TY - JOUR
T1 - Detection of Low Blood Hemoglobin Levels on Pulmonary CT Angiography
T2 - A Feasibility Study Combining Dual-Energy CT and Machine Learning
AU - Kay, Fernando U.
AU - Lumby, Cynthia
AU - Tanabe, Yuki
AU - Abbara, Suhny
AU - Rajiah, Prabhakar
N1 - Publisher Copyright:
© 2023 by the authors.
PY - 2023/8
Y1 - 2023/8
N2 - Objectives: To evaluate if dual-energy CT (DECT) pulmonary angiography (CTPA) can detect anemia with the aid of machine learning. Methods: Inclusion of 100 patients (mean age ± SD, 51.3 ± 14.8 years; male-to-female ratio, 42/58) who underwent DECT CTPA and hemoglobin (Hb) analysis within 24 h, including 50 cases with Hb below and 50 controls with Hb ≥ 12 g/dL. Blood pool attenuation was assessed on virtual noncontrast (VNC) images at eight locations. A classification model using extreme gradient-boosted trees was developed on a training set (n = 76) for differentiating cases from controls. The best model was evaluated in a separate test set (n = 24). Results: Blood pool attenuation was significantly lower in cases than controls (p-values < 0.01), except in the right atrium (p = 0.06). The machine learning model had sensitivity, specificity, and accuracy of 83%, 92%, and 88%, respectively. Measurements at the descending aorta had the highest relative importance among all features; a threshold of 43 HU yielded sensitivity, specificity, and accuracy of 68%, 76%, and 72%, respectively. Conclusion: VNC imaging and machine learning shows good diagnostic performance for detecting anemia on DECT CTPA.
AB - Objectives: To evaluate if dual-energy CT (DECT) pulmonary angiography (CTPA) can detect anemia with the aid of machine learning. Methods: Inclusion of 100 patients (mean age ± SD, 51.3 ± 14.8 years; male-to-female ratio, 42/58) who underwent DECT CTPA and hemoglobin (Hb) analysis within 24 h, including 50 cases with Hb below and 50 controls with Hb ≥ 12 g/dL. Blood pool attenuation was assessed on virtual noncontrast (VNC) images at eight locations. A classification model using extreme gradient-boosted trees was developed on a training set (n = 76) for differentiating cases from controls. The best model was evaluated in a separate test set (n = 24). Results: Blood pool attenuation was significantly lower in cases than controls (p-values < 0.01), except in the right atrium (p = 0.06). The machine learning model had sensitivity, specificity, and accuracy of 83%, 92%, and 88%, respectively. Measurements at the descending aorta had the highest relative importance among all features; a threshold of 43 HU yielded sensitivity, specificity, and accuracy of 68%, 76%, and 72%, respectively. Conclusion: VNC imaging and machine learning shows good diagnostic performance for detecting anemia on DECT CTPA.
KW - anemia
KW - computed tomography angiography
KW - machine learning
KW - pulmonary embolism
UR - http://www.scopus.com/inward/record.url?scp=85168751707&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85168751707&partnerID=8YFLogxK
U2 - 10.3390/tomography9040123
DO - 10.3390/tomography9040123
M3 - Article
C2 - 37624116
AN - SCOPUS:85168751707
SN - 2379-1381
VL - 9
SP - 1538
EP - 1550
JO - Tomography
JF - Tomography
IS - 4
ER -