness of Kep & Ktrans. ADC mean showed initial response & then stabilized, while 10th %ile showed a trend through the 2 month follow up. In Ktrans & Kep, the skewness indicated a histogram peak shift from right to left over the course of tx, which agrees with a decreasing 90th %ile values. Conclusion: This study identified 7 features that correlate to radiation dose over tx. The strongest trends were discovered in the Ktrans histogram skewness & Kep 90th %ile metrics. These metrics show a lower contrast inflow & outflow over the course of tx from the DIL relative to NP. This work successfully demonstrated metrics to track DIL changes with radiation dose, and subsequent post tx stability. The changes of these features over the tx course have potential to predict DIL recurrence or progression in the prostate. [Figure presented]