Discussion

Previous studies have reported on reperfusion outcomes in patients with hematologic malignancies and ischemic stroke. ^{Reference Rael, Webb, Brown, Ruff, Keser and Sener11} Our study expands upon this prior work by using a large, nationally representative cohort to compare outcomes across specific hematologic malignancy subtypes and by evaluating post-discharge stroke readmissions. In this nationwide study of 1.3 million hospitalizations with AIS, nearly 1% of patients had a comorbid hematological malignancy, most commonly leukemia. We found that acute leukemias were associated with the highest risk of in-hospital mortality, SAH and adverse discharge disposition. Hodgkin lymphoma was also associated with increased in-hospital mortality, whereas non-Hodgkin lymphoma and other hematological malignancies showed no significant excess risk. We observed a significant disparity in the use of tPA, where patients with malignancy were 23% less likely to receive tPA but the rates of MT were comparable between patients with and without cancer. Our findings are consistent with recent literature showing significantly greater utilization of reperfusion therapies in teaching (or academic) hospitals than in non-teaching/community hospitals. ^{Reference Moradiya, Crystal, Valsamis and Levine12,Reference Qureshi, Bhatti and Gomez13} We found that when indicated, the use of tPA or MT was safe, and among the subgroups of patients treated with these reperfusion therapies, the in-hospital outcomes were comparable between patients with and without cancer, except for SAH, which remained significantly more likely in patients with acute leukemias. Finally, we found that the early risk of recurrent stroke was similar between patients with and without hematological malignancies.

There are important differences between hematologic and solid organ malignancies in the setting of AIS. Prior studies in solid tumors have largely attributed stroke risk to tumor-associated hypercoagulability, systemic inflammation and treatment-related vascular injury. ^{Reference Navi, Reiner and Kamel14–Reference Lau, Wong and Teo16} In contrast, hematologic cancers directly alter circulating blood components and coagulation pathways, producing a unique risk profile that includes both bleeding and thrombotic complications. For example, patients with leukemia may present with hyperleukocytosis, leukostasis or disseminated intravascular coagulation, while those with lymphoma or plasma cell disorders may develop hyperviscosity or therapy-related cytopenias. ^{Reference Ferro and Infante3,Reference Adams17,Reference Del Prete, Kim, Lansigan, Shatzel and Friedman18} These distinctions highlight why hematologic malignancies warrant separate investigation, as stroke mechanisms, treatment eligibility and outcomes may differ substantially from those observed in patients with solid tumors.

Several studies have suggested that hematological cancers, including leukemias, lymphomas and plasma cell neoplasms, are all associated with an increased risk of ischemic stroke. ^{Reference Adelborg, Corraini and Darvalics2,Reference Navi, Reiner and Kamel14,Reference Wei, Chen and Wu19–Reference Del Prete, Kim, Lansigan, Shatzel and Friedman22} This association is likely multifactorial and presumably driven by cancer-associated hypercoagulability, hyperviscosity, cancer therapies (such as radiation, chemotherapy and stem cell transplantation), shared risk factors between cancer and stroke, infections and vascular involvement. ^{Reference Adams17} Acute management of ischemic stroke centers on restoring perfusion with tPA and/or MT. Major clinical trials that established the role of these acute reperfusion therapies in ischemic stroke either excluded patients with active cancer or did not provide sufficient data regarding the safety and effectiveness of their use in the setting of malignancy. ^{23,Reference Goyal, Menon and van Zwam24} Therefore, there is limited evidence that guides the treatment decisions in these patients. In our study, patients with hematological cancers were less likely to receive tPA but not MT. This finding is in line with results of prior studies in patients with solid organ malignancies. ^{Reference Selvik, Naess and Kvistad15,Reference Owusu-Guha, Guha and Miller25,Reference Pana, Mohamed, Mamas and Myint26} The NRD does not provide information on clinical decision-making or specific contraindications for thrombolysis; therefore, the reasons for lower tPA use in patients with hematologic malignancies cannot be directly determined. Potential explanations include higher rates of thrombocytopenia or coagulopathy, provider concern for hemorrhagic complications, concurrent chemotherapy or invasive procedures and institutional practice variation. These factors likely contribute to the observed differences in the utilization of tPA in these patients and warrant further prospective study. Our findings highlight the need for close collaboration between hematology/oncology and vascular neurology teams when evaluating patients with hematologic malignancies for reperfusion therapies. Multidisciplinary decision-making may help address underuse of tPA by balancing the risks of coagulopathy against the potential benefits of early reperfusion. Future prospective studies are needed to define evidence-based eligibility criteria for IV thrombolysis in patients with hematologic malignancies. In particular, determining safe platelet count thresholds, coagulation parameter cut-offs and the role of transfusion support strategies would help guide individualized decision-making and reduce unnecessary exclusion from reperfusion therapies.

