Imaging
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.
Original Research n Gastrointestinal
Incremental Value of Liver MR Imaging in Patients with Potentially Curable Colorectal Hepatic Metastasis Detected at CT: A Prospective Comparison of Diffusion-weighted Imaging, Gadoxetic Acid–enhanced MR Imaging, and a Combination of Both MR Techniques1 Hye Jin Kim, MD2 Seung Soo Lee, MD Jae Ho Byun, MD Jin Cheon Kim, MD Chang Sik Yu, MD Seong Ho Park, MD Ah Young Kim, MD Hyun Kwon Ha, MD
h1 From the Department of Radiology and Research Institute of Radiology (H.J.K., S.S.L., J.H.B., S.H.P., A.Y.K., H.K.H.) and Department of Surgery (J.C.K., C.S.Y.), University of Ulsan College of Medicine, Asan Medical Center, 86 Asanbyeongwon-Gil, Songa-Gu, Seoul 138-736, Korea. From the 2009 RSNA Annual Meeting. Received February 14, 2014; revision requested April 9; final revision received July 16; final version accepted August 12. Supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), which is funded by the Ministry of Education, Science and Technology (grants 2012R1A1A1005326 and 2011-0011544). Address correspondence to S.S.L. (e-mail: [email protected]).
Current address: Department of Radiology, Eulji University School of Medicine, Eulji General Hospital, Seoul, Korea. 2
RSNA, 2014
q
712
Purpose:
To prospectively compare diagnostic performance of diffusionweighted (DW) imaging, gadoxetic acid–enhanced magnetic resonance (MR) imaging, both techniques combined (combined MR imaging), and computed tomography (CT) for detecting colorectal hepatic metastases and evaluate incremental value of MR for patients with potentially curable colorectal hepatic metastases detected with CT.
Materials and Methods:
In this institutional review board–approved prospective study, with informed consent, 51 patients (39 men, 12 women; mean age, 62 years) with potentially resectable hepatic metastases detected with CT underwent liver MR, including DW imaging and gadoxetic acid–enhanced MR. Two independent readers reviewed DW, gadoxetic acid–enhanced, combined MR, and CT image sets to detect hepatic metastases. The figure-of-merit (FOM) value representing overall diagnostic performance, sensitivity, and positive predictive value (PPV) for each image set were analyzed by using free-response receiver operating characteristic analysis and generalized estimating equations.
Results:
There were 104 hepatic metastases in 47 patients. The pooled FOM values, sensitivities, and PPVs of combined MR (FOM value, 0.93; sensitivity, 98%; and PPV, 88%) and gadoxetic acid–enhanced MR (FOM value, 0.92; sensitivity, 95%; and PPV, 90%) were significantly higher than those of CT (FOM value, 0.82; sensitivity, 85%; and PPV, 73%) (P , .006). The pooled FOM value and sensitivity of combined MR (FOM value, 0.92; sensitivity, 95%) was also significantly higher than that of DW imaging (FOM value, 0.82; sensitivity, 79%) for metastases (1-cm diameter) (P .003). DW imaging showed significantly higher pooled sensitivity (79%) and PPV (60%) than CT (sensitivity, 50%; PPV, 33%) for the metastases (1-cm diameter) (P .004). In 47 patients with hepatic metastases, combined MR depicted more metastases than CT in 10 and 14 patients, respectively, according to both readers.
Conclusion:
Gadoxetic acid–enhanced MR and combined MR are more accurate than CT in detecting colorectal hepatic metastases, have an incremental value when added to CT alone for detecting additional metastases, and can be routinely performed in patients with potentially curable hepatic metastases detected with CT. RSNA, 2014
q
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
T
he liver is the most common site for colorectal cancer metastases (1). In patients with colorectal hepatic metastases, meticulous pretherapeutic assessment of the hepatic tumor burden is important for selecting the optimal treatment method among surgical resection, which is the current treatment of choice, and alternative therapeutic methods, including local ablation therapies and adjuvant chemotherapy (1–4).
Advances in Knowledge nn Gadoxetic acid–enhanced MR imaging alone (figure-of-merit [FOM] value, 0.92; sensitivity, 95%; and positive predictive value [PPV], 90%) or combined with diffusion-weighted (DW) imaging (FOM value, 0.93; sensitivity, 98%; and PPV, 88%) resulted in significantly higher reader-averaged, overall diagnostic performance, sensitivity, and PPV for detection of colorectal hepatic metastases than CT (FOM value, 0.82; sensitivity, 85%; and PPV, 73%) (P , .006). nn Gadoxetic acid–enhanced MR imaging combined with DW imaging had significantly higher reader-averaged sensitivity (98% for all metastases and 95% for metastases 1 cm in diameter) for detection of colorectal hepatic metastases than DW imaging alone (88% for all metastases and 79% for metastases 1 cm in diameter) (P .003). nn DW imaging showed a significantly higher sensitivity (79%) and PPV (60%) for detection of small hepatic metastases 1 cm or smaller in diameter than CT (sensitivity, 50%; PPV, 33%) (P .004). nn Gadoxetic acid–enhanced MR imaging helped identify additional hepatic metastases in at least 10 (20%) of 51 patients with potentially curable hepatic metastasis detected with routine staging CT.
Among recent advances in magnetic resonance (MR) imaging techniques, diffusion-weighted (DW) imaging and gadoxetic acid (Primovist, Bayer Schering Pharma, Berlin, Germany; Eovist, Bayer Healthcare Pharmaceuticals, Whippany, NJ)–enhanced MR imaging improved the effectiveness of liver MR imaging for evaluating focal hepatic lesions by providing images with high lesion-to-liver contrast, as well as information in regard to the lesion characteristics (5–11). As a result, these techniques have been increasingly used for clinical liver MR imaging. Despite their frequent use in current clinical practice, there have been only limited data in regard to the diagnostic accuracy of gadoxetic acid–enhanced MR imaging and DW imaging for the evaluation of colorectal hepatic metastases (12–16). All of the previous research on this topic is in retrospective studies, which may be subject to verification and selection bias (12–16). In patients with colorectal cancer, computed tomography (CT) is usually performed for the initial staging workup. Most clinical practice guidelines for colorectal cancer recommend only liver MR imaging as one of the optional worku-ps for patients with potentially surgically curable liver metastases detected with routine staging CT (17,18). However, to our knowledge, no previous study has demonstrated the incremental value of liver MR imaging performed in the clinical setting in which liver MR imaging is indicated. Therefore, the actual clinical benefit of liver MR imaging for the evaluation of colorectal cancer has not yet been well determined. Therefore, the purpose of our study was to prospectively compare the
Implication for Patient Care nn Gadoxetic acid–enhanced MR imaging adds an incremental value to that of CT for detection of additional metastases and can thus be routinely performed in patients with colorectal cancer who have potentially curable hepatic metastases detected with CT.
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
Kim et al
diagnostic performance of DW imaging, gadoxetic acid–enhanced MR imaging, both techniques combined, and CT for detection of colorectal hepatic metastases and to evaluate the incremental value of MR imaging for patients with potentially curable colorectal hepatic metastases detected with CT.
Materials and Methods This prospective study was approved by our institutional review board. All patients provided written informed consent before being enrolled in the study.
Study Population From March 2008 to March 2009, a panel of abdominal radiologists (H.J.K. and A.Y.K., with 5 and 13 years of experience, respectively) and colorectal surgeons (J.C.K. and C.S.Y., with 22 and 15 years of experience, respectively) prospectively evaluated the patients who underwent contrast material–enhanced multidetector CT for the routine preoperative work-up or postoperative follow-up of their colorectal cancer. They also determined each patient’s eligibility for this study according to the following inclusion criteria: (a) The patient was suspected of having one or more hepatic metastases seen on a CT scan, (b)
Published online before print 10.1148/radiol.14140390 Content codes: Radiology 2015; 274:712–722 Abbreviations: DW = diffusion weighted FOM = figure of merit PPV = positive predictive value Author contributions: Guarantors of integrity of entire study, S.S.L., J.C.K.; 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; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, H.J.K., S.S.L., J.H.B., S.H.P.; clinical studies, H.J.K., S.S.L., J.H.B., J.C.K., C.S.Y., A.Y.K.; statistical analysis, S.S.L., J.C.K.; and manuscript editing, H.J.K., S.S.L., J.H.B., J.C.K., S.H.P., H.K.H. Conflicts of interest are listed at the end of this article.
713
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
the patient was a potential candidate for curative hepatic resection, and (c) the patient was willing to participate in this study. Of 501 patients who were evaluated during this period, 97 patients were suspected of having hepatic metastases seen on a CT scan. Among them, 39 patients were excluded because they were not candidates for curative hepatic resection owing to (a) the presence of too many (more than 10) hepatic metastases (n = 20) and (b) the presence of distant metastases other than metastases to the liver (n = 19). A total of 58 eligible patients underwent MR imaging within 30 days after their CT scan (mean interval, 4 days; range, 0–28 days). Among them, seven patients were excluded from our study because of the unavailability of the reference standard required to definitely identify their hepatic lesions (n = 3), the lesions that were suspected of being liver metastases proven to be primary hepatic malignancies at subsequent hepatic resection (n = 3, with hepatocellular carcinoma in one and cholangiocarcinoma in two), and failure to complete the MR examination (n = 1). The remaining 51 patients (39 men, 12 women; mean age, 62 years; range, 41–85 years) constituted the study population. At the time of the study enrollment, 30 of 51 patients were undergoing work-up for already diagnosed rectal (n = 16) or colon (n = 14) cancer; among them, five patients with rectal cancer were being treated with neoadjuvant chemoradiotherapy. Twenty-one of 51 patients were undergoing follow-up after curative resection of primary colorectal cancer, and 18 of them had received adjuvant chemotherapy 1–35 months prior to their enrollment. The flowchart of the study population is presented in Figure 1.
