2004•Zhonghua heyixue yu fenzi yingxiang zazhiRequires access

Accuracy of image coregistration of pulmonary carcinoma using integrated PET/CT system

Jinming Yu

Open publisher page 0 citations

Abstract

Objective To evaluate the accuracy of coregistration of PET/CT images in patients with pulmonary carcinoma and the affecting factors, and to investigate the difference between two methods for measuring coregistration accuracy. Methods Twenty-four patients with well-circumscribed pulmonary carcinoma were examined by integrated PET/CT. Thirty-one lesions with the diameter of 1.0~5.0 cm were analyzed. The registration accuracy of PET/CT was measured with two methods, volume measure and center of geometry (COG) measure. The volume of interest of each lesion was defined separately on PET, CT and the images were fused (V_~PET , V_~CT , V_~PET+CT ). The overlap volume of PET and CT images was calculated with the formula: V_~PET/CT =V_~PET +V_~CT -V_~PET+CT . The percentage of V_~PET/CT on V_~PET+CT , P_~PET/CT , was used to assess the fusion accuracy of PET and CT images. The COG was assessed. The distance of COGs between the PET image and the CT image, D_~COG , was measured. The coregistration accuracy was compared between lesions classified according to the lesion location and volume. Results The mean D_~COG was (4.39±2.33) mm (range, 1.00~8.68 mm). P_~PET/CT was (36.18±22.66)% (range, 0%~80.53%). Misregistration tended to be profounder in the lower lungs [D_~COG =(4.63±2.37)mm, P_~PET/CT =(32.25±18.40)%] than in the upper lungs [D_~COG =(4.11±2.36) mm, P_~PET/CT =(40.95±26.91)%, P=0.30]. Misregistration also tended to be slightly profounder in the left lung [D_~COG =(5.18±3.08) mm, P_~PET/CT =(33.01±20.21)%] than in the right lung [D_~COG =(4.02±1.87) mm, P_~PET/CT =(37.68±24.07)%, P=0.78]. There was no statistical difference among lesions located at different lung regions (the apex, peripheral, central, and base of lung) while using two indexes. As for the lesions with different volumes as ≤2 cm3, 2.1~9.9 cm3 or ≥10 cm3, the P_~PET/CT were (14.11±10.64)%, (41.64±16.22)% and (53.77±17.71)%, respectively (P0.01). Conclusions The registration of the pulmonary lesion in PET/CT images is usually accurate, but spatial misregistration does occur, especially for the smaller lesions. The combination of different methods for measuring coregistration accuracy may be more valuable in defining this kind of misregistration.

About this research paper

What this paper is about

Objective To evaluate the accuracy of coregistration of PET/CT images in patients with pulmonary carcinoma and the affecting factors, and to investigate the difference between two methods for measuring coregistration accuracy. Methods Twenty-four patients with well-circumscribed pulmonary carcinoma were examined by integrated PET/CT. Thirty-one lesions with the diameter of 1.0~5.0 cm were analyzed. The registration accuracy of PET/CT was measured with two methods, volume measure and center of geometry (COG) measure. The volume of interest of each lesion was defined separately on PET, CT and the images were fused (V_~PET , V_~CT , V_~PET+CT ). The overlap volume of PET and CT images was calculated with the formula: V_~PET/CT =V_~PET +V_~CT -V_~PET+CT . The percentage of V_~PET/CT on V_~PET+CT , P_~PET/CT , was used to assess the fusion accuracy of PET and CT images. The COG was assessed. The distance of COGs between the PET image and the CT image, D_~COG , was measured. The coregistration accuracy was compared between lesions classified according to the lesion location and volume. Results The mean D_~COG was (4.39±2.33) mm (range, 1.00~8.68 mm). P_~PET/CT was (36.18±22.66)% (range, 0%~80.53%). Misregistration tended to be profounder in the lower lungs [D_~COG =(4.63±2.37)mm, P_~PET/CT =(32.25±18.40)%] than in the upper lungs [D_~COG =(4.11±2.36) mm, P_~PET/CT =(40.95±26.91)%, P=0.30]. Misregistration also tended to be slightly profounder in the left lung [D_~COG =(5.18±3.08) mm, P_~PET/CT =(33.01±20.21)%] than in the right lung [D_~COG =(4.02±1.87) mm, P_~PET/CT =(37.68±24.07)%, P=0.78]. There was no statistical difference among lesions located at different lung regions (the apex, peripheral, central, and base of lung) while using two indexes. As for the lesions with different volumes as ≤2 cm3, 2.1~9.9 cm3 or ≥10 cm3, the P_~PET/CT were (14.11±10.64)%, (41.64±16.22)% and (53.77±17.71)%, respectively (P0.01). Conclusions The registration of the pulmonary lesion in PET/CT images is usually accurate, but spatial misregistration does occur, especially for the smaller lesions. The combination of different methods for measuring coregistration accuracy may be more valuable in defining this kind of misregistration.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Objective To evaluate the accuracy of coregistration of PET/CT images in patients with pulmonary carcinoma and the affecting factors, and to investigate the difference between two methods for measuring coregistration accuracy. Methods Twenty-four patients with well-circumscribed pulmonary carcinoma were examined by integrated PET/CT. Thirty-one lesions with the diameter of 1.0~5.0 cm were analyzed. The registration accuracy of PET/CT was measured with two methods, volume measure and center of geometry (COG) measure. The volume of interest of each lesion was defined separately on PET, CT and the images were fused (V_~PET , V_~CT , V_~PET+CT ). The overlap volume of PET and CT images was calculated with the formula: V_~PET/CT =V_~PET +V_~CT -V_~PET+CT . The percentage of V_~PET/CT on V_~PET+CT , P_~PET/CT , was used to assess the fusion accuracy of PET and CT images. The COG was assessed. The distance of COGs between the PET image and the CT image, D_~COG , was measured. The coregistration accuracy was compared between lesions classified according to the lesion location and volume. Results The mean D_~COG was (4.39±2.33) mm (range, 1.00~8.68 mm). P_~PET/CT was (36.18±22.66)% (range, 0%~80.53%). Misregistration tended to be profounder in the lower lungs [D_~COG =(4.63±2.37)mm, P_~PET/CT =(32.25±18.40)%] than in the upper lungs [D_~COG =(4.11±2.36) mm, P_~PET/CT =(40.95±26.91)%, P=0.30]. Misregistration also tended to be slightly profounder in the left lung [D_~COG =(5.18±3.08) mm, P_~PET/CT =(33.01±20.21)%] than in the right lung [D_~COG =(4.02±1.87) mm, P_~PET/CT =(37.68±24.07)%, P=0.78]. There was no statistical difference among lesions located at different lung regions (the apex, peripheral, central, and base of lung) while using two indexes. As for the lesions with different volumes as ≤2 cm3, 2.1~9.9 cm3 or ≥10 cm3, the P_~PET/CT were (14.11±10.64)%, (41.64±16.22)% and (53.77±17.71)%, respectively (P0.01). Conclusions The registration of the pulmonary lesion in PET/CT images is usually accurate, but spatial misregistration does occur, especially for the smaller lesions. The combination of different methods for measuring coregistration accuracy may be more valuable in defining this kind of misregistration.

Key concepts: Nuclear medicine, Medicine, PET-CT, Positron emission tomography, Cog, Image registration, Computed tomography, Radiology

Related papers

Back to paper searchBrowse research topicsOriginal source
Accuracy of image coregistration of pulmonary carcinoma using integrated PET/CT system — Research Paper | ScholarLens