Differential Diagnostics of Soft Tissue Sarcoma Recurrence and Postoperative Changes: A Comparative Efficacy of Contrast-Enhanced Ultrasound versus Magnetic Resonance Imaging

Abstract

Introduction. Soft tissue sarcomas (STS) present a considerable clinical challenge due to their high risk of local recurrence (LR) and the absence of standardized surveillance protocols. Although MRI remains the gold standard for diagnosis, its utility is limited by high costs and reduced diagnostic accuracy in patients following radiation therapy. Contrast-enhanced ultrasound (CEUS), which demonstrates greater accuracy than conventional ultrasound, represents a promising alternative.

Aim. To evaluate the diagnostic performance of CEUS compared to MRI in differentiating postoperative changes from LR diagnostics.

Materials and Methods. This study included 43 patients with suspected LR who underwent both CEUS and MRI. Histopathological analysis served as the reference standard. Statistical analyses were conducted using specialized software packages, with a significance threshold set at p < 0.05. Diagnostic performance metrics were calculated, and receiver operating characteristic (ROC) curves were generated.

Results. We identified statistically significant (p < 0.05) ultrasonographic and MRI criteria for distinguishing LR from postoperative changes (PO). LR was characterized by structural inclusions with hyperechoic halos, hypervascular or mixed flow patterns, and spiral-shaped contrast enhancement (CE) with type 3 kinetic curves. Conversely, PO typically demonstrated homogeneous hyperechoic structure, cystic components, avascularity, and absence of contrast enhancement without kinetic curves. Notably, lesion location, shape, margins, and stiffness showed no diagnostic value for differentiating LR from PO (p > 0.05). MRI findings further revealed that LR predominantly presented as solid or cystic-solid structures with necrotic areas and exhibited pronounced, heterogeneous CE. PO, however, was characterized by fluid-filled structures, fibrosis, anatomical connections to vascular or bony structures, and weak linear or homogeneous CE (p < 0.05). Diagnostic performance metrics demonstrated high accuracy for both modalities: CEUS achieved 90.0 % sensitivity, 92.3 % specificity, and an AUC of 0.911, while MRI showed 93.3 % sensitivity, 92.3 % specificity, and an AUC of 0.928.

Conclusion. CEUS and MRI should be regarded as complementary rather than competing modalities. Their strategic integration, guided by clinical context and resource availability, facilitates optimal LR detection. Continued technical refinement of ultrasound methodologies and accumulation of clinical experience will further optimize surveillance protocols, enabling earlier LR detection and improving patient prognosis.