49.1 Introduction

The approach to imaging for surveillance following resection of vestibular schwannoma (VS) varies based on the experience of the center, preoperative tumor size and geometry, intraoperative findings, evolution of symptoms, and other patient-specific factors. The frequency and modality of imaging can be effectively tailored to the particular clinical scenario. Postoperative imaging is used to establish the degree of tumor resection, assess for complications, and evaluate for tumor regrowth, which may necessitate additional therapy. Surveillance to assess the fate of a tumor remnant has become increasingly important with the trend toward more conservative VS surgical management and there has been additional focus on facial nerve functional outcomes.s. Literatur With the advent of radiosurgery as an option to manage residual disease, and an increasing appreciation for the potential for tumor remnants to remain stable over time, surgeons are increasingly leaving some tumor behind to optimize postoperative function. With more conservative tumor resection comes additional need for ongoing vigilance. In this chapter, we will focus on imaging of VS following surgical resection, and will provide guidelines for its use, although there is no established paradigm and clinical practice varies widely. The topic of initial diagnostic imaging of VS is discussed separately in Chapter 8. In addition, the topic of VS surveillance following radiation treatment is discussed in Chapter 25.

49.2 Surveillance Modalities

Initial modes of detecting VS with imaging employed plain film radiography. It was common in the 1900s to obtain radiographs of the skull in the workup for brain tumors, but it was not until 1912 that dilation of the porus acusticus was noted to help localize VSs.s. Literatur Cushing used preoperative plain films but did not mention the use of postoperative surveillance imaging in his seminal work, instead focusing on clinical signs and symptoms to indicate tumor growth. Intracranial tumor localization, and the potential to survey tumors postoperatively, became possible with contrast agents using angiography and cisternography. However, it was not until 1974 that posterior fossa cisternography was modified to become an outpatient procedure.s. Literatur Commercial computed tomography (CT) first became available in 1974 which made possible axial cross-sectional visualization of bony changes of the temporal bone and in 1975 advances allowed visualization of soft-tissue intracranial structures. CT affords higher resolution of bone than magnetic resonance imaging (MRI), and excels when bony anatomic definition is helpful in surgical planning, but it is less useful in tumor surveillance. Both CT and MRI can provide useful complementing information for operative planning and intraoperative guidance. The limitations of CT in postoperative surveillance are similar to those seen in its application in the diagnosis of primary VS, where tumors are difficult to detect until they are of sufficient size to be seen with contrast administration, or they erode adjacent bone. There is also ionizing radiation with CTs, which can be detrimental, especially for use in serial exams in younger patients.s. Literatur Current MRI sequences can provide excellent resolution of soft-tissue structures and have the ability to discriminate between different soft tissues, differentiating tumor from adjacent anatomic structures. MR has now become the primary surveillance modality replacing CT after first being proposed in 1985 by DeFilipp and Buchheit.s. Literatur

49.3 MRI Protocols

Advances in MRI have revolutionized the imaging of VS for postoperative surveillance, just as it has for initial diagnosis. Contemporary MRI scanners with high-resolution sequences and gadolinium enhancement can diagnose intracranial tumors as small as 1 mm. Similar accuracy can be seen in postresection surveillance, and smallest detectable differences between serial scans can be made down to 21% of either linear difference or volume.s. Literatur However, the presence of adjacent postoperative changes related to inflammation, scarring, and reconstructive materials can make identification of residual disease more challenging and can persist for years.s. Literatur An internal auditory canal (IAC) MRI protocol is typically employed for postoperative surveillance, which includes contrast-enhanced thin (<5 mm) and often submillimeter (0.5–1 mm) slices through regions most affected by the tumor and resection (temporal bone, IAC, and cerebellopontine angle [CPA]). Additional whole head sequences can be included especially if there is any evidence of intracranial complications or concurrent lesions.s. Literatur ^,​ s. Literatur A standard protocol would include some variant of the following sequences:

1. 
   

   T1-weighted, axial and coronal, pre- and postcontrast sequences for the detection of enhancement due to the relative vascularity of VS. It is essential to include precontrast T1 sequences to differentiate inherently bright signal from enhancement, such as from fat. Gadolinium uptake can also be increased by local inflammation, which can persist for 2 years postoperatively, and some enhancement may be permanent.s. Literatur ^,​ s. Literatur ^,​ s. Literatur This is the gold standard for tumor detection and postoperative surveillance.

   
   
2. 
   

   Fat-saturated T1-weighted postgadolinium sequences reduce the inherently bright signal of fat. Comparing these images with precontrast T1 sequences can help avoid the confusion of fat with enhancing tumor. It should be noted that muscle grafts, however, are isointense or hypointense on T1 and often take up gadolinium, making them difficult to distinguish from tumor based solely on signal characteristics.

   
   
3. 
   

