Automated Deep Learning Artificial Intelligence Tool for Spleen Segmentation on CT: Defining Volume-Based Thresholds for Splenomegaly.

Abstract

BACKGROUND: Splenomegaly has historically been assessed on imaging using potentially inaccurate linear measurements. Prior work tested a deep-learning artificial intelligence (AI) tool that automatically segments the spleen to determine splenic volume. OBJECTIVE: To apply the deep-learning AI tool in a large screening population to establish volume-based splenomegaly thresholds. METHODS: This retrospective study included a primary (screening) sample of 8901 patients (mean age, 56±10 years; 4235 men, 4666 women) who underwent CT colonoscopy (n=7736) or renal-donor CT (n=1165) from April 2004 to January 2017, and a secondary sample of 104 patients (mean age, 56±8 years; 62 men, 42 women) with end-stage liver disease (ESLD) who underwent pre-liver transplant CT from January 2011 to May 2013. The automated deep-learning AI tool was used for spleen segmentation, to determine splenic volumes. Two radiologists independently reviewed a subset of segmentations. Weight-based volume thresholds for splenomegaly were derived using regression analysis. Performance of linear measurements was assessed. Frequency of splenomegaly using weight-based volumetric thresholds in the secondary sample was determined. RESULTS: In the primary sample, both observers confirmed splenectomy in 20 patients with automated splenic volume of zero; confirmed incomplete splenic coverage in 28 patients with a tool output error; and confirmed adequate segmentation in 21 patients with low (<50 cc) volume, 49 patients with high (>600 cc) volume, and 200 additional randomly selected patients. In 8853 patients included in analysis of splenic volumes (i.e., excluding zero or error values), the mean automated splenic volume was 216±100 ml. Weight-based volumetric threshold (in milliliters) for splenomegaly was defined as 3.0*(weight[kg])+127; for weight >125 kg, splenomegaly threshold was constant (503 ml). Sensitivity and specificity for volume-defined splenomegaly were 13% and 100% at true-craniocaudal length of 13 cm, and 78% and 88% for maximum-3D length of 13 cm. In the secondary sample, both observers identified segmentation failure in one patient. Mean automated splenic volume in the 103 remaining patients was 796±457 ml; 84% (87/103) of patients met the weight-based volume-defined splenomegaly threshold. CONCLUSION: We derived a weight-based volumetric threshold for splenomegaly using an automated AI-based tool. CLINICAL IMPACT: The AI tool could facilitate large-scale opportunistic screening for splenomegaly.