Proof of Hepatic Lesions Forty-seven (92%) of 51 patients in our study population had a total of 104 hepatic metastases, whereas four (8%) patients had no hepatic metastases. The hepatic metastases were proven with histologic analysis for 77 lesions and with imaging findings and follow-up 714
Kim et al
Figure 1
Figure 1: Flowchart of the study population. F/U = follow-up, IOUS = intraoperative ultrasonography (US), RFA = radiofrequency ablation.
for 27 lesions. The mean number of metastases per patient was 2.2 (range, 1–10), with a mean lesion diameter of 1.9 cm 6 1.4 (standard deviation) and a range of 0.3–8.5 cm. Twenty-eight (27%) of the 104 metastases were 1 cm or smaller in diameter. Forty-three of 51 patients underwent curative-intent surgery with manual exploration and intraoperative US of the liver within 30 days (mean interval, 4 days; range, 1–30 days) after MR imaging. One abdominal radiologist (H.J.K., with 5 years of experience in intraoperative US) performed intraoperative US and evaluated the entire liver with a special focus on the lesions that were suspected of being hepatic metastases and that were seen on CT
scans and MR images. In all lesions that were suspected of being hepatic metastases, samples were then removed for pathologic analysis by means of hepatic resection (n = 67) or biopsy (n = 13) with (n = 11) or without (n = 2) radiofrequency ablation. Pathologic examination of surgical or biopsy specimens revealed a total of 77 metastases in 40 patients. In three patients, the lesions that were suspected of being hepatic metastases were proven to be benign, including focal eosinophilic infiltration (n = 2) and focal fibrosis (n = 1). Seven of 51 patients did not undergo resection of their hepatic lesions because of the final clinical decision of their unsuitability for curative hepatic resection that was made on the basis
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
of the hepatic tumor burden, each patient’s general condition, and hepatic reservoir (n = 4); unexpected peritoneal tumor seeding detected at surgery (n = 2); and the patient’s refusal to undergo liver surgery (n = 1). In these patients, a total of 27 hepatic metastases were diagnosed nonpathologically on the basis of the imaging findings and substantial tumor growth (ie, increase in the longest tumor diameter 30%) on follow-up CT scans or MR images (mean follow-up, 3 months; range, 1–9 months). In the remaining one of 51 patients, the lesions that were suspected of being hepatic metastases seen on a CT scan were proven to be focal fat deposition on the basis of the MR imaging findings and their disappearance as seen on a follow-up CT scan obtained within 4 months following the initial CT scan. Thirty-one of 47 patients with hepatic metastases also had a total of 85 benign hepatic lesions, including hepatic cyst (n = 74), hemangioma (n = 10), and focal fat deposition (n = 1). One hemangioma was histologically proven by means of hepatic resection. The other 84 benign lesions in 46 patients were diagnosed on the basis of the MR imaging findings in all patients, as well as on the basis of US findings in 25 patients and/or positron emission tomography (PET)/CT in 24 patients and were reconfirmed by their stability (n = 83) or disappearance (n = 1) on follow-up CT or MR imaging (mean follow-up, 25 months; range, 12–48 months).
CT Examination CT examinations were performed by using 16– or 64–section multidetector CT scanners (Somatom Sensation 16 or 64; Siemens, Erlangen, Germany). The scanning parameters were beam collimation, 16 3 0.75 or 64 3 0.6 mm; 120 kVp; automated dose modulation (CARE dose 4D; Siemens) with the reference tube current set at 200 effective mAs; gantry rotation time, 0.5 seconds; beam pitch, 1.2; and field of view to fit. Portal venous phase images were obtained at a 72-second delay after the initiation of intravenous administration
of 150 mL of iopromide (Ultravist 300; Bayer Schering Pharma) at a rate of 3.0 mL/sec by using an automated power injector (Optivantage DH; Mallinckrodt Imaging Solutions, Hazelwood, Mo). Both axial and coronal images were reconstructed to 5-mm thickness at 5-mm intervals.
MR Imaging MR imaging was performed by using a 1.5-T system (Magnetom Avanto; Siemens) with a dedicated, six-channel, torso-array coil. All patients underwent baseline, nonenhanced MR imaging, including breath-hold T1-weighted two-dimensional dual gradient-echo in-phase and opposed-phase imaging, breathhold heavily T2-weighted half-Fourier single-shot turbo spin-echo imaging, and navigator-triggered intermediate T2-weighted turbo spin-echo imaging. DW imaging was performed by using a navigator-triggered single-shot echoplanar sequence with DW gradients (ie, b factors of 50, 300, 600, and 900 sec/ mm2, applied in three orthogonal directions). After nonenhanced MR imaging, gadoxetic acid–enhanced dynamic T1weighted images were obtained by using a fat-suppressed three-dimensional spoiled gradient-echo sequence, volumetric interpolated breath-hold examination (VIBE; Siemens) after a bolus injection of 10 mL of gadoxetic acid (Primovist; Bayer Schering Pharma) at a rate of 1.0 mL/sec, followed by a 30mL saline flush using a power injector (Spectris Solaris; Medrad, Warrendale, Pa). The images were obtained before contrast agent injection and in the arterial phase, which was determined by using a real-time bolus display method (CARE bolus; Siemens) in the venous phase (50 seconds after contrast agent injection), in the delayed phase (3 minutes following contrast agent injection), and in the hepatobiliary phase (10 minutes and 20 minutes after contrast agent injection). The detailed parameters for the MR sequences are summarized in Table 1. Image Analysis Image interpretation was performed in a retrospective manner after completion
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
Kim et al
of patient enrollment. Two board-certified radiologists (S.S.L. and J.H.B., with 8 and 10 years of experience in abdominal imaging, respectively), who were not involved in the patient selection, intraoperative US, or the construction of the reference standard, independently evaluated the CT and MR image sets. Image analysis was performed during four separate sessions; the readers interpreted the CT images during the first session, DW images and baseline nonenhanced T1- and T2-weighted images (hereafter referred to as DW MR images or imaging) during the second session, gadoxetic acid–enhanced MR images and baseline nonenhanced MR images (hereafter referred to as gadoxetic acid MR images or imaging) during the third session, and all of the MR images (hereafter referred to as combined MR images or imaging) during the fourth session. To reduce recall bias, each image review session was separated by at least 4 weeks, and the images were made anonymous and were randomized. The readers knew that all of the patients had colorectal cancer, although they were unaware of the results of the other review sessions, the surgical findings, pathologic examination results, and the follow-up imaging findings. During each review session, the readers were asked to identify all possible metastatic lesions and to record their size, segmental location (19), and the number of the image where the lesion was contained, and then to grade their confidence level for each lesion on a three-point scale as follows: score 1, equivocal but possibly metastatic; score 2, probably metastatic; and score 3, definitely metastatic. Definite or probable benign lesions were not recorded. The readers characterized focal hepatic lesions by using the known imaging criteria on CT and gadoxetic acid MR images (12,13,20–25). On DW MR images, lesions were characterized on the basis of their signal intensities on T2-weighted images and on DW images obtained with a b value of 50 sec/mm2 and the changes in their signal intensities on DW images obtained with b values of 600 and 900 sec/mm2, according to the previously reported criteria (5,26–29). On combined MR images, 715
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
Table 1 Parameters of MR Pulse Sequences Parameter Repetition time (msec) Echo time (msec) Flip angle (degrees) Field of view Matrix Section thickness (mm) Intersection gap (mm) Echo train length No. of acquisitions Parallel imaging factor‡ Fat suppression used§
T1 Dual Gradient-Echo Imaging
Heavily T2 Half-Fourier Turbo SE Imaging
Intermediate T2 Turbo SE Imaging
DW Imaging
Contrast-enhanced Dynamic T1 and Hepatobiliary Phase Imaging
164 2.3, opposed phase; 4.6, in phase 70 350 3 260 256 3 192 6 1.2 NA 1 2 No
1100 151 150 350 3 260 256 3 192 6 1.2 134 1 0 Yes
4023–5755* 85 150 350 3 260 384 3 288 6 1.2 9 2 2 Yes
4800 73 180 350 3 300 150 3 128 6 1.2 76† 5 2 Yes
4.1 1.5 10 350 3 284 320 3 260 3 0 NA 1 2 Yes
Note.—T1 dual gradient-echo imaging refers to dual gradient-echo in-phase and opposed-phase imaging. NA = not applicable, SE = spin echo, T1 = T1 weighted, T2 = T2 weighted. * Repetition time varied, depending on the patient’s respiratory cycle. †
Echo-planar imaging factor.
‡
Parallel imaging was performed by using a k-space–based technique (GRAPPA; Siemens).
§
Fat saturation was achieved with the chemical shift-selective fat-suppression technique.
hepatic lesions were characterized by using a combination of the aforementioned criteria for DW MR images and gadoxetic acid MR images.
Statistical Analysis The overall diagnostic performance for detecting hepatic metastases was compared among CT, DW MR imaging, gadoxetic acid MR imaging, and combined MR imaging for all metastatic lesions and for metastatic lesions of 1 cm or smaller in diameter by using an alternative free-response receiver operating characteristic analysis (JAFROC1 version 1.0; http://www.devchakraborty. com/index.php) (30). The figure-ofmerit (FOM) value, defined as a probability of a lesion being rated higher than the highest rated nonlesions on normal images (30–32), was used to indicate the overall diagnostic performance of each image set, where the value of one indicated the highest performance and zero indicated the lowest performance. The sensitivities and positive predictive values (PPVs) for the four image sets were compared by using the generalized estimating equations with software (IBM SPSS statistics 19; IBM, New York, NY) to address data clustering (ie, multiple lesions in a single patient). 716
For all statistical analysis in our study, the differences with P values less than the Bonferroni-corrected value of .008 (calculated as .05/6) were considered significant to account for the increased type I error in multiple comparisons.
Results Overall Diagnostic Performance The results of alternative free-response receiver operating characteristic analysis in regard to the overall diagnostic performance for detecting hepatic metastases are summarized in Table 2. For detecting all hepatic metastases, the pooled FOM value was the highest for combined MR imaging (FOM value = 0.93), followed by gadoxetic acid MR imaging (FOM value = 0.92), DW MR imaging (FOM value = 0.89), and CT (FOM value = 0.82). The FOM values of combined MR imaging and gadoxetic acid MR imaging were significantly higher than the FOM value of CT for the pooled data and for both readers (P .006). For detecting hepatic metastases of 1 cm or smaller in diameter, the pooled FOM values of combined MR imaging, gadoxetic acid MR imaging, and DW MR imaging were significantly
higher than the pooled FOM value of CT (P , .001). The pooled FOM value of combined MR imaging was also significantly higher than that of DW MR imaging (P , .001). The FOM value of combined MR imaging for detecting hepatic metastases of 1 cm or smaller in diameter was higher than that of gadoxetic acid MR imaging, with a marginal significant difference only for reader 2 (P = .03).