   Cisternography sequences with high-resolution slices (i.e., Constructive Interference in Steady State [CISS] or Fast Imaging Employing Steady-state Acquisition [FIESTA] from GE [General Electric, Milwaukee]) are heavily T2-weighted series that allow the visualization of low-signal neurovascular anatomy in comparison with surrounding bright cerebrospinal fluid (CSF). By showing exclusion of CSF, these sequences can help differentiate between a true space-occupying structure such as residual tumor and inflammation. With submillimeter slice thickness, these sequences can provide adequate resolution for making fine measurements for comparing residual tumor size in sequential studies. This may offer a viable alternative to the need for sequential gadolinium administration when following selected patients.

   
   
4. 
   

   Diffusion-weighted imaging (DWI)/Fluid-attenuated inversion recovery (FLAIR) sequences allow examination of central nervous system (CNS) changes. In particular, DWI excels at the detection of even subtle central ischemia. FLAIR sequences are sensitive in detecting edema, proteinaceous fluids, and demyelination. These sequences are more critical in perioperative imaging, especially with larger tumors, and cases in which CNS dysfunction is clinically apparent. These sequences may not be required in long-term surveillance of stable patients, but often are included in institutional protocols without additional cost, despite an increase in total scan time.

   
   
5. 
   

   T2-weighted axial and coronal sequences can help provide additional information regarding central ischemia and edema with a higher resolution than DWI but a lower resolution than cisternography.

49.4 CT Protocols

CT can be highly effective in defining the extent of bone removal as well as in assessing the integrity of the inner ear in hearing-preservation attempts. When conducting CT for VS studies, it is typical to use submillimeter slices in a temporal bone/IAC protocol that can be later combined in thicker slices for decreased noise artifact. Contrast is not necessary to assess bony structures. However, for tumor detection in patients with adequate renal function, CT with contrast can be used to identify VS tumors down to 2 cm. With the use of gas cisternography CT, intracanalicular tumors can be detected but may lead to over diagnosis, with a 22% false-positive rate when described in the 1980s.s. Literatur The role for cisternography in postresection surveillance has not been defined. CT becomes the surveillance modality of choice in patients unable to undergo MRI due to the presence of MRI-incompatible devices/prostheses or other ferromagnetic implants. This has additional significance as more VS patients are undergoing cochlear implantation (CI) or auditory brainstem implant (ABI), which may adversely affect the ability to image areas of interest with MRI (refer to Chapter 10 for further discussion).s. Literatur ^,​ s. Literatur Cone-beam CT, an increasingly popular in-office technology, can offer additional bone definition with less radiation, although these cannot be used with contrast, show little or no soft-tissue definition, and can be limited by the relatively small anatomic area covered.

49.5 Specific Considerations

Special considerations can be given for reconstructive materials, postoperative changes, metallic implants/prostheses, neurofibromatosis type 2 (NF2) patients, and reducing the risk of gadolinium exposure as reviewed later.

Tissue grafts used for CSF leak prevention have different imaging characteristics that can affect the appearance of postresection MRI, and may be a factor in selecting tissues used. Muscle grafts, for instance, can appear very similar on imaging sequences to residual tumor, including the uptake of gadolinium.s. Literatur This has resulted in a reduction of the use of muscle grafts in anatomic areas where tumor recurrence may occur.s. Literatur Adipose tissue is inherently bright on T1-weighted sequences, which can obscure adjacent enhancement with the administration of gadolinium, unless fat-saturated sequences are obtained. Often enhancement from postoperative scar and inflammation will be linear or diffuse, and can be differentiated from the nodular appearance of residual tumor.s. Literatur However, these often cannot be definitively diagnosed without the benefit of serial studies that demonstrate either stability or growth.

The use of bone-anchored auditory prostheses, metallic mesh or plates, CIs, and ABIs can all introduce artifact that can make imaging of the ipsilateral skull base challenging.s. Literatur ^,​ s. Literatur Despite this, using a combination of sequences can usually yield sufficient diagnostic accuracy in most cases (Fig. 49‑1 ). There are now auditory prostheses with MRI-conditional magnets. The presence of such an internal magnet produces a significantly larger artifact than does the electronic device alone, and the implication of this potentially obscuring tumor recurrence/growth needs to be carefully considered in device selection and positioning.

Patients with NF2 also require special consideration since they typically require more frequent surveillance with the risk of bilateral VS, more aggressive growth patterns, and the potential for new and multifocal cranial and spinal neoplasms. There are some NF2 patients with relatively indolent disease that can be followed up in a manner similar to that for sporadic VSs, while more aggressive forms of the disease require more frequent imaging to guide intervention.s. Literatur Surveillance MRI should not be limited to the temporal bone and CPA, but needs to include the entire head, and in most cases the spine as well. It is reasonable to anticipate the need for lifelong surveillance in patients with NF2.

The use of MRI cisternography sequences alone can be considered to avoid repeated contrast administration in stable patients. Gadolinium and other MRI contrast agents have been shown to deposit in the dentate nucleus and globus pallidus, even in patients with normal renal function—the clinical significance of which is not well known.s. Literatur After baseline and serial images have demonstrated stability on both enhanced and cisternography sequences, it is reasonable to consider using noncontrast MRI cisternography alone in future studies.