Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Ming-Hua Li, MD, PhD Yong-Dong Li, MD, PhD Bin-Xian Gu, MD Ying-Sheng Cheng, MD, PhD Wu Wang, MD Hua-Qiao Tan, MD, PhD Yuan-Chang Chen, MD, PhD
Purpose:
To evaluate the diagnostic accuracy of three-dimensional (3D) time-of-flight (TOF) magnetic resonance (MR) angiography at 3.0 T in the detection of small cerebral aneurysms.
Materials and Methods:
The institutional review board approved the study protocol, and patients or qualifying family members provided informed consent. A total of 403 consecutive patients undergoing 3D TOF MR angiography and digital subtraction angiography (DSA) were prospectively enrolled. Small aneurysms were those 5 mm in diameter or smaller. DSA served as the reference standard. Three observers were blinded to clinical and DSA results, and they independently analyzed all 3D TOF MR angiographic data sets. Interobserver agreement was expressed in terms of Cohen k value for categorical variables. Accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 3D TOF MR angiography in the detection of cerebral aneurysms were determined by using patient-, aneurysm-, and location-based evaluations.
Results:
Of 403 patients, 273 aneurysms were detected with DSA in 230 patients. Patient-based evaluation with 3D TOF MR angiography at 3.0 T yielded an accuracy of 96%– 97%, a sensitivity of 98.2%–98.7%, a specificity of 93.2% –94.8%, a PPV of 94.9%–96.2%, and an NPV of 97.6%– 98.2% in the detection of cerebral aneurysms. Aneurysmbased evaluation yielded an accuracy of 96.4%–97.3%, a sensitivity of 98.5%–98.9%, a specificity of 93.2%–94.9%, a PPV of 95.7%–96.8%, and an NPV of 97.6%–98.2%. Aneurysm-location evaluations yielded similar results.
Conclusion:
Three-dimensional TOF MR angiography is a noninvasive method that shows promising diagnostic accuracy in the detection of small cerebral aneurysms. RSNA, 2014
q
1
From the Institute of Diagnostic and Interventional Radiology, The Sixth Affiliated People’s Hospital, Shanghai Jiao Tong University, 600 Yi Shan Rd, Shanghai 200233, China. Received December 15, 2012; revision requested January 22, 2013; revision received June 17; accepted July 25; final version accepted October 8. Supported by the National Natural Scientific Fund of China (contract no. 81201199 and 81171440). Address correspondence to M.H.L. (e-mail: [email protected]). RSNA, 2014
q
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
553
Original Research n Neuroradiology
Accurate Diagnosis of Small Cerebral Aneurysms 5 mm in Diameter with 3.0-T MR Angiography1
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
B
ecause of rapid improvements in imaging techniques, increasing numbers of cerebral aneurysms are now detected with computed tomographic (CT) angiography and magnetic resonance (MR) angiography (1–8). The reported accuracy of MR angiography in the detection of cerebral aneurysms is approximately 90%, with better results for larger aneurysms and poorer results for smaller aneurysms (1,9,10). Small cerebral aneurysms are generally classified as those with a diameter of 5 mm or less (11), and they present particular technical challenges for the neuroradiologist. Early MR angiography studies have shown limited diagnostic accuracy in the detection of small cerebral aneurysms (12,13). With advanced modern MR imaging systems, increasing observer experience, and the introduction and widespread adoption of improved postprocessing techniques, several single-center studies
Advances in Knowledge nn Three-dimensional time-of-flight (TOF) MR angiography has a high accuracy (range over reader: 96%–97%), sensitivity (98.2%–98.7%), specificity (93.2%–94.8%), positive predictive value (PPV) (94.9%–96.2%), and negative predictive value (NPV) (97.6%–98.2%) in the detection of small cerebral aneurysms in patient-based evaluations. nn Three-dimensional TOF MR angiography has a high accuracy (range over reader: 96.4%– 97.3%), sensitivity (98.5%– 98.9%), specificity (93.2%– 94.9%), PPV (95.7%–96.8%), and NPV (97.6%–98.2%) in the detection of small cerebral aneurysms in aneurysm-based evaluations. nn Three-dimensional TOF MR angiography with specialized postprocessing techniques, including volume rendering and a singleartery highlighting approach, may improve the detection of small cerebral aneurysms. 554
Li et al
have reported high rates of accuracy and sensitivity in the diagnosis of small cerebral aneurysms (1,2,9,14–16). In an attempt to improve our understanding of efficacy profiles associated with the diagnosis of small cerebral aneurysms, we report our experience with three-dimensional (3D) time-of-flight (TOF) MR angiography at 3.0 T in the evaluation of the diagnostic accuracy of small (5 mm) cerebral aneurysms, with digital subtraction angiography (DSA) as the reference standard.
Materials and Methods Patients The institutional review board approved the study protocol, and patients or qualifying family members provided informed consent before participation. From June 2007 to February 2012, patients with subarachnoid hemorrhage (SAH) or other cerebrovascular disease underwent both MR angiography and DSA at our institution. DSA was regarded as the reference standard. The inclusion criteria for this prospective study were as follows: (a) patients who had SAH with a Glasgow coma scale of 15, (b) patients who were suspected of having other cerebrovascular disease, and (c) patients whose cerebral aneurysm was smaller than or equal to 5 mm in maximal diameter
at DSA. The exclusion criteria were as follows: (a) patients who had SAH with a Glasgow coma scale of less than 15 (n = 58), (b) patients who had undergone DSA prior to MR angiography (n = 6), (c) patients who had not undergone DSA due to an allergy to contrast material or renal dysfunction that made them unable to tolerate the contrast material load associated with DSA (n = 5), and (d) patients whose cerebral aneurysms were larger than 5 mm in maximal diameter at DSA (n = 110). We considered a small cerebral aneurysm to be a saccular protrusion from the side wall or bifurcation of the cerebral arteries, without the artery emerging at its top. We defined a negative case as one in which a patient had no cerebral aneurysm regardless of the presence of other cerebral vascular diseases. A total of 513 consecutive patients (183 patients with SAH, 330 patients suspected of having other cerebrovascular diseases) underwent both 3D TOF MR angiography and DSA. DSA revealed 273 small cerebral aneurysms in 230 of 513 patients. No cerebral aneurysm was seen in 173 patients. A total of 110 patients with cerebral aneurysms larger than 5 mm in maximal diameter Published online before print 10.1148/radiol.14122770 Content code: Radiology 2014; 271:553–560
Implications for Patient Care nn Three-dimensional TOF MR angiography offers a high degree of diagnostic accuracy in the detection of small cerebral aneurysms. nn Three-dimensional TOF MR angiography can safely replace invasive digital subtraction angiography in the diagnosis of small cerebral aneurysms. nn Because three-dimensional TOF MR angiography is noninvasive and does not require contrast material administration or radiation, it could become the preferred modality in the screening of unruptured cerebral aneurysms.
Abbreviations: DSA = digital subtraction angiography ICA = internal carotid artery NPV = negative predictive value PPV = positive predictive value SAH = subarachnoid hemorrhage 3D = three-dimensional TOF = time of flight VR = volume rendering Author contributions: Guarantors of integrity of entire study, M.H.L., Y.D.L., B.X.G., Y.S.C., W.W.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, M.H.L., Y.D.L., W.W., H.Q.T.; clinical studies, all authors; statistical analysis, M.H.L., Y.D.L., W.W.; and manuscript editing, M.H.L., W.W., H.Q.T. Conflicts of interest are listed at the end of this article. radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
were excluded. Thus, 230 patients with small aneurysms and 173 patients with no aneurysm (ie, negative control subjects), were included in the final analysis. This study group comprised 213 men and 190 women, with a mean age of 54.30 years 6 12.73 (standard deviation) (age range, 15–83 years).
