Venous wash-out at contrast-enhanced ultrasound for the identification of hematologic splenic lesions: an exploratory real-world substudy.

Abstract

Purpose: Splenic lesions are a relatively rare but clinically relevant finding in internal medicine practice, particularly when hematologic or lymphoproliferative disorders are suspected. The aim of this exploratory subanalysis was to evaluate the role of CEUS perfusion patterns, with particular attention to venous wash-out and its timing, in discriminating between hematologic and non-hematologic disease in a real-world Internal Medicine cohort.

Methods: An exploratory retrospective subanalysis was conducted on patients included in the intervention arm of a multicenter Quality Improvement study evaluating a structured contrast-enhanced ultrasound (CEUS)-based diagnostic pathway in newly identified splenic lesions. Consecutive adult patients with newly detected splenic lesions identified on B-mode ultrasound who underwent CEUS as part of the initial diagnostic evaluation were included. CEUS perfusion patterns were qualitatively classified according to arterial wash-in, venous wash-out, and wash-out timing. The diagnostic reference standard consisted of histologic and/or cytologic examination when available, or integration of clinical, laboratory, and imaging follow-up data within 90 days. In patients without histologic confirmation, non-hematologic classification required clinical and laboratory stability or improvement, stability or resolution of the splenic lesion on follow-up imaging, identification of an alternative non-hematologic diagnosis when present, and absence of a hematologic diagnosis during the 90-day follow-up. In line with the parent study, one patient in the intervention-arm cohort did not receive a definitive lesion-specific etiologic diagnosis within 90 days; however, no patient remained indeterminate for the prespecified binary endpoint because classification as hematologic versus non-hematologic was possible in all cases. The study endpoint was binary diagnostic classification (hematologic vs non-hematologic).

Results: Fifty patients were included, of whom 11 (22%) had a hematologic diagnosis and 39 (78%) had a non-hematologic diagnosis. In univariable analysis, multiple lesions, systemic symptoms, and especially venous wash-out were associated with hematologic diagnosis. Venous wash-out was present in 10/11 hematologic cases and in 3/39 non-hematologic cases. In hematologic cases, wash-out was almost always early (< 60 s), whereas in the few non-hematologic cases with wash-out it was late. In the exploratory multivariable model, venous wash-out was the only variable independently associated with hematologic diagnosis. Because of the small number of hematologic events, this association should be interpreted as statistically unstable and hypothesis-generating. Venous wash-out showed a sensitivity of 90.9%, specificity of 92.3%, positive predictive value of 76.9%, negative predictive value of 97.3%, and accuracy of 92.0%.

Conclusion: In patients with newly identified splenic lesions, venous wash-out at CEUS, particularly if early, was the pattern most strongly associated with hematologic diagnosis in this cohort. The absence of wash-out was associated with a low probability of hematologic disease, suggesting a possible utility of CEUS in diagnostic rule-out in real-world Internal Medicine settings. However, diagnostic performance estimates, particularly negative predictive value, should be interpreted cautiously in view of the low prevalence of hematologic disease and the exploratory sample size. These findings require confirmation in larger cohorts.