xRead - September 2022

Cerebrospinal fluid rhinorrhea diagnosis

Magnetic resonance cisternography MRC is a common imaging modality used for CSF leak diagnosis. It is a noninvasive study capable of both leak confirmation and site localization by the inherent bright signal of CSF on T2-weighted images passing from intracra nial into paranasal sinuses, as well as by the identification of herniated soft tissue in skull-base defects. The database searches found 21 studies that discuss use of MRC for CSF leak identification and met inclusion criteria. These are all level 3 evidence cross-sectional studies. Collectively, the studies reported a sensitivity of MRC in detecting CSF leak site of 56% to 94%, 24, 38, 42, 44–47, 51–56 a specificity of 57% to 100%, 44, 46, 55, 56 a positive predictive value of 64% to 92%, 52, 55 a negative predictive value of 71%, 55 and an overall accuracy of 78% to 96%. 24, 43, 44, 53, 56 In the Zapalac et al. 24 study, the cost of MRC was re ported to be $913. They recommended it be used only after HRCT has failed. Other studies recommend they be used in combination to improve accuracy. The combined accu racy of these 2 modalities is 92% to 100%, 42, 44, 45 and they remain noninvasive techniques. A related imaging technique that is also described in many of these studies is intrathecal gadolinium-enhanced MRC (CE-MRC). Although this study foregoes the noninvasive benefits of standardMRC, many feel it is a safe and effective alternative for identifying CSF fistulae. These studies report a sensitivity ranging from 61% to 100%, 38, 46, 51, 55, 57–59 a specificity of 66% to 80%, 46, 55 a positive predictive value of 76%, 55 and a negative predictive value of 93%. 55 Four of these studies also evaluated the safety of this method by performing frequent neurologic checks on their pa tients following the procedure and at follow-up visits over the next several months. Although intrathecal injection of contrast agents involves risks such as behavioral or neu rological disturbance, seizure, intracerebral hemorrhage, and allergic reaction, no adverse effects were reported in any of these studies, with the exception of a headache in up to 24% of patients, which resolved with conserva tive management. 57, 58, 60, 61 The studies that compared CE MRC with standard T2-weighted MRC reported CE-MRC to be more effective at identifying leaks 38, 51, 55 ; however, given that it is an invasive study, reserving its use for more complicated cases may be more prudent. According to these studies, MRC is a legitimate op tion for CSF leak identification. One may use it in com bination with other imaging techniques or choose to add intrathecal gadolinium, although these variables do affect the cost-effectiveness and noninvasive nature of this method (Supporting material 8). 24, 38, 40, 42–47, 51–59, 62–64 Aggregate grade of evidence: C (Level 3: 21 studies); Benefit: Noninvasive (unless combined with intrathecal contrast) with good accuracy; Harm: None (unless combined with intrathecal contrast); Cost: $807.34; Benefits-harm assessment: Preponderance of benefit over harm when diagnosis or site of leak is in doubt;

Value judgments: Highly accurate test for the localization of CSF leak is important for operative planning; patients may prefer a test that does not involve radiation; nearly 3 times more expensive than HRCT; Recommendation level: Recommendation for cases where cheaper or less invasive studies have failed to diagnose or localize the site of a leak; Intervention: Using MRC for localizing CSF leaks. Computed tomography cisternography CTC involves the intrathecal injection of a contrast medium through a lumbar puncture, followed by CT scanning to identify the CSF fistula site where contrast passes extracra nially. Similar to MRC, CTC can be used for confirmation of a leak as well as localization. There were 12 studies that included CTC imaging as a diagnostic technique for CSF leaks, all of which were cross-sectional studies. According to these studies, the sensitivity of CTC ranges from 33% to 100%, 24, 34, 42, 46, 51, 54, 65, 66 the specificity is 94%, 46 and the accuracy ranges from 33% to 63%. 24, 43 In 3 studies, the efficacy of single photon emission CT (SPECT) cisternography was assessed as an alternative lo calization option. SPECT planar imaging sensitivity and ac curacy for localization was 94% and 79%, 67 respectively, and for fusion imaging sensitivity was 94% to 100% 67, 68 and accuracy was 91%. 67 When compared to MRC, CTC was noted to have an inferior sensitivity of 33% to 72% vs 67% to 93% for standard MRC and 80% for intrathecal gadolinium-enhanced MRC. 51, 54 Mostafa et al. 42 found that the combination of HRCT and MRI correlated 100% with the sensitivity of CTC, but these options were less invasive. CTC is a valid option for CSF leak localization, although the ideal indications for it and its accuracy are not entirely clear (Supporting material 9). 24, 34, 42, 43, 46, 51, 54, 65–69 Aggregate grade of evidence: C (Level 3: 12 studies); Benefit: Relatively lower accuracy than other available modalities; Harm: Invasive study–potential risk of lumbar puncture and intrathecal contrast injection; Cost: $542.76; Benefits-harm assessment: Potential harm; Value judgments: Low level of evidence; lower cost com pared to MRC; lower sensitivity and accuracy compared to other modalities and similar if not higher risks due to intrathecal injection; Recommendation level: Recommendation against; Intervention: Avoid routine use of CTC for localization of CSF leaks. Summary of intrathecal fluorescein Injecting fluorescein dye into the CSF via lumbar puncture can allow for accurate identification of a skull-base fistula site during endoscopy. The use of IF for this purpose is off label and lacks clear guidelines as to the best dosage and

International Forum of Allergy & Rhinology, Vol. 6, No. 1, January 2016

13