Sensitivity and PPV For all metastases and metastases of 1 cm or smaller in diameter, the pooled sensitivities were significantly higher with combined MR imaging (98% for all metastases, 95% for metastases 1 cm in diameter) than with CT (85% for all metastases, 50% for metastases 1 cm in diameter) and DW MR imaging (88% for all metastases, 79% for metastases 1 cm in diameter) (P .003). The pooled sensitivities of gadoxetic acid MR imaging (95% for all metastases, 86% for metastases 1 cm in diameter) were also significantly higher than with CT (P , .001) (Table 3). DW MR imaging resulted in a significantly higher pooled sensitivity than CT only for metastases of 1 cm or
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
smaller in diameter (P = .004). The sensitivities of combined MR imaging for all metastases and metastases of 1 cm or smaller in diameter were slightly higher than those of gadoxetic acid MR imaging, with a marginal significant difference for pooled data (P .027) and for reader 2 (P .011) (Figs 2, 3). The pooled PPVs of combined MR imaging, gadoxetic acid MR imaging,
Kim et al
and DW MR imaging were similar to each other, ranging from 84% to 90% for all metastases and from 60% to 72% for the metastases of 1 cm or smaller in diameter, although they were significantly higher than those of CT (73% for all lesions, 33% for lesions 1 cm in diameter) for both all metastases and the metastases that were 1 cm or smaller in diameter (P , .001) (Table 4).
Table 2 FOM Values Indicating the Diagnostic Performance for the Detection of Hepatic Metastases Lesion Group and Imaging Modality All metastases (n = 104) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging Metastases 1.0 cm in diameter (n = 28) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging
Reader 1
Reader 2
Pooled Data
0.80 (0.73, 0.86) 0.88 (0.82, 0.93) 0.93 (0.87, 0.97)* 0.93 (0.87, 0.96)*
0.83 (0.76, 0.89) 0.89 (0.83, 0.93) 0.92 (0.86, 0.95)* 0.94 (0.88, 0.97)*
0.82 (0.75, 0.87) 0.89 (0.84, 0.92) 0.92 (0.88, 0.96)* 0.93 (0.89, 0.96)*
0.59 (0.46, 0.71) 0.82 (0.71, 0.91) 0.92 (0.83, 0.97)* 0.92 (0.84, 0.97)*
0.56 (0.45, 0.68) 0.82 (0.70, 0.91)* 0.85 (0.74, 0.93)* 0.92 (0.83, 0.97)*
0.58 (0.46, 0.68) 0.82 (0.73, 0.90)* 0.89 (0.80, 0.94)* 0.92 (0.85, 0.96)†
Note.—Numbers in parentheses are 95% confidence intervals. * The FOM value is significantly higher than that of CT (P .006). †
The FOM value is significantly higher than that of CT (P < .001). The FOM value is significantly higher than that of DW MR
imaging (P , .001).
Per-Patient Analysis and the Potential Effect of MR Imaging on Patient Treatment In all 47 patients with hepatic metastasis, the presence of hepatic metastasis was correctly diagnosed with CT, gadoxetic acid MR imaging, and combined MR imaging by both readers 1 and 2 and with DW MR imaging by reader 2, thus resulting in a per-patient sensitivity of 100% (47 of 47), while diagnosis with DW MR imaging resulted in a false-negative result of hepatic metastasis in one patient for reader 1. Among 47 patients with hepatic metastases, DW MR imaging, gadoxetic acid MR imaging, and combined MR imaging depicted more metastases than CT in eight, 10, and 10 patients, respectively, for reader 1 and in eight, 10, and 14 patients, respectively, for reader 2 (Figs 2, 3). The detection of additional metastases with combined MR imaging would have actually changed the treatment of at least four patients; in three patients, hepatic wedge resection (n = 2) and intraoperative radiofrequency ablation (n = 1) were performed to treat the hepatic metastases that were detected with combined MR imaging but were missed with CT by both readers, in addition to the hepatic resection (n = 2) or intraoperative
Table 3 Sensitivities for the Detection of Hepatic Metastases Reader 1 Lesion Group and Imaging Modality All metastases CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging Metastases 1.0 cm in diameter CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging
Sensitivity (%)
95% Confidence Interval
87 (90/104) 88 (92/104) 98 (102/104)* 99 (103/104)* 57 (16/28) 75 (21/28) 93 (26/28)* 96 (27/28)*
Reader 2
Pooled Data
Sensitivity (%)
95% Confidence Interval
Sensitivity (%)
95% Confidence Interval
79, 92 81, 93 93, 100 93, 100
84 (87/104) 88 (92/104) 92 (96/104) 97 (101/104)*
75, 90 81, 93 85, 96 92, 99
85 88 95* 98†
80, 89 83, 92 91, 97 95, 99
39, 74 56, 88 76, 99 81,100
43 (12/28) 82 (23/28) 79 (22/28)* 93 (26/28)*
26, 61 64, 93 60, 90 76, 99
50 79* 86* 95†
37, 63 66, 87 74, 93 85, 99
Note.—Numbers in parentheses are the numbers of lesions, which were used for calculating the percentages. Percentages were rounded. * The sensitivity is significantly higher than that of CT (P .003). †
The sensitivity is significantly higher than that of CT (P < .001). The sensitivity is significantly higher than that of DW MR imaging for both all metastases and metastases of 1 cm or smaller in diameter (P .003).
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
717
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
without hepatic metastasis, true-negative interpretations were made with CT, DW MR imaging, gadoxetic acid MR imaging, and combined MR imaging in zero, two, zero, and zero patients, respectively, for reader 1 and zero, one, one, and two patients, respectively, for reader 2.
Figure 2
Discussion
Figure 2: Multisegmental involvement of hepatic metastases from rectal cancer not seen on a CT scan in a 68-year-old man. A, Transverse contrast-enhanced portal venous phase CT images show a hypoattenuating lesion (arrow) in hepatic segment IV on middle image. No metastatic lesions were observed on left and right images. B, Transverse gadoxetic acid–enhanced hepatobiliary phase images and, C, DW images (b = 600 sec/mm2) demonstrate two lesions in hepatic segments VII (solid arrow) on left images and III (arrowhead) on right images, in addition to the lesion in segment IV (open arrow), which was detected on the CT scan, on middle images. With gadoxetic acid–enhanced MR images alone (B), the readers missed the tiny lesion in segment III (arrowhead) on right image. With DW images alone and with combined DW images and gadoxetic acid–enhanced MR images, all three lesions were identified by both readers.
radiofrequency ablation to cover the lesions detected with CT (Fig 2). In one patient, the additional metastases depicted with combined MR imaging 718
precluded surgical resection, leading to a change in the treatment plan from surgical resection to chemotherapy for this patient. In four patients
Our study results demonstrated that gadoxetic acid–enhanced MR imaging alone or in combination with DW imaging had a significantly higher reader-averaged, overall diagnostic performance, sensitivity, and PPV for depicting colorectal hepatic metastases, compared with CT. Unlike previous studies, which reported a comparable or even higher diagnostic performance of DW imaging for depicting hepatic metastases compared with gadopentetate dimeglumine–enhanced or superparamagnetic iron oxide–enhanced MR imaging (5,24), our study demonstrated somewhat different results; DW MR imaging showed lower overall diagnostic performance, sensitivity, and PPV compared with combined MR imaging and gadoxetic acid MR imaging, with a significant difference in reader-averaged FOM values and sensitivities for depicting metastases of 1 cm or smaller in diameter, compared with combined MR imaging. Investigators in several retrospective studies (12,13,25) also reported findings consistent with those of our study. The addition of DW imaging to gadoxetic acid–enhanced MR imaging slightly improved the overall diagnostic performance and sensitivity for detecting colorectal hepatic metastases, especially for lesions of 1 cm or smaller in diameter, compared with gadoxetic acid–enhanced MR imaging alone, although these improvements were not significantly different. Our results, taken together with the results of previous studies, suggest that combined gadoxetic acid–enhanced MR imaging and DW imaging would be the most preferred imaging technique for the evaluation of colorectal hepatic metastases among the techniques evaluated in our study.
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
Figure 3
Figure 3: Multiple hepatic metastases from colon cancer in a 68-year-old man. A, Transverse contrast-enhanced portal venous phase CT image shows a hypoattenuating metastatic lesion in hepatic segment VIII (arrow) but no focal lesion in the left hepatic lobe. B, Transverse DW images (b = 900 sec/mm2) demonstrate a tiny hyperintense lesion in hepatic segment III (arrowhead) in addition to the lesion in hepatic segment VIII (arrow) seen on the CT scan. C, Transverse gadoxetic acid–enhanced MR images show two hypointense lesions (arrowhead and arrow), which correspond to the lesions shown on DW images (B). D, Follow-up contrast-enhanced CT image obtained 2 months following the initial CT scan demonstrates obvious growth of the two hepatic lesions (arrowhead and arrow), confirming the diagnosis of hepatic metastases.
Although it is well known that contrast-enhanced liver MR imaging is superior to CT for detection of colorectal liver metastases, especially for detection of small metastases that are 1 cm or smaller in diameter (11,20,33), it has not been clearly determined whether the improved accuracy of liver MR imaging leads to an actual clinical benefit for treating patients with colorectal cancer. Our results suggest that gadoxetic acid–enhanced MR imaging may provide substantial clinical benefits for
patients with potentially curable hepatic metastases depicted with CT by depicting a greater number of hepatic metastases and thus validating the current guidelines, which recommend liver MR imaging for these patients (17,18). In our study, combined MR imaging allowed readers 1 and 2 to identify 23 and 24, respectively, more hepatic metastases (including 14 and 11, respectively, more liver metastases 1 cm in diameter) than they did with CT, while this protocol resulted in fewer false-positive lesions (12 and
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
six, respectively, for readers 1 and 2) than CT (41 and 26, respectively, for readers 1 and 2). With per-patient– based analyses, combined MR imaging allowed the detection of more hepatic metastases in at least 10 (20%) of 51 patients, and which would have resulted in actual change in treatment in four of them. However, liver MR imaging may not be effective for proving the benign nature of lesions suspected of being hepatic metastases that are seen on CT scans. In four patients whose lesions were suspected 719
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
Table 4 PPVs for the Detection of Hepatic Metastases Reader 1 Lesion Group and Imaging Modality All metastases (n = 104) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging Metastases 1.0 cm in diameter (n = 28) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging
PPV (%)
95% Confidence Interval
69 (90/131) 87 (92/106)* 92 (102/111)* 90 (103/115)* 31 (16/51) 66 (21/32)* 79 (27/34)* 76 (28/37)*
Reader 2
Pooled Data
PPV (%)
95% Confidence Interval
PPV (%)
95% Confidence Interval
60, 76 79, 92 85, 96 83, 94
77 (87/113) 81 (92/113) 87 (96/110) 86 (101/117)
68, 84 73, 88 80, 92 79, 92
73 84* 90* 88*
67, 78 79, 88 85, 93 83, 93
20, 45 48, 80 63, 80 60, 87
36 (13/36) 55 (23/42) 65 (22/34)* 65 (26/40)*
22, 52 40, 69 48, 79 49, 78
33 60* 72* 70*
24, 44 49, 71 60, 81 59, 79
Note.—Numbers in parentheses are the numbers of lesions, which were used for calculating the percentages. Percentages were rounded. * The PPV is significantly higher than that of CT (P .001).