Image Acquisition MR angiography.—All MR angiographic examinations were performed with a 3.0-T system (Achieva X Series; Philips Medical Systems, Best, the Netherlands) with a Sense-Head-8 receiver head coil. The 3D TOF MR angiograms were obtained by using 3D T1-weighted fast field-echo sequences (repetition time msec/echo time msec, 35/7; flip angle, 20°; field of view, 250 3 190 3 108; four slabs [180 sections]; section thickness, 0.8 mm; matrix, 732 3 1024; acquisition time, 8 minutes 56 seconds). The acquired image data sets were transferred to an extended workstation (EWS; Philips Medical Systems) and then were processed to a volume-rendering (VR) image at 3D volume inspection (Philips Medical Systems). The maximum intensity projection image was also reconstructed for the evaluation. To reduce arterial overlay and to identify cerebral aneurysms effectively, we used the single-artery highlighting method, which is also referred to as catheter cerebral angiography (15,17,18). For the left or right internal carotid artery (ICA), we removed the right or left ICA system, respectively, together with the posterior circulation system. For the posterior circulation system, we removed the anterior circulation system. We analyzed three vessels in each patient from six basic views (anterior-posterior projection, posterior-anterior projection, bilateral projections, and bioblique projections) and from arbitrary angles to clearly depict the aneurysm origins and courses. DSA.—In patients with SAH, DSA was performed as soon as possible after MR angiography, whereas in patients suspected of having either an unruptured cerebral aneurysm
or another cerebrovascular disease, DSA was performed within 2 weeks after MR angiography. A catheter was placed in four vessels, including the bilateral ICA and the bilateral vertebral artery, for DSA (including posterior-anterior and lateral projections) with a monoplanar unit (Axiom Arits VB22N; Siemens Healthcare, Forchheim, Germany) with a 1024 3 1024 matrix and a 17–20-cm field of view in all patients. Rotational angiography was performed with an 8-second 200° rotational run, acquiring 200 images; further 3D images with VR reconstruction were produced for affected arteries on a workstation with a 128 3 128 3 128 to 512 3 512 3 512 matrix (Syngo XWP VA70B; Siemens Healthcare). Contrast medium was injected for a total of 10 mL for the ICA (rate, 4–5 mL/sec), 7 mL for the vertebral artery (rate, 2–3 mL/ sec), and 16–20 mL per artery for rotational angiography (rate, 3–4 mL/sec). Two experienced observers (B.X.G., W.W.; 15 and 8 years of experience, respectively, with interventional neuroradiology) evaluated the cerebral aneurysms together. Image review.—Three experienced observers (observers A, B, and C [M.H.L., Y.D.L., H.Q.T., respectively]; 17, 8, and 7 years of experience with interventional neuroradiology, respectively) who had previously tested common standard interpretation techniques were blinded to all clinical and DSA results. They analyzed all 3D TOF MR angiography with VR image data sets independently at an offline workstation from multiple on-screen viewing angles by using the single-artery highlighting approach. The source images and maximum intensity projection images were presented on screen, allowing for adjustment of the appropriate threshold of the window width and window level to diagnose or differentiate small aneurysms with infundibula. Cases with one or more aneurysms detected were considered positive; all others were considered negative. Aneurysm locations were classified as the ICA, including the posterior communicating artery; the anterior cerebral artery, including the anterior
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
Li et al
communicating artery; the middle cerebral artery, including the M1–2 segment; and the vertebral and basilar arteries, including the vertebral, basilar, cerebellar, and posterior cerebral arteries. Aneurysm size was recorded as the maximum two-dimensional angiographic dimension (,3 mm or 3–5 mm).
Statistical Analysis The categorical demographic and basic characteristic variables are summarized as numbers and percentages, while continuous variables are summarized as means 6 standard deviations. The results from the individual readers were analyzed separately. Interobserver agreement was expressed in terms of the Cohen k value for categorical variables. Descriptive statistical analyses were performed on three levels: patient-by-patient (no or any cerebral aneurysm per patient), aneurysm-by-aneurysm, and locationby-location. The diagnostic performance parameters of 3D TOF MR angiography at 3.0 T in the diagnosis of cerebral aneurysms relative to DSA as the reference standard is summarized in terms of overall accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating characteristic curve. Results DSA images revealed 273 aneurysms in 230 patients (101 men, 129 women; mean age, 56.10 years 6 10.41). One patient had four aneurysms, five patients had three aneurysms, 30 patients had two aneurysms, and 194 patients had one aneurysm. Overall, 161 aneurysms were located at the ICA (Fig 1), 59 were located at the ACA (Fig 2), 34 were located at the middle cerebral artery (Fig 3), and 19 were located at the vertebral and basilar arteries. The mean maximal diameter of the aneurysm sac was 3.17 mm 6 0.97. A total of 142 aneurysms were smaller than 3 mm in maximum diameter, and 131 were 3–5 mm in maximum diameter. Because of the long acquisition time and patient intolerance to the 555
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Figure 1
Figure 1: An unruptured aneurysm in the junction of the C5-6 segment of the left ICA in a 57-year-old female patient. A, This 3D TOF MR angiogram obtained with MR imaging reveals a small aneurysm (arrow) at the junction of the C5-6 segment of the ICA. B, Rotational DSA image shows an aneurysm (arrow) at the junction of the C5-6 segment of the ICA. C, DSA image with VR reconstruction shows an aneurysm (arrow) as seen on a 3D TOF MR angiogram with VR reconstruction.
Figure 2
Figure 2: Ruptured aneurysm on the anterior communicating artery in a 33-year-old male patient with SAH and a Glasgow coma scale value of 15 before MR angiography. A, This 3D TOF MR angiogram with VR reconstruction shows a small aneurysm (arrow) on the anterior communicating artery. B, DSA image also shows a small aneurysm (arrow) on the anterior communicating artery. C, DSA image shows total occlusion of the aneurysm (arrow) after embolization with coils.
procedure, motion artifacts occurred in seven patients, which initially led to uninterpretable image quality. All of these patients were asked to cooperate with technicians to undergo a second examination, which subsequently improved the image quality so that the case was eligible for diagnosis. We found 3D TOF MR angiography with maximum intensity projection imaging 556
alone enabled us to confirm the diagnosis in only 227 of 276 small cerebral aneurysms and that the remaining 49 lesions were not clearly delineated on the maximum intensity projection images. This was more frequently encountered at the ICA segment, with torturous courses due to overlapping. The final evaluation for small cerebral aneurysms in this study was made according to the
findings of 3D TOF MR angiography with VR images by using views from arbitrary angles and a single-artery highlighting approach. The diagnostic performance of 3D TOF MR angiography at 3.0 T in the detection of small cerebral aneurysms compared with DSA in patient-, aneurysm-, and location-based evaluation is detailed in Tables 1 and 2. When
radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Figure 3
Figure 3: Ruptured aneurysm on the M1-2 segment of the right middle cerebral artery in a 47-year-old female patient with SAH and a Glasgow coma scale value of 15 before MR angiography. A, This 3D TOF MR angiogram with VR reconstruction shows a small aneurysm (arrow) at the M1-2 segment of the right middle cerebral artery. B, DSA image also shows a small aneurysm (arrow) at the M1-2 segment of the right cerebral artery. C, DSA image shows total occlusion of the aneurysm (arrow) after embolization with coils.
Table 1 Diagnostic Performance of 3D TOF MR Angiography in Patient- and Aneurysm-based Evaluations Evaluation and Observer
AUC (%)*
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
Accuracy (%)
97.5 (95.6, 99.4) 97.2 (95.2, 99.2) 96.4 (94.2, 98.6)
98.7 (226/229) 98.7 (225/228) 98.2 (223/227)
94.8 (165/174) 94.3 (165/175) 93.2 (164/176)
96.2 (226/235) 95.8 (225/235) 94.9 (223/235)
98.2 (165/168) 98.2 (165/168) 97.6 (164/168)
97.0 (391/403) 96.8 (390/403) 96.0 (387/403)
97.5 (95.7, 99.4) 97.2 (95.3, 99.2) 96.5 (94.3, 98.7)
98.9 (270/273) 98.9 (269/272) 98.5 (267/271)
94.9 (166/175) 94.3 (166/176) 93.2 (165/177)
96.8 (270/279) 96.4 (269/279) 95.7 (267/279)
98.2 (166/169) 98.2 (166/169) 97.6 (165/169)
97.3 (435/447) 97.1 (434/447) 96.4 (431/447)
†
Patient-based evaluation Observer A Observer B Observer C Aneurysm-based evaluation‡ Observer A Observer B Observer C
Note.—Unless otherwise indicated, data in parentheses are raw data. AUC = area under the receiver operating characteristic curve. * Data in parentheses are 95% confidence intervals. †
The k values range from 0.98 to 0.995 and are coefficients of interreader agreement.