Patients with hematological malignancies are at an increased risk of hemorrhagic and thrombotic complications after AIS. In a nationwide survey of approximately 11,000 patients with acute myeloid leukemia, the authors showed that patients with leukemia had a 5.5-fold increased risk of in-hospital mortality after AIS as compared to those without leukemia. ^{Reference Del Prete, Kim, Lansigan, Shatzel and Friedman22} Hematological malignancies comprise a heterogeneous group of cancers and similar data for other hematological malignancies are lacking. It is also unknown if the utilization of acute reperfusion therapies affects these outcomes. In our study, we found that patients with leukemia had higher in-hospital mortality and adverse discharge disposition. Acute leukemias frequently present with severe cytopenias, DIC or hyperleukocytosis, and therapies such as L-asparaginase further increase thrombotic risk. Infections and treatment-related complications are also disproportionately common in leukemia. Our subgroup analysis indicated that acute leukemias were the primary drivers of this excess mortality. These pathophysiologic and treatment-related factors likely explain the worse outcomes observed in leukemia compared with other hematologic malignancies. The specific causes of in-hospital mortality could not be determined from the NRD, as cause-of-death detail is not available. It is therefore unclear whether the excess mortality observed in leukemia was directly attributable to stroke, stroke complications or malignancy-related factors such as infection or treatment toxicity. Notably, the rates of coded intracranial hemorrhage were not significantly higher in the leukemia group, suggesting that hemorrhagic events were unlikely to fully explain the mortality difference. The increased risk is likely multifactorial, reflecting both the acute neurological insult and the systemic vulnerability conferred by leukemia. We showed that the use of tPA or MT appeared to attenuate the association between leukemia and adverse outcomes. Furthermore, the use of acute reperfusion was not associated with increased risk of intracranial hemorrhage in patients with cancer. These findings suggest that when indicated, the use of tPA and MT is safe in patients with hematological malignancies.

Patients with certain solid organ malignancies are at a higher risk of ischemic stroke recurrence as compared to patients without cancer. ^{Reference Lau, Wong and Teo16} A study including pediatric patients also showed that patients with cancer (including 40% hematological malignancies) had an increased risk of stroke recurrence, and this risk was higher in children with arteriopathies. ^{Reference Sun, Linds and Sharma27} The potential mechanisms of stroke recurrence in the setting of hematological malignancies could be associated with cancer-associated coagulopathy or the use of cancer therapies such as tyrosine kinase inhibitors or other chemotherapeutic agents, radiation therapy and stem cell transplantation. ^{Reference Adams17} Furthermore, patients with cancer are less likely to receive antithrombotic medications after AIS, ^{Reference Sun, Linds and Sharma27} and inadequate use of appropriate secondary stroke prevention measures may also contribute to stroke recurrence in these patients. In this study, we found that the rates of stroke recurrence were comparable between patients with and without hematological malignancies during the early follow-up period. These results have important clinical implications as the optimal treatment of cancer-associated stroke remains unknown, ^{Reference Zhang, Wang and Du1} and establishing the risk of recurrent cerebrovascular events may assist with the treatment decisions in select cases.

In our study, obesity was independently associated with a higher risk of recurrent AIS readmission. This may reflect clustering of vascular risk factors such as hypertension, diabetes and sleep apnea, which frequently co-occur with obesity and elevate stroke risk. ^{Reference Kernan, Inzucchi, Sawan, Macko and Furie28} Obesity may also complicate cancer treatment and secondary stroke prevention through altered drug metabolism and dosing. Additionally, obese patients often face greater functional limitations after stroke, which may predispose them to rehospitalization. ^{Reference Albitar, Harun and Abidin29} However, residual confounding and limitations of administrative coding cannot be excluded, and this association warrants further validation in studies with detailed clinical and lifestyle data.

Major strengths of our study include the use of a large, nationally representative patient cohort and evaluation of outcomes after comprehensive adjustments for the differences in baseline comorbidities and stroke severity, which increases the validity and generalizability of our findings. However, our study has several limitations, mostly inherent to the use of an administrative database. First, we lacked pertinent clinical information regarding the stage or treatment status of cancer, baseline laboratory data, location or etiology of stroke (and whether the stroke was precipitated by cancer-related complications) and the use of stroke prevention measures, all of which may influence treatment decisions and outcomes in patients with hematologic malignancy. Future datasets linking oncologic treatment exposures with cerebrovascular outcomes will be critical to address this gap. Second, as mortality data after hospital discharge are not available in the NRD, we could not perform competing risk analyses. Instead, we used Kaplan–Meier methods, which assume non-informative censoring and may overestimate stroke recurrence by not accounting for death as a competing risk, particularly in patients with hematologic malignancy who are at higher baseline risk of mortality. Third, due to lack of long-term follow-up, we could only assess the early risk of recurrent stroke. Fourth, we could not differentiate between symptomatic and asymptomatic IPH as the NRD does not contain radiographic or clinical adjudication data necessary for making that distinction. Also, we were unable to adjust for key clinical and imaging factors such as onset-to-treatment times or infarct core versus penumbra, as these data are not available in the NRD. Although we included the validated SASI score to approximate stroke severity, this may not fully capture the prognostic impact of treatment timing and radiographic profile. Fifth, we were unable to assess workflow metrics such as door-to-imaging or door-to-needle time. Patients with hematologic malignancies may undergo additional laboratory or clinical evaluations before treatment, which could prolong decision-making and potentially reduce eligibility for thrombolysis.

Despite these limitations, our study findings address an important knowledge gap regarding the outcomes of AIS in the setting of hematological malignancies. Our findings apply primarily to acute, in-hospital outcomes and do not address long-term functional status or mortality beyond the hospitalization and short-term readmission period. Also, our findings on the safety of reperfusion therapies apply at the population level and do not extend to unmeasured high-risk subgroups, such as patients with thrombocytopenia, severe coagulopathy or recent procedures, in whom thrombolysis remains contraindicated by current guidelines. This study emphasizes the urgent need for prospective studies to develop decision frameworks that integrate standard stroke reperfusion criteria with hematology-specific risks. Such frameworks, ideally informed by multidisciplinary collaboration, may help guide clinicians in determining when reperfusion therapy is appropriate for patients with hematologic malignancies.