ABSTRACT
BACKGROUND AND OBJECTIVES: Incidental diffuse-weighted imaging (DWI)-positive subcortical and cortical lesions, or acute incidental cerebral microinfarcts (CMIs), are a common type of brain ischemia, which can be detected on magnetic resonance DWI for approximately 2 weeks after occurrence. Acute incidental CMI was found to be more common in patients with cancer. Whether acute incidental CMI predicts future ischemic stroke is still unknown. We aimed to examine the association between acute incidental CMI in patients with cancer and subsequent ischemic stroke or transient ischemic attack (TIA). METHODS: This is a retrospective cohort study. We used Clalit Health Services records, representing over half of the Israeli population, to identify adults with lung, breast, pancreatic, or colon cancer who underwent brain MRI between January 2014 and April 2020. We included patients who underwent scan between 1 year before cancer diagnosis and 1 year after diagnosis. Primary outcome was ischemic stroke or TIA using International Classification of Diseases, Ninth Revision codes. Secondary outcomes were intracranial hemorrhage (ICH) and mortality. Records were followed from first MRI until primary outcome, death, or end of follow-up (January 2023). Cox proportional hazards models were used to calculate hazard ratio (HR) for patients with and without acute incidental CMI, as a time-dependent covariate. RESULTS: The study cohort included 1,618 patients with cancer, among whom, 59 (3.6%) had acute incidental CMI on at least 1 brain MRI. The median (interquartile range) time from acute incidental CMI to stroke or TIA was 26 days (14-84). On multivariable analysis, patients with acute incidental CMI had a higher stroke or TIA risk (HR 2.97, 95% CI 1.08-8.18, p = 0.035) compared with their non-CMI counterparts. Acute incidental CMIs were also associated with mortality after multivariable analysis (HR 2.76, 95% CI 2.06-3.71, p < 0.001); no association with ICH was found. DISCUSSION: Acute incidental CMI on brain MRI in patients with active cancer is associated with an increased risk of near-future ischemic stroke or TIA and mortality. This finding might suggest that randomly detected acute incidental CMI in patients with cancer may guide primary cerebrovascular risk prevention and etiologic workup.
Subject(s)
Incidental Findings , Ischemic Stroke , Neoplasms , Humans , Female , Male , Aged , Middle Aged , Ischemic Stroke/epidemiology , Ischemic Stroke/diagnostic imaging , Ischemic Stroke/complications , Retrospective Studies , Neoplasms/complications , Neoplasms/epidemiology , Neoplasms/diagnostic imaging , Ischemic Attack, Transient/diagnostic imaging , Ischemic Attack, Transient/epidemiology , Ischemic Attack, Transient/complications , Israel/epidemiology , Cerebral Infarction/diagnostic imaging , Cerebral Infarction/epidemiology , Cohort Studies , Aged, 80 and over , AdultABSTRACT
BACKGROUND AND OBJECTIVES: Cerebral microinfarcts (CMIs) are the most common type of brain ischemia; however, they are extremely rare in the general population. CMIs can be detected by magnetic resonance diffusion-weighted imaging (MRI-DWI) only for a very short period of approximately 2 weeks after their formation and are associated with an increased stroke risk and cognitive impairment. We aimed to examine CMI detection rate in patients with lung cancer (LC), which is strongly associated with ischemic stroke risk relative to other cancer types. METHODS: We used the Clalit Health Services record (representing more than 5 million patients) to identify adults with LC and breast, pancreatic, or colon cancer (non-lung cancer, NLC) who underwent brain magnetic resonance diffusion (MRI) scan within 5 years following cancer diagnosis. All brain MRI scans were reviewed, and CMIs were documented, as well as cardiovascular risk factors. RESULTS: Our cohort contained a total of 2056 MRI scans of LC patients and 1598 of NLC patients. A total of 143 CMI were found in 73/2056 (3.5%) MRI scans of LC group compared to a total of 29 CMI in 22/1598 (1.4%) MRI scans of NLC (p < 0.01). Cancer type (e.g. LC vs NLC) was the only associated factor with CMI incidence on multivariate analysis. After calculating accumulated risk, we found an incidence of 2.5 CMI per year in LC patients and 0.5 in NLC. DISCUSSION: CMIs are common findings in cancer patients, especially in LC patients and therefore might serve as a marker for occult brain ischemia, cognitive decline, and cancer-related stroke (CRS) risk.
