Novel Biomarkers Based on Dual-Energy Computed Tomography for Risk Stratification of Very Early Distant Metastasis in Colorectal Cancer after Surgery.
Wenjing Peng,Lijuan Wan,Rui Zhao,Shuang Chen,Shushan Dong,Lin Li,Hongmei Zhang
DOI: https://doi.org/10.21037/qims-23-861
2024-01-01
Quantitative Imaging in Medicine and Surgery
Abstract:Background: Very early distant metastasis (VEDM) for patients with colorectal cancer (CRC) following surgery suggests failure of local treatment strategy and few biomarkers are available for its effective risk stratification. This study aimed to explore the potential of quantitative dual-energy computed tomography (DECT) spectral parameters and build models to predict VEDM. Methods: Consecutive patients suspected of having CRC and with a clinical indication for enhanced CT from April 2021 to July 2022 at a single institution were prospectively enrolled to undertake spectral CT scanning. The spectral features were extracted by two reviewers and intraclass correlation coefficient (ICC) was used for interobserver agreement evaluation. A total of 16 spectral parameters, including unenhanced effective atomic number, triphasic iodine concentrations (ICs)/normalized ICs (NICs)-(A/V/E)/1/NIC-(A/V/E)/spectral curve slopes (lambda-(A/V/E)), two arterial enhancement fractions (AEFs), and venous enhancement fraction (VEF), were determined for analysis. Patients with and without VEDM after surgery were matched using propensity score matching (PSM). The diagnostic performance was assessed using the area under the curve (AUC). Models of multiple modalities were generated. Results: In total, 222 patients were included (141 males, age range, 32-83 years) and 13 patients developed VEDM. Interobserver agreement ranged from good to excellent (ICC, 0.773-0.964). A total of three spectral parameters (VEF, lambda-(V), and 1/NIC-(V)) exhibited significant discriminatory ability (P<0.05) in predicting VEDM, with AUCs of 0.822 [95% confidence interval (CI): 0.667-0.926], 0.738 (95% CI: 0.573-0.866), and 0.713 (95% CI: 0.546-0.846) and optimal cutoff points of 67.16%, 2.46, and 2.44, respectively. The performance of these spectral parameters was validated in the entire cohort; the combined spectral model showed comparable efficiency to the combined clinical model [AUC, 0.771 (95% CI: 0.622-0.919) vs. 0.779 (95% CI: 0.663-0.894), P>0.05]; the clinical-spectral model achieved further improved AUC of 0.887 (95% CI: 0.812-0.962), which was significantly higher than the combined clinical model (P=0.015), yet not superior to the combined spectral model (P=0.078). Conclusions: Novel spectral parameters showed potential in predicting VEDM in CRC following surgery in this preliminary study, which were closely related with spectral perfusion in the venous phase. However, further studies with larger samples are warranted.