Splenic Irradiation Preceding Allogeneic Hematopoietic Stem Cell Transplantation as a Possible Risk Factor of Sinusoidal Obstruction Syndrome: A Report of Three Cases

Sinusoidal obstruction syndrome (SOS), previously termed veno-occlusive disease, is a life-threatening complication targeting the liver. SOS is characterized by obstruction of small hepatic venous vessels, leading to liver damage and potentially progressing to liver failure. SOS subsequent to allogeneic hematopoietic stem cell transplantation (allo-HSCT) has been reported to occur at a rate of 13.7%, of which 25–33% will be accompanied by multiorgan dysfunction (MOD). SOS with MOD has a mortality rate of >80%, highlighting the clinical importance of this condition [1, 2]. Identifying SOS risk factors not only facilitates early diagnosis but may also offer the option of preventive interventions for patients at high risk.

Splenomegaly is a frequently encountered trait in patients with myeloproliferative neoplasms (MPNs). In a large series of 1,282 patients with primary myelofibrosis, palpable splenomegaly was noted in 72% of patients [3]. Splenomegaly is associated with an increased incidence of engraftment failure, and some centers perform splenectomy or splenic irradiation prior to allo-HSCT [4]. We report three consecutive cases undergoing low-dose splenic irradiation (LDSI) prior to allo-HSCT that subsequently developed SOS (Table 1). Although abdominal irradiation including hepatic irradiation has been identified as a risk factor of SOS, to what extent splenic irradiation contributes to development of SOS has not been elucidated to date [2]. The CARE Checklist has been completed by the authors for this case report, attached as online supplementary material (for all online suppl. material, see https://doi.org/10.1159/000542608).

A 53-year-old woman was diagnosed with MPN, unclassifiable, followed by transformation to acute myeloid leukemia 18 months later. Low-dose cytarabine in conjunction with aclarubicin was administered, but an increase of blasts and progressive splenomegaly were observed. A human leukocyte antigen (HLA)-matched donor was not available, and cord blood transplantation (CBT) in non-remission was scheduled. LDSI was performed from day −14 to day −11 (1.5 Gy/day for 4 days). CBT was carried out with conditioning regimen consisting of fludarabine (Flu)/busulfan (Bu)/total body irradiation (TBI) 4 Gy (O+ to AB+, female to female, nucleated cells 3.92 × 10⁷ cells/kg, CD34-positive cells 1.19 × 10⁵ cells/kg, 5/8 HLA allele-matched [mismatch at the HLA-A, B, C locus]). Tacrolimus and mycophenolate mofetil were administered as graft-versus-host disease (GVHD) prophylaxis. The patient showed sudden weight gain of more than 5% (41.7 kg to 44.0 kg) on day 2, total bilirubin rose to 2.1 mg/dL, and painful enlargement of the liver developed; she was diagnosed with SOS on day 6 by both the modified Seattle criteria and Baltimore criteria, and also fulfilled the European Society for Blood and Marrow Transplantation (EBMT) criteria by retrospective evaluation. There was no history of prior hepatic disease or iron overload. Dry concentrated human antithrombin III preparation of 1,500 units/day was administered on days 6–8, but the SOS did not improve. Recombinant thrombomodulin (rTM) 380 U/kg was started on day 8, total bilirubin levels started to improve after peaking out at 12.8 mg/dL on day 12, and body weight also decreased after day 26. rTM was discontinued on day 34, and normalization of total bilirubin was observed on day 55. However, the acute myeloid leukemia relapsed on day 51 and the patient succumbed to the disease on day 83. Engraftment was never accomplished.

A 51-year-old man with a 9-year history of essential thrombocythemia presented with progressive anemia and splenomegaly. Bone marrow examination revealed evolution to secondary myelofibrosis. Ruxolitinib 25 mg/day was initiated but showed no efficacy, and allo-HSCT from an HLA-matched unrelated donor was planned. LDSI was performed from day −11 to day −8 (1.5 Gy/day for 4 days). Allo-HSCT with conditioning consisting of Flu/melphalan/TBI 4 Gy was performed (A+ to O+, male to male, nucleated cells 2.72 × 10⁸ cells/kg). Tacrolimus and methotrexate were administered as GVHD prophylaxis. Weight gain of more than 5% (71.0 kg to 75.2 kg) was confirmed on day 13, total bilirubin rose to 2.2 mg/dL, ascites and painful enlargement of the liver developed, and he was diagnosed with SOS on day 19 by the EBMT criteria. There was no history of prior hepatic disease or iron overload. Defibrotide was initiated on day 20, and total bilirubin started to decline after peaking at 2.8 mg/dL on day 21. Engraftment was confirmed on day 24. Defibrotide was discontinued from day 27 because he developed orbital apex syndrome due to sinusitis and required surgery. SOS did not recur, but he developed chronic GVHD of the skin and liver and died of septic shock on day 268.

