MRI-defined treatment margins, urinary toxicity, and PSA response in LDR prostate brachytherapy

Implant quality metrics as measured by D90 and V100 do not address the adequacy of periprostatic margins. Relative margin deficiencies may relate to efficacy and margin excesses to post-implant toxicity. Our purpose is to determine MRI-defined treatment margins on prostate achieved with LDR brachytherapy.

Post implant CT and MR images from 487 consecutive patients who received LDR brachytherapy from 2010 to 14 were co-registered. Four prostate quadrants were defined; anterior-superior (AS), posterior-superior (PS), anterior-inferior (AI), posterior-inferior (PI). Dosimetric variables were measured for prostate, and for each quadrant with a 0-, 2-, 3-, and 5-mm margin, as well as for the membranous urethra defined on MRI.

Prostate D90 (no margin) was associated with D90 to the volume enclosed by 2 mm, 3 mm and 5 mm margins (R2 = 0.9 - 1.0) with an average 7.1% decrease in dose per mm of margin. Mean D90 for PS, AI and PI quadrants were > 110% of prescription dose for margins of 2-, 3-, and 5-mm. AS quadrant mean D90s were generally lower (83.2% for 2 mm, 76.4% for 3 mm and 62.2% for 5 mm). Urethral strictures (n = 9) were associated with higher doses in the AI quadrant, and higher membranous urethral V125 (51 vs. 32%, p 0.013) and V150 (34.5 vs. 17.6%, p 0.01).

Using MR-CT post implant dosimetry, margin coverage up to 5 mm was excellent with less margin coverage in the AS quadrant. Late ≥ grade 2 toxicity and urinary strictures are more likely to occur with relative margin excess in the anterior-inferior quadrant and higher doses caudal to the prostate apex. Within this analysis, there was no relationship between treatment margins, and PSA outcome.