of being hepatic metastases and were depicted on CT scans were proven to be benign lesions, including eosinophilic infiltration (n = 2), focal fibrosis (n = 1), and focal fat deposition (n = 1), combined MR imaging, gadoxetic acid MR imaging, and DW MR imaging could not consistently help diagnose these lesions as benign. In our study, we did not compare the diagnostic performance of liver MR imaging techniques with the diagnostic performance of other imaging examinations, such as dynamic contrastenhanced CT and PET/CT. Therefore, our results cannot directly address the question of which examination would be the most appropriate as a further work-up for patients with potentially curable hepatic metastases. However, the results of previous studies showing higher sensitivities of gadoxetic acid– enhanced MR imaging in the detection of hepatic metastases compared with those of dynamic contrast-enhanced CT (7) or PET/CT (15) suggest that gadoxetic acid–enhanced MR imaging would be preferable to these examinations for the evaluation of hepatic metastases. In addition to its higher diagnostic performance, gadoxetic acid–enhanced MR imaging has an additional advantage in its lack of radiation exposure compared with dynamic contrast-enhanced CT and PET/CT, 720
both of which are subject to the potential hazards of ionizing radiation. Our study had several limitations. First, as it included only patients who were suspected of having colorectal hepatic metastases as seen on CT scans, our results cannot be generalized for all patients with colorectal cancer. For those patients with colorectal cancer in whom CT shows no focal hepatic lesion or only hepatic lesions deemed too small to be characterized on a CT scan, it is still unknown whether further evaluation of these patients with the use of gadoxetic acid–enhanced liver MR imaging can offer an important clinical benefit. Second, our study population may have had a higher prevalence of hepatic metastases than the population of all patients with colorectal cancer. Therefore, the PPVs of the imaging examinations may have been overestimated in our study. Although a study population enrolled from among general colorectal cancer patients may have led to unbiased results, this would require a much larger number of patients than we included in our study to comprise a sufficient number of hepatic metastases for statistical analysis. Therefore, we instead decided to focus our study on the clinical setting in which liver MR imaging is recommended according to the current guidelines (17,18). Third,
not all of the hepatic metastases that were detected in our study were pathologically proven, and most of the benign hepatic lesions were nonpathologically diagnosed on the basis of the imaging findings and follow-up results. However, except for the lesions in patients who were determined to be unsuitable for liver resection after their study enrollment, most (74%, 77 of 104) of the hepatic metastases were pathologically proven in our study. As the pathologic confirmation of all focal hepatic lesions for research purposes would not be practical or ethical, this limitation may have been unavoidable in our study. Finally, there may be a possibility that our CT protocol using a greater section thickness (5 mm) than the section thickness for gadoxetic acid–enhanced MR images (3 mm) may have been partially responsible for a worse diagnostic performance of CT than MR imaging. However, the researchers in a recent study (34) with the use of 3-mm-thick CT images also reported results consistent with those of our study (ie, a higher sensitivity of gadoxetic acid–enhanced MR imaging than CT in the detection of colorectal hepatic metastases). In conclusion, gadoxetic acid–enhanced MR imaging alone or in combination with DW imaging is more accurate than CT for detection of
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
colorectal hepatic metastases. The addition of DW imaging to gadoxetic acid–enhanced MR imaging slightly improves the sensitivity for detection of small metastases. Gadoxetic acid–enhanced MR imaging adds incremental value to CT for detection of additional metastases and thus can be routinely performed for patients with potentially curable hepatic metastases seen at CT. Disclosures of Conflicts of Interest: H.J.K. disclosed no relevant relationships. S.S.L. disclosed no relevant relationships. J.H.B. disclosed no relevant relationships. J.C.K. disclosed no relevant relationships. C.S.Y. disclosed no relevant relationships. S.H.P. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: received grants from DongKook Pharm and GE Healthcare. Other relationships: disclosed no relevant relationships. A.Y.K. disclosed no relevant relationships. H.K.H. disclosed no relevant relationships.
References 1. Wong SL, Mangu PB, Choti MA, et al. American Society of Clinical Oncology 2009 clinical evidence review on radiofrequency ablation of hepatic metastases from colorectal cancer. J Clin Oncol 2010;28(3):493–508. 2. Adam R, Aloia T, Lévi F, et al. Hepatic resection after rescue cetuximab treatment for colorectal liver metastases previously refractory to conventional systemic therapy. J Clin Oncol 2007;25(29):4593–4602. 3. Chong G, Cunningham D. Improving longterm outcomes for patients with liver metastases from colorectal cancer. J Clin Oncol 2005;23(36):9063–9066. 4. Zech CJ, Grazioli L, Jonas E, et al. Healtheconomic evaluation of three imaging strategies in patients with suspected colorectal liver metastases: Gd-EOB-DTPA-enhanced MRI vs. extracellular contrast media-enhanced MRI and 3-phase MDCT in Germany, Italy and Sweden. Eur Radiol 2009;19(suppl 3):S753–S763.
Kim et al
7. Motosugi U, Ichikawa T, Morisaka H, et al. Detection of pancreatic carcinoma and liver metastases with gadoxetic acid-enhanced MR imaging: comparison with contrastenhanced multi-detector row CT. Radiology 2011;260(2):446–453.
16. Kim YK, Lee MW, Lee WJ, et al. Diagnostic accuracy and sensitivity of diffusion-weighted and of gadoxetic acid-enhanced 3-T MR imaging alone or in combination in the detection of small liver metastasis ( 1.5 cm in diameter). Invest Radiol 2012;47(3):159–166.
8. Ichikawa T, Saito K, Yoshioka N, et al. Detection and characterization of focal liver lesions: a Japanese phase III, multicenter comparison between gadoxetic acid disodium-enhanced magnetic resonance imaging and contrast-enhanced computed tomography predominantly in patients with hepatocellular carcinoma and chronic liver disease. Invest Radiol 2010;45(3):133–141.
17. Glimelius B, Påhlman L, Cervantes A; ESMO Guidelines Working Group. Rectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2010;21(suppl 5):v82–v86.
9. Hammerstingl R, Huppertz A, Breuer J, et al. Diagnostic efficacy of gadoxetic acid (Primovist)-enhanced MRI and spiral CT for a therapeutic strategy: comparison with intraoperative and histopathologic findings in focal liver lesions. Eur Radiol 2008;18(3): 457–467. 10. Huppertz A, Balzer T, Blakeborough A, et al. Improved detection of focal liver lesions at MR imaging: multicenter comparison of gadoxetic acid-enhanced MR images with intraoperative findings. Radiology 2004;230(1): 266–275. 11. Niekel MC, Bipat S, Stoker J. Diagnostic imaging of colorectal liver metastases with CT, MR imaging, FDG PET, and/or FDG PET/CT: a meta-analysis of prospective studies including patients who have not previously undergone treatment. Radiology 2010;257(3):674–684. 12. Löwenthal D, Zeile M, Lim WY, et al. Detection and characterisation of focal liver lesions in colorectal carcinoma patients: comparison of diffusion-weighted and GdEOB-DTPA enhanced MR imaging. Eur Radiol 2011;21(4):832–840. 13. Chung WS, Kim MJ, Chung YE, et al. Comparison of gadoxetic acid-enhanced dynamic imaging and diffusion-weighted imaging for the preoperative evaluation of colorectal liver metastases. J Magn Reson Imaging 2011;34(2):345–353.
5. Nasu K, Kuroki Y, Nawano S, et al. Hepatic metastases: diffusion-weighted sensitivityencoding versus SPIO-enhanced MR imaging. Radiology 2006;239(1):122–130.
14. Muhi A, Ichikawa T, Motosugi U, et al. Diagnosis of colorectal hepatic metastases: comparison of contrast-enhanced CT, contrast-enhanced US, superparamagnetic iron oxide-enhanced MRI, and gadoxetic acid-enhanced MRI. J Magn Reson Imaging 2011;34(2):326–335.
6. Parikh T, Drew SJ, Lee VS, et al. Focal liver lesion detection and characterization with diffusion-weighted MR imaging: comparison with standard breath-hold T2weighted imaging. Radiology 2008;246(3): 812–822.
15. Donati OF, Hany TF, Reiner CS, et al. Value of retrospective fusion of PET and MR images in detection of hepatic metastases: comparison with 18F-FDG PET/CT and Gd-EOB-DTPA-enhanced MRI. J Nucl Med 2010;51(5):692–699.
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
18. Garden OJ, Rees M, Poston GJ, et al. Guidelines for resection of colorectal cancer liver metastases. Gut 2006;55(suppl 3): iii1–iii8. 19. Dodd GD 3rd. An American’s guide to Couinaud’s numbering system. AJR Am J Roentgenol 1993;161(3):574–575. 20. Onishi H, Murakami T, Kim T, et al. He patic metastases: detection with multi-detector row CT, SPIO-enhanced MR imaging, and both techniques combined. Radiology 2006;239(1):131–138. 21. Rappeport ED, Loft A, Berthelsen AK, et al. Contrast-enhanced FDG-PET/CT vs. SPIOenhanced MRI vs. FDG-PET vs. CT in patients with liver metastases from colorectal cancer: a prospective study with intraoperative confirmation. Acta Radiol 2007;48(4): 369–378. 22. Ward J, Robinson PJ, Guthrie JA, et al. Liver metastases in candidates for hepatic resection: comparison of helical CT and gadolinium- and SPIO-enhanced MR imaging. Radiology 2005;237(1):170–180. 23. Goshima S, Kanematsu M, Watanabe H, et al. Hepatic hemangioma and metastasis: differentiation with gadoxetate disodiumenhanced 3-T MRI. AJR Am J Roentgenol 2010;195(4):941–946. 24. Hardie AD, Naik M, Hecht EM, et al. Diagnosis of liver metastases: value of diffusion-weighted MRI compared with gadolinium-enhanced MRI. Eur Radiol 2010;20(6): 1431–1441. 25. Shimada K, Isoda H, Hirokawa Y, Ari zono S, Shibata T, Togashi K. Comparison of gadolinium-EOB-DTPA-enhanced and diffusion-weighted liver MRI for detection of small hepatic metastases. Eur Radiol 2010;20(11):2690–2698. 26. Coenegrachts K, ter Beek L, Haspeslagh M, Bipat S, Stoker J, Rigauts H. Comparison of respiratory-triggered T2-weighted turbo spin-echo imaging versus breath-hold T2weighted turbo spin-echo imaging: distinguishing benign from malignant liver lesions in patients with colorectal cancer. JBR-BTR 2009;92(4):195–201.