‡
The k values range from 0.981 to 0.995 and are coefficients of interreader agreement.
compared with DSA, observer A diagnosed nine false-positive aneurysms with 3D TOF MR angiography, observer B diagnosed 10, and observer C diagnosed 12. Observers A and B missed three aneurysms with 3D TOF MR angiography, and observer C missed four (Table 3).
Discussion We used patient-, aneurysm-, and location-based evaluations and compared the results with DSA findings; our
results show high accuracy and sensitivity of more than 95% when 3D TOF MR angiography at 3.0 T is used to detect small cerebral aneurysms. These results indicate that the current technology for 3D TOF MR angiography can safely replace intraarterial DSA in the diagnostic work-up of patients with small cerebral aneurysms. To our knowledge, this study represents the largest cohort of patients to date in which 3D TOF MR angiography at 3.0 T has been investigated in the detection of small cerebral aneurysms.
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
Previous studies have suggested that TOF MR angiography in the diagnosis of cerebral aneurysms larger than 5 mm is highly sensitive (up to 95%); however, for lesions 5 mm in diameter or smaller, sensitivity has been shown to be limited (10,12,13). Aprile (19) reported that the sensitivity of MR angiography in the detection of cerebral aneurysms smaller than 3 mm (25%) was much lower than that for cerebral aneurysms larger than 3 mm (92%). More recently, Hiratsuka et al (9) reported that mean sensitivity, specificity, and 557
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Table 2 Diagnostic Performance of 3D TOF MR Angiography in Location-based Evaluations Location and Observer
AUC (%)*
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
Accuracy (%)
97.3 (95.3, 99.3) 96.4 (94.1, 98.7) 95.6 (93.0, 98.1)
98.8 (160/162) 98.8 (159/161) 98.1 (159/162)
95.9 (164/171) 95.4 (164/172) 95.4 (163/171)
95.8 (160/167) 95.2 (159/167) 95.2 (159/167)
98.8 (164/166) 98.8 (164/166) 98.2 (163/166)
97.3 (324/333) 97.0 (323/333) 96.7 (322/333)
99.1 (97.9, 100) 98.2 (96.6, 99.9) 97.7 (95.8, 99.6)
98.3 (58/59) 98.3 (58/59) 98.3 (57/58)
100 (164/164) 100 (164/164) 99.4 (164/165)
100 (58/58) 100 (58/58) 98.3 (57/58)
99.4 (164/165) 99.4 (164/165) 99.4 (164/165)
99.6 (222/223) 99.6 (222/223) 99.1 (221/223)
98.6 (95.4, 102) 97.9 (94.4, 101) 97.8 (94.2, 101)
100 (34/34) 100 (34/34) 100 (33/33)
98.8 (165/167) 98.8 (165/167) 98.2 (165/168)
94.4 (34/36) 94.4 (34/36) 91.7 (33/36)
100 (165/165) 100 (165/165) 100 (165/165)
99.0 (199/201) 99.0 (199/201) 98.5 (198/201)
100 (100, 100) 100 (100, 100) 100 (100, 100)
100 (19/19) 100 (19/19) 100 (19/19)
100 (165/165) 100 (165/165) 100 (165/165)
100 (19/19) 100 (19/19) 100 (19/19)
100 (165/165) 100 (165/165) 100 (165/165)
100 (184/184) 100 (184/184) 100 (184/184)
†
ICA Observer A Observer B Observer C ACA‡ Observer A Observer B Observer C MCA§ Observer A Observer B Observer C VBA|| Observer A Observer B Observer C
Note.—Unless otherwise indicated, data in parentheses are raw data. ACA = anterior cerebral artery, AUC = area under the receiver operating characteristic curve, MCA = middle cerebral artery, VBA = vertebral and basilar arteries. * Data in parentheses are 95% confidence intervals. †
The k values range from 0.988 to 0.994 and are coefficients of interreader agreement.
‡
The k values range from 0.988 to 1.0 and are coefficients of interreader agreement.
§
The k values range from 0.983 to 1.0 and are coefficients of interreader agreement.
||
The k value is 1.0 for all, and all are coefficients of interreader agreement.
accuracy of 3D TOF MR angiography on a per-aneurysm basis were 89%, 76%, and 87%, respectively, for both independent readers in the study. The peraneurysm sensitivity of all readers was greater for depicting aneurysms larger than 3 mm than for depicting aneurysms 3 mm or smaller (92% vs 67%). An analysis of the methods of previous investigators with poor reported detection rates for cerebral aneurysms reveals the lack of not only adequately advanced MR imaging systems but also the lack of optimal MR angiography imaging parameters and use of VR images (20). The reliance on 0.5- or 1.5-T MR imaging systems alone is prone to error, as aneurysms can demonstrate a confusing variability in signal intensities on MR images, including high signal intensity on T1-weighted images due to flow-related enhancement, signal void due to turbulence and circumferential calcification, and miscellaneous signal intensities produced by blood clots in different states (17,18). 558
In addition, small aneurysms are often misinterpreted on maximum intensity projection images, and false-positive diagnoses are occasionally unavoidable in these imaging studies as a result of loop formation, vessel overlap, atherosclerotic plaques, turbulent flow, or a combination thereof (15,20). Finally, infundibula can mimic aneurysms at MR angiography if the vessel emerging at the infundibula apex is not seen (21). Application of the advanced 3.0 T-MR imaging systems with parallel imaging and a multichannel phasedarray head coil, which has an increased signal-to-noise ratio and improved background suppression, facilitates enhanced delineation of vessels and visualization of a small cerebral aneurysm, as well as the relative position between the aneurysm and the adjacent vessels (1,9,15,21–25). On the basis of our experience, use of VR imaging combined with the single-artery highlighting method and views from arbitrary angles may improve accuracy in the
detection of small cerebral aneurysms. In addition, a trained and experienced observer was also an important factor that contributed to diagnostic accuracy. White et al (26) showed that the sensitivity and accuracy attained by experienced observers were much higher than those achieved by less experienced observers. The published data indicate the site of the aneurysm was a major factor that influenced its detection (9,10,13). Common locations at which to detect false-positive aneurysms are the origins of the posterior communicating artery, anterior choroidal artery, ophthalmic artery, and other small vessel branches from the ICA. An infundibulum at the previously mentioned site can mimic a small aneurysm because the origin of the vessels is very tortuous and overlapping or because the vessel emerging at the infundibulum apex may not be clearly visualized (9,13). In this study, there were no significant differences in accuracy, sensitivity, or
radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Table 3 Details of All False-Positive and False-Negative Results with 3D TOF MR Angiography Finding and Lesion No. False-positive 1 2 3 4 5 6 7 8 9 10 11 12 False-negative 13 14 15 16
Age (y)
Sex
Location
Size (mm)
Reason
Observer
59 58 54 67 63 45 64 48 54 73 60 68
M F F F M F M M F F F M
Right C4 Right PCA Right PCA Left PCA Right ACA Right M1 Right C4 Right C4 Right PCA Left M1–2 Left C6 Right M1–2
3.6 2.6 2.3 2.7 2.3 2.4 2.5 2.6 2.2 2.2 2.7 2.1
Acute turn of the siphon at C4 Infundibula Infundibula Excessive tortuosity at the origin of the PCA Infundibula Excessive tortuosity at the origin of the lenticulostriate arteries Infundibula Infundibula Excessive tortuosity at the origin of the PCA Excessive tortuosity at the origin of M1-2 Infundibula Excessively tortuosi at the origin of M1-2
A, B, and C A, B, and C A, B, and C A, B, and C C A, B, and C A, B, and C A, B, and C A, B, and C A, B, and C B and C C
64 42 65 49
F M M F
Right C6 Left ACA Right C5 Right C7
2.0 3.5 2.8 2.4
Infundibula Artifact Infundibula Infundibula
A, B, and C A, B, and C A, B, and C C
Note.—ACA = anterior cerebral artery, PCA = posterior communicating artery.
specificity between the four locations (middle cerebral artery, ICA, anterior cerebral artery, and vertebral and basilar arteries), which should be considered the direct result of the technical advances in image acquisition and postprocessing algorithms mentioned previously. There were limitations to this study. First, this was a single-center study. Second, false-positive results were seen in both small infundibula, with no demonstration of the artery emerging at its top and a tortuous origin of the vessel branches due to the limitation of spatial resolution at 3D TOF MR angiography. Third, MR angiographic examinations require a longer examination time to obtain precise images, and this is not usually tolerated well by patients with SAH and impaired Glasgow coma scale values (,15). In this situation, CT angiography may be a better diagnostic method because of the shorter imaging time involved. Finally, MR angiographic examinations were sometimes incomplete as a result of the presence of metal implants or pacemakers. In conclusion, 3D TOF MR angiography shows high diagnostic accuracy
in the detection of small cerebral aneurysms, and this accuracy appears to be similar to that obtained with DSA. We recommend that 3D TOF MR angiography would be a safe replacement for invasive DSA in the diagnostic workup of patients with small cerebral aneurysms due to its noninvasive nature and high degree of accuracy and sensitivity. Disclosures of Conflicts of Interest: M.H.L. No relevant conflicts of interest to disclose. Y.D.L. No relevant conflicts of interest to disclose. B.X.G. No relevant conflicts of interest to disclose. Y.S.C. No relevant conflicts of interest to disclose. W.W. No relevant conflicts of interest to disclose. H.Q.T. No relevant conflicts of interest to disclose. Y.C.C. No relevant conflicts of interest to disclose.
References 1. Igase K, Matsubara I, Igase M, Miyazaki H, Sadamoto K. Initial experience in evaluating the prevalence of unruptured intracranial aneurysms detected on 3-tesla MRI. Cerebrovasc Dis 2012;33(4):348–353. 2. Inoue T, Shimizu H, Fujimura M, Saito A, Tominaga T. Annual rupture risk of growing unruptured cerebral aneurysms detected by magnetic resonance angiography. J Neurosurg 2012;117(1):20–25.
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
3. UCAS Japan Investigators, Morita A, Kirino T, et al. The natural course of unruptured cerebral aneurysms in a Japanese cohort. N Engl J Med 2012;366(26):2474–2482. 4. Lu L, Zhang LJ, Poon CS, et al. Digital subtraction CT angiography for detection of intracranial aneurysms: comparison with three-dimensional digital subtraction angiography. Radiology 2012;262(2):605–612. 5. Luo Z, Wang D, Sun X, et al. Comparison of the accuracy of subtraction CT angiography performed on 320-detector row volume CT with conventional CT angiography for diagnosis of intracranial aneurysms. Eur J Radiol 2012;81(1):118–122. 6. Westerlaan HE, van Dijk JM, Jansen-van der Weide MC, et al. Intracranial aneurysms in patients with subarachnoid hemorrhage: CT angiography as a primary examination tool for diagnosis—systematic review and metaanalysis. Radiology 2011;258(1):134–145. 7. Menke J, Larsen J, Kallenberg K. Diagnosing cerebral aneurysms by computed tomographic angiography: meta-analysis. Ann Neurol 2011;69(4):646–654. 8. Donmez H, Serifov E, Kahriman G, Mavili E, Durak AC, Menkü A. Comparison of 16row multislice CT angiography with conventional angiography for detection and evaluation of intracranial aneurysms. Eur J Radiol 2011;80(2):455–461.
559
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
9. Hiratsuka Y, Miki H, Kiriyama I, et al. Diagnosis of unruptured intracranial aneurysms: 3T MR angiography versus 64-channel multi-detector row CT angiography. Magn Reson Med Sci 2008;7(4):169–178. 10. Korogi Y, Takahashi M, Mabuchi N, et al. Intracranial aneurysms: diagnostic accuracy of MR angiography with evaluation of maximum intensity projection and source images. Radiology 1996;199(1):199–207. 11. Chen W, Wang J, Xin W, Peng Y, Xu Q. Accuracy of 16-row multislice computed tomographic angiography for assessment of small cerebral aneurysms. Neurosurgery 2008; 62(1):113–121; discussion 121–122. 12. Okahara M, Kiyosue H, Yamashita M, et al. Diagnostic accuracy of magnetic resonance angiography for cerebral aneurysms in correlation with 3D-digital subtraction angiographic images: a study of 133 aneurysms. Stroke 2002;33(7):1803–1808. 13. White PM, Teasdale EM, Wardlaw JM, Easton V. Intracranial aneurysms: CT angiography and MR angiography for detection prospective blinded comparison in a large patient cohort. Radiology 2001;219(3):739–749. 14. Tang PH, Hui F, Sitoh YY. Intracranial aneurysm detection with 3T magnetic resonance angiography. Ann Acad Med Singapore 2007; 36(6):388–393.
560
Li et al
15. Li MH, Li YD, Tan HQ, et al. Contrast-free MRA at 3.0 T for the detection of intracranial aneurysms. Neurology 2011;77(7):667– 676. 16. Numminen J, Tarkiainen A, Niemelä M, Porras M, Hernesniemi J, Kangasniemi M. Detection of unruptured cerebral artery aneurysms by MRA at 3.0 tesla: comparison with multislice helical computed tomographic angiography. Acta Radiol 2011; 52(6):670–674. 17. Li MH, Cheng YS, Li YD, et al. Large-cohort comparison between three-dimensional timeof-flight magnetic resonance and rotational digital subtraction angiographies in intracranial aneurysm detection. Stroke 2009;40(9): 3127–3129. 18. Li H, Yan L, Li MH, et al. Evaluation of intracranial aneurysms with high-resolution MR angiography using single-artery highlighting technique: correlation with digital subtraction angiography. Radiol Med (Torino) 2013;18(8):1379–1387. 19. Aprile I. Evaluation of cerebral aneurysms with MR-angiography. Rev Neuroradiol 1996;9:541–550. 20. Schwab KE, Gailloud P, Wyse G, Tamargo RJ. Limitations of magnetic resonance imaging and magnetic resonance angiography in the diagnosis of intracranial aneurysms.
Neurosurgery 2008;63(1):29–34; discussion 34–35. 21. Lu H, Nagae-Poetscher LM, Golay X, Lin D, Pomper M, van Zijl PC. Routine clinical brain MRI sequences for use at 3.0 Tesla. J Magn Reson Imaging 2005;22(1):13–22. 22. Hiratsuka Y, Miki H, Kikuchi K, et al. Sensitivity of an eight-element phased array coil in 3 Tesla MR imaging: a basic analysis. Magn Reson Med Sci 2007;6(3):177–181. 23. Willinek WA, Born M, Simon B, et al. Timeof-flight MR angiography: comparison of 3.0-T imaging and 1.5-T imaging—initial experience. Radiology 2003;229(3):913–920. 24. Fushimi Y, Miki Y, Kikuta K, et al. Com parison of 3.0- and 1.5-T three-dimensional time-of-flight MR angiography in moyamoya disease: preliminary experience. Radiology 2006;239(1):232–237. 25. de Zwart JA, Ledden PJ, van Gelderen P, Bodurka J, Chu R, Duyn JH. Signal-to-noise ratio and parallel imaging performance of a 16-channel receive-only brain coil array at 3.0 Tesla. Magn Reson Med 2004;51(1):22–26. 26. White PM, Wardlaw JM, Lindsay KW, Sloss S, Patel DK, Teasdale EM. The non-invasive detection of intracranial aneurysms: are neuroradiologists any better than other observers? Eur Radiol 2003;13(2):389–396.
radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
Ming-Hua Li, MD, PhD Yong-Dong Li, MD, PhD Bin-Xian Gu, MD Ying-Sheng Cheng, MD, PhD Wu Wang, MD Hua-Qiao Tan, MD, PhD Yuan-Chang Chen, MD, PhD
Purpose:
To evaluate the diagnostic accuracy of three-dimensional (3D) time-of-flight (TOF) magnetic resonance (MR) angiography at 3.0 T in the detection of small cerebral aneurysms.
