Nomograms based on clinical characteristics and baseline dual-layer spectral detector computed tomography imaging parameters: potential clinical utility in predicting the efficacy of first-line lymphoma treatment

Abstract Background Spectral computed tomography (CT) is considered a useful tool for predicting tumor treatment response and survival. However, its role in lymphoma treatment remains unclear. The purpose of this study was to develop spectral CT-based nomogram for predicting response and prognosis in lymphoma patients after first-line therapy. Methods The baseline clinical and spectral CT data from 91 patients with lymphoma were retrospectively analyzed. Treatment response was assessed according to the Lugano criteria, with patients classified into complete response (CR) and non-CR groups. Spectral CT parameters were measured from the largest lesion in the arterial and venous phases. Univariable and multivariable logistic regression analyses were performed to identify predictors of non-CR and to construct clinical, spectral CT, and combined models. The final combined model was internally validated using 1000 bootstrap resamples without repeating variable selection. Model performance was assessed using ROC analysis, calibration analysis, and decision curve analysis. Median follow up was estimated using the reverse Kaplan Meier method, while Kaplan Meier analysis and Cox regression were used to evaluate progression-free survival (PFS) and overall survival (OS). Results In the multivariable analysis, B symptoms, Ki67, and arterial enhancement fraction based on IC (AEF IC) were identified as predictors of non-CR. The combined nomogram showed an apparent AUC of 0.866 (95% CI: 0.791, 0.938), which was higher than that of the clinical model but not significantly different from that of the spectral CT model. Risk stratification based on the clinical, spectral CT, and combined models was associated with PFS, with the combined model showing the strongest association with shorter PFS (HR: 4.05; 95% CI: 1.85, 8.88). Conclusions Spectral CT and clinical models showed potential for predicting treatment response and PFS in patients with lymphoma. Integrating spectral CT parameters with clinical characteristics in a nomogram may improve response prediction and support exploratory risk stratification. Further external validation is required.