721
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
27. Low RN, Gurney J. Diffusion-weighted MRI (DWI) in the oncology patient: value of breathhold DWI compared to unenhanced and gadolinium-enhanced MRI. J Magn Reson Imaging 2007;25(4):848–858. 28. Miller FH, Hammond N, Siddiqi AJ, et al. Utility of diffusion-weighted MRI in distinguishing benign and malignant hepatic lesions. J Magn Reson Imaging 2010;32(1):138–147. 29. Taouli B, Koh DM. Diffusion-weighted MR imaging of the liver. Radiology 2010;254(1): 47–66.
722
Kim et al
30. Chakraborty DP. Recent advances in observer performance methodology: jackknife free-response ROC (JAFROC). Radiat Prot Dosimetry 2005;114(1-3):26–31. 31. Chakraborty DP. New developments in observer performance methodology in medical imaging. Semin Nucl Med 2011;41(6): 401–418. 32. Chakraborty DP, Berbaum KS. Observer studies involving detection and localization: modeling, analysis, and validation. Med Phys 2004;31(8):2313–2330.
33. Bartolozzi C, Donati F, Cioni D, et al. Detection of colorectal liver metastases: a prospective multicenter trial comparing unenhanced MRI, MnDPDP-enhanced MRI, and spiral CT. Eur Radiol 2004;14(1):14–20. 34. Scharitzer M, Ba-Ssalamah A, Ringl H, et al. Preoperative evaluation of colorectal liver metastases: comparison between gadoxetic acid-enhanced 3.0-T MRI and contrast-enhanced MDCT with histopathological correlation. Eur Radiol 2013;23(8): 2187–2196.
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
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.
Original Research n Gastrointestinal
Incremental Value of Liver MR Imaging in Patients with Potentially Curable Colorectal Hepatic Metastasis Detected at CT: A Prospective Comparison of Diffusion-weighted Imaging, Gadoxetic Acid–enhanced MR Imaging, and a Combination of Both MR Techniques1 Hye Jin Kim, MD2 Seung Soo Lee, MD Jae Ho Byun, MD Jin Cheon Kim, MD Chang Sik Yu, MD Seong Ho Park, MD Ah Young Kim, MD Hyun Kwon Ha, MD
h1 From the Department of Radiology and Research Institute of Radiology (H.J.K., S.S.L., J.H.B., S.H.P., A.Y.K., H.K.H.) and Department of Surgery (J.C.K., C.S.Y.), University of Ulsan College of Medicine, Asan Medical Center, 86 Asanbyeongwon-Gil, Songa-Gu, Seoul 138-736, Korea. From the 2009 RSNA Annual Meeting. Received February 14, 2014; revision requested April 9; final revision received July 16; final version accepted August 12. Supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), which is funded by the Ministry of Education, Science and Technology (grants 2012R1A1A1005326 and 2011-0011544). Address correspondence to S.S.L. (e-mail: [email protected]).
Current address: Department of Radiology, Eulji University School of Medicine, Eulji General Hospital, Seoul, Korea. 2
RSNA, 2014
q
712
Purpose:
To prospectively compare diagnostic performance of diffusionweighted (DW) imaging, gadoxetic acid–enhanced magnetic resonance (MR) imaging, both techniques combined (combined MR imaging), and computed tomography (CT) for detecting colorectal hepatic metastases and evaluate incremental value of MR for patients with potentially curable colorectal hepatic metastases detected with CT.
Materials and Methods:
In this institutional review board–approved prospective study, with informed consent, 51 patients (39 men, 12 women; mean age, 62 years) with potentially resectable hepatic metastases detected with CT underwent liver MR, including DW imaging and gadoxetic acid–enhanced MR. Two independent readers reviewed DW, gadoxetic acid–enhanced, combined MR, and CT image sets to detect hepatic metastases. The figure-of-merit (FOM) value representing overall diagnostic performance, sensitivity, and positive predictive value (PPV) for each image set were analyzed by using free-response receiver operating characteristic analysis and generalized estimating equations.
Results:
There were 104 hepatic metastases in 47 patients. The pooled FOM values, sensitivities, and PPVs of combined MR (FOM value, 0.93; sensitivity, 98%; and PPV, 88%) and gadoxetic acid–enhanced MR (FOM value, 0.92; sensitivity, 95%; and PPV, 90%) were significantly higher than those of CT (FOM value, 0.82; sensitivity, 85%; and PPV, 73%) (P , .006). The pooled FOM value and sensitivity of combined MR (FOM value, 0.92; sensitivity, 95%) was also significantly higher than that of DW imaging (FOM value, 0.82; sensitivity, 79%) for metastases (1-cm diameter) (P .003). DW imaging showed significantly higher pooled sensitivity (79%) and PPV (60%) than CT (sensitivity, 50%; PPV, 33%) for the metastases (1-cm diameter) (P .004). In 47 patients with hepatic metastases, combined MR depicted more metastases than CT in 10 and 14 patients, respectively, according to both readers.
Conclusion:
Gadoxetic acid–enhanced MR and combined MR are more accurate than CT in detecting colorectal hepatic metastases, have an incremental value when added to CT alone for detecting additional metastases, and can be routinely performed in patients with potentially curable hepatic metastases detected with CT. RSNA, 2014
q
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
T
he liver is the most common site for colorectal cancer metastases (1). In patients with colorectal hepatic metastases, meticulous pretherapeutic assessment of the hepatic tumor burden is important for selecting the optimal treatment method among surgical resection, which is the current treatment of choice, and alternative therapeutic methods, including local ablation therapies and adjuvant chemotherapy (1–4).
Advances in Knowledge nn Gadoxetic acid–enhanced MR imaging alone (figure-of-merit [FOM] value, 0.92; sensitivity, 95%; and positive predictive value [PPV], 90%) or combined with diffusion-weighted (DW) imaging (FOM value, 0.93; sensitivity, 98%; and PPV, 88%) resulted in significantly higher reader-averaged, overall diagnostic performance, sensitivity, and PPV for detection of colorectal hepatic metastases than CT (FOM value, 0.82; sensitivity, 85%; and PPV, 73%) (P , .006). nn Gadoxetic acid–enhanced MR imaging combined with DW imaging had significantly higher reader-averaged sensitivity (98% for all metastases and 95% for metastases 1 cm in diameter) for detection of colorectal hepatic metastases than DW imaging alone (88% for all metastases and 79% for metastases 1 cm in diameter) (P .003). nn DW imaging showed a significantly higher sensitivity (79%) and PPV (60%) for detection of small hepatic metastases 1 cm or smaller in diameter than CT (sensitivity, 50%; PPV, 33%) (P .004). nn Gadoxetic acid–enhanced MR imaging helped identify additional hepatic metastases in at least 10 (20%) of 51 patients with potentially curable hepatic metastasis detected with routine staging CT.
Among recent advances in magnetic resonance (MR) imaging techniques, diffusion-weighted (DW) imaging and gadoxetic acid (Primovist, Bayer Schering Pharma, Berlin, Germany; Eovist, Bayer Healthcare Pharmaceuticals, Whippany, NJ)–enhanced MR imaging improved the effectiveness of liver MR imaging for evaluating focal hepatic lesions by providing images with high lesion-to-liver contrast, as well as information in regard to the lesion characteristics (5–11). As a result, these techniques have been increasingly used for clinical liver MR imaging. Despite their frequent use in current clinical practice, there have been only limited data in regard to the diagnostic accuracy of gadoxetic acid–enhanced MR imaging and DW imaging for the evaluation of colorectal hepatic metastases (12–16). All of the previous research on this topic is in retrospective studies, which may be subject to verification and selection bias (12–16). In patients with colorectal cancer, computed tomography (CT) is usually performed for the initial staging workup. Most clinical practice guidelines for colorectal cancer recommend only liver MR imaging as one of the optional worku-ps for patients with potentially surgically curable liver metastases detected with routine staging CT (17,18). However, to our knowledge, no previous study has demonstrated the incremental value of liver MR imaging performed in the clinical setting in which liver MR imaging is indicated. Therefore, the actual clinical benefit of liver MR imaging for the evaluation of colorectal cancer has not yet been well determined. Therefore, the purpose of our study was to prospectively compare the
Implication for Patient Care nn Gadoxetic acid–enhanced MR imaging adds an incremental value to that of CT for detection of additional metastases and can thus be routinely performed in patients with colorectal cancer who have potentially curable hepatic metastases detected with CT.
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
Kim et al
diagnostic performance of DW imaging, gadoxetic acid–enhanced MR imaging, both techniques combined, and CT for detection of colorectal hepatic metastases and to evaluate the incremental value of MR imaging for patients with potentially curable colorectal hepatic metastases detected with CT.
Materials and Methods This prospective study was approved by our institutional review board. All patients provided written informed consent before being enrolled in the study.
Study Population From March 2008 to March 2009, a panel of abdominal radiologists (H.J.K. and A.Y.K., with 5 and 13 years of experience, respectively) and colorectal surgeons (J.C.K. and C.S.Y., with 22 and 15 years of experience, respectively) prospectively evaluated the patients who underwent contrast material–enhanced multidetector CT for the routine preoperative work-up or postoperative follow-up of their colorectal cancer. They also determined each patient’s eligibility for this study according to the following inclusion criteria: (a) The patient was suspected of having one or more hepatic metastases seen on a CT scan, (b)
Published online before print 10.1148/radiol.14140390 Content codes: Radiology 2015; 274:712–722 Abbreviations: DW = diffusion weighted FOM = figure of merit PPV = positive predictive value Author contributions: Guarantors of integrity of entire study, S.S.L., J.C.K.; 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; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, H.J.K., S.S.L., J.H.B., S.H.P.; clinical studies, H.J.K., S.S.L., J.H.B., J.C.K., C.S.Y., A.Y.K.; statistical analysis, S.S.L., J.C.K.; and manuscript editing, H.J.K., S.S.L., J.H.B., J.C.K., S.H.P., H.K.H. Conflicts of interest are listed at the end of this article.