Materials and Methods:
The institutional review board approved the study protocol, and patients or qualifying family members provided informed consent. A total of 403 consecutive patients undergoing 3D TOF MR angiography and digital subtraction angiography (DSA) were prospectively enrolled. Small aneurysms were those 5 mm in diameter or smaller. DSA served as the reference standard. Three observers were blinded to clinical and DSA results, and they independently analyzed all 3D TOF MR angiographic data sets. Interobserver agreement was expressed in terms of Cohen k value for categorical variables. Accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 3D TOF MR angiography in the detection of cerebral aneurysms were determined by using patient-, aneurysm-, and location-based evaluations.
Results:
Of 403 patients, 273 aneurysms were detected with DSA in 230 patients. Patient-based evaluation with 3D TOF MR angiography at 3.0 T yielded an accuracy of 96%– 97%, a sensitivity of 98.2%–98.7%, a specificity of 93.2% –94.8%, a PPV of 94.9%–96.2%, and an NPV of 97.6%– 98.2% in the detection of cerebral aneurysms. Aneurysmbased evaluation yielded an accuracy of 96.4%–97.3%, a sensitivity of 98.5%–98.9%, a specificity of 93.2%–94.9%, a PPV of 95.7%–96.8%, and an NPV of 97.6%–98.2%. Aneurysm-location evaluations yielded similar results.
Conclusion:
Three-dimensional TOF MR angiography is a noninvasive method that shows promising diagnostic accuracy in the detection of small cerebral aneurysms. RSNA, 2014
q
1
From the Institute of Diagnostic and Interventional Radiology, The Sixth Affiliated People’s Hospital, Shanghai Jiao Tong University, 600 Yi Shan Rd, Shanghai 200233, China. Received December 15, 2012; revision requested January 22, 2013; revision received June 17; accepted July 25; final version accepted October 8. Supported by the National Natural Scientific Fund of China (contract no. 81201199 and 81171440). Address correspondence to M.H.L. (e-mail: [email protected]). RSNA, 2014
q
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
553
Original Research n Neuroradiology
Accurate Diagnosis of Small Cerebral Aneurysms 5 mm in Diameter with 3.0-T MR Angiography1
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
B
ecause of rapid improvements in imaging techniques, increasing numbers of cerebral aneurysms are now detected with computed tomographic (CT) angiography and magnetic resonance (MR) angiography (1–8). The reported accuracy of MR angiography in the detection of cerebral aneurysms is approximately 90%, with better results for larger aneurysms and poorer results for smaller aneurysms (1,9,10). Small cerebral aneurysms are generally classified as those with a diameter of 5 mm or less (11), and they present particular technical challenges for the neuroradiologist. Early MR angiography studies have shown limited diagnostic accuracy in the detection of small cerebral aneurysms (12,13). With advanced modern MR imaging systems, increasing observer experience, and the introduction and widespread adoption of improved postprocessing techniques, several single-center studies
Advances in Knowledge nn Three-dimensional time-of-flight (TOF) MR angiography has a high accuracy (range over reader: 96%–97%), sensitivity (98.2%–98.7%), specificity (93.2%–94.8%), positive predictive value (PPV) (94.9%–96.2%), and negative predictive value (NPV) (97.6%–98.2%) in the detection of small cerebral aneurysms in patient-based evaluations. nn Three-dimensional TOF MR angiography has a high accuracy (range over reader: 96.4%– 97.3%), sensitivity (98.5%– 98.9%), specificity (93.2%– 94.9%), PPV (95.7%–96.8%), and NPV (97.6%–98.2%) in the detection of small cerebral aneurysms in aneurysm-based evaluations. nn Three-dimensional TOF MR angiography with specialized postprocessing techniques, including volume rendering and a singleartery highlighting approach, may improve the detection of small cerebral aneurysms. 554
Li et al
have reported high rates of accuracy and sensitivity in the diagnosis of small cerebral aneurysms (1,2,9,14–16). In an attempt to improve our understanding of efficacy profiles associated with the diagnosis of small cerebral aneurysms, we report our experience with three-dimensional (3D) time-of-flight (TOF) MR angiography at 3.0 T in the evaluation of the diagnostic accuracy of small (5 mm) cerebral aneurysms, with digital subtraction angiography (DSA) as the reference standard.
Materials and Methods Patients The institutional review board approved the study protocol, and patients or qualifying family members provided informed consent before participation. From June 2007 to February 2012, patients with subarachnoid hemorrhage (SAH) or other cerebrovascular disease underwent both MR angiography and DSA at our institution. DSA was regarded as the reference standard. The inclusion criteria for this prospective study were as follows: (a) patients who had SAH with a Glasgow coma scale of 15, (b) patients who were suspected of having other cerebrovascular disease, and (c) patients whose cerebral aneurysm was smaller than or equal to 5 mm in maximal diameter
at DSA. The exclusion criteria were as follows: (a) patients who had SAH with a Glasgow coma scale of less than 15 (n = 58), (b) patients who had undergone DSA prior to MR angiography (n = 6), (c) patients who had not undergone DSA due to an allergy to contrast material or renal dysfunction that made them unable to tolerate the contrast material load associated with DSA (n = 5), and (d) patients whose cerebral aneurysms were larger than 5 mm in maximal diameter at DSA (n = 110). We considered a small cerebral aneurysm to be a saccular protrusion from the side wall or bifurcation of the cerebral arteries, without the artery emerging at its top. We defined a negative case as one in which a patient had no cerebral aneurysm regardless of the presence of other cerebral vascular diseases. A total of 513 consecutive patients (183 patients with SAH, 330 patients suspected of having other cerebrovascular diseases) underwent both 3D TOF MR angiography and DSA. DSA revealed 273 small cerebral aneurysms in 230 of 513 patients. No cerebral aneurysm was seen in 173 patients. A total of 110 patients with cerebral aneurysms larger than 5 mm in maximal diameter Published online before print 10.1148/radiol.14122770 Content code: Radiology 2014; 271:553–560
Implications for Patient Care nn Three-dimensional TOF MR angiography offers a high degree of diagnostic accuracy in the detection of small cerebral aneurysms. nn Three-dimensional TOF MR angiography can safely replace invasive digital subtraction angiography in the diagnosis of small cerebral aneurysms. nn Because three-dimensional TOF MR angiography is noninvasive and does not require contrast material administration or radiation, it could become the preferred modality in the screening of unruptured cerebral aneurysms.
Abbreviations: DSA = digital subtraction angiography ICA = internal carotid artery NPV = negative predictive value PPV = positive predictive value SAH = subarachnoid hemorrhage 3D = three-dimensional TOF = time of flight VR = volume rendering Author contributions: Guarantors of integrity of entire study, M.H.L., Y.D.L., B.X.G., Y.S.C., W.W.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, M.H.L., Y.D.L., W.W., H.Q.T.; clinical studies, all authors; statistical analysis, M.H.L., Y.D.L., W.W.; and manuscript editing, M.H.L., W.W., H.Q.T. Conflicts of interest are listed at the end of this article. radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
were excluded. Thus, 230 patients with small aneurysms and 173 patients with no aneurysm (ie, negative control subjects), were included in the final analysis. This study group comprised 213 men and 190 women, with a mean age of 54.30 years 6 12.73 (standard deviation) (age range, 15–83 years).