A 62-year-old man was referred to our institution due to a 2-year history of leukocytosis. Bone marrow examination revealed atypical chronic myeloid leukemia. Leukocytosis was controlled with hydroxyurea for 5 months. An HLA-matched donor was not available, and CBT was planned. Splenomegaly was evident, and LDSI was performed from day −14 to day −11 (1.5 Gy/day for 4 days). CBT with conditioning consisting of Flu/Bu/melphalan/TBI 4 Gy was performed (O+ to B+, female to male, nucleated cells 3.12 × 10⁷ cells/kg, CD34-positive cells 0.77 × 10⁵ cells/kg, 5/8 HLA allele-matched [mismatch at the HLA-B, C, DRB1 locus]). Tacrolimus and mycophenolate mofetil were administered as GVHD prophylaxis. Weight gain of more than 5% (54.1 kg to 57.1 kg) was observed on day 9, followed by elevation of total bilirubin to 2.1 mg/dL, development of ascites and painful enlargement of the liver, and he was diagnosed with SOS on day 11 according to the EBMT criteria. There was no prior history of hepatic disease or iron overload. Defibrotide was started on day 11, but jaundice and ascites did not improve. On day 22, the patient was engrafted, but a subdural hematoma developed, and defibrotide was discontinued. On day 26, defibrotide was resumed after confirmation that the hematoma had not expanded, but the SOS did not improve, and the patient died of MOD on day 49.

SOS is a potentially fatal complication of allo-HSCT, and the mortality rates of severe cases are reported to be over 80% [5]. Defibrotide and rTM are both effective for treating SOS, but the 100-day overall survival rate remains poor at approximately 50% even with utilization of such therapeutics [6]. The EBMT criteria are currently widely accepted for diagnosis of SOS and were utilized in cases 2 and 3, but case 1 was at a time before the EBMT criteria existed, and SOS was diagnosed based on the modified Seattle criteria and Baltimore criteria. Prediction of disease onset and early therapeutic intervention are important, and various risk factors have been reported for the development of SOS including older age, females receiving norethisterone, certain chemotherapeutic agents (e.g., Bu, gemtuzumab ozogamicin, inotuzumab ozogamicin), serum bilirubin >1.5 mg/L, preexisting liver disease, unrelated or HLA-mismatched donor source, and high-dose TBI [2]. All 3 cases had several risk factors for developing SOS, including older age, unrelated donors, or HLA-mismatched donors. However, the presence of these risk factors alone is not enough to estimate the incidence of SOS. Recently, the Endothelial Activation and Stress Index (EASIX) score has been reported to be useful in predicting the development of SOS [7]. EASIX was tested in a validation cohort including 380 patients undergoing allo-HSCT of which 32 patients developed SOS, and scores for patients with SOS were significantly higher compared to patients without SOS (8.64 vs. 2.28, respectively). EASIX scores of the 3 presented cases were relatively low ranging from 0.40 to 4.40 (Table 1). Splenomegaly has been reported to be a risk factor for SOS in patients with colorectal cancer receiving oxaliplatin-based chemotherapy [8], but splenomegaly has not been recognized as a risk factor of SOS during allo-HSCT. LDSI can ameliorate symptoms of splenomegaly and also may contribute to prevention of engraftment failure by reducing spleen size. Splenomegaly has been reported to be associated with longer engraftment times, but Kalman et al. [9] demonstrated that LDSI prior to allo-HSCT in 8 myelofibrosis patients with splenomegaly led to similar engraftment times compared to patients without splenomegaly, and the authors concluded that LDSI prior to allo-HSCT was both safe and effective. We report conflicting results. At our hospital, a total of 26 allo-HSCT patients presented with splenomegaly (mainly MPN) of which 3 patients (11.5%) underwent pretransplant LDSI. All 3 patients with LDSI developed SOS, whereas all 23 patients without LDSI did not develop SOS. The underlying mechanisms of how LDSI may evoke SOS are unknown. Although all 3 patients underwent LDSI, the irradiation fields basically did not include the liver, suggesting an indirect mechanism of SOS development such as LDSI-induced blood flow changes and vascular endothelial cell damage. An alternative method for managing splenomegaly during allo-HSCT may be ruxolitinib administration prior to conditioning. A large retrospective study including 551 patients undergoing allo-HSCT for primary or secondary myelofibrosis that compared patients pretreated with (n = 277) or without (n = 274) ruxolitinib showed that in patients experiencing spleen response with ruxolitinib, the neutrophil engraftment on day 45 was higher compared to patients who had no or lost response to ruxolitinib (94% vs. 85%). Furthermore, ruxolitinib-pretreated patients with ongoing spleen response at time of allo-HSCT had significantly lower risk of relapse (8.1% vs. 19.1%) and better 2-year event-free survival (68.9% vs. 53.7%) compared to patients without ruxolitinib pretreatment [10].

In conclusion, we report three consecutive cases undergoing LDSI prior to allo-HSCT that developed SOS. EASIX scores of the 3 patients were 0.40, 3.82, and 4.40, respectively. LDSI may not be as safe as previously reported, but our findings need to be confirmed in larger studies. For patients with myelofibrosis, ruxolitinib may be a safer alternative for treating splenomegaly prior to allo-HSCT.

This study was reviewed and approved by the Institutional Review Board of Juntendo University, Approval No. JHS24-007. Written informed consent was obtained from all patients for publication of the details of their medical case and any accompanying images prior to their passing away.

All authors have no conflicts of interest to declare.

There is no funding source for this work.

Writing the manuscript: S.S. and H.Y. Data collection, analysis, and interpretation: A.U., Y.M., and T.I. Literature research: M.T. and Y.H. Revising the manuscript: M.A.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from S.S. or H.Y. upon reasonable request.