713
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
the patient was a potential candidate for curative hepatic resection, and (c) the patient was willing to participate in this study. Of 501 patients who were evaluated during this period, 97 patients were suspected of having hepatic metastases seen on a CT scan. Among them, 39 patients were excluded because they were not candidates for curative hepatic resection owing to (a) the presence of too many (more than 10) hepatic metastases (n = 20) and (b) the presence of distant metastases other than metastases to the liver (n = 19). A total of 58 eligible patients underwent MR imaging within 30 days after their CT scan (mean interval, 4 days; range, 0–28 days). Among them, seven patients were excluded from our study because of the unavailability of the reference standard required to definitely identify their hepatic lesions (n = 3), the lesions that were suspected of being liver metastases proven to be primary hepatic malignancies at subsequent hepatic resection (n = 3, with hepatocellular carcinoma in one and cholangiocarcinoma in two), and failure to complete the MR examination (n = 1). The remaining 51 patients (39 men, 12 women; mean age, 62 years; range, 41–85 years) constituted the study population. At the time of the study enrollment, 30 of 51 patients were undergoing work-up for already diagnosed rectal (n = 16) or colon (n = 14) cancer; among them, five patients with rectal cancer were being treated with neoadjuvant chemoradiotherapy. Twenty-one of 51 patients were undergoing follow-up after curative resection of primary colorectal cancer, and 18 of them had received adjuvant chemotherapy 1–35 months prior to their enrollment. The flowchart of the study population is presented in Figure 1.
Proof of Hepatic Lesions Forty-seven (92%) of 51 patients in our study population had a total of 104 hepatic metastases, whereas four (8%) patients had no hepatic metastases. The hepatic metastases were proven with histologic analysis for 77 lesions and with imaging findings and follow-up 714
Kim et al
Figure 1
Figure 1: Flowchart of the study population. F/U = follow-up, IOUS = intraoperative ultrasonography (US), RFA = radiofrequency ablation.
for 27 lesions. The mean number of metastases per patient was 2.2 (range, 1–10), with a mean lesion diameter of 1.9 cm 6 1.4 (standard deviation) and a range of 0.3–8.5 cm. Twenty-eight (27%) of the 104 metastases were 1 cm or smaller in diameter. Forty-three of 51 patients underwent curative-intent surgery with manual exploration and intraoperative US of the liver within 30 days (mean interval, 4 days; range, 1–30 days) after MR imaging. One abdominal radiologist (H.J.K., with 5 years of experience in intraoperative US) performed intraoperative US and evaluated the entire liver with a special focus on the lesions that were suspected of being hepatic metastases and that were seen on CT
scans and MR images. In all lesions that were suspected of being hepatic metastases, samples were then removed for pathologic analysis by means of hepatic resection (n = 67) or biopsy (n = 13) with (n = 11) or without (n = 2) radiofrequency ablation. Pathologic examination of surgical or biopsy specimens revealed a total of 77 metastases in 40 patients. In three patients, the lesions that were suspected of being hepatic metastases were proven to be benign, including focal eosinophilic infiltration (n = 2) and focal fibrosis (n = 1). Seven of 51 patients did not undergo resection of their hepatic lesions because of the final clinical decision of their unsuitability for curative hepatic resection that was made on the basis
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
of the hepatic tumor burden, each patient’s general condition, and hepatic reservoir (n = 4); unexpected peritoneal tumor seeding detected at surgery (n = 2); and the patient’s refusal to undergo liver surgery (n = 1). In these patients, a total of 27 hepatic metastases were diagnosed nonpathologically on the basis of the imaging findings and substantial tumor growth (ie, increase in the longest tumor diameter 30%) on follow-up CT scans or MR images (mean follow-up, 3 months; range, 1–9 months). In the remaining one of 51 patients, the lesions that were suspected of being hepatic metastases seen on a CT scan were proven to be focal fat deposition on the basis of the MR imaging findings and their disappearance as seen on a follow-up CT scan obtained within 4 months following the initial CT scan. Thirty-one of 47 patients with hepatic metastases also had a total of 85 benign hepatic lesions, including hepatic cyst (n = 74), hemangioma (n = 10), and focal fat deposition (n = 1). One hemangioma was histologically proven by means of hepatic resection. The other 84 benign lesions in 46 patients were diagnosed on the basis of the MR imaging findings in all patients, as well as on the basis of US findings in 25 patients and/or positron emission tomography (PET)/CT in 24 patients and were reconfirmed by their stability (n = 83) or disappearance (n = 1) on follow-up CT or MR imaging (mean follow-up, 25 months; range, 12–48 months).
CT Examination CT examinations were performed by using 16– or 64–section multidetector CT scanners (Somatom Sensation 16 or 64; Siemens, Erlangen, Germany). The scanning parameters were beam collimation, 16 3 0.75 or 64 3 0.6 mm; 120 kVp; automated dose modulation (CARE dose 4D; Siemens) with the reference tube current set at 200 effective mAs; gantry rotation time, 0.5 seconds; beam pitch, 1.2; and field of view to fit. Portal venous phase images were obtained at a 72-second delay after the initiation of intravenous administration
of 150 mL of iopromide (Ultravist 300; Bayer Schering Pharma) at a rate of 3.0 mL/sec by using an automated power injector (Optivantage DH; Mallinckrodt Imaging Solutions, Hazelwood, Mo). Both axial and coronal images were reconstructed to 5-mm thickness at 5-mm intervals.
MR Imaging MR imaging was performed by using a 1.5-T system (Magnetom Avanto; Siemens) with a dedicated, six-channel, torso-array coil. All patients underwent baseline, nonenhanced MR imaging, including breath-hold T1-weighted two-dimensional dual gradient-echo in-phase and opposed-phase imaging, breathhold heavily T2-weighted half-Fourier single-shot turbo spin-echo imaging, and navigator-triggered intermediate T2-weighted turbo spin-echo imaging. DW imaging was performed by using a navigator-triggered single-shot echoplanar sequence with DW gradients (ie, b factors of 50, 300, 600, and 900 sec/ mm2, applied in three orthogonal directions). After nonenhanced MR imaging, gadoxetic acid–enhanced dynamic T1weighted images were obtained by using a fat-suppressed three-dimensional spoiled gradient-echo sequence, volumetric interpolated breath-hold examination (VIBE; Siemens) after a bolus injection of 10 mL of gadoxetic acid (Primovist; Bayer Schering Pharma) at a rate of 1.0 mL/sec, followed by a 30mL saline flush using a power injector (Spectris Solaris; Medrad, Warrendale, Pa). The images were obtained before contrast agent injection and in the arterial phase, which was determined by using a real-time bolus display method (CARE bolus; Siemens) in the venous phase (50 seconds after contrast agent injection), in the delayed phase (3 minutes following contrast agent injection), and in the hepatobiliary phase (10 minutes and 20 minutes after contrast agent injection). The detailed parameters for the MR sequences are summarized in Table 1. Image Analysis Image interpretation was performed in a retrospective manner after completion
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
Kim et al
of patient enrollment. Two board-certified radiologists (S.S.L. and J.H.B., with 8 and 10 years of experience in abdominal imaging, respectively), who were not involved in the patient selection, intraoperative US, or the construction of the reference standard, independently evaluated the CT and MR image sets. Image analysis was performed during four separate sessions; the readers interpreted the CT images during the first session, DW images and baseline nonenhanced T1- and T2-weighted images (hereafter referred to as DW MR images or imaging) during the second session, gadoxetic acid–enhanced MR images and baseline nonenhanced MR images (hereafter referred to as gadoxetic acid MR images or imaging) during the third session, and all of the MR images (hereafter referred to as combined MR images or imaging) during the fourth session. To reduce recall bias, each image review session was separated by at least 4 weeks, and the images were made anonymous and were randomized. The readers knew that all of the patients had colorectal cancer, although they were unaware of the results of the other review sessions, the surgical findings, pathologic examination results, and the follow-up imaging findings. During each review session, the readers were asked to identify all possible metastatic lesions and to record their size, segmental location (19), and the number of the image where the lesion was contained, and then to grade their confidence level for each lesion on a three-point scale as follows: score 1, equivocal but possibly metastatic; score 2, probably metastatic; and score 3, definitely metastatic. Definite or probable benign lesions were not recorded. The readers characterized focal hepatic lesions by using the known imaging criteria on CT and gadoxetic acid MR images (12,13,20–25). On DW MR images, lesions were characterized on the basis of their signal intensities on T2-weighted images and on DW images obtained with a b value of 50 sec/mm2 and the changes in their signal intensities on DW images obtained with b values of 600 and 900 sec/mm2, according to the previously reported criteria (5,26–29). On combined MR images, 715
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
Table 1 Parameters of MR Pulse Sequences Parameter Repetition time (msec) Echo time (msec) Flip angle (degrees) Field of view Matrix Section thickness (mm) Intersection gap (mm) Echo train length No. of acquisitions Parallel imaging factor‡ Fat suppression used§
T1 Dual Gradient-Echo Imaging
Heavily T2 Half-Fourier Turbo SE Imaging
Intermediate T2 Turbo SE Imaging
DW Imaging
Contrast-enhanced Dynamic T1 and Hepatobiliary Phase Imaging
164 2.3, opposed phase; 4.6, in phase 70 350 3 260 256 3 192 6 1.2 NA 1 2 No
1100 151 150 350 3 260 256 3 192 6 1.2 134 1 0 Yes
4023–5755* 85 150 350 3 260 384 3 288 6 1.2 9 2 2 Yes
4800 73 180 350 3 300 150 3 128 6 1.2 76† 5 2 Yes
4.1 1.5 10 350 3 284 320 3 260 3 0 NA 1 2 Yes
Note.—T1 dual gradient-echo imaging refers to dual gradient-echo in-phase and opposed-phase imaging. NA = not applicable, SE = spin echo, T1 = T1 weighted, T2 = T2 weighted. * Repetition time varied, depending on the patient’s respiratory cycle. †
Echo-planar imaging factor.
‡
Parallel imaging was performed by using a k-space–based technique (GRAPPA; Siemens).