Image Acquisition MR angiography.—All MR angiographic examinations were performed with a 3.0-T system (Achieva X Series; Philips Medical Systems, Best, the Netherlands) with a Sense-Head-8 receiver head coil. The 3D TOF MR angiograms were obtained by using 3D T1-weighted fast field-echo sequences (repetition time msec/echo time msec, 35/7; flip angle, 20°; field of view, 250 3 190 3 108; four slabs [180 sections]; section thickness, 0.8 mm; matrix, 732 3 1024; acquisition time, 8 minutes 56 seconds). The acquired image data sets were transferred to an extended workstation (EWS; Philips Medical Systems) and then were processed to a volume-rendering (VR) image at 3D volume inspection (Philips Medical Systems). The maximum intensity projection image was also reconstructed for the evaluation. To reduce arterial overlay and to identify cerebral aneurysms effectively, we used the single-artery highlighting method, which is also referred to as catheter cerebral angiography (15,17,18). For the left or right internal carotid artery (ICA), we removed the right or left ICA system, respectively, together with the posterior circulation system. For the posterior circulation system, we removed the anterior circulation system. We analyzed three vessels in each patient from six basic views (anterior-posterior projection, posterior-anterior projection, bilateral projections, and bioblique projections) and from arbitrary angles to clearly depict the aneurysm origins and courses. DSA.—In patients with SAH, DSA was performed as soon as possible after MR angiography, whereas in patients suspected of having either an unruptured cerebral aneurysm
or another cerebrovascular disease, DSA was performed within 2 weeks after MR angiography. A catheter was placed in four vessels, including the bilateral ICA and the bilateral vertebral artery, for DSA (including posterior-anterior and lateral projections) with a monoplanar unit (Axiom Arits VB22N; Siemens Healthcare, Forchheim, Germany) with a 1024 3 1024 matrix and a 17–20-cm field of view in all patients. Rotational angiography was performed with an 8-second 200° rotational run, acquiring 200 images; further 3D images with VR reconstruction were produced for affected arteries on a workstation with a 128 3 128 3 128 to 512 3 512 3 512 matrix (Syngo XWP VA70B; Siemens Healthcare). Contrast medium was injected for a total of 10 mL for the ICA (rate, 4–5 mL/sec), 7 mL for the vertebral artery (rate, 2–3 mL/ sec), and 16–20 mL per artery for rotational angiography (rate, 3–4 mL/sec). Two experienced observers (B.X.G., W.W.; 15 and 8 years of experience, respectively, with interventional neuroradiology) evaluated the cerebral aneurysms together. Image review.—Three experienced observers (observers A, B, and C [M.H.L., Y.D.L., H.Q.T., respectively]; 17, 8, and 7 years of experience with interventional neuroradiology, respectively) who had previously tested common standard interpretation techniques were blinded to all clinical and DSA results. They analyzed all 3D TOF MR angiography with VR image data sets independently at an offline workstation from multiple on-screen viewing angles by using the single-artery highlighting approach. The source images and maximum intensity projection images were presented on screen, allowing for adjustment of the appropriate threshold of the window width and window level to diagnose or differentiate small aneurysms with infundibula. Cases with one or more aneurysms detected were considered positive; all others were considered negative. Aneurysm locations were classified as the ICA, including the posterior communicating artery; the anterior cerebral artery, including the anterior
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
Li et al
communicating artery; the middle cerebral artery, including the M1–2 segment; and the vertebral and basilar arteries, including the vertebral, basilar, cerebellar, and posterior cerebral arteries. Aneurysm size was recorded as the maximum two-dimensional angiographic dimension (,3 mm or 3–5 mm).
Statistical Analysis The categorical demographic and basic characteristic variables are summarized as numbers and percentages, while continuous variables are summarized as means 6 standard deviations. The results from the individual readers were analyzed separately. Interobserver agreement was expressed in terms of the Cohen k value for categorical variables. Descriptive statistical analyses were performed on three levels: patient-by-patient (no or any cerebral aneurysm per patient), aneurysm-by-aneurysm, and locationby-location. The diagnostic performance parameters of 3D TOF MR angiography at 3.0 T in the diagnosis of cerebral aneurysms relative to DSA as the reference standard is summarized in terms of overall accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating characteristic curve. Results DSA images revealed 273 aneurysms in 230 patients (101 men, 129 women; mean age, 56.10 years 6 10.41). One patient had four aneurysms, five patients had three aneurysms, 30 patients had two aneurysms, and 194 patients had one aneurysm. Overall, 161 aneurysms were located at the ICA (Fig 1), 59 were located at the ACA (Fig 2), 34 were located at the middle cerebral artery (Fig 3), and 19 were located at the vertebral and basilar arteries. The mean maximal diameter of the aneurysm sac was 3.17 mm 6 0.97. A total of 142 aneurysms were smaller than 3 mm in maximum diameter, and 131 were 3–5 mm in maximum diameter. Because of the long acquisition time and patient intolerance to the 555
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Figure 1
Figure 1: An unruptured aneurysm in the junction of the C5-6 segment of the left ICA in a 57-year-old female patient. A, This 3D TOF MR angiogram obtained with MR imaging reveals a small aneurysm (arrow) at the junction of the C5-6 segment of the ICA. B, Rotational DSA image shows an aneurysm (arrow) at the junction of the C5-6 segment of the ICA. C, DSA image with VR reconstruction shows an aneurysm (arrow) as seen on a 3D TOF MR angiogram with VR reconstruction.
Figure 2
Figure 2: Ruptured aneurysm on the anterior communicating artery in a 33-year-old male patient with SAH and a Glasgow coma scale value of 15 before MR angiography. A, This 3D TOF MR angiogram with VR reconstruction shows a small aneurysm (arrow) on the anterior communicating artery. B, DSA image also shows a small aneurysm (arrow) on the anterior communicating artery. C, DSA image shows total occlusion of the aneurysm (arrow) after embolization with coils.
procedure, motion artifacts occurred in seven patients, which initially led to uninterpretable image quality. All of these patients were asked to cooperate with technicians to undergo a second examination, which subsequently improved the image quality so that the case was eligible for diagnosis. We found 3D TOF MR angiography with maximum intensity projection imaging 556
alone enabled us to confirm the diagnosis in only 227 of 276 small cerebral aneurysms and that the remaining 49 lesions were not clearly delineated on the maximum intensity projection images. This was more frequently encountered at the ICA segment, with torturous courses due to overlapping. The final evaluation for small cerebral aneurysms in this study was made according to the
findings of 3D TOF MR angiography with VR images by using views from arbitrary angles and a single-artery highlighting approach. The diagnostic performance of 3D TOF MR angiography at 3.0 T in the detection of small cerebral aneurysms compared with DSA in patient-, aneurysm-, and location-based evaluation is detailed in Tables 1 and 2. When
radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Figure 3
Figure 3: Ruptured aneurysm on the M1-2 segment of the right middle cerebral artery in a 47-year-old female patient with SAH and a Glasgow coma scale value of 15 before MR angiography. A, This 3D TOF MR angiogram with VR reconstruction shows a small aneurysm (arrow) at the M1-2 segment of the right middle cerebral artery. B, DSA image also shows a small aneurysm (arrow) at the M1-2 segment of the right cerebral artery. C, DSA image shows total occlusion of the aneurysm (arrow) after embolization with coils.
Table 1 Diagnostic Performance of 3D TOF MR Angiography in Patient- and Aneurysm-based Evaluations Evaluation and Observer
AUC (%)*
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
Accuracy (%)
97.5 (95.6, 99.4) 97.2 (95.2, 99.2) 96.4 (94.2, 98.6)
98.7 (226/229) 98.7 (225/228) 98.2 (223/227)
94.8 (165/174) 94.3 (165/175) 93.2 (164/176)
96.2 (226/235) 95.8 (225/235) 94.9 (223/235)
98.2 (165/168) 98.2 (165/168) 97.6 (164/168)
97.0 (391/403) 96.8 (390/403) 96.0 (387/403)
97.5 (95.7, 99.4) 97.2 (95.3, 99.2) 96.5 (94.3, 98.7)
98.9 (270/273) 98.9 (269/272) 98.5 (267/271)
94.9 (166/175) 94.3 (166/176) 93.2 (165/177)
96.8 (270/279) 96.4 (269/279) 95.7 (267/279)
98.2 (166/169) 98.2 (166/169) 97.6 (165/169)
97.3 (435/447) 97.1 (434/447) 96.4 (431/447)
†
Patient-based evaluation Observer A Observer B Observer C Aneurysm-based evaluation‡ Observer A Observer B Observer C
Note.—Unless otherwise indicated, data in parentheses are raw data. AUC = area under the receiver operating characteristic curve. * Data in parentheses are 95% confidence intervals. †
The k values range from 0.98 to 0.995 and are coefficients of interreader agreement.
