Virtual monoenergetic imaging predicting Ki-67 expression in lung cancer

Area of Science:

• Radiology and Imaging
• Oncology
• Pathology

Background:

• Ki-67 expression is a crucial biomarker for assessing proliferation in lung cancer.
• Virtual monoenergetic imaging (VMI) offers potential for optimizing computed tomography (CT) spectral analysis.
• Accurate assessment of Ki-67 expression is vital for treatment planning and prognosis in lung cancer.

Purpose of the Study:

• To optimize energy levels and spectral slopes (λHU) for evaluating Ki-67 expression in lung cancer using VMI.
• To compare the predictive efficiency of different VMI parameters for Ki-67 expression.
• To identify the optimal CT energy levels for assessing Ki-67 in lung cancer.

Main Methods:

• Forty-three lung cancer patients underwent arterial-phase (AP) and venous-phase (VP) dual-energy CT scanning.
• VMI was performed across a range of 40-190 keV.
• Immunohistochemistry was used to determine Ki-67 expression; receiver operating characteristic (ROC) curves analyzed predictive performance of λHU values.

Main Results:

• Significant differences in CT values between high- and low-Ki-67 expression groups were observed at specific energy levels (e.g., 40 keV in AP, 40, 60, 70 keV in VP).
• λHU values from the energy spectrum curve differed significantly between expression groups in both AP and VP.
• Venous-phase (VP) data demonstrated superior predictive value for Ki-67 expression, with areas under the ROC curve reaching 0.859.

Conclusions:

• The 40-keV single-energy VMI sequence is optimal for evaluating Ki-67 expression in lung cancer.
• Venous-phase spectral data, particularly at 40 keV, provides the best diagnostic efficiency for Ki-67 prediction.
• Optimized VMI parameters enhance the diagnostic accuracy of CT for assessing Ki-67 expression in lung cancer.