Spectral computed tomography multi-parameter images for differentiating between tumor tissue and inflammatory tissue...

Pneumonic-type lung carcinoma (PTLC) is a special type of lung cancer with imaging manifestations similar to pneumonia. Due to the intralesional heterogeneity of PTLC, it is difficult to accurately select the optimal target for needle biopsy. Differentiating between tumor tissue and inflammatory tissue within the lesion before puncture could significantly increase the biopsy rate and facilitate early diagnosis. This study aimed to investigate the utility of spectral computed tomography (CT) parameters in differentiating between tumor tissue and inflammatory tissue in PTLC lesions, providing a novel method for target selection before needle biopsy.

A total of 102 patients with PTLC and 96 patients with pneumonia were retrospectively enrolled in the study between September 2020 and April 2024 to evaluate the differentiation of tumor and inflammatory tissue components in PTLC lesions. An additional 30 patients with high suspicion of PTLC were prospectively enrolled to validate the diagnostic accuracy of tumor tissue localization. All patients underwent contrast-enhanced spectral CT examinations, and final diagnoses were confirmed based on pathological findings and follow-up CT evaluations. The ability to visually distinguish tumor tissue from inflammatory tissue in PTLC lesions using conventional CT imaging and spectral CT imaging was evaluated. Conventional CT (CT_conventional) values, iodine density (ID), 40-keV monochromatic energy CT (CT_40keV) values, and Z-effective (Zeff) values were measured in the suspicious tumor area (STA), suspicious inflammatory area (SIA), tumor area (TA), and inflammatory area (IA). The Mann-Whitney U test was used to compare spectral CT parameters among STA, TA, SIA, and IA, and the diagnostic efficiency of each parameter for differentiating tumor from inflammatory tissues was evaluated. The cut-off value of arterial-phase (AP) ID was applied to guide targeted biopsy in 30 prospectively enrolled patients with PTLC, and the diagnostic accuracy was evaluated.

Among the 102 patients with PTLC, the numbers of cases in which tumor tissue and inflammatory tissue could be differentiated on AP/venous phase (VP) images were 34/27 for conventional CT images, 90/81 for ID images, 87/71 on 40-keV monoenergetic images, and 95/84 for Zeff images, respectively. Spectral CT images exhibited significantly higher discriminatory efficiency than conventional CT images (Q =157.48 for AP and Q =138.87 for VP; both P<0.001). In both the AP and VP, the SIAs presented with significantly higher CT_conventional, CT_40keV, ID, and Zeff values than the STAs (all P<0.001). The AP ID value demonstrated good diagnostic performance, with a cut-off value of 1.47 mg/mL [area under the curve (AUC) =0.913, specificity =94.10%]. In the prospective validation cohort of 30 patients, an accuracy of 96.7% was achieved.

Spectral CT parametric imaging can effectively differentiate tumor tissue from inflammatory tissue in PTLC lesions and may facilitate accurate target selection for biopsy.