Application of the advanced lung cancer inflammation index for patients with coronavirus disease 2019 pneumonia: Combined risk prediction model with lung cancer inflammation index, computed tomography and chest radiograph

Autor: Akitoshi Inoue, Hiroaki Takahashi, Tatsuya Ibe, Hisashi Ishii, Yuhei Kurata, Yoshikazu Ishizuka, Bolorkhand Batsaikhan, Yoichiro Hamamoto
Rok vydání: 2022
Předmět:
Zdroj: Experimental and Therapeutic Medicine. 23
ISSN: 1792-1015
1792-0981
DOI: 10.3892/etm.2022.11315
Popis: The purpose of the present study was to evaluate the feasibility of applying the advanced lung cancer inflammation index (ALI) in patients with coronavirus disease 2019 (COVID-19) and to establish a combined ALI and radiologic risk prediction model for disease exacerbation. The present study included patients diagnosed with COVID-19 infection in our single institution from March to October 2020. Patients without clinical information and/or chest computed tomography (CT) upon admission were excluded. A radiologist assessed the CT severity score and abnormality on chest radiograph. The combined ALI and radiologic risk prediction model was developed via random forest classification. Among 79 patients (age, 43±19 years; male/female, 45:34), 72 experienced improvement and seven patients experienced exacerbation after admission. Significant differences were observed between the improved and exacerbated groups in the ALI (median, 47.6 vs. 13.2; P=0.011), frequency of chest radiograph abnormality (24.7 vs. 83.3%; P0.001), and chest CT score (CCTS; median, 1 vs. 9; P0.001). For the accuracy of predicting exacerbation, the receiver-operating characteristic curve analysis demonstrated an area under the curve of 0.79 and 0.92 for the ALI and CCTS, respectively. The combined ALI and radiologic risk prediction model had a sensitivity of 1.00 and a specificity of 0.81. Overall, ALI alone and CCTS alone modestly predicted the exacerbation of COVID-19, and the combined ALI and radiologic risk prediction model exhibited decent sensitivity and specificity.
Databáze: OpenAIRE
Externí odkaz: