MRI and ADC histogram features as predictors of distant metastasis and prognosis in alveolar soft tissue sarcomas

Abstract

Objective

To evaluate the value of pre-treatment MRI features and whole-volume apparent diffusion coefficient (ADC) histogram parameters in predicting distant metastasis and prognosis in alveolar soft part sarcoma (ASPS).

Methods

A retrospective analysis was performed on 43 patients with histologically confirmed ASPS, including 25 with metastasis and 18 without at the time of diagnosis. All patients underwent routine MRI, including diffusion-weighted imaging (DWI) and enhanced T1-weighted imaging. Clinical data and MRI features of the primary tumor were recorded. MATLAB software was used to analyze ADC histogram parameters. Differences in clinical data and MRI features between the metastasis and non-metastasis groups were assessed using t-tests or chi-square tests. Histogram parameters were compared using independent t-tests or Mann-Whitney U tests. Multivariate binary logistic regression analysis was used to identify independent factors associated with metastasis and to establish a combined model. Receiver operating characteristic (ROC) analysis evaluated the diagnostic performance of independent factors and the combined model. Kaplan-Meier analysis and Cox proportional hazards models assessed the risk of distant metastasis and overall survival (OS).

Results

Significant differences were found between the metastasis and non-metastasis groups in tumor size, heterogeneous signal intensity on T2WI, peritumoral edema, and peritumoral enhancement (P = 0.014, 0.001, 0.019, and < 0.001, respectively). The metastasis group had lower ADCmean and ADC50th values (P = 0.001, 0.046, respectively). Logistic regression identified tumor size, peritumoral enhancement, and ADCmean as independent predictors of metastasis. The combined model showed the highest AUC (0.872, 95 % CI: 0.735–0.954). Size and ADCmean were independent factors for OS.

Conclusion

MRI features and ADC histogram parameters can effectively predict distant metastasis and prognosis in ASPS.