2264 - Renal Dosimetry for ¹⁷⁷Lu-DOTATATE: Impact of Region of Interest Selection

B. Driscoll^1,2, M. Zhou³, R. K. Wong¹, A. Shessel⁴, and I. W. T. Yeung⁵; ¹Radiation Medicine Program, Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada, ²Techna Institute, University Health Network, Toronto, ON, Canada, ³University Health Network, Toronto, ON, Canada, ⁴Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada, ⁵Southlake Regional Health Centre, Newmarket, ON, Canada

Purpose/Objective(s): OZM-067 trial is a multi-institution prospective single-arm study of ¹⁷⁷Lu-DOTATATE in patients with ⁶⁸Ga-DOTATATE positive, progressive metastatic disease. Patients received 4 cycles of therapy with individualized dosimetry as assessed using SPECT at 4, 24 and 72 hours post treatment. Dose escalation or de-escalation was determined based on renal dosimetry with cumulative threshold of 22 Gy over the 4 cycles. Purpose of this sub-study was to compare the method of dosimetry utilized in the trial, a 2-cm diameter spherical region of interest (ROI), with using a full kidney ROI. Materials/

Methods: The sub-study was performed using data from a random subset of 33 OZM-067 patients (50 infusions) all treated at the Princess Margaret Cancer Centre. Dosimetry for the study was performed by placing 2-cm diameter spheres on the kidney from each of the 3 SPECT-CTs and fitting a Bi-Exponential model to the activity values which could then be converted to absorbed dose (Gy) by MIRD formalism. For comparison, automated kidney segmentations were produced via Limbus AI Segmentation based on the CT from the higher quality pre-treatment ⁶⁸Ga-DOTATATE PET-CT in lieu of the SPECT-CT. These contours were then refined to remove any erroneously segmented regions and the renal pelvis. The segmented left and right kidney masks were then co-registered to the 4, 24 and 72 hour SPECT-CTs individually. In order to adjust for partial volume averaging, a 5 mm contraction of the kidney segmentation was performed prior to fitting the curve for dosimetry.
Results: When comparing the mean activity (MBq/mL) in the spherical ROI vs the segmented kidney (n=100, 50 cycles x 2 kidneys, injected activity - median 7718 MBq - range 5180-11877 MBq) the average value was found to be 0.04±0.08 MBq/mL higher for the 4-hour scan with a Pearson correlation of 0.93. Similar trends were observed at 24 hours (0.03±0.08; r=0.92) and 72 hours (0.01±0.05; r=0.89). When comparing the impact of these measurements on the absorbed dose the mean difference was slightly increased for the spherical ROIs with a mean error of 0.08±0.87 Gy and a Pearson correlation (r) of 0.87. When using only the single 72 hour time point for dosimetry following the method of Hanscheid et al the impact of ROI on absorbed dose was 0.17±0.86 Gy with a pearson correlation of 0.89.
Conclusion: Placing spherical ROIs within the kidney provided an efficient means to measure the activity in kidneys in the study when compared to hybrid (AI+manual correction) segmentation. Given the fact that both the activity measurements and resulting absorbed dose were highly correlated (r>0.89 and r=0.87) and had minimal bias (<0.044 MBq per mL and 0.083 Gy), our data suggest that the spherical ROI method may be used for renal dosimetry. (NCT02743741). Abstract 2264 – Table 1

Activity	4 hr	24 hr	72 hr
? Sphere – Whole Kidney (MBq/cc)	0.04±0.08	0.03±0.08	0.01±0.05
Pearson correlation (r)	0.93	0.92	0.89
Absorbed Dose	3 TP Bi-Exp	One TP
? Sphere – Whole Kidney (Gy)	0.08±0.87	0.17±0.86
Pearson correlation (r)	0.87	0.89