‡
The k values range from 0.981 to 0.995 and are coefficients of interreader agreement.
compared with DSA, observer A diagnosed nine false-positive aneurysms with 3D TOF MR angiography, observer B diagnosed 10, and observer C diagnosed 12. Observers A and B missed three aneurysms with 3D TOF MR angiography, and observer C missed four (Table 3).
Discussion We used patient-, aneurysm-, and location-based evaluations and compared the results with DSA findings; our
results show high accuracy and sensitivity of more than 95% when 3D TOF MR angiography at 3.0 T is used to detect small cerebral aneurysms. These results indicate that the current technology for 3D TOF MR angiography can safely replace intraarterial DSA in the diagnostic work-up of patients with small cerebral aneurysms. To our knowledge, this study represents the largest cohort of patients to date in which 3D TOF MR angiography at 3.0 T has been investigated in the detection of small cerebral aneurysms.
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
Previous studies have suggested that TOF MR angiography in the diagnosis of cerebral aneurysms larger than 5 mm is highly sensitive (up to 95%); however, for lesions 5 mm in diameter or smaller, sensitivity has been shown to be limited (10,12,13). Aprile (19) reported that the sensitivity of MR angiography in the detection of cerebral aneurysms smaller than 3 mm (25%) was much lower than that for cerebral aneurysms larger than 3 mm (92%). More recently, Hiratsuka et al (9) reported that mean sensitivity, specificity, and 557
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Table 2 Diagnostic Performance of 3D TOF MR Angiography in Location-based Evaluations Location and Observer
AUC (%)*
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
Accuracy (%)
97.3 (95.3, 99.3) 96.4 (94.1, 98.7) 95.6 (93.0, 98.1)
98.8 (160/162) 98.8 (159/161) 98.1 (159/162)
95.9 (164/171) 95.4 (164/172) 95.4 (163/171)
95.8 (160/167) 95.2 (159/167) 95.2 (159/167)
98.8 (164/166) 98.8 (164/166) 98.2 (163/166)
97.3 (324/333) 97.0 (323/333) 96.7 (322/333)
99.1 (97.9, 100) 98.2 (96.6, 99.9) 97.7 (95.8, 99.6)
98.3 (58/59) 98.3 (58/59) 98.3 (57/58)
100 (164/164) 100 (164/164) 99.4 (164/165)
100 (58/58) 100 (58/58) 98.3 (57/58)
99.4 (164/165) 99.4 (164/165) 99.4 (164/165)
99.6 (222/223) 99.6 (222/223) 99.1 (221/223)
98.6 (95.4, 102) 97.9 (94.4, 101) 97.8 (94.2, 101)
100 (34/34) 100 (34/34) 100 (33/33)
98.8 (165/167) 98.8 (165/167) 98.2 (165/168)
94.4 (34/36) 94.4 (34/36) 91.7 (33/36)
100 (165/165) 100 (165/165) 100 (165/165)
99.0 (199/201) 99.0 (199/201) 98.5 (198/201)
100 (100, 100) 100 (100, 100) 100 (100, 100)
100 (19/19) 100 (19/19) 100 (19/19)
100 (165/165) 100 (165/165) 100 (165/165)
100 (19/19) 100 (19/19) 100 (19/19)
100 (165/165) 100 (165/165) 100 (165/165)
100 (184/184) 100 (184/184) 100 (184/184)
†
ICA Observer A Observer B Observer C ACA‡ Observer A Observer B Observer C MCA§ Observer A Observer B Observer C VBA|| Observer A Observer B Observer C
Note.—Unless otherwise indicated, data in parentheses are raw data. ACA = anterior cerebral artery, AUC = area under the receiver operating characteristic curve, MCA = middle cerebral artery, VBA = vertebral and basilar arteries. * Data in parentheses are 95% confidence intervals. †
The k values range from 0.988 to 0.994 and are coefficients of interreader agreement.
‡
The k values range from 0.988 to 1.0 and are coefficients of interreader agreement.
§
The k values range from 0.983 to 1.0 and are coefficients of interreader agreement.
||
The k value is 1.0 for all, and all are coefficients of interreader agreement.
accuracy of 3D TOF MR angiography on a per-aneurysm basis were 89%, 76%, and 87%, respectively, for both independent readers in the study. The peraneurysm sensitivity of all readers was greater for depicting aneurysms larger than 3 mm than for depicting aneurysms 3 mm or smaller (92% vs 67%). An analysis of the methods of previous investigators with poor reported detection rates for cerebral aneurysms reveals the lack of not only adequately advanced MR imaging systems but also the lack of optimal MR angiography imaging parameters and use of VR images (20). The reliance on 0.5- or 1.5-T MR imaging systems alone is prone to error, as aneurysms can demonstrate a confusing variability in signal intensities on MR images, including high signal intensity on T1-weighted images due to flow-related enhancement, signal void due to turbulence and circumferential calcification, and miscellaneous signal intensities produced by blood clots in different states (17,18). 558
In addition, small aneurysms are often misinterpreted on maximum intensity projection images, and false-positive diagnoses are occasionally unavoidable in these imaging studies as a result of loop formation, vessel overlap, atherosclerotic plaques, turbulent flow, or a combination thereof (15,20). Finally, infundibula can mimic aneurysms at MR angiography if the vessel emerging at the infundibula apex is not seen (21). Application of the advanced 3.0 T-MR imaging systems with parallel imaging and a multichannel phasedarray head coil, which has an increased signal-to-noise ratio and improved background suppression, facilitates enhanced delineation of vessels and visualization of a small cerebral aneurysm, as well as the relative position between the aneurysm and the adjacent vessels (1,9,15,21–25). On the basis of our experience, use of VR imaging combined with the single-artery highlighting method and views from arbitrary angles may improve accuracy in the
detection of small cerebral aneurysms. In addition, a trained and experienced observer was also an important factor that contributed to diagnostic accuracy. White et al (26) showed that the sensitivity and accuracy attained by experienced observers were much higher than those achieved by less experienced observers. The published data indicate the site of the aneurysm was a major factor that influenced its detection (9,10,13). Common locations at which to detect false-positive aneurysms are the origins of the posterior communicating artery, anterior choroidal artery, ophthalmic artery, and other small vessel branches from the ICA. An infundibulum at the previously mentioned site can mimic a small aneurysm because the origin of the vessels is very tortuous and overlapping or because the vessel emerging at the infundibulum apex may not be clearly visualized (9,13). In this study, there were no significant differences in accuracy, sensitivity, or
radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
Li et al
Table 3 Details of All False-Positive and False-Negative Results with 3D TOF MR Angiography Finding and Lesion No. False-positive 1 2 3 4 5 6 7 8 9 10 11 12 False-negative 13 14 15 16
Age (y)
Sex
Location
Size (mm)
Reason
Observer
59 58 54 67 63 45 64 48 54 73 60 68
M F F F M F M M F F F M
Right C4 Right PCA Right PCA Left PCA Right ACA Right M1 Right C4 Right C4 Right PCA Left M1–2 Left C6 Right M1–2
3.6 2.6 2.3 2.7 2.3 2.4 2.5 2.6 2.2 2.2 2.7 2.1
Acute turn of the siphon at C4 Infundibula Infundibula Excessive tortuosity at the origin of the PCA Infundibula Excessive tortuosity at the origin of the lenticulostriate arteries Infundibula Infundibula Excessive tortuosity at the origin of the PCA Excessive tortuosity at the origin of M1-2 Infundibula Excessively tortuosi at the origin of M1-2
A, B, and C A, B, and C A, B, and C A, B, and C C A, B, and C A, B, and C A, B, and C A, B, and C A, B, and C B and C C
64 42 65 49
F M M F
Right C6 Left ACA Right C5 Right C7
2.0 3.5 2.8 2.4
Infundibula Artifact Infundibula Infundibula
A, B, and C A, B, and C A, B, and C C
Note.—ACA = anterior cerebral artery, PCA = posterior communicating artery.
