Virtual monoenergetic imaging predicting Ki-67 expression in lung cancer

Abstract

Abstract This study aimed to optimize slope and energy levels for evaluating Ki-67 expression in lung cancer using virtual monoenergetic imaging and compare the predictive efficiency of different energy spectrum slopes (λHU) for Ki-67. Forty-three patients with primary lung cancer confirmed via pathological examination were enrolled in this study. They underwent baseline arterial-phase (AP) and venous-phase (VP) energy spectrum computed tomography (CT) scanning before surgery. The CT values were 40–190 keV, with 40–140 keV indicating pulmonary lesions at AP and VP, and P < 0.05 indicating a statistically significant difference. An immunohistochemical examination was conducted, and receiver operating characteristic curves were used to analyze the prediction performance of λHU for Ki-67 expression. SPSS Statistics 22.0 (IBM Corp., NY, USA) was used for statistical analysis, and χ 2, t, and Mann–Whitney U tests were used for quantitative and qualitative analyses of data. Significant differences were observed at the corresponding CT values of 40 keV (as 40-keV is considered the best for single-energy image for evaluating Ki-67 expression) and 50 keV in AP and at 40, 60, and 70 keV in VP between high- and low-Ki-67 expression groups (P < 0.05). In addition, the λHU values of three-segment energy spectrum curve in both AP and VP were quite different between two groups (P < 0.05). However, the VP data had greater predictive values for Ki-67. The areas under the curve were 0.859, 0.856, and 0.859, respectively. The 40-keV single-energy sequence was the best single-energy sequence to evaluate the expression of Ki-67 in lung cancer and to obtain λHU values using the energy spectrum curve in the VP. The CT values had better diagnostic efficiency.