The Diagnostic Accuracy of the Refined EBMT Criteria 2023 for SOS

Introduction：

Sinusoidal obstruction syndrome (SOS) is a rare but serious complication of allogeneic hematopoietic stem cell transplantation (allo-HSCT). An early diagnosis and intervention are crucial to improve patient outcomes. The gold standard for the diagnosis is a liver biopsy, but this can be challenging due to technical difficulties and complications. In 2023, the European Society for Blood and Marrow Transplantation (EBMT) proposed new diagnostic criteria for SOS (refined EBMT criteria 2023). The new diagnostic criteria define “proven” cases by histological findings or hepatic venous pressure gradient (HVPG) ≥ 10 mmHg, while “probable” and “clinical” criteria are based on clinical signs. However, the association between the histological findings and these clinical signs in recent allo-HSCT procedures remain unverified, and the diagnostic performance of new the criteria is therefore uncertain. We retrospectively investigated cases in which a histological examination of the liver was performed after transplant-related liver injury to evaluate the diagnostic performance of the new SOS clinical criteria.

Method：

In this single-center and retrospective study, we investigated cases in which a histological examination of the liver, including biopsy, necropsy, or autopsy, was performed to examine the cause of liver disorder or unexplained ascites within 180 days after allo-HSCT performed between January 2010 and March 2022. We collected laboratory data, clinical observations, imaging, and histological findings from medical records. In this study, we defined true SOS based on histological findings and evaluated the diagnostic accuracy for the “probable” criteria, “clinical” criteria, and HVPG ≥ 10 mmHg, one of the conditions of the “proven” criteria, in the refined EBMT criteria 2023.

Results：

Fifty cases were included in this study. The median age at allo-HSCT was 54.5 (range: 24-66) years. Diseases included AML (n = 19), ALL (n = 10), MDS (n = 8), ML (n = 7), MM (n = 1), CML (n = 2), AA (n = 2), and ATLL (n = 1). The graft sources included 12 related donors and 38 unrelated donors. Haploidentical transplantation using PTCy was performed in 3 cases. The number of bone marrow, peripheral blood, and cord blood sources was 10, 15, and 25, respectively. The conditioning regimens consisted of myeloablative (n = 31) and reduced-intensity (n = 19) regimens. There were 22 cases of “probable” SOS and 18 cases of “clinical” SOS. The median time to the diagnosis of “probable” SOS and clinical “SOS” was 19 days (range: 0-95) and 27 days (range: 8-151), respectively. Twelve cases of “probable” SOS and 5 cases of “clinical” SOS were classical SOS with onset before day21. A histological examination of the liver was performed by a transjugular biopsy in 29 cases, percutaneous biopsy in 9, autopsy in 10, and necropsy in 2. Fifteen patients were diagnosed with SOS on the basis of histology. The HVPG was measured in 21 patients, with a median of 9 mmHg (range: 2–16). The sensitivity / specificity of “probable” SOS, “clinical” SOS, and HVPG ≥ 10 mm Hg were 93.3% / 77.1%, 80% / 82.8%, and 85.7% / 78.6%, respectively.

Conclusion：

The diagnostic performance of “probable” and “clinical” criteria for SOS was found to be comparable to that of a diagnosis defined by HVPG ≥ 10 mmHg. “Probable” criteria reached an SOS diagnosis earlier when comparing the median time to diagnosis and it was also more sensitive than the “clinical” criteria. On the other hand, the “clinical” criteria had higher specificity than the “probable” criteria and showed acceptable sensitivity as a result of not missing late onset SOS. In the clinical setting, the “probable” and “clinical” criteria are accurate enough to diagnose true SOS.