Many conditions resulting in false positives overlap with false-negative “missed” findings, highlighting the difficulty of calibrating our threshold: when is a dense MCA dense enough and when is the dense superior sagittal sinus too dense to be normal? Therefore, the preparation of residents to take call needs to be extensive and focused on perceptual manifestations of neuroradiology in order to have optimal impact (Funaki et al. 1997). One important maneuver during softcopy viewing on picture archiving and communication system (PACS) workstations would be to vary the window width and center level (for stroke, suggest 8 and 32, respectively, rather than the preset “brain” settings of 80 and 20) in order to improve contrast and detection of subtle signs (Lev et al. 1999; Turner and Holdworth 2011) (Fig. 10.3a). Direct measurements of the density in Hounsfield units may also be helpful, especially to differentiate small incidental falcine lipomas from pneumocephalus in traumatic injury. It is also necessary to assess the bones for abnormalities, not only in trauma but also destructive processes in suspected infection and neoplasm. Finally, CT suffers from attenuation artifacts caused by beam hardening in the posterior fossa (Fig. 10.3b), and, despite improvements in multi-detector CT (MDCT), ideal assessment of the brainstem and cerebellum often requires MRI: this and other anatomical blind spots are potential pitfalls to bear in mind (Bahrami and Yim 2009) (see below Sect. 10.4).

With invention and widespread adoption of MDCT, volume three-dimensional (3D) reconstruction of exquisite bony and vascular detail becomes commonplace after post-processing using multiplanar reconstruction (MPR) and surface rendering. High-resolution intravenous CT angiography (CTA) allows immediate, noninvasive detection of aneurysms in patients with SAH, replacing invasive and logistically intensive digital subtraction angiography (DSA) in many centers, especially during on call after hours with comparable accuracy, at the same time decreasing the thick-section reconstruction pitfalls and stair-step artifacts (Villablanca et al. 2002; Takhtani 2005). In the assessment of CTA for aneurysms in patients with SAH, ideal viewing conditions would include removing overlapping vessels and a good bone-removal algorithm that eliminates potential pitfalls such as anatomical “blind spots” in the distal internal carotid artery (ICA), posterior communicating artery, ophthalmic artery, and superior hypophyseal artery (Sakamoto et al. 2006; Hochberg et al. 2011). Review of the source dataset is essential. The lack of availability of post-processing and oblique curved reformats (especially the cavernous segment of the ICA) due to inadequate technologist staffing may explain the higher (13 % discrepancy in one study) false-negative rates of on-call residents, especially to distinguish infundibular dilatation from aneurysm (Fig. 10.4) and to detect small (<3 mm) and multiple aneurysms (Villablanca et al. 2002; Li et al. 2009; Meyer et al. 2009; Hochberg et al. 2011).

Among patients with SAH, 15–35 % will harbor multiple aneurysms (Kaminogo et al. 2003): the pitfall that these may be missed should be emphasized in resident education (Hochberg et al. 2011). Peri-aneurysmal blood and intra-aneurysmal thrombosis may also reduce conspicuity. Distal branch aneurysms farther from the circle of Willis in the pericallosal and posterior inferior cerebellar arteries may also be missed (Seo et al. 2009). False-positive findings from enhancing tumors, contamination by opacified venous structures, dilated infundibula, and vascular loops are also potential pitfalls, especially in small aneurysm detection (Takhtani 2005). In large aneurysms, the “kissing vessel” partial volume artifacts, where an adjacent vessel is not well discriminated from the aneurysm, can result in the erroneous impression that there is a connection between them (Tomandl et al. 2004). Patients with clips or coils and the beam hardening obscuring nearby and parent vessels are also prone to misinterpretation.

In CTA for steno-occlusive disease in patients with ischemia, source images are also important to avoid pitfalls of diagnosis (Fig. 10.5). Maximum intensity projection (MIP) shows the calcification and stenosis better than volume-rendered surface 3D images, which may fail to show the extent of stenosis. MPR is particularly useful in dissection, plaques, and intraluminal defects (Bash et al. 2005) (Fig. 10.6). CTA is especially valuable in vertebrobasilar ischemia, where CT and perfusion CT (PCT) are hampered by beam-hardening artifacts and limited coverage (Sylaja et al. 2008). However, the false-positive rate for CTA for occlusion should be borne in mind when heavy atheromatous calcifications are present (Walker et al. 2002). Reflux of contrast material into jugular veins during CTA injection may mimic thrombus (Takhtani 2005).

Multiple MRI pulse sequences provide different information about physical properties (such as the amount of free water, blood flow, water diffusion) in normal anatomy and pathology. Hence, synthesizing sometimes conflicting data from multiple sources is a challenge inherent in neurological MRI. The trend has been to develop pulse sequences with better and better lesion-to-normal contrast, with each new technique accentuating pathology, highlighting and hopefully at the same time decreasing ambiguity: for instance, T2-weighted images, fluid-attenuated inversion-recovery (FLAIR), and diffusion-weighted imaging (DWI) to improve contrast-to-noise ratio and sensitivity in stroke (Gonzalez et al. 1999). The neuroradiologist’s job also involves awareness of the limitations and pitfalls, suggesting confirmatory tests (sometimes while the patient in still in the MRI scanner, such as vascular or advanced techniques including perfusion or spectroscopy) and expanding or reducing the appropriate differential diagnosis. Each pulse sequence has slightly different pitfalls.

Certain anatomic sites deserve special attention because of the difficulty in analyzing them, and for neuroimaging, there should be a checklist of important review areas (similar to the familiar retrocardiac, lung apex, and other review areas for chest radiography interpretation) (Raskin 2006; Bahrami and Yim 2009). These anatomic “blind spots” of cranial CT and MRI include the cerebral sulci, dural sinuses, cavernous sinuses, orbits, clivus, Meckel cave, brainstem, skull base, and nasopharyngeal soft tissues (Bahrami and Yim 2009) (Fig. 10.1c). Perhaps the most important point to be aware is the fact that pathology outside the “brain” may be visible on cranial imaging and must be sought, with a special attention to the sellar/pituitary region, skull base, and venous sinus, which have been identified as most frequently missed. Specifically, we will focus on the cerebral sulci, dural sinuses, and bilaterally symmetrical abnormalities. Using a comprehensive checklist may be helpful to avoid blind spots and trigger the appropriate focused study of the pituitary, orbit, internal auditory canal, etc., using reduced field of view and slice thickness (Bahrami and Yim 2009).