specificity between the four locations (middle cerebral artery, ICA, anterior cerebral artery, and vertebral and basilar arteries), which should be considered the direct result of the technical advances in image acquisition and postprocessing algorithms mentioned previously. There were limitations to this study. First, this was a single-center study. Second, false-positive results were seen in both small infundibula, with no demonstration of the artery emerging at its top and a tortuous origin of the vessel branches due to the limitation of spatial resolution at 3D TOF MR angiography. Third, MR angiographic examinations require a longer examination time to obtain precise images, and this is not usually tolerated well by patients with SAH and impaired Glasgow coma scale values (,15). In this situation, CT angiography may be a better diagnostic method because of the shorter imaging time involved. Finally, MR angiographic examinations were sometimes incomplete as a result of the presence of metal implants or pacemakers. In conclusion, 3D TOF MR angiography shows high diagnostic accuracy
in the detection of small cerebral aneurysms, and this accuracy appears to be similar to that obtained with DSA. We recommend that 3D TOF MR angiography would be a safe replacement for invasive DSA in the diagnostic workup of patients with small cerebral aneurysms due to its noninvasive nature and high degree of accuracy and sensitivity. Disclosures of Conflicts of Interest: M.H.L. No relevant conflicts of interest to disclose. Y.D.L. No relevant conflicts of interest to disclose. B.X.G. No relevant conflicts of interest to disclose. Y.S.C. No relevant conflicts of interest to disclose. W.W. No relevant conflicts of interest to disclose. H.Q.T. No relevant conflicts of interest to disclose. Y.C.C. No relevant conflicts of interest to disclose.
References 1. Igase K, Matsubara I, Igase M, Miyazaki H, Sadamoto K. Initial experience in evaluating the prevalence of unruptured intracranial aneurysms detected on 3-tesla MRI. Cerebrovasc Dis 2012;33(4):348–353. 2. Inoue T, Shimizu H, Fujimura M, Saito A, Tominaga T. Annual rupture risk of growing unruptured cerebral aneurysms detected by magnetic resonance angiography. J Neurosurg 2012;117(1):20–25.
Radiology: Volume 271: Number 2—May 2014 n radiology.rsna.org
3. UCAS Japan Investigators, Morita A, Kirino T, et al. The natural course of unruptured cerebral aneurysms in a Japanese cohort. N Engl J Med 2012;366(26):2474–2482. 4. Lu L, Zhang LJ, Poon CS, et al. Digital subtraction CT angiography for detection of intracranial aneurysms: comparison with three-dimensional digital subtraction angiography. Radiology 2012;262(2):605–612. 5. Luo Z, Wang D, Sun X, et al. Comparison of the accuracy of subtraction CT angiography performed on 320-detector row volume CT with conventional CT angiography for diagnosis of intracranial aneurysms. Eur J Radiol 2012;81(1):118–122. 6. Westerlaan HE, van Dijk JM, Jansen-van der Weide MC, et al. Intracranial aneurysms in patients with subarachnoid hemorrhage: CT angiography as a primary examination tool for diagnosis—systematic review and metaanalysis. Radiology 2011;258(1):134–145. 7. Menke J, Larsen J, Kallenberg K. Diagnosing cerebral aneurysms by computed tomographic angiography: meta-analysis. Ann Neurol 2011;69(4):646–654. 8. Donmez H, Serifov E, Kahriman G, Mavili E, Durak AC, Menkü A. Comparison of 16row multislice CT angiography with conventional angiography for detection and evaluation of intracranial aneurysms. Eur J Radiol 2011;80(2):455–461.
559
NEURORADIOLOGY: Accurate Diagnosis of Small Cerebral Aneurysms
9. Hiratsuka Y, Miki H, Kiriyama I, et al. Diagnosis of unruptured intracranial aneurysms: 3T MR angiography versus 64-channel multi-detector row CT angiography. Magn Reson Med Sci 2008;7(4):169–178. 10. Korogi Y, Takahashi M, Mabuchi N, et al. Intracranial aneurysms: diagnostic accuracy of MR angiography with evaluation of maximum intensity projection and source images. Radiology 1996;199(1):199–207. 11. Chen W, Wang J, Xin W, Peng Y, Xu Q. Accuracy of 16-row multislice computed tomographic angiography for assessment of small cerebral aneurysms. Neurosurgery 2008; 62(1):113–121; discussion 121–122. 12. Okahara M, Kiyosue H, Yamashita M, et al. Diagnostic accuracy of magnetic resonance angiography for cerebral aneurysms in correlation with 3D-digital subtraction angiographic images: a study of 133 aneurysms. Stroke 2002;33(7):1803–1808. 13. White PM, Teasdale EM, Wardlaw JM, Easton V. Intracranial aneurysms: CT angiography and MR angiography for detection prospective blinded comparison in a large patient cohort. Radiology 2001;219(3):739–749. 14. Tang PH, Hui F, Sitoh YY. Intracranial aneurysm detection with 3T magnetic resonance angiography. Ann Acad Med Singapore 2007; 36(6):388–393.
560
Li et al
15. Li MH, Li YD, Tan HQ, et al. Contrast-free MRA at 3.0 T for the detection of intracranial aneurysms. Neurology 2011;77(7):667– 676. 16. Numminen J, Tarkiainen A, Niemelä M, Porras M, Hernesniemi J, Kangasniemi M. Detection of unruptured cerebral artery aneurysms by MRA at 3.0 tesla: comparison with multislice helical computed tomographic angiography. Acta Radiol 2011; 52(6):670–674. 17. Li MH, Cheng YS, Li YD, et al. Large-cohort comparison between three-dimensional timeof-flight magnetic resonance and rotational digital subtraction angiographies in intracranial aneurysm detection. Stroke 2009;40(9): 3127–3129. 18. Li H, Yan L, Li MH, et al. Evaluation of intracranial aneurysms with high-resolution MR angiography using single-artery highlighting technique: correlation with digital subtraction angiography. Radiol Med (Torino) 2013;18(8):1379–1387. 19. Aprile I. Evaluation of cerebral aneurysms with MR-angiography. Rev Neuroradiol 1996;9:541–550. 20. Schwab KE, Gailloud P, Wyse G, Tamargo RJ. Limitations of magnetic resonance imaging and magnetic resonance angiography in the diagnosis of intracranial aneurysms.
Neurosurgery 2008;63(1):29–34; discussion 34–35. 21. Lu H, Nagae-Poetscher LM, Golay X, Lin D, Pomper M, van Zijl PC. Routine clinical brain MRI sequences for use at 3.0 Tesla. J Magn Reson Imaging 2005;22(1):13–22. 22. Hiratsuka Y, Miki H, Kikuchi K, et al. Sensitivity of an eight-element phased array coil in 3 Tesla MR imaging: a basic analysis. Magn Reson Med Sci 2007;6(3):177–181. 23. Willinek WA, Born M, Simon B, et al. Timeof-flight MR angiography: comparison of 3.0-T imaging and 1.5-T imaging—initial experience. Radiology 2003;229(3):913–920. 24. Fushimi Y, Miki Y, Kikuta K, et al. Com parison of 3.0- and 1.5-T three-dimensional time-of-flight MR angiography in moyamoya disease: preliminary experience. Radiology 2006;239(1):232–237. 25. de Zwart JA, Ledden PJ, van Gelderen P, Bodurka J, Chu R, Duyn JH. Signal-to-noise ratio and parallel imaging performance of a 16-channel receive-only brain coil array at 3.0 Tesla. Magn Reson Med 2004;51(1):22–26. 26. White PM, Wardlaw JM, Lindsay KW, Sloss S, Patel DK, Teasdale EM. The non-invasive detection of intracranial aneurysms: are neuroradiologists any better than other observers? Eur Radiol 2003;13(2):389–396.
radiology.rsna.org n Radiology: Volume 271: Number 2—May 2014
