Spectral computed tomography in the assessment of metastatic lymph nodes in cervical cancer patients treated with definitive radiotherapy: a single-center, prospective study

Identifying metastatic lymph nodes (LNs) in patients with cervical cancer treated with definitive radiotherapy may inform treatment strategy and determine prognosis, but available methods have limitations, especially in developing regions. Herein, we aimed to evaluate the performance of quantitative parameters in spectral computed tomography (CT) scanning in this context, focusing on its complementary role alongside conventional diagnostic approaches like 18-fluorine-fuorodeoxyglucose positron emission tomography computed tomography (18 F FDG-PET/CT). Patients with cervical cancer, who underwent pretreatment spectral CT simulation scanning and planned radiotherapy, were enrolled in this prospective study. The LNs were categorized as "metastatic" and "non-metastatic", based on a procedure that included 18 F FDG-PET/CT as well as CT, magnetic resonance imaging, Node Reporting and Data System and follow-up results. Iodine concentrations (IC), normalized IC (NIC), effective atom number (effZ), and spectral curve slope (lambda HU) in the arterial (AP) and venous (VP) phases, were compared between metastatic and non-metastatic LNs. IC were derived from iodine-based material decomposition through manual delineation and normalized to the iodine concentration in the adjacent artery (NIC). effZ and lambda HU were calculated based on the effective atom number image and virtual monochromatic images. Univariate and multivariate logistic regression analyses were used to determine spectral CT factors independently associated with LNs metastasis, and their diagnostic efficacies were assessed using the area under the curve (AUC) analysis. The diagnostic efficiency of 18 F FDG-PET/CT and spectral CT was compared. A total of 115 metastatic and 97 non-metastatic LNs were detected, and spectral CT parameters (IC, NIC, effZ, lambda HU) significantly differed between the two groups. In univariate and multivariable logistic regression analysis, lambda HU in the AP and NIC in the VP were independent predictors for metastatic LNs and their combination improved AUC to 0.923, with a sensitivity of 84.4%, and a specificity of 85.6%. Spectral CT could achieve similar sensitivity as 18 FFDG-PET/CT in total LNs, and, more importantly, a higher sensitivity (95.5% vs. 59.1%) and diagnostic accuracy (92.9% vs. 67.9%) for para-aortic LNs. Quantitative spectral CT parameters can help distinguish metastatic from non-metastatic LNs in patients with cervical cancer treated with definitive radiotherapy. Combination of lambda HU in AP and NIC in VP further improves diagnostic performance. Spectral CT, while promising, complements rather than replaces PET/CT, especially for diagnosing para-aortic LNs, where PET/CT may have limitations. It could be a valuable adjunct to conventional imaging, particularly in settings with limited access to advanced tools.