Diagnosis and differentiation of Kaposiform hemangioendothelioma based on ultrasound radiomics.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
Kaposiform hemangioendothelioma (KHE) is a rare neoplasm of the newborn, but has a very high mortality rate. In this study, we explore the application value of ultrasound radiomics in the differential diagnosis of KHE so as to provide reference for early diagnosis of KHE. We selected 194 cases of children with suspected KHE admitted to Henan Provincial People's Hospital from March 2016 to April 2024 for this retrospective analysis. All children completed ultrasound examinations in our hospital. After pathological biopsy, 132 cases were diagnosed with KHE. Taking pathological biopsy as the gold standard, the diagnostic rate of ultrasound examination was determined. Our results showed that ultrasound examination diagnosed 124 cases with KHE. Compared with pathological biopsy, the diagnostic sensitivity, specificity, and accuracy of ultrasound were 81.82, 77.42, and 80.41%, respectively (Kappa = 0.725). Most of the children had a single lesion, often involving the skin. The tumor was hard in texture and red or purple in color and did not fade when pressed. The two-dimensional sonogram showed a solid heterogeneous echo mass in the soft tissue. The lesion was generally large, irregular in shape, and unclear in boundaries, extending to the adipose layer and forming a "tree root-like" change. Microscopically, crossed spindle-shaped cell bundles could be seen, showing diffuse multinodular infiltrative growth. Children with the Kasabach-Merritt phenomenon (KMP) generally had larger lesions. Subsequently, the children with KHE were assigned to a training set and a validation set in a ratio of 7:3. High-throughput data acquisition of the regions of interest (ROIs) on the ultrasound images was carried out to construct a KEH prediction model based on ultrasound radiomics, and validation analysis was conducted. We found the diagnostic accuracy rate of ultrasound radiomics was 91.41%, the sensitivity was 77.44%, and the specificity was 97.21%, which were better than those of the conventional ultrasound. In conclusion, ultrasound radiomics analysis is highly effective in the diagnosis of KHE, which can contribute to the early diagnosis rate of KHE.