Semi-automated assessment of tumor burden in neuroendocrine liver metastases using 68Ga-DOTATATE

V. Perim; K. Gultekin; S. Reddy; V. Dudeja; J.R. Howe; J.E. McConathy; J.B. Rose

doi:10.1016/j.hpb.2025.03.124

Background : Despite high recurrence rates, surgery remains the best approach for select patients with neuroendocrine tumor liver metastases (NETLM), improving survival and symptom control. Traditional tumor burden estimation relies on the surgeon's judgment using perioperative imaging. Given the routine use 68Ga-DOTATATE PET/CT for assessing somatostatin receptor expression and staging in NET patients, it could also serve to objectively calculate tumor burden. This pilot study seeks to develop a semi-automated method for calculating tumor volume using 68Ga-DOTATATE PET/CT imaging in NETLM patients before and after cytoreductive surgery. Methods : We retrospectively analyzed data from seven patients with NETLM who underwent 68Ga-DOTATATE PET/CT scans before and after cytoreductive surgery at a single institution over the last ten years. Liver contours were generated automatically using a validated deep learning algorithm, and tumor volumes were delineated using MIM Encore 7.3 software. The absolute SUV cut-off of 7 was chosen as threshold after experimentation with different fixed values. After excluding physiologic uptake areas and refining tumor delineation, pre- and post-operative total tumor volumes (TTV), as well as TTV change post resection ( Δ TTV) were calculated. Results : The cohort included three women and four men, aged 45-78 years. The primary tumor sites included the pancreas (2), rectum (1), terminal ileum (1), and unknown primary locations (3). Preoperative imaging was done within a median of 69 days (30-190 days) prior to surgery and showed a median TTV of 712.7 mL (98.6–1572.7 mL). Postoperative imaging was done within a median of 98 days (60-1036 days) after the procedure and showed a median TTV of 33.7 mL (2.6–517.9 mL). The median Δ TTV was 73.0% (21.7%–99.8%), indicating a significant reduction in tumor burden in most cases after surgery. According to the our analysis, only one of the estimates was considered to be innacurate, mainly due to influence of background activity on imaging. Conclusion : This proof-of-concept study shows that a semi-automated method using 68Ga-DOTATATE PET/CT can objectively assess tumor burden in NETLM patients pre- and post-surgery. Future studies should include larger cohorts, standardized imaging timeframes, and varied thresholds for improved accuracy.

Semi-automated assessment of tumor burden in neuroendocrine liver metastases using 68Ga-DOTATATE

View Online

Abstract

Details

Metrics