This study complied with the Declaration of Helsinki and received approval from our institutional review board. Written informed consent was acquired from all the participants and their guardians.

This prospective cohort study included consecutive pediatric patients who underwent hematopoietic stem cell transplantation at a single tertiary hospital from March 2021 to April 2022. Considering the documented prevalence of veno-occlusive disease in hematopoietic stem cell transplantation patients receiving myeloablative conditioning regimens at this institution, which exceeds 20% [25], we had planned to enroll 40 participants in this feasibility study.

Patients younger than 2 years were excluded because the SWE and SWD imaging require sufficient intercostal space to access the liver parenchyma. Other exclusion criteria included patients with pre-existing chronic liver disease at the pre-hematopoietic stem cell transplantation assessment to avoid potential bias, incomplete follow-up, and patients older than the age of 18 years. One patient was excluded because hematopoietic stem cell transplantation was canceled due to the worsening of the underlying disease after baseline US examination, and the other patient was excluded from the analysis due to being older than the age of 18 years.

The study protocol consisted of four sessions: a baseline US before conditioning chemotherapy and three follow-up US after hematopoietic stem cell transplantation. The baseline US was performed 7 ± 3 days before hematopoietic stem cell transplantation (baseline session), followed by subsequent examinations at s7 ± 3 days (1st follow-up), 14 ± 3 days (2nd follow-up), and 28 ± 3 days (3rd follow-up) after hematopoietic stem cell transplantation. In addition, clinical data from all participants were collected on the date of the US examination.

The US examinations included grayscale, color Doppler, SWE, and SWD imaging techniques. All US examinations were conducted by an experienced pediatric radiologist (S.L. with 12 years of experience in pediatric US examination) using the same machine (Aplio i800, Canon Medical Systems, Otawara, Japan) with a 1–6 MHz convex transducer. The 1st and 2nd post-transplant follow-up US examinations were performed in a sterile room, considering the patient’s neutropenia, in contrast to other follow-up US examinations which were performed in the US room.

Following the criteria outlined in previous studies [9], the collected grayscale US findings encompassed: (1) liver size, determined by the craniocaudal length in the mid-axillary level sagittal scan (hepatomegaly defined as a size increase >1 cm compared to the baseline), (2) spleen size (splenomegaly defined as a size increase >1 cm compared to the baseline), (3) gallbladder wall thickness (gallbladder edema defined as thickness >6 mm), and (4) presence of ascites. Evaluated Doppler US findings encompassed (1) portal vein velocity (reversed or weak portal vein flow defined as velocity <10 cm/s), (2) hepatic vein flow phasicity (presence of monophasic flow), and (3) hepatic arterial resistive index (index >0.75).

SWE and SWD measurements were conducted in the intercostal view under gentle free-breath conditions. Quad-view mode, including SWE, SWD, propagation map, and corresponding grayscale map, can simultaneously depict four maps; we placed a 2 × 2 cm-sized sample box at a consistent depth below the liver capsule and obtained values within a uniform color-coded map with a parallel propagation area. The median stiffness (kPa) and viscosity (m/s/kHz) values, obtained from ten measurements on the SWE map and SWD map, were utilized in the analysis to yield reliable measurements [26].

For veno-occlusive disease patients, additional US follow-up examinations outside the study protocol were performed at the clinician’s request until there was no further clinical evidence of veno-occlusive disease. To assess serial changes in parameters before and after the diagnosis of veno-occlusive disease, we analyzed the changes in US data based on the time of diagnosis, including the additional US. In each patient, we reordered and named the US sessions as follows: the session just before diagnosis (− 1 session), the session at diagnosis (0 session), the first follow-up session after diagnosis (+ 1 session), and the second follow-up session after diagnosis (+ 2 session). The parameter values corresponding to each reordered US session were averaged across patients.

The clinical diagnosis of veno-occlusive disease was based on the European Society for Blood and Marrow Transplantation criteria for pediatric patients [5]. Veno-occlusive disease was clinically diagnosed if two or more criteria were present: transfusion-refractory thrombocytopenia, weight gain (≥5% above baseline), bilirubin ≥2 mg/dL within 3 days, hepatomegaly, and ascites.

The severity of veno-occlusive disease was classified into three categories, ranging from mild to severe, based on clinical parameters assessed on the day of each US session. The severity of veno-occlusive disease by the European Society for Blood and Marrow Transplantation criteria included the following conditions: serological liver function abnormalities, such as increased transaminases to twice the normal range (mild) or less than five times (moderate), lasting days of refractory thrombocytopenia, hyperbilirubinemia, ascites, glomerular filtration rate, pulmonary, and coagulation status [5].

For veno-occlusive disease prophylaxis, all patients received a continuous infusion of lipo-prostaglandin E1 (Dongkook Pharm., Seoul, Korea) at a dose of 1 µg/kg per day with or without the addition of low-molecular-weight heparin (Handok Inc., Seoul, Korea) according to institutional guidelines for hematopoietic stem cell transplantation. Patients diagnosed with veno-occlusive disease were treated with defibrotide (Handok Inc.) until the resolution of veno-occlusive disease.

The distribution of clinical variables was summarized, and a comparison was conducted between the clinical factors and US parameters of the veno-occlusive disease and non-veno-occlusive disease groups during the surveillance period following hematopoietic stem cell transplantation. Continuous variables between the groups were compared using either the independent t-test or the Wilcoxon rank-sum test. Categorical variables were analyzed using the Fisher’s exact test.

The diagnostic performance of the US parameters for detecting veno-occlusive disease was evaluated by constructing receiver operating characteristic (ROC) curves. Cutoff values for the US parameters were determined using the Youden J index to optimize sensitivity and specificity. A generalized estimating equation analysis was carried out to assess the predictors of veno-occlusive disease with the clinical and US parameters. This model accounted for the correlation between repeated measurements within an individual. Significant factors identified in the univariate analysis (those with P<0.20) were included in the multivariate analysis. In the multivariate analysis, statistical significance was defined as P<0.05. In addition, the descriptive analysis was only performed to compare the temporal changes in US data before/after clinical veno-occlusive disease diagnosis due to the limited number of patients in each session of veno-occlusive disease patients. All statistical analyses were conducted using SAS statistical software (SAS system for Windows, version 9.4; SAS Institute, Cary, NC).