Background
Ischemic stroke in people with active cancer is often difficult to classify because more than one stroke mechanism may be present at the same time. Traditional stroke systems such as TOAST were designed for the general stroke population and frequently label these events as cryptogenic, meaning no single clear cause is identified. In cancer patients, however, hidden cancer-related mechanisms such as hypercoagulability, disseminated intravascular coagulation, nonbacterial thrombotic endocarditis, treatment-related thrombosis, and tumor-associated embolism may be the true drivers of stroke. To address this gap, the American Heart Association recently proposed a cancer-related ischemic stroke classification, or CRIS, aimed at identifying strokes more likely linked to cancer itself.
This study evaluated how the CRIS framework changes stroke classification in real-world patients and whether it helps predict prognosis more accurately than the traditional TOAST approach.
Study Design and Patients
The investigators analyzed data from the SCAN study, a prospective Japanese cohort focused on ischemic stroke in patients with cancer and neoplasia. The study included patients with acute ischemic stroke and active cancer. Among the registered participants, only those with available D-dimer data were included in the present analysis, since D-dimer is a marker of coagulation activation and is often elevated in cancer-associated thrombosis.
A total of 135 patients were enrolled, and 132 were included in the final analysis. The median age was 75 years, and 37.9% were women. Stroke subtype was first assigned using TOAST criteria, then reclassified according to the CRIS framework. Survival analysis was performed using Kaplan-Meier curves, with group differences tested by the log-rank method.
How the New Classification Changed Stroke Labels
Under the TOAST system, the 132 patients were distributed as follows: 9 had small vessel occlusion, 20 had large artery atherosclerosis, 28 had cardioembolism, 10 had another determined cause, and 65 were classified as cryptogenic stroke.
After applying the CRIS framework, the picture changed meaningfully. Two patients originally placed in the other determined etiology group because they had disseminated intravascular coagulation were reclassified as CRIS. In addition, 46 patients previously labeled as cryptogenic were also reclassified as CRIS. In other words, the new cancer-focused framework identified many strokes that likely had an underlying cancer-related mechanism but were missed by standard classification.
This is an important clinical point. A large cryptogenic category can obscure the true burden of cancer-associated stroke and make it harder to estimate prognosis, choose the most appropriate workup, and consider anticoagulation or cancer-directed management when indicated.
Prognostic Impact
The most striking finding of the study was that patients classified as CRIS had substantially worse survival than patients with conventional stroke mechanisms or those remaining in the reclassified cryptogenic group. The difference in 1-year survival was statistically significant, with a global log-rank P value of less than 0.001.
At 3 months, survival was only 37.5% in the CRIS group, compared with 89.2% in the reclassified cryptogenic stroke group. This suggests that CRIS does not merely improve diagnostic labeling; it also identifies a subgroup with a far more aggressive clinical course.
The poor prognosis likely reflects the biology of advanced or active cancer and its systemic effects on coagulation and inflammation. Cancer-related stroke is often associated with very high D-dimer levels, multiple vascular territory infarcts, recurrent embolic events, and concurrent cancer progression. In such patients, stroke may be a marker of severe systemic disease rather than an isolated neurologic event.
Clinical Meaning of CRIS
The CRIS framework appears to provide two major benefits. First, it reduces the number of strokes left unexplained after routine evaluation by recognizing cancer-specific etiologies. Second, it improves prognostic stratification by separating patients with likely cancer-driven stroke from those with cryptogenic stroke for other reasons.
This distinction matters in everyday practice. Patients with suspected CRIS may need a more cancer-focused evaluation, including review of tumor status, coagulation markers, evidence of systemic embolism, and possible cardiac sources such as nonbacterial thrombotic endocarditis. Management may differ as well. Standard secondary stroke prevention strategies still apply, but the optimal choice between antiplatelet therapy and anticoagulation remains uncertain and should be individualized based on bleeding risk, cancer type, extent of disease, and likely stroke mechanism.
The study also reinforces the value of D-dimer as a supportive biomarker. Although D-dimer is not specific for cancer-related stroke, marked elevation can strengthen suspicion of a hypercoagulable state and may help identify patients who fit the CRIS phenotype.
Limitations
Several limitations should be considered. This was a cohort study from a Japanese population, so findings may not generalize fully to other ethnic groups or health-care settings. The analysis included only patients with available D-dimer data, which may introduce selection bias. The sample size was modest, and the number of events in some stroke categories was small.
In addition, reclassification based on a proposed system does not by itself prove that the assigned mechanism is always correct. Some patients may have multiple competing causes, such as both atherosclerosis and cancer-related hypercoagulability. Cancer stage, treatment status, and pathology details may also influence stroke risk and prognosis, and these factors can be difficult to capture completely in a registry-based study.
Finally, the study shows prognostic separation but does not establish the best treatment strategy for CRIS. Prospective trials are still needed to determine whether anticoagulation, antiplatelet therapy, cancer treatment optimization, or a combined approach offers the best outcomes.
Clinical Takeaway
This study supports the use of the AHA-proposed CRIS classification as a practical tool for patients with active cancer and ischemic stroke. By reclassifying many cases previously labeled cryptogenic, the system better reflects the underlying role of cancer and identifies a group with markedly poor short-term and long-term survival.
For clinicians, the key message is that ischemic stroke in a patient with active cancer should not automatically be treated as a routine cryptogenic event. A cancer-specific evaluation may uncover a hypercoagulable or embolic mechanism with important implications for prognosis and management.
Conclusion
In this prospective Japanese cohort, the CRIS classification reduced the proportion of cryptogenic strokes under the TOAST system and revealed a subgroup with significantly worse outcomes. The findings suggest that cancer-related ischemic stroke is not only a distinct clinical entity but also a strong marker of poor prognosis. As the framework is validated in larger and more diverse populations, it may become an important part of stroke evaluation in patients with active malignancy.
