Abstract：Objective To explore the role of the discrimination model based on dual energy calcium map/virtual non-calcium image in
differentiating intracranial hemorrhage foci from calcification foci, to clarify the value of differential diagnosis. Methods A total of
189 moderate hyperdense foci in 157 patients with suspected stroke in our hospital (94 in the training set and 95 in the test set) were
collected retrospectively. A classification model based on dual-energy calcium /virtual non-calcium measurements was established
and compared with the diagnoses based on CT values to analyze its benefits in terms of diagnostic efficacy. Meanwhile, subgroups
of different lesion size were also studied to analyze the stability of the diagnostic model in large/small lesions. Results In the test
set, the AUC of the CT-based model was 0.817 (95% CI: 0.724~0.890), while the AUC obtained using calcium/virtual non-calcium
was 0.994 (95% CI: 0.951~1.000). Subjective assessment accuracy using calcium/virtual non-calcium and CT images was 82.1% and
68.4%, respectively. Meanwhile, the detection rate was also consistent in the <5 mm or >5 mm subgroup with no statistical difference
(P>0.05). Conclusion The diagnostic model based on dual energy calcium /virtual non-calcium can accurately identify the different
manifestations of hemorrhagic and calcification foci in patients with suspected stroke and outperforms the diagnostic results of CT
values.
修超a，于洋b，姜倩男b，王贺b，林锐b，张铎a. 双能量钙图/虚拟去钙诊断模型在颅内出血与钙化灶鉴别诊断中的价值[J]. 中国医疗设备, 2021, 36(10): 108-112.
XIU Chaoa，YU Yangb，JIANG Qiannanb，WANG Heb，LIN Ruib，ZHANG Duoa. Value of Discrimination Model Based on Dual Energy Calcium/Virtual Non-Calcium in
Differentiating Intracranial Hemorrhage and Calcified Foci. China Medical Devices, 2021, 36(10): 108-112.
