Journal of Neuro-Ophthalmology, June 2011, Volume 31, Issue 2

Pitfalls in Imaging

Only a devout Luddite could fail to acknowledge the enormous contribution that technology makes to medical diagnosis. One can probably date the modern era of medical technology to Roentgen's discovery of X-rays that were immediately adopted and adapted by physicians as an aid to diagnosis (1). Plain radiographs, the first fruits of Roentgen's discovery, literally provided revelations, but their value for neurological diagnosis was limited since the location and nature of the brain, the spinal cord, and nerve diseases could be made only if there were alterations in bone or abnormal calcification of soft tissues. With the introduction of contrast techniques, the ability to evaluate the brain and spinal cord was greatly increased. The current imaging techniques-CT, MRI, and ultrasound-have become the clinician's indispensable handmaiden. Imaging is invaluable for neuroophthalmic diagnosis. While the history and physical examination remain sovereign, imaging makes an enormous contribution often forming part of the matrix of information that leads to the correct diagnosis. A major advance in imaging occurred in the 1990s with the introduction of optical coherence tomography (OCT), which provides a view of the retina in cross section (2). OCT took retinal imaging from a macro- to a microlevel, and with the refinements and modifications that are certain to occur, we can look forward to a time when OCT will show the details of individual cells. Little wonder that we have become enamored with technology. However, so enamored are we that we are in danger of accepting test results uncritically. This poses potential problems for the clinician and for the clinical investigator. In this issue of the Journal of Neuro-Ophthalmology, 2 interesting and well-written publications warn of some limitations of imaging: the problem of instrument-dependent variation among OCT results in one report and the problem of ‘‘pilot error'' in neuroimaging in the other.