§
Fat saturation was achieved with the chemical shift-selective fat-suppression technique.
hepatic lesions were characterized by using a combination of the aforementioned criteria for DW MR images and gadoxetic acid MR images.
Statistical Analysis The overall diagnostic performance for detecting hepatic metastases was compared among CT, DW MR imaging, gadoxetic acid MR imaging, and combined MR imaging for all metastatic lesions and for metastatic lesions of 1 cm or smaller in diameter by using an alternative free-response receiver operating characteristic analysis (JAFROC1 version 1.0; http://www.devchakraborty. com/index.php) (30). The figure-ofmerit (FOM) value, defined as a probability of a lesion being rated higher than the highest rated nonlesions on normal images (30–32), was used to indicate the overall diagnostic performance of each image set, where the value of one indicated the highest performance and zero indicated the lowest performance. The sensitivities and positive predictive values (PPVs) for the four image sets were compared by using the generalized estimating equations with software (IBM SPSS statistics 19; IBM, New York, NY) to address data clustering (ie, multiple lesions in a single patient). 716
For all statistical analysis in our study, the differences with P values less than the Bonferroni-corrected value of .008 (calculated as .05/6) were considered significant to account for the increased type I error in multiple comparisons.
Results Overall Diagnostic Performance The results of alternative free-response receiver operating characteristic analysis in regard to the overall diagnostic performance for detecting hepatic metastases are summarized in Table 2. For detecting all hepatic metastases, the pooled FOM value was the highest for combined MR imaging (FOM value = 0.93), followed by gadoxetic acid MR imaging (FOM value = 0.92), DW MR imaging (FOM value = 0.89), and CT (FOM value = 0.82). The FOM values of combined MR imaging and gadoxetic acid MR imaging were significantly higher than the FOM value of CT for the pooled data and for both readers (P .006). For detecting hepatic metastases of 1 cm or smaller in diameter, the pooled FOM values of combined MR imaging, gadoxetic acid MR imaging, and DW MR imaging were significantly
higher than the pooled FOM value of CT (P , .001). The pooled FOM value of combined MR imaging was also significantly higher than that of DW MR imaging (P , .001). The FOM value of combined MR imaging for detecting hepatic metastases of 1 cm or smaller in diameter was higher than that of gadoxetic acid MR imaging, with a marginal significant difference only for reader 2 (P = .03).
Sensitivity and PPV For all metastases and metastases of 1 cm or smaller in diameter, the pooled sensitivities were significantly higher with combined MR imaging (98% for all metastases, 95% for metastases 1 cm in diameter) than with CT (85% for all metastases, 50% for metastases 1 cm in diameter) and DW MR imaging (88% for all metastases, 79% for metastases 1 cm in diameter) (P .003). The pooled sensitivities of gadoxetic acid MR imaging (95% for all metastases, 86% for metastases 1 cm in diameter) were also significantly higher than with CT (P , .001) (Table 3). DW MR imaging resulted in a significantly higher pooled sensitivity than CT only for metastases of 1 cm or
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
smaller in diameter (P = .004). The sensitivities of combined MR imaging for all metastases and metastases of 1 cm or smaller in diameter were slightly higher than those of gadoxetic acid MR imaging, with a marginal significant difference for pooled data (P .027) and for reader 2 (P .011) (Figs 2, 3). The pooled PPVs of combined MR imaging, gadoxetic acid MR imaging,
Kim et al
and DW MR imaging were similar to each other, ranging from 84% to 90% for all metastases and from 60% to 72% for the metastases of 1 cm or smaller in diameter, although they were significantly higher than those of CT (73% for all lesions, 33% for lesions 1 cm in diameter) for both all metastases and the metastases that were 1 cm or smaller in diameter (P , .001) (Table 4).
Table 2 FOM Values Indicating the Diagnostic Performance for the Detection of Hepatic Metastases Lesion Group and Imaging Modality All metastases (n = 104) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging Metastases 1.0 cm in diameter (n = 28) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging
Reader 1
Reader 2
Pooled Data
0.80 (0.73, 0.86) 0.88 (0.82, 0.93) 0.93 (0.87, 0.97)* 0.93 (0.87, 0.96)*
0.83 (0.76, 0.89) 0.89 (0.83, 0.93) 0.92 (0.86, 0.95)* 0.94 (0.88, 0.97)*
0.82 (0.75, 0.87) 0.89 (0.84, 0.92) 0.92 (0.88, 0.96)* 0.93 (0.89, 0.96)*
0.59 (0.46, 0.71) 0.82 (0.71, 0.91) 0.92 (0.83, 0.97)* 0.92 (0.84, 0.97)*
0.56 (0.45, 0.68) 0.82 (0.70, 0.91)* 0.85 (0.74, 0.93)* 0.92 (0.83, 0.97)*
0.58 (0.46, 0.68) 0.82 (0.73, 0.90)* 0.89 (0.80, 0.94)* 0.92 (0.85, 0.96)†
Note.—Numbers in parentheses are 95% confidence intervals. * The FOM value is significantly higher than that of CT (P .006). †
The FOM value is significantly higher than that of CT (P < .001). The FOM value is significantly higher than that of DW MR
imaging (P , .001).
Per-Patient Analysis and the Potential Effect of MR Imaging on Patient Treatment In all 47 patients with hepatic metastasis, the presence of hepatic metastasis was correctly diagnosed with CT, gadoxetic acid MR imaging, and combined MR imaging by both readers 1 and 2 and with DW MR imaging by reader 2, thus resulting in a per-patient sensitivity of 100% (47 of 47), while diagnosis with DW MR imaging resulted in a false-negative result of hepatic metastasis in one patient for reader 1. Among 47 patients with hepatic metastases, DW MR imaging, gadoxetic acid MR imaging, and combined MR imaging depicted more metastases than CT in eight, 10, and 10 patients, respectively, for reader 1 and in eight, 10, and 14 patients, respectively, for reader 2 (Figs 2, 3). The detection of additional metastases with combined MR imaging would have actually changed the treatment of at least four patients; in three patients, hepatic wedge resection (n = 2) and intraoperative radiofrequency ablation (n = 1) were performed to treat the hepatic metastases that were detected with combined MR imaging but were missed with CT by both readers, in addition to the hepatic resection (n = 2) or intraoperative
Table 3 Sensitivities for the Detection of Hepatic Metastases Reader 1 Lesion Group and Imaging Modality All metastases CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging Metastases 1.0 cm in diameter CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging
Sensitivity (%)
95% Confidence Interval
87 (90/104) 88 (92/104) 98 (102/104)* 99 (103/104)* 57 (16/28) 75 (21/28) 93 (26/28)* 96 (27/28)*
Reader 2
Pooled Data
Sensitivity (%)
95% Confidence Interval
Sensitivity (%)
95% Confidence Interval
79, 92 81, 93 93, 100 93, 100
84 (87/104) 88 (92/104) 92 (96/104) 97 (101/104)*
75, 90 81, 93 85, 96 92, 99
85 88 95* 98†
80, 89 83, 92 91, 97 95, 99
39, 74 56, 88 76, 99 81,100
43 (12/28) 82 (23/28) 79 (22/28)* 93 (26/28)*
26, 61 64, 93 60, 90 76, 99
50 79* 86* 95†
37, 63 66, 87 74, 93 85, 99
Note.—Numbers in parentheses are the numbers of lesions, which were used for calculating the percentages. Percentages were rounded. * The sensitivity is significantly higher than that of CT (P .003). †
The sensitivity is significantly higher than that of CT (P < .001). The sensitivity is significantly higher than that of DW MR imaging for both all metastases and metastases of 1 cm or smaller in diameter (P .003).
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
717
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
without hepatic metastasis, true-negative interpretations were made with CT, DW MR imaging, gadoxetic acid MR imaging, and combined MR imaging in zero, two, zero, and zero patients, respectively, for reader 1 and zero, one, one, and two patients, respectively, for reader 2.
Figure 2
Discussion
Figure 2: Multisegmental involvement of hepatic metastases from rectal cancer not seen on a CT scan in a 68-year-old man. A, Transverse contrast-enhanced portal venous phase CT images show a hypoattenuating lesion (arrow) in hepatic segment IV on middle image. No metastatic lesions were observed on left and right images. B, Transverse gadoxetic acid–enhanced hepatobiliary phase images and, C, DW images (b = 600 sec/mm2) demonstrate two lesions in hepatic segments VII (solid arrow) on left images and III (arrowhead) on right images, in addition to the lesion in segment IV (open arrow), which was detected on the CT scan, on middle images. With gadoxetic acid–enhanced MR images alone (B), the readers missed the tiny lesion in segment III (arrowhead) on right image. With DW images alone and with combined DW images and gadoxetic acid–enhanced MR images, all three lesions were identified by both readers.
radiofrequency ablation to cover the lesions detected with CT (Fig 2). In one patient, the additional metastases depicted with combined MR imaging 718
precluded surgical resection, leading to a change in the treatment plan from surgical resection to chemotherapy for this patient. In four patients
Our study results demonstrated that gadoxetic acid–enhanced MR imaging alone or in combination with DW imaging had a significantly higher reader-averaged, overall diagnostic performance, sensitivity, and PPV for depicting colorectal hepatic metastases, compared with CT. Unlike previous studies, which reported a comparable or even higher diagnostic performance of DW imaging for depicting hepatic metastases compared with gadopentetate dimeglumine–enhanced or superparamagnetic iron oxide–enhanced MR imaging (5,24), our study demonstrated somewhat different results; DW MR imaging showed lower overall diagnostic performance, sensitivity, and PPV compared with combined MR imaging and gadoxetic acid MR imaging, with a significant difference in reader-averaged FOM values and sensitivities for depicting metastases of 1 cm or smaller in diameter, compared with combined MR imaging. Investigators in several retrospective studies (12,13,25) also reported findings consistent with those of our study. The addition of DW imaging to gadoxetic acid–enhanced MR imaging slightly improved the overall diagnostic performance and sensitivity for detecting colorectal hepatic metastases, especially for lesions of 1 cm or smaller in diameter, compared with gadoxetic acid–enhanced MR imaging alone, although these improvements were not significantly different. Our results, taken together with the results of previous studies, suggest that combined gadoxetic acid–enhanced MR imaging and DW imaging would be the most preferred imaging technique for the evaluation of colorectal hepatic metastases among the techniques evaluated in our study.
