Personalized dosimetry for Re-188-SIRT: a comparison of gafchromic film and commercial tools—are we delivering what we intent?

<i>Objective.</i>Personalized dosimetry in radiopharmaceutical therapy offers superior accuracy over generic methods, though absorbed dose (AD) estimates vary due to differing techniques. While Monte Carlo simulation (MCS) is precise but complex, gafchromic film dosimetry offers a simpler alternative with strong agreement to MCS for<i>β</i>-AD. We present a cost-effective, high-resolution gafchromic film dosimetry using a custom polymethyl methacrylate (PMMA) phantom for Rhenium-188. A Dosimetric Volume Unit (DVU) was defined and a conversion factor (J GBq^-1) was used to estimate the AD to liver tumors of varying volumes via mono-compartmental method in¹⁸⁸Re-Selective Intra-arterial Radionuclide Therapy (SIRT) patients. Results were compared with three commercial softwares: OLINDA/EXM® V2.0, MIRDcalc, and QThera.AI.<i>Approach.</i>A tissue-equivalent PMMA phantom incorporating a central cavity for EBT3 gafchromic film placement was developed for high-resolution dosimetry of¹⁸⁸Re. Optical density measurements from gafchromic film exposed to varying¹⁸⁸Re activities were utilized to quantify AD as a function of radioactivity. Percentage depth dose curves were generated using 25<i>µ</i>Ci of¹⁸⁸Re for AD estimates at every 0.275 mm depth increments along the entire path length. Dose fall-off profiles were used to define the DVU, from which a conversion factor (J GBq^-1) was derived and applied for patient-specific AD estimation using mono-compartmental method in 25¹⁸⁸Re-SIRT patients. The results were compared with dosimetric outputs from three commonly-used softwares.<i>Main results.</i>A perfect linear correlation (<i>R</i>²= 1) was observed between AD and radioactivity. In a homogeneous distribution, 1<i>µ</i>Ci of¹⁸⁸Re delivered 3.6 Gy mm^-3in the first half-life and a cumulative 7.3 Gy mm^-3over ten half-lives, with a maximum tissue penetration of 3.7 mm. The DVU dimensions for¹⁸⁸Re were estimated as 4.7 × 4.7 × 7.4 mm³, and the conversion factor was calculated as 34 J GBq^-1. In 25 patients, the median AD calculated using mono-compartmental model was 77.5 Gy (Interquartile range IQR: 57.1-120.2 Gy), closely aligning with QThera.AI estimates [87.1 (IQR: 43.0-127.9) Gy;<i>p</i>= 0.968] and showing a strong correlation (<i>r</i>= 0.88). In contrast, MIRDcalc produced significantly lower ADs [47.6 (IQR: 29.4-84.0) Gy;<i>p</i>= 0.000] with a moderate correlation (<i>r =</i> 0.62), while OLINDA/EXM® V2.0 yielded the lowest values [13.5 (IQR: 7.0-24.) Gy;<i>p</i>= 0.000] and exhibited a weak correlation<i>(r</i>= 0.58).<i>Significance.</i>DVU based mono-compartmental method showed strong concordance with QThera.AI estimates, supporting its use for accurate, empirical AD calculation in¹⁸⁸Re-SIRT, similar to⁹⁰Y dosimetry. Given the high cost of commercial softwares, this low-cost approach may offer a practical alternative, promoting broader adoption of precision dosimetry, particularly in resource-limited settings.