Pitfalls in Imaging Simmons Lessell, MD Only a devout Luddite could fail to acknowledge the enormous contribution that technology makes to medical diagnosis. One can probably date the modern era of medical technology to Roentgen's discovery of X-rays that were immediately adopted and adapted by physicians as an aid to diagnosis (1). Plain radiographs, the first fruits of Roentgen's discovery, literally provided revelations, but their value for neurological diagnosis was limited since the location and nature of the brain, the spinal cord, and nerve diseases could be made only if there were alterations in bone or abnormal calcification of soft tissues. With the introduction of contrast techniques, the ability to evaluate the brain and spinal cord was greatly increased. The current imaging techniques-CT, MRI, and ultrasound-have become the clinician's indispensable handmaiden. Imaging is invaluable for neuroophthalmic diagnosis. While the history and physical examination remain sovereign, imaging makes an enormous contribution often forming part of the matrix of information that leads to the correct diagnosis. A major advance in imaging occurred in the 1990s with the introduction of optical coherence tomography (OCT), which provides a view of the retina in cross section (2). OCT took retinal imaging from a macro- to a microlevel, and with the refinements and modifications that are certain to occur, we can look forward to a time when OCT will show the details of individual cells. Little wonder that we have become enamored with technology. However, so enamored are we that we are in danger of accepting test results uncritically. This poses potential problems for the clinician and for the clinical investigator. In this issue of the Journal of Neuro-Ophthalmology, 2 interesting and well-written publications warn of some limitations of imaging: the problem of instrument-dependent variation among OCT results in one report and the problem of ‘‘pilot error'' in neuroimaging in the other. The value of OCT in glaucoma and primarily retinal disorders is already well established, but OCT is also a valuable resource for the neuroophthalmologist and neurologist. Measurements of central macular thickness and nerve fiber layer thickness are especially useful because these layers are thinned following the death of axons in the optic nerve. After optic neuritis and even in some cases of multiple sclerosis without other evidence of optic neuritis, OCT shows depletion of the nerve fiber layer and thinning of the macula (3). In some situations, these measurements will be made serially or the clinician will wish to compare his or her result with those of OCT performed on a different instrument. OCT can also be used as a surrogate for more expensive and time-consuming tests to assess the effect of drugs on multiple sclerosis. The technology has evolved from image resolution based on time and distance (time-domain OCT [TD-OCT]) to image resolution based on spectral data using Fourier transforms (spectral-domain OCT [SD-OCT]). Several SD-OCT instruments are currently available.Watson et al (4) selected 1 TD-OCT device and 4 Fourier-domain OCT devices to compare the results of nerve fiber layer and macular thickness among this group of instruments. Their 25 study subjects (50 eyes) were patients with multiple sclerosis, optic neuritis, or both, precisely the group of greatest interest to neuroophthalmologists. Each patient's testing on the 5 instruments was completed on the same day. Watson et al (4) found poor agreement among the results from the 5 devices. One would not have expected this a priori, but another investigation has shown similar interinstrument disagreement in patients with glaucoma (5). The authors suspected that the variations could be from differences in the way the various devices acquire and analyze data. For example, while all the techniques use the internal limiting membrane as the inner boundary for measuring macular thickness, 4 different outer boundaries are used. Whatever the reason for the variation, the Neuro-Ophthalmology Unit, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts. Address correspondence to Simmons Lessell, MD, Neuro-Ophthalmology Unit, Massachusetts Eye and Ear Infirmary and Harvard Medical School, 243 Charles Street, Boston, MA 02114; E-mail: [e-mail] Lessell: J Neuro-Ophthalmol 2011; 31: 101-102 101 Editorial Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. implications are clear. First, if one wishes to serially measure retinal nerve fiber layer and macular thickness in patients with multiple sclerosis or optic neuritis, use the same device each time. Second, beware of extrapolating the measurements obtained from one device to another; they are not necessarily interchangeable. Every July, we emphasize to the new ophthalmology residents that while there is a myriad of neuroophthalmic pseudoemergencies, there are very few true neuro-ophthalmic emergencies, disorders that require immediate attention. Of these, the most important is the patient with an acute painful third nerve palsy since it may announce the presence of a life-threatening aneurysm that is enlarging and has the potential to rupture. Missing an aneurysm in that setting can result in an otherwise avoidable tragedy. For-tunately, nearly all aneurysms have the potential to be detected by noninvasive means using computerized tomo-graphic angiography or magnetic resonance angiography. The diagnosis, or more exactly the misdiagnosis, of in-tracranial aneurysms from imaging studies in patients with third nerve palsies is the subject of the report by Elmalem et al (6) at the Emory University. They review instances in which a patient presented to the proverbial ‘‘outside hos-pital'' with an isolated third nerve palsy from an aneurysm that went unrecognized despite being evident on the orig-inal scans. The palsies were each accompanied by pain or headache (V. Biousse,MD, personal communication, February 2011). There were 8 patients, all of whom had aneurysms of the posterior communicating artery. In each case, the reading radiologist missed the aneurysm when interpreting the initial scans. The authors obtained the original studies in all but 1 case, and in each of them, the aneurysm was easily recognized. One of the missed aneurysms measured 12 mm! One might be inclined to blame these failures on the equipment, the technique, or the technologist. However, since the aneurysms were demonstrated on the original scans, these can be exonerated. Failure to detect the aneurysms resulted from what I'll call ‘‘pilot error'' by the reading radiologist. What caused the pilot error in these cases? The authors identified 2 possible factors. One is the expertise of the reading radiologist. Only 2 of the initial reading radiologists were neuroradiologists, but one of them had trained in neuroradiology 20 years earlier and ‘‘had been practicing mostly as a general radiologist for many years.'' The simple fact is that nonneuroradiologists may lack the training and experience necessary to detect aneurysms on CT and MRI. The superiority of neuro-radiologists over general radiologists in detecting in-tracranial aneurysms on CT or MR images has been documented in the literature (7). There can be no doubt that as the authors remind us that ‘‘. . . the most important step in imaging remains interpretation, which is entirely dependent on the training, skills and experience of the radiologist.'' The second factor contributing to non-diagnosis was the poor quality of the information provided to the radiologist. Elmalem et al (6) were able to obtain the information provided to the radiologist in 7 of the 8 cases. Only 2 of the radiologists were advised that aneurysm was a consideration, and in one of them, the side of the sus-pected aneurysm was not specified. The other requests were vague or downright misleading. Perhaps if the physician ordering the study provided an accurate description of the clinical findings and mentioned concern for the possibility of an aneurysm, the radiologist might have been more likely to recognize the aneurysm. Unfortunately, it seems to be the rule rather than the exception that referring physicians provide the radiologist with inadequate information about what the clinician wants to learn from the radiologist. If you ask the right question, you have a better chance of getting the right answer. One can assume that there are cases of patients with acute third nerve palsies in which the radiologist failed to detect the aneurysm present on images and who later suf-fered a subarachnoid hemorrhage that might have been avoided. Elmalem et al (6) are not the first to point out the problem, but it is so important that it is worth reinforcing the message articulated by Chaudhary et al (8). ‘‘To avoid diagnostic mishaps, noninvasive studies should be reviewed by at least 1 neuroradiologist before aneurysm is rejected as the cause or before the patient undergoes CCA [catheter cerebral angiography].'' One can say of technology what has been said of fire and the x2 test: it is a wonderful servant but a cruel master (9). REFERENCES 1. Lessell S. One hundred years of X-rays. Arch Ophthalmol. 1995;113:1107. 2. Swanson EA, Izatt JA, Hee MR, Huang D, Lin CP, Schuman JS, Puliafito CA, Fujimoto JG. In vivo retinal imaging by optical coherence tomography. Opt Lett. 1993;18:1864-1866. 3. Calabresi PA, Balcer LJ, Frohman EM. Retinal pathology in multiple sclerosis. Brain. 2010;133:1575-1577. 4. Watson GM, Keltner JL, Chin EK, Harvey D, Nguyen A, Park SS. Comparison of retinal nerve fiber layer and central macular thicknessmeasurements among five different optical coherence tomography instruments in patients with multiple sclerosis and optic neuritis. J Neuroophthalmol. 2011;31:110-116. 5. Leite MT, Rao HL, Weinreb RN, Zangwill LM, Bowd W, Sample PA, Tafreshi A, Medeiros FA. Agreement among spectral domain optic coherence tomography instruments for assessing retinal nerve fiber layer thickness. Am J Ophthalmol. 2011;151:85-92. 6. Elmalem VI, Hudgins PA, Bruce BB, Newman NJ, Biousse V. Underdiagnosis of posterior communicating artery aneurysm in noninvasive brain vascular studies. J Neuroophthalmol. 2011;31:103-109. 7. White PM, Wardlaw JM, Lindsay KW, Sloss S, Patel DK, Teasdale EM. The non-invasive detection of intracranial aneurysms: are neuro-radiologists any better than other observers? Eur Radiol. 2003;13:389-396. 8. Chaudhary N, DavagnanamI,AnsariSA,PandeyA,ThompsonBG, Gemmete JJ. Imaging of intracranial aneurysms causing isolated third cranial nerve palsy. J Neuroophthalmol. 2009; 29:238-244. 9. Hill AB. The environment and disease: association or causation. Proc R Soc Med. 1965;58:295-300. 102 Lessell: J Neuro-Ophthalmol 2011; 31: 101-102 Editorial Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.