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
Figure 3
Figure 3: Multiple hepatic metastases from colon cancer in a 68-year-old man. A, Transverse contrast-enhanced portal venous phase CT image shows a hypoattenuating metastatic lesion in hepatic segment VIII (arrow) but no focal lesion in the left hepatic lobe. B, Transverse DW images (b = 900 sec/mm2) demonstrate a tiny hyperintense lesion in hepatic segment III (arrowhead) in addition to the lesion in hepatic segment VIII (arrow) seen on the CT scan. C, Transverse gadoxetic acid–enhanced MR images show two hypointense lesions (arrowhead and arrow), which correspond to the lesions shown on DW images (B). D, Follow-up contrast-enhanced CT image obtained 2 months following the initial CT scan demonstrates obvious growth of the two hepatic lesions (arrowhead and arrow), confirming the diagnosis of hepatic metastases.
Although it is well known that contrast-enhanced liver MR imaging is superior to CT for detection of colorectal liver metastases, especially for detection of small metastases that are 1 cm or smaller in diameter (11,20,33), it has not been clearly determined whether the improved accuracy of liver MR imaging leads to an actual clinical benefit for treating patients with colorectal cancer. Our results suggest that gadoxetic acid–enhanced MR imaging may provide substantial clinical benefits for
patients with potentially curable hepatic metastases depicted with CT by depicting a greater number of hepatic metastases and thus validating the current guidelines, which recommend liver MR imaging for these patients (17,18). In our study, combined MR imaging allowed readers 1 and 2 to identify 23 and 24, respectively, more hepatic metastases (including 14 and 11, respectively, more liver metastases 1 cm in diameter) than they did with CT, while this protocol resulted in fewer false-positive lesions (12 and
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
six, respectively, for readers 1 and 2) than CT (41 and 26, respectively, for readers 1 and 2). With per-patient– based analyses, combined MR imaging allowed the detection of more hepatic metastases in at least 10 (20%) of 51 patients, and which would have resulted in actual change in treatment in four of them. However, liver MR imaging may not be effective for proving the benign nature of lesions suspected of being hepatic metastases that are seen on CT scans. In four patients whose lesions were suspected 719
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
Kim et al
Table 4 PPVs for the Detection of Hepatic Metastases Reader 1 Lesion Group and Imaging Modality All metastases (n = 104) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging Metastases 1.0 cm in diameter (n = 28) CT DW MR imaging Gadoxetic acid MR imaging Combined MR imaging
PPV (%)
95% Confidence Interval
69 (90/131) 87 (92/106)* 92 (102/111)* 90 (103/115)* 31 (16/51) 66 (21/32)* 79 (27/34)* 76 (28/37)*
Reader 2
Pooled Data
PPV (%)
95% Confidence Interval
PPV (%)
95% Confidence Interval
60, 76 79, 92 85, 96 83, 94
77 (87/113) 81 (92/113) 87 (96/110) 86 (101/117)
68, 84 73, 88 80, 92 79, 92
73 84* 90* 88*
67, 78 79, 88 85, 93 83, 93
20, 45 48, 80 63, 80 60, 87
36 (13/36) 55 (23/42) 65 (22/34)* 65 (26/40)*
22, 52 40, 69 48, 79 49, 78
33 60* 72* 70*
24, 44 49, 71 60, 81 59, 79
Note.—Numbers in parentheses are the numbers of lesions, which were used for calculating the percentages. Percentages were rounded. * The PPV is significantly higher than that of CT (P .001).
of being hepatic metastases and were depicted on CT scans were proven to be benign lesions, including eosinophilic infiltration (n = 2), focal fibrosis (n = 1), and focal fat deposition (n = 1), combined MR imaging, gadoxetic acid MR imaging, and DW MR imaging could not consistently help diagnose these lesions as benign. In our study, we did not compare the diagnostic performance of liver MR imaging techniques with the diagnostic performance of other imaging examinations, such as dynamic contrastenhanced CT and PET/CT. Therefore, our results cannot directly address the question of which examination would be the most appropriate as a further work-up for patients with potentially curable hepatic metastases. However, the results of previous studies showing higher sensitivities of gadoxetic acid– enhanced MR imaging in the detection of hepatic metastases compared with those of dynamic contrast-enhanced CT (7) or PET/CT (15) suggest that gadoxetic acid–enhanced MR imaging would be preferable to these examinations for the evaluation of hepatic metastases. In addition to its higher diagnostic performance, gadoxetic acid–enhanced MR imaging has an additional advantage in its lack of radiation exposure compared with dynamic contrast-enhanced CT and PET/CT, 720
both of which are subject to the potential hazards of ionizing radiation. Our study had several limitations. First, as it included only patients who were suspected of having colorectal hepatic metastases as seen on CT scans, our results cannot be generalized for all patients with colorectal cancer. For those patients with colorectal cancer in whom CT shows no focal hepatic lesion or only hepatic lesions deemed too small to be characterized on a CT scan, it is still unknown whether further evaluation of these patients with the use of gadoxetic acid–enhanced liver MR imaging can offer an important clinical benefit. Second, our study population may have had a higher prevalence of hepatic metastases than the population of all patients with colorectal cancer. Therefore, the PPVs of the imaging examinations may have been overestimated in our study. Although a study population enrolled from among general colorectal cancer patients may have led to unbiased results, this would require a much larger number of patients than we included in our study to comprise a sufficient number of hepatic metastases for statistical analysis. Therefore, we instead decided to focus our study on the clinical setting in which liver MR imaging is recommended according to the current guidelines (17,18). Third,
not all of the hepatic metastases that were detected in our study were pathologically proven, and most of the benign hepatic lesions were nonpathologically diagnosed on the basis of the imaging findings and follow-up results. However, except for the lesions in patients who were determined to be unsuitable for liver resection after their study enrollment, most (74%, 77 of 104) of the hepatic metastases were pathologically proven in our study. As the pathologic confirmation of all focal hepatic lesions for research purposes would not be practical or ethical, this limitation may have been unavoidable in our study. Finally, there may be a possibility that our CT protocol using a greater section thickness (5 mm) than the section thickness for gadoxetic acid–enhanced MR images (3 mm) may have been partially responsible for a worse diagnostic performance of CT than MR imaging. However, the researchers in a recent study (34) with the use of 3-mm-thick CT images also reported results consistent with those of our study (ie, a higher sensitivity of gadoxetic acid–enhanced MR imaging than CT in the detection of colorectal hepatic metastases). In conclusion, gadoxetic acid–enhanced MR imaging alone or in combination with DW imaging is more accurate than CT for detection of
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
colorectal hepatic metastases. The addition of DW imaging to gadoxetic acid–enhanced MR imaging slightly improves the sensitivity for detection of small metastases. Gadoxetic acid–enhanced MR imaging adds incremental value to CT for detection of additional metastases and thus can be routinely performed for patients with potentially curable hepatic metastases seen at CT. Disclosures of Conflicts of Interest: H.J.K. disclosed no relevant relationships. S.S.L. disclosed no relevant relationships. J.H.B. disclosed no relevant relationships. J.C.K. disclosed no relevant relationships. C.S.Y. disclosed no relevant relationships. S.H.P. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: received grants from DongKook Pharm and GE Healthcare. Other relationships: disclosed no relevant relationships. A.Y.K. disclosed no relevant relationships. H.K.H. disclosed no relevant relationships.
References 1. Wong SL, Mangu PB, Choti MA, et al. American Society of Clinical Oncology 2009 clinical evidence review on radiofrequency ablation of hepatic metastases from colorectal cancer. J Clin Oncol 2010;28(3):493–508. 2. Adam R, Aloia T, Lévi F, et al. Hepatic resection after rescue cetuximab treatment for colorectal liver metastases previously refractory to conventional systemic therapy. J Clin Oncol 2007;25(29):4593–4602. 3. Chong G, Cunningham D. Improving longterm outcomes for patients with liver metastases from colorectal cancer. J Clin Oncol 2005;23(36):9063–9066. 4. Zech CJ, Grazioli L, Jonas E, et al. Healtheconomic evaluation of three imaging strategies in patients with suspected colorectal liver metastases: Gd-EOB-DTPA-enhanced MRI vs. extracellular contrast media-enhanced MRI and 3-phase MDCT in Germany, Italy and Sweden. Eur Radiol 2009;19(suppl 3):S753–S763.
Kim et al
7. Motosugi U, Ichikawa T, Morisaka H, et al. Detection of pancreatic carcinoma and liver metastases with gadoxetic acid-enhanced MR imaging: comparison with contrastenhanced multi-detector row CT. Radiology 2011;260(2):446–453.
16. Kim YK, Lee MW, Lee WJ, et al. Diagnostic accuracy and sensitivity of diffusion-weighted and of gadoxetic acid-enhanced 3-T MR imaging alone or in combination in the detection of small liver metastasis ( 1.5 cm in diameter). Invest Radiol 2012;47(3):159–166.
8. Ichikawa T, Saito K, Yoshioka N, et al. Detection and characterization of focal liver lesions: a Japanese phase III, multicenter comparison between gadoxetic acid disodium-enhanced magnetic resonance imaging and contrast-enhanced computed tomography predominantly in patients with hepatocellular carcinoma and chronic liver disease. Invest Radiol 2010;45(3):133–141.
17. Glimelius B, Påhlman L, Cervantes A; ESMO Guidelines Working Group. Rectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2010;21(suppl 5):v82–v86.
9. Hammerstingl R, Huppertz A, Breuer J, et al. Diagnostic efficacy of gadoxetic acid (Primovist)-enhanced MRI and spiral CT for a therapeutic strategy: comparison with intraoperative and histopathologic findings in focal liver lesions. Eur Radiol 2008;18(3): 457–467. 10. Huppertz A, Balzer T, Blakeborough A, et al. Improved detection of focal liver lesions at MR imaging: multicenter comparison of gadoxetic acid-enhanced MR images with intraoperative findings. Radiology 2004;230(1): 266–275. 11. Niekel MC, Bipat S, Stoker J. Diagnostic imaging of colorectal liver metastases with CT, MR imaging, FDG PET, and/or FDG PET/CT: a meta-analysis of prospective studies including patients who have not previously undergone treatment. Radiology 2010;257(3):674–684. 12. Löwenthal D, Zeile M, Lim WY, et al. Detection and characterisation of focal liver lesions in colorectal carcinoma patients: comparison of diffusion-weighted and GdEOB-DTPA enhanced MR imaging. Eur Radiol 2011;21(4):832–840. 13. Chung WS, Kim MJ, Chung YE, et al. Comparison of gadoxetic acid-enhanