The treatment of acute stroke now includes recombinant tissue plasminogen activator (TPA), approved for use in select patients within three hours of acute ischemic stroke. The presence of intracranial blood on computed tomographic (CT) scan of the head excludes patients from this therapy. It is generally advised that patients with an increased risk of hemorrhage (such as those with sulcal effacement, mass effect or edema) not receive recombinant TPA. Schriger and associates evaluated physician accuracy in interpreting cranial CT scans to determine eligibility for thrombolytic therapy in patients with acute stroke.

The authors of the study and a neuroradiologist reviewed all cranial CT scans performed at a university teaching hospital during a one-year period and selected head CTs that showed parenchymal hemorrhage, early infarction, calcification and old infarction. Using a consensus process, they classified each scan that demonstrated acute infarction as easy, intermediate or difficult to interpret, and each scan that demonstrated hemorrhage as easy or difficult to interpret.

Physicians participating in the study were asked to assume that each scan was of a patient who arrived at the hospital within hours after the onset of a neurologic deficit. Each physician was asked to decide if the patient was eligible for thrombolytic therapy. Possible answers were (1) yes; (2) no, because of hemorrhage; and (3) no, because of acute infarction. Each physician was given a list of findings that were contraindications to thrombolytic therapy (hemorrhage, early hypodensity, mass effect and shift) and findings that were not contraindications (calcification, atrophy and old infarction) to remind them of the criteria for this study. Physicians were shown two difficult hemorrhages, one intermediate acute infarction, one normal CT and one “imposter” scan. Physicians who correctly interpreted all of these scans were then placed in the advanced track, where they interpreted 10 more scans at a higher level of difficulty. Those who did not read all of the initial CT scans correctly were placed in the standard track and read 10 more scans.

A total of 29 neurologists, 36 radiologists and 38 emergency physicians participated in the study. Most (78 percent) failed to achieve correct readings on at least one of the five initial scans. Even the 21 physicians who were placed in the advanced track had a substantial number of incorrect readings overall. Neurologists and radiologists were 100 percent accurate in identifying easy hemorrhages; 94 percent of emergency physicians were also able to correctly interpret these scans. For scans classified as difficult hemorrhages, 80 percent of radiologists were correct in their interpretation compared with 78 percent of neurologists and 56 percent of emergency physicians. The average correct score by all three groups of physicians on all scans was 77 percent. The overall score for radiologists was 83 percent; the overall score was 83 percent for neurologists and 67 percent for emergency physicians.

The authors conclude that many physicians who are routinely called on to determine whether a CT scan precludes the use of recombinant TPA do not possess the skills needed to recognize hemorrhage on cranial CT scans. They note that the three-hour time limit imposed on administering recombinant TPA to a patient with an acute ischemic stroke mandates skillful interpretation of cranial CT scans. The authors suggest advanced training in CT scan interpretation or more widespread use of teleradiography to improve the accuracy of cranial CT scan interpretations.