While spatial and contrast resolution has greatly advanced in HN MR imaging and extraordinary detail can be seen with cranial nerve and inner ear imaging, our greatest problem in HN imaging remains the routine neck scan. Obtaining at least two-plane imaging of the neck with T1, T2 FS and T1 post contrast FS sequences is necessary. Diffusion weighted imaging has become an essential component of neck and skull base imaging and should also be included in a HN protocol. Trying to achieve at least 6 sequences in a reasonable time frame is problematic for HN MR imaging, and trying to achieve it with patients who have difficulty with secretion management after chemoradiation is even more complex.
In the past 30 years, head and neck (HN) MR imaging has evolved enormously and our ability to image both the bulk of the neck and also the exquisite anatomical detail of the inner ear and the cranial nerves has vastly improved. While some MR imaging problems have been ameliorated with improved imaging techniques, the greatest problem with neck MR imaging remains unsolved.
Neck anatomy is not getting any simpler. While no one expects changes in MR imaging to enable this, structural complexity is one of the reasons cited by neuroradiologists as to why HN imaging is so difficult to do well. After surgery and radiation, neck structures look asymmetric, of different signal intensity and frequently swollen, to resemble as so eloquently described by Larry Ginsberg a “big mess”. Increased spatial and contrast resolution with high field strength imaging developments over the last 15 years particularly, do enable greater distinction of tissues, which is extremely helpful in complex post-surgical and post-radiation cases. Excellent spatial resolution improves detection of the subtle findings, such as perineural tumor, that can be critical to a surgical or radiation plan.
Physicists have also largely solved one of the major problems that hampered HN imaging for a long time, the problem of uniform fat saturation in the neck. Current DIXON techniques for fat suppression and STIR technique improvements have resolved many of the issues with bulk susceptibility artifact and heterogeneous fat saturation.
The good news with neck imaging is that ‘plain vanilla’ T1, T2 FS, post gad T1 FS sequences –and perhaps DWI - will allow characterization and mapping of the majority of pathologies in the skull base and neck. There are advanced techniques with DCE-MRI, BOLD-MRI and MRS which, while still largely research tools, offer potential for predicting tumor responsiveness to therapy and a better understanding of tumor pathophysiology, but for most HN imaging we don’t need these advanced techniques. There does need to be a minimum of three sequences for good HN imaging and most radiologists will also have diffusion at least through the area of greatest concern. In addition, however we want in at least two imaging planes: HN radiologists rely on axial with coronal, particularly for the evaluation of the orbits and the oral cavity. Our sinonasal tumor surgeons like the sagittal plane also. This leads us to at least 7 sequences per patient, which from a time and patient throughput perspective is less than ideal. As with all MR imaging, we want physicists to reduce scan time. This will also solve for radiologists our current most frustrating problem with MR imaging, neck motion and swallowing artifacts. Clearly this problem does not exist in the same way for brain and spine imaging. Swallowing artifacts make scans more difficult to review, are less pleasing to the eye, and allow distraction from subtle abnormalities, which may be critical to diagnostic care. Patients who have undergone HN radiation frequently have difficulty in managing oral secretions and are often less able than untreated patients to remain comfortably still in the magnet. These are also often the early post-treatment patients where it is critical to detect subclinical recurrences to allow salvage treatment.
While spatial and contrast resolution has been greatly advanced in HN MR imaging and extraordinary detail can be seen with cranial nerve and inner ear imaging, our greatest problem in HN imaging remains the routine neck scan and obtaining multiplanar, 3-sequence imaging with minimal motion. The ideal HN imaging MR imaging protocol that we want physicists to create would be a rapidly acquired (!), high spatial resolution technique with a 3D T1, 3D T2 FS and 3D T1 FS. And always add DWI to a HN protocol.
1: Dai YL, King AD. State of the art MRI in head and neck cancer. Clin Radiol. 2018 Jan;73(1):45-59. Review.
2: Widmann G, Henninger B, Kremser C, Jaschke W. MRI Sequences in Head & Neck Radiology - State of the Art. Rofo. 2017 May;189(5):413-422. Review.
3: King AD, Thoeny HC. Functional MRI for the prediction of treatment response in head and neck squamous cell carcinoma: potential and limitations. Cancer Imaging. 2016 Aug 19;16(1):23. Review.
4: Gaddikeri S, Gaddikeri RS, Tailor T, Anzai Y. Dynamic Contrast-Enhanced MR Imaging in Head and Neck Cancer: Techniques and Clinical Applications. AJNR Am J Neuroradiol. 2016 Apr;37(4):588-95. Review.
5: Dickerson E, Srinivasan A. Advanced Imaging Techniques of the Skull Base. Radiol Clin North Am. 2017 Jan;55(1):189-200. Review.