Is the Modified Fisher Scale Insufficient or Is Better Training Needed?

Four days after being successfully treated for subarachnoid hemorrhage, a 38-year-old patient in your Neuro ICU declines neurologically and loses her hearing. An MRI reveals severe, bilateral vasospasm and delayed cerebral ischemia (DCI) in her right auditory cortex.

DCI after SAH is complex and dynamic. While the Modified Fisher Scale (mFS) is one of the standard predictive tools used for DCI after SAH, its limitations fuel frustrations for neurosurgeons and neurocritical care physicians charged with caring for these patients.

Is the scale lacking or is physician training? And are there better tools on the horizon?

Researchers recently highlighted the need to standardize the definition and training for the mFS via a cross-sectional survey and literature review that reveals how the widely used scale lacks in reliability.

In the survey, 46 attending physicians who assign mFS from 32 institutions were asked to grade the mFS of randomly selected SAH patients’ admission CT scans.

Melinosky et. al found what they consider only “moderate” interrater reliability of mFS scoring, noting specifically:

• • Only 24% of participants correctly identified that there is no clear measurement of thick or thin SAH
  • 52% correctly identified that any blood in any ventricle is scored as intraventricular blood
  • Half were not aware that Claassen’s scale was unique from the mFS
  • Participants who could properly define the mFS score components showed good agreement
  • 72% of the respondents said they would like more formal training on the scale

The study notes that the survey participants, who assign mFS at institutions with high numbers of SAH patients, were mostly neurologists with neurocritical care certification with a median of 5 years of neurocritical care practice.

As part of their literature review, Melinosky et. analyzed 164 original research papers utilizing the mFS. Of the 37 studies that included some definition of the mFS, they found that only 17 listed the correct criteria and 20 listed incorrect or incomplete criteria. Melinosky et. al also note that the study that demonstrated the highest reliability of the Fisher scale provided each rater with a detailed description of the scale. This led them to conclude that simply providing the criteria could be enough to improve the reliability of mFS.

The Need for Better Tools

The Fisher grade was developed in 1980, when Canadian neurologist C. Miller Fisher investigated the relationship between subarachnoid blood detected by CT scans and the development of cerebral vasospasm in 47 cases of verified ruptured saccular aneurysm.

In 2001, Claassen et. al proposed a modified scale to account for thick SAH and bilateral intraventricular hemorrhage as risk factors for DCI. Five years later, Frontera et. al confirmed that the modified scale predicts symptomatic vasospasm after subarachnoid hemorrhage more accurately than original Fisher scale.

There have been strides to develop more accurate and timely prediction models. Recently, researchers in Texas analyzed whether machine learning could predict DCI more accurately than both the mFS and Hunt-Hess scale in 399 SAH patients. In the end, the machine learning model was 36% more accurate in predicting DCI than both scales, yet on par when compared to physician’s ability to predict DCI.

Another study sought to test whether machine learning could predict DCI earlier than physicians in 310 SAH patients. The hourly DCI risk scores, generated via machine learning, were able to correctly predict between 64% and 91% of DCI events as soon as 12 hours before clinicians.

Lastly, transcranial dopplers have also been used to diagnose vasospasms but recent research has shown the tool’s limited ability to help identify patients who may develop ischemia from vasospasm. One study also found that such dopplers only able to obtain the sensitivity and specificity needed to detect DCI on day 8 of SAH.