Feasibility study of discriminating different solutions using spectral CT imaging / 实用放射学杂志

ABSTRACT

Objective To investigate the feasibility of discriminating different solutions and solutions with different concentrations using gemstone spectral imaging (GSI).Methods Glucose solutions (GSs)with different concentrations (5%,10%,15%,20% g/mL)and 0.9% normal saline(NS)were scanned at a CT scanner with GSI model.Spectral analysis software was used to generate spectral curves of different solutions,representing as CT values on monochromatic images(40~140keV).The same procedure was repeated 3 days later.Reliability analysis with intra-class correlation coefficient (ICC)was used to evaluate the agreement between the twice CT scans.Nonlinear regression analysis was used to generate the regression equation of spectral curves of each solution.One-way ANO-VA was performed to compare the slopes,asymptotes and intercepts of each solution.Results Reliability analysis showed excellent agreements between the twice CT scans for each solution (ICC>0.9).The regression equation was expressed asf(keV-40)=α-(α-β)×e [-(keV-40)×ρ],where f(keV-40)represented CT values on each monochromatic images,αrepresented asymptotes,βrepre-sented intercepts,ρrepresented slopes of each solution.Statistical differences were found among the slopes of NS and those of GSs with different concentrations (P 0.05). The result was highly suggestive that different ingredient solutions demonstrated different shapes of spectral curves.The asymptotes and intercepts of GSs with different concentrations were statistically different (P <0.05).The heights of the spectral curves of GSs elevated with the increase of the concentration.Conclusion Spectral curve with GSI model can be used to discriminate between NS and GSs,or even GSs with different concentrations.It provided a promising potential for in vivo body fluid analyses.