Fully Automatic Volume Measurement of the Spleen at CT Using Deep Learning

Abstract

To develop a fully automated algorithm for spleen segmentation and to assess the performance of this algorithm in a large dataset. Materials and Methods:In this retrospective study, a three-dimensional deep learning network was developed to segment the spleen on thorax-abdomen CT scans. Scans were extracted from patients undergoing oncologic treatment from 2014 to 2017. A total of 1100 scans from 1100 patients were used in this study, and 400 were selected for development of the algorithm. For testing, a dataset of 50 scans was annotated to assess the segmentation accuracy and was compared against the splenic index equation. In a qualitative observer experiment, an enriched set of 100 scan-pairs was used to evaluate whether the algorithm could aid a radiologist in assessing splenic volume change. The reference standard was set by the consensus of two other independent radiologists. A Mann-Whitney U test was conducted to test whether there was a performance difference between the algorithm and the independent observer. Results:The algorithm and the independent observer obtained comparable Dice scores (P = .834) on the test set of 50 scans of 0.962 and 0.964, respectively. The radiologist had an agreement with the reference standard in 81% (81 of 100) of the cases after a visual classification of volume change, which increased to 92% (92 of 100) when aided by the algorithm. Conclusion:A segmentation method based on deep learning can accurately segment the spleen on CT scans and may help radiologists to detect abnormal splenic volumes and splenic volume changes.

Prevalence and clinical significance of clinically evident portal hypertension in patients with hepatocellular carcinoma undergoing transarterial chemoembolization

“…In comparison, measuring the axial plane diameter is easy to implement in clinical routines. Nevertheless, the first results from a fully automated spleen segmentation with the use of artificial intelligence methods have shown promising results and it is likely that those tools will become increasingly available in the daily clinical routine 32,33 . These novel methods offer an automated report of the splenic volume, ad hoc after imaging, which will tremendously reduce the time investment.…”

Radiomics Features of the Spleen as Surrogates for CT-Based Lymphoma Diagnosis and Subtype Differentiation

“…Secondly, the spleen is a parenchymatous organ surrounded by peri-splenic fat and therefore easy to automatically segment. Humpire-Mamani et al have shown that a neural network can be trained to segment the spleen in CT scans with an accuracy that is comparable to an experienced radiologist [34,35]. Such a performance would not be expected for involved lymph node sites, which can occur on any site within the scan volume.…”

“…In this study, a U-Net for automatic spleen segmentation was successfully trained with high accuracy. The U-Net is currently the state of the art for automatic image segmentation and was also used by Humpire-Mamani et al for spleen segmentation [34]. When a U-Net is trained on scans from different scanners, with different slice thicknesses and so on, it learns a general concept of the appearance of a spleen that is independent of the actual imaging settings.…”

AbdomenCT-1K: Is Abdominal Organ Segmentation a Solved Problem?

“…For instance, from a recently presented review work, liver segmentation can reach accuracy of 95% in terms of Dice coefficients [36]. In a recent work for spleen segmentation [65], 96.2% Dice score is reported. However, it is worth re-thinking that is abdominal organ segmentation a solved problem?…”