- Email: [email protected]
Comparison of magnetic resonance angiography with conventional angiography in the detection of intracranial aneurysms in patients presenting with subarachnoid haemorrhage
Clinical Radiology (1996) 51, 330-334
Comparison of Magnetic Resonance Angiography With Conventional Angiography in the Detection of Intracranial Aneurysms in Patients Presenting With Subarachnoid Haemorrhage D. W I L C O C K , T. JASPAN*, I. H O L L A N D * , G. C H E R R Y M A N t and B. W O R T H I N G T O N
Sub-department of Academic Radiology, *Department of Radiology, University Hospital, Nottingham and f"University Department of Radiology, Leicester Royal Infirmary, Leicester, UK Thirty-nine patients admitted with proven subarachnoid haemorrhage were imaged both with 3-D time of flight (TOF) magnetic resonance angiography (MRA) and conventional angiography. As the definitive examination, catheter angiography demonstrated 37 aneurysms; ten patients had no aneurysm, the remaining 29 patients had 37 aneurysms. We found the sensitivity of 3-D TOF MRA for the detection of aneurysms to be 81% and specificity to be 100% when the reporting radiologist inspects not only the MIP reconstructions but also the MRA source data and the axial spin-echo images. The investigation is less accurate if all the available imaging data is not considered. Wilcock, D., Jaspan, T., Holland, I., Cherryman, G. & Worthington, B. (1996). Clinical Radiology 51, 330-334. Comparison of Magnetic Resonance Angiography With Conventional Angiography in the Detection of Intracranial Aneurysms in Patients Presenting With Subarachnoid Haemorrhage
Accepted for Publication 6 December 1995
M R A is now widely used in neuroradiology. It is a safe technique when compared t o the small but definite morbidity and mortality of catheter angiography [1]. In asymptomatic patients, M R A is known to be highly sensitive in the detection of intracranial aneurysms [26]. We have assessed the use of 3-D time of flight (TOF) magnetic resonance angiography (MRA) in the clinical setting of subarachnoid haemorrhage to evaluate its sensitivity and specificity. Our findings were compared with conventional angiography. P A T I E N T S AND M E T H O D S All patients presenting to University Hospital Nottingham between January 1994 and June 1994 with recent onset of sub-arachnoid haemorrhage (8-96 h post ictus) confirmed by CT or lumbar puncture and who had a relatively stable clinical condition were included in the study. Thirty-nine patients were recruited (19 males, 20 females; age range 19-65 years). All patients were investigated both by conventional angiography and MRA. The M R A examinations were performed using a 1.5 Tesla superconducting system. A transverse axial study, using a fast spin-echo T2-weighted sequence, was performed (TR = 6000ms, effective T E = 90ms, field of view = 230 mm, matrix = 256/512/oversampled, number of acquisitions = 1, acquisition time = 2 rain 45 s). The axial sequences used a slice thickness of 3 mm with an interslice gap of 1 mm. Subsequently, a 3-D T O F M R angiogram was performed with the following parameters. (TR = 43 ms, TE = 8 ms, flip angle ----20 °, field of Correspondence to: Dr D. J. Wilcock, Sub-department of Academic Radiology, University Hospital, Nottingham NG7 2UH, UK. © 1996 The Royal College of Radiologists.
view = 200mm, volume thickness = 52mm, matrix = 256/512/oversampled, number of a c q u i s i t i o n s : 1, acquisition t i m e - 11 min 47 s). Our protocol was initially to image a 52 mm thick transverse axial slab which includes the cavernous carotid arteries, circle of Willis, anterior communicating artery, and the M 1 segment of the middle cerebral arteries and the basilar trunk. Then, if this was negative, to proceed to a lower 52 mm thick transverse axial slab at the level of the origins of the posterior inferior cerebellar arteries. Later in the study, the protocol was altered to use a .72ram slab which includes all the above vessels. The M R data were analysed prospectively by one radiologist (DJW). Maximal Intensity Projection (MIP) reconstructions were performed at the time of imaging. The data was reconstructed around both a head to foot axis and a right to left axis. In each reconstruction, nine views were acquired at 20 ° intervals. The reporting radiologist recorded whether he thought that an aneurysm was present first on inspection of the MIP reconstruction alone, then with the benefit of the MIP reconstruction and the M R A source data, and finally using the additional help of the transverse axial spin-echo images. This allowed the calculation of sensitivity and specificity values for each situation. The catheter angiograms were performed employing a conventional angiographic screening room and 100mm cut film. They were reported by a consultant neuroradiologist who was unaware of the M R A findings.
RESULTS As the gold standard, catheter angiography demonstrated 37 aneurysms in the 39 patients included in the
331
DETECTION OF INTRACRANIAL ANEURYSMS Table 1 - Results of the 39 patients
Patients
Catheter angiography Number of aneurysms
Site(s) of aneurysms
1 2 3 5 5 6 7 8 9 10 l1 12 13 14 15
1 1 1 1 1 1 0 0 1 1 0 1 1 0 2
16
3
Anterior communicating artery Vertebro-basilar junction Right middle cerebral artery Anterior cerebral artery Basilar tip Anterior communicating artery Absent Absent Anterior communicating artery Anterior communicating artery Absent Pericallosal artery Anterior communicating artery Absent Basilar tip Right posterior communicating artery Basilar tip
Patients age
36 29 35 47 51 37 54 39 33 39 19 30 35 44 66 65
Right posterior communicating artery Left posterior communicating artery 17 18
1 2
Anterior communicating artery Left middle cerebral artery
19 20 21 22 23 24 25
0 1 0 1 1 1 2
Absent Right middle cerebral artery Absent Anterior communicating artery Right middle cerebral artery Anterior communicating artery Anterior communicating artery
26
1
27 28 29
1 1 2
30
3
Right posterior inferior cerebellar •artery Left middle cerebral artery Left posterior communicating artery Anterior communicating artery Carotid bifurcation Anterior communicating artery
Anterior cerebral artery
38
Absent Absent Absent
Left superior cerebellar artery Anterior communicating artery Anterior communicating artery Absent Left ophthalmic artery Basilar tip
Magnetic resonance angiography (aneurysm identified on) MRA
MRA and source data
MRAand s~in-echodata
Present Present Present Absent Present Absent Absent Absent Present Present Absent Absent Absent Absent Present Absent Present
Present Present Present Absent Present Absent Absent Absent Present Present Absent Absent Absent Absent Present Absent Present
Present Present Present Present Present Present Absent Absent Present Present Absent Present Present Absent Present Present Present
Absent Absent
68
0
Present
Present
Present
51 58 58 58 62
0 0 ÷ + + + 0
Present Present Absent Present Present
Present Present Absent Present Present
Present Present Present Present Present
0 0 51 39 40 65 60 53 69 36 39
+ + 0 0 0 0 0 + 0 +
Absent Absent Present Absent Absent Present Present Present
Absent
Absent Absent Absent Absent Absent
Absent Present Present Absent Present Present
Present Present
Absent Absent
65 72 27 61 44 57 45 64
0
Present Present
Absent Absent
+ 0 0 0 0 0 0 ÷ + 0 0
Left posterior communicating artery Right posterior communicating artery 0 0 0 1 1 1 0 1 1
0 0 0 + 0 + + 0 0 0 0 0 0 + + 0 0 0 0
0 0
Right middle cerebral artery
31 32 33 34 35 36 37 38 39
Local haematoma
Absent Absent Present Absent Absent Present Present Present
Absent
Absent Absent Absent Absent Absent
Absent Present Present Absent Present Present
Present Present
Absent Absent Present Absent Absent Present Present Present
Absent
Absent Absent Absent Absent Absent
Absent Present Present Absent Present Present
Italics indicate aneurysm not detected on magnetic resonance. 0, none; + , moderate; + +, severe.
study (Table 1). Ten patients had no aneurysm, 23 patients had one aneurysm, four patients had two aneurysms and two patients had three aneurysms. We found the sensitivity of 3-D T O F M R A for the detection of cerebral aneurysms after subarachnoid haemorrhage to be 81% and specificity to be 100% when the examining radiologist inspects not only the MIP reconstructions but also the M R A source data and the axial spin-echo images. However, without the information from the axial spin-echo images, the sensitivity of the technique falls to 65%. This can be attributed to the detection of aneurysms as flow voids on the T2 images which are not visualized on the MIP reconstructions. No false positives were encountered and the specificity was therefore 100%. A total of seven aneurysms seen on catheter angiography were missed, all were small © 1996 The Royal College of Radiologists, ClinicalRadiology, 51,330-334.
( < 5mm) and six were from patients with multiple aneurysms and not thought to be responsible for the subarachnoid haemorrhage (Table 2). The other was a single aneurysm that had bled but the images in that patient had been markedly degraded by motion artefact. A severe haematoma in the region of the aneurysm was present in three patients and a moderate haematoma was present in six further patients. M R demonstrated the aneurysm in all of these patients although in four cases, it was only visible on the spin-echo images.
DISCUSSION Cerebral aneurysms are found at autopsy in 5% of the population [1]. The most usual presentation is that of a
332
CLINICAL RADIOLOGY
(a)
(b) Fig. 1 - A 27-year-old m a n who presented with a subarachnoid haemorrhage. (a, b) M I P reconstructions demonstrate a small aneurysm arising at the bifurcation of the right middle cerebral artery (arrow).
subarachnoid haemorrhage which carries a high mortality and morbidity. Only one third of patients who present with a subarachnoid haemorrhage will survive without significant morbidity [1]. The high sensitivity of 3-D T O F M R A in the detection of asymptomatic cerebral aneurysms has been demonstrated in patients with presentations other than acute subarachnoid haemorrhage [2-5]. In these reports, M R has a high sensitivity and specificity when all components of the examination are inspected, i.e. the spin-echo images and the M R A source data as well as the MIP reconstructions. A further paper suggested that in a prospective evaluation the threshold size with an 87% sensitivity for perception of cerebral aneurysms is 5 mm [6]. The role of M R A in the evaluation of cerebral aneurysms in patients who have suffered an acute subarachnoid haemorrhage has not been established. A
non-invasive method of detecting and characterizing cerebral aneurysms without the risks of catheter angiography may be helpful in this group of acutely unwell patients. These patients present additional problems performing M R I and MRA. They often do not tolerate the close confinement and high noise level of the magnet and there is a high incidence of motion artefacts. These patients also require close monitoring. One report suggests, however, that even in these conditions M R A may have a high sensitivity to detection of aneurysms [7]. But does M R A provide the neurosurgeon with adequate morphological information about the aneurysm? Before possible surgery, the number, size and site of aneurysms needs to be known. Information about morphology of the aneurysm neck in relation to parent vessels is essential. These factors influence the decision about the therapeutic options between surgery or endovascular occlusion or conservative treatment. There is one group in the U K who are sufficiently confident in some cases with the results of M R A to proceed to surgery without a catheter angiographic study [8]. In our study, a major consideration was the time that the patient spent within the mag~net and therefore, we developed a compromise protocol to include the largest yield in a reasonable time. The patients usually spent less than 20min in the M R examination room. We performed a spin-echo sequence (2min 45 s) followed by the M R angiogram (11 min 47s). The transverse axial imaging employed a fast spin-echo T2 weighted sequence. We preferred a slice thickness of 3 mm with an interslice gap of 1 mm as this enabled optimal visualization of the aneurysm as a flow void. Due to time constraints, we did not include a T1 sequence in our routine protocol, although this might have provided useful information. We realise that our M R A acquisition protocol is not a full examination of the intracranial arterial circulation as it does not include the peri-callosal arteries. However, time constraints forced this concession to be made and the incidence of aneurysms at such distal sites is very low (less than 1%). Missing the lesion, because it is not included in the angiographic volume is a rare problem but it did occur in one patient. However, the lesion was identified as a flow void on the spin-echo images. Following inspection of the M R A reconstructions, the aneurysm could be clearly identified in many patients. However, there were several instances in which the MIP data alone was inconclusive. In these occasions, a diagnosis could usually be made from visualization of a flow void from the spin-echo images. This is reflected in the results demonstrating a much higher sensitivity if all the available data is examined. In the majority of our patients, the aneurysm was well visualized with 3-D T O F MRA. In several cases, the aneurysm was more easily identified on M R A than with conventional angiography. Aneurysms visualized by M R but not conventional angiography have been previously reported and the most probable explanation is the increased number of projections available from the MIP reconstruction [9]. In our series, there was one patient in whom an aneurysm was only perceived on the catheter angiogram after it had been recognized on MRA. The results demonstrate that M R A is a sensitive and specific test for the depiction and characterization of © 1996 The Royal College of Radiologists, Clinical Radiology, 51, 330-334.
333
DETECTION OF INTRACRANIAL ANEURYSMS
(a)
(b)
(d) Fig. 2 - A 64-year-old man who presented with a subarachnoid haemorrhage. The anterior communicating artery aneurysm can be identified as a flow void on the spin-echo images (a, arrow), and high signal on the MRA source images (b, arrow). (c,d) The anterior communicating artery aneurysm is seen as high signal on the MIP reconstruction (straight arrows). However, the small right middle cerebral artery aneurysm was only appreciated on selective MIP reconstruction (curved arrow Fig. 2d). Haemorrhage had occurred from the anterior communicating artery aneurysm.
(c) aneurysms after SAH. However, sensitivity is not as high as in some reported series of MRA evaluation of berry aneurysms in high risk but unruptured cases. The resolution of MRA is not high enough to allow planning of
interventional procedures. In the endovascular treatment of aneurysms, the ratio of the width of the fundus to the width of the neck is critically important but this was not reliably visualized by the MRA images.
Table 2 - Aneurysms not detected with magnetic resonance
1 2 3 4 5 6 7
Right posterior communicating artery Left posterior communicating artery Anterior cerebral artery Right middle cerebral artery . Left posterior communicating artery Right posterior communicating artery Left superior cerebellar artery
Was aneurysm the source o f subarachnoid haemorrhage
Single or multiple aneurysms
Is the aneurysm visible on M R A in retrospect
No No No No No No Yes
Multiple Multiple Multiple Multiple Multiple Multiple Single
No No No Yes No No No
© 1996 The Royal College of Radiologists, Clinical Radiology, 51, 330-334.
334
CLINICAL RADIOLOGY
(a)
(b)
Fig. 3 - A 36-year-old woman who presented with a subarachnoid haemorrhage. (a) T1 weighted spin-echo images demonstrated a large subacute haematoma in the left frontal region. (b) Despite this, the MR angiogram was able to show the 6ram aneurysm arising from the anterior communicating artery (arrow).
CONCLUSION 3-D T O F M R A is a sensitive and specific test for the depiction and characterization of aneurysms. In this study of 39 patients, only one aneurysm responsible for subarachnoid h a e m o r r h a g e was not detected by M R imaging. However, in this case, the images were so degraded by m o t i o n artefact that it was not possible to say with any confidence that an aneurysm was not present, and such a patient would still need catheter angiography. The specificity of the technique was 100% and we can therefore be confident that no unnecessary surgical procedures would have been p e r f o r m e d on the basis of the M R images alone. Problems contributing to p o o r or non-visualization of aneurysms included significant local h a e m a t o m a and motion artefact. REFERENCES
1 Perniicone JR. Intracranial Aneurysms. In: Potchen E, Haacke EM,
2 3 4 5 6
7 8 9
Siebert JE & Gottshalk A, eds. Magnetic Resonance Angiography: Concepts and Applications. Mosby, 1993:405 431. Sevick RJ, Tsuruda JS, Schmalbrock P. Three-dimensional time-offlight MR angiography in the evaluation of cerebral aneurysms. Journal of Computer Assisted Tomography 1990;14:874 881. Ross JS, Masaryk TJ, Modic MT et al. lntracranial aneurysms: Evaluation by MR Angiography. American Journal of Neuroradiology 1990; 11:449-456. Schuierer G, Huk WJ, Laub G. Magnetic resonance angiography of intracranial aneurysms: comparison with intra-arterial digital subtraction angiography. Neuroradiology 1992;35:50-54. Litt AW. MR Angiography of intracranial aneurysms: proceed, but with caution. American Journal of Neuroradiology 1994;15: 1615-1616. Huston J, Nichols DA, Luetmer PH et al. Blinded prospective evaluation of sensitivity of M R angiography to known intracranial aneurysms: importance of aneurysm size. American Journal of Neuroradiology 1994;15:1607 1614. Gouliamos A, Gotsis E, Vlahos L e t al. Magnetic resonance angiography compared to digital subtraction arteriography in patients with subarachnoid haemorrhage. Neuroradiology 1992;35:46 49. Sankla S, Gunawardena WJ, Keogh AJ. MRA in the Management of Aneurysmal Subarachnoid Haemorrhage. 1994, 125th Meeting of the Society of British Neurological Surgeons. Curnes JT, Shogry ME, Clark DC et al. MR angiographic demonstration of an intracranial aneurysm not seen on conventional angiography. American Journal of Neuroradiology 1993;14:971 973.
© 1996 The Royal College of Radiologists, ClinicalRadiology, 51, 330 334.
Comparison of Magnetic Resonance Angiography With Conventional Angiography in the Detection of Intracranial Aneurysms in Patients Presenting With Subarachnoid Haemorrhage D. W I L C O C K , T. JASPAN*, I. H O L L A N D * , G. C H E R R Y M A N t and B. W O R T H I N G T O N
Sub-department of Academic Radiology, *Department of Radiology, University Hospital, Nottingham and f"University Department of Radiology, Leicester Royal Infirmary, Leicester, UK Thirty-nine patients admitted with proven subarachnoid haemorrhage were imaged both with 3-D time of flight (TOF) magnetic resonance angiography (MRA) and conventional angiography. As the definitive examination, catheter angiography demonstrated 37 aneurysms; ten patients had no aneurysm, the remaining 29 patients had 37 aneurysms. We found the sensitivity of 3-D TOF MRA for the detection of aneurysms to be 81% and specificity to be 100% when the reporting radiologist inspects not only the MIP reconstructions but also the MRA source data and the axial spin-echo images. The investigation is less accurate if all the available imaging data is not considered. Wilcock, D., Jaspan, T., Holland, I., Cherryman, G. & Worthington, B. (1996). Clinical Radiology 51, 330-334. Comparison of Magnetic Resonance Angiography With Conventional Angiography in the Detection of Intracranial Aneurysms in Patients Presenting With Subarachnoid Haemorrhage
Accepted for Publication 6 December 1995
M R A is now widely used in neuroradiology. It is a safe technique when compared t o the small but definite morbidity and mortality of catheter angiography [1]. In asymptomatic patients, M R A is known to be highly sensitive in the detection of intracranial aneurysms [26]. We have assessed the use of 3-D time of flight (TOF) magnetic resonance angiography (MRA) in the clinical setting of subarachnoid haemorrhage to evaluate its sensitivity and specificity. Our findings were compared with conventional angiography. P A T I E N T S AND M E T H O D S All patients presenting to University Hospital Nottingham between January 1994 and June 1994 with recent onset of sub-arachnoid haemorrhage (8-96 h post ictus) confirmed by CT or lumbar puncture and who had a relatively stable clinical condition were included in the study. Thirty-nine patients were recruited (19 males, 20 females; age range 19-65 years). All patients were investigated both by conventional angiography and MRA. The M R A examinations were performed using a 1.5 Tesla superconducting system. A transverse axial study, using a fast spin-echo T2-weighted sequence, was performed (TR = 6000ms, effective T E = 90ms, field of view = 230 mm, matrix = 256/512/oversampled, number of acquisitions = 1, acquisition time = 2 rain 45 s). The axial sequences used a slice thickness of 3 mm with an interslice gap of 1 mm. Subsequently, a 3-D T O F M R angiogram was performed with the following parameters. (TR = 43 ms, TE = 8 ms, flip angle ----20 °, field of Correspondence to: Dr D. J. Wilcock, Sub-department of Academic Radiology, University Hospital, Nottingham NG7 2UH, UK. © 1996 The Royal College of Radiologists.
view = 200mm, volume thickness = 52mm, matrix = 256/512/oversampled, number of a c q u i s i t i o n s : 1, acquisition t i m e - 11 min 47 s). Our protocol was initially to image a 52 mm thick transverse axial slab which includes the cavernous carotid arteries, circle of Willis, anterior communicating artery, and the M 1 segment of the middle cerebral arteries and the basilar trunk. Then, if this was negative, to proceed to a lower 52 mm thick transverse axial slab at the level of the origins of the posterior inferior cerebellar arteries. Later in the study, the protocol was altered to use a .72ram slab which includes all the above vessels. The M R data were analysed prospectively by one radiologist (DJW). Maximal Intensity Projection (MIP) reconstructions were performed at the time of imaging. The data was reconstructed around both a head to foot axis and a right to left axis. In each reconstruction, nine views were acquired at 20 ° intervals. The reporting radiologist recorded whether he thought that an aneurysm was present first on inspection of the MIP reconstruction alone, then with the benefit of the MIP reconstruction and the M R A source data, and finally using the additional help of the transverse axial spin-echo images. This allowed the calculation of sensitivity and specificity values for each situation. The catheter angiograms were performed employing a conventional angiographic screening room and 100mm cut film. They were reported by a consultant neuroradiologist who was unaware of the M R A findings.
RESULTS As the gold standard, catheter angiography demonstrated 37 aneurysms in the 39 patients included in the
331
DETECTION OF INTRACRANIAL ANEURYSMS Table 1 - Results of the 39 patients
Patients
Catheter angiography Number of aneurysms
Site(s) of aneurysms
1 2 3 5 5 6 7 8 9 10 l1 12 13 14 15
1 1 1 1 1 1 0 0 1 1 0 1 1 0 2
16
3
Anterior communicating artery Vertebro-basilar junction Right middle cerebral artery Anterior cerebral artery Basilar tip Anterior communicating artery Absent Absent Anterior communicating artery Anterior communicating artery Absent Pericallosal artery Anterior communicating artery Absent Basilar tip Right posterior communicating artery Basilar tip
Patients age
36 29 35 47 51 37 54 39 33 39 19 30 35 44 66 65
Right posterior communicating artery Left posterior communicating artery 17 18
1 2
Anterior communicating artery Left middle cerebral artery
19 20 21 22 23 24 25
0 1 0 1 1 1 2
Absent Right middle cerebral artery Absent Anterior communicating artery Right middle cerebral artery Anterior communicating artery Anterior communicating artery
26
1
27 28 29
1 1 2
30
3
Right posterior inferior cerebellar •artery Left middle cerebral artery Left posterior communicating artery Anterior communicating artery Carotid bifurcation Anterior communicating artery
Anterior cerebral artery
38
Absent Absent Absent
Left superior cerebellar artery Anterior communicating artery Anterior communicating artery Absent Left ophthalmic artery Basilar tip
Magnetic resonance angiography (aneurysm identified on) MRA
MRA and source data
MRAand s~in-echodata
Present Present Present Absent Present Absent Absent Absent Present Present Absent Absent Absent Absent Present Absent Present
Present Present Present Absent Present Absent Absent Absent Present Present Absent Absent Absent Absent Present Absent Present
Present Present Present Present Present Present Absent Absent Present Present Absent Present Present Absent Present Present Present
Absent Absent
68
0
Present
Present
Present
51 58 58 58 62
0 0 ÷ + + + 0
Present Present Absent Present Present
Present Present Absent Present Present
Present Present Present Present Present
0 0 51 39 40 65 60 53 69 36 39
+ + 0 0 0 0 0 + 0 +
Absent Absent Present Absent Absent Present Present Present
Absent
Absent Absent Absent Absent Absent
Absent Present Present Absent Present Present
Present Present
Absent Absent
65 72 27 61 44 57 45 64
0
Present Present
Absent Absent
+ 0 0 0 0 0 0 ÷ + 0 0
Left posterior communicating artery Right posterior communicating artery 0 0 0 1 1 1 0 1 1
0 0 0 + 0 + + 0 0 0 0 0 0 + + 0 0 0 0
0 0
Right middle cerebral artery
31 32 33 34 35 36 37 38 39
Local haematoma
Absent Absent Present Absent Absent Present Present Present
Absent
Absent Absent Absent Absent Absent
Absent Present Present Absent Present Present
Present Present
Absent Absent Present Absent Absent Present Present Present
Absent
Absent Absent Absent Absent Absent
Absent Present Present Absent Present Present
Italics indicate aneurysm not detected on magnetic resonance. 0, none; + , moderate; + +, severe.
study (Table 1). Ten patients had no aneurysm, 23 patients had one aneurysm, four patients had two aneurysms and two patients had three aneurysms. We found the sensitivity of 3-D T O F M R A for the detection of cerebral aneurysms after subarachnoid haemorrhage to be 81% and specificity to be 100% when the examining radiologist inspects not only the MIP reconstructions but also the M R A source data and the axial spin-echo images. However, without the information from the axial spin-echo images, the sensitivity of the technique falls to 65%. This can be attributed to the detection of aneurysms as flow voids on the T2 images which are not visualized on the MIP reconstructions. No false positives were encountered and the specificity was therefore 100%. A total of seven aneurysms seen on catheter angiography were missed, all were small © 1996 The Royal College of Radiologists, ClinicalRadiology, 51,330-334.
( < 5mm) and six were from patients with multiple aneurysms and not thought to be responsible for the subarachnoid haemorrhage (Table 2). The other was a single aneurysm that had bled but the images in that patient had been markedly degraded by motion artefact. A severe haematoma in the region of the aneurysm was present in three patients and a moderate haematoma was present in six further patients. M R demonstrated the aneurysm in all of these patients although in four cases, it was only visible on the spin-echo images.
DISCUSSION Cerebral aneurysms are found at autopsy in 5% of the population [1]. The most usual presentation is that of a
332
CLINICAL RADIOLOGY
(a)
(b) Fig. 1 - A 27-year-old m a n who presented with a subarachnoid haemorrhage. (a, b) M I P reconstructions demonstrate a small aneurysm arising at the bifurcation of the right middle cerebral artery (arrow).
subarachnoid haemorrhage which carries a high mortality and morbidity. Only one third of patients who present with a subarachnoid haemorrhage will survive without significant morbidity [1]. The high sensitivity of 3-D T O F M R A in the detection of asymptomatic cerebral aneurysms has been demonstrated in patients with presentations other than acute subarachnoid haemorrhage [2-5]. In these reports, M R has a high sensitivity and specificity when all components of the examination are inspected, i.e. the spin-echo images and the M R A source data as well as the MIP reconstructions. A further paper suggested that in a prospective evaluation the threshold size with an 87% sensitivity for perception of cerebral aneurysms is 5 mm [6]. The role of M R A in the evaluation of cerebral aneurysms in patients who have suffered an acute subarachnoid haemorrhage has not been established. A
non-invasive method of detecting and characterizing cerebral aneurysms without the risks of catheter angiography may be helpful in this group of acutely unwell patients. These patients present additional problems performing M R I and MRA. They often do not tolerate the close confinement and high noise level of the magnet and there is a high incidence of motion artefacts. These patients also require close monitoring. One report suggests, however, that even in these conditions M R A may have a high sensitivity to detection of aneurysms [7]. But does M R A provide the neurosurgeon with adequate morphological information about the aneurysm? Before possible surgery, the number, size and site of aneurysms needs to be known. Information about morphology of the aneurysm neck in relation to parent vessels is essential. These factors influence the decision about the therapeutic options between surgery or endovascular occlusion or conservative treatment. There is one group in the U K who are sufficiently confident in some cases with the results of M R A to proceed to surgery without a catheter angiographic study [8]. In our study, a major consideration was the time that the patient spent within the mag~net and therefore, we developed a compromise protocol to include the largest yield in a reasonable time. The patients usually spent less than 20min in the M R examination room. We performed a spin-echo sequence (2min 45 s) followed by the M R angiogram (11 min 47s). The transverse axial imaging employed a fast spin-echo T2 weighted sequence. We preferred a slice thickness of 3 mm with an interslice gap of 1 mm as this enabled optimal visualization of the aneurysm as a flow void. Due to time constraints, we did not include a T1 sequence in our routine protocol, although this might have provided useful information. We realise that our M R A acquisition protocol is not a full examination of the intracranial arterial circulation as it does not include the peri-callosal arteries. However, time constraints forced this concession to be made and the incidence of aneurysms at such distal sites is very low (less than 1%). Missing the lesion, because it is not included in the angiographic volume is a rare problem but it did occur in one patient. However, the lesion was identified as a flow void on the spin-echo images. Following inspection of the M R A reconstructions, the aneurysm could be clearly identified in many patients. However, there were several instances in which the MIP data alone was inconclusive. In these occasions, a diagnosis could usually be made from visualization of a flow void from the spin-echo images. This is reflected in the results demonstrating a much higher sensitivity if all the available data is examined. In the majority of our patients, the aneurysm was well visualized with 3-D T O F MRA. In several cases, the aneurysm was more easily identified on M R A than with conventional angiography. Aneurysms visualized by M R but not conventional angiography have been previously reported and the most probable explanation is the increased number of projections available from the MIP reconstruction [9]. In our series, there was one patient in whom an aneurysm was only perceived on the catheter angiogram after it had been recognized on MRA. The results demonstrate that M R A is a sensitive and specific test for the depiction and characterization of © 1996 The Royal College of Radiologists, Clinical Radiology, 51, 330-334.
333
DETECTION OF INTRACRANIAL ANEURYSMS
(a)
(b)
(d) Fig. 2 - A 64-year-old man who presented with a subarachnoid haemorrhage. The anterior communicating artery aneurysm can be identified as a flow void on the spin-echo images (a, arrow), and high signal on the MRA source images (b, arrow). (c,d) The anterior communicating artery aneurysm is seen as high signal on the MIP reconstruction (straight arrows). However, the small right middle cerebral artery aneurysm was only appreciated on selective MIP reconstruction (curved arrow Fig. 2d). Haemorrhage had occurred from the anterior communicating artery aneurysm.
(c) aneurysms after SAH. However, sensitivity is not as high as in some reported series of MRA evaluation of berry aneurysms in high risk but unruptured cases. The resolution of MRA is not high enough to allow planning of
interventional procedures. In the endovascular treatment of aneurysms, the ratio of the width of the fundus to the width of the neck is critically important but this was not reliably visualized by the MRA images.
Table 2 - Aneurysms not detected with magnetic resonance
1 2 3 4 5 6 7
Right posterior communicating artery Left posterior communicating artery Anterior cerebral artery Right middle cerebral artery . Left posterior communicating artery Right posterior communicating artery Left superior cerebellar artery
Was aneurysm the source o f subarachnoid haemorrhage
Single or multiple aneurysms
Is the aneurysm visible on M R A in retrospect
No No No No No No Yes
Multiple Multiple Multiple Multiple Multiple Multiple Single
No No No Yes No No No
© 1996 The Royal College of Radiologists, Clinical Radiology, 51, 330-334.
334
CLINICAL RADIOLOGY
(a)
(b)
Fig. 3 - A 36-year-old woman who presented with a subarachnoid haemorrhage. (a) T1 weighted spin-echo images demonstrated a large subacute haematoma in the left frontal region. (b) Despite this, the MR angiogram was able to show the 6ram aneurysm arising from the anterior communicating artery (arrow).
CONCLUSION 3-D T O F M R A is a sensitive and specific test for the depiction and characterization of aneurysms. In this study of 39 patients, only one aneurysm responsible for subarachnoid h a e m o r r h a g e was not detected by M R imaging. However, in this case, the images were so degraded by m o t i o n artefact that it was not possible to say with any confidence that an aneurysm was not present, and such a patient would still need catheter angiography. The specificity of the technique was 100% and we can therefore be confident that no unnecessary surgical procedures would have been p e r f o r m e d on the basis of the M R images alone. Problems contributing to p o o r or non-visualization of aneurysms included significant local h a e m a t o m a and motion artefact. REFERENCES
1 Perniicone JR. Intracranial Aneurysms. In: Potchen E, Haacke EM,
2 3 4 5 6
7 8 9
Siebert JE & Gottshalk A, eds. Magnetic Resonance Angiography: Concepts and Applications. Mosby, 1993:405 431. Sevick RJ, Tsuruda JS, Schmalbrock P. Three-dimensional time-offlight MR angiography in the evaluation of cerebral aneurysms. Journal of Computer Assisted Tomography 1990;14:874 881. Ross JS, Masaryk TJ, Modic MT et al. lntracranial aneurysms: Evaluation by MR Angiography. American Journal of Neuroradiology 1990; 11:449-456. Schuierer G, Huk WJ, Laub G. Magnetic resonance angiography of intracranial aneurysms: comparison with intra-arterial digital subtraction angiography. Neuroradiology 1992;35:50-54. Litt AW. MR Angiography of intracranial aneurysms: proceed, but with caution. American Journal of Neuroradiology 1994;15: 1615-1616. Huston J, Nichols DA, Luetmer PH et al. Blinded prospective evaluation of sensitivity of M R angiography to known intracranial aneurysms: importance of aneurysm size. American Journal of Neuroradiology 1994;15:1607 1614. Gouliamos A, Gotsis E, Vlahos L e t al. Magnetic resonance angiography compared to digital subtraction arteriography in patients with subarachnoid haemorrhage. Neuroradiology 1992;35:46 49. Sankla S, Gunawardena WJ, Keogh AJ. MRA in the Management of Aneurysmal Subarachnoid Haemorrhage. 1994, 125th Meeting of the Society of British Neurological Surgeons. Curnes JT, Shogry ME, Clark DC et al. MR angiographic demonstration of an intracranial aneurysm not seen on conventional angiography. American Journal of Neuroradiology 1993;14:971 973.
© 1996 The Royal College of Radiologists, ClinicalRadiology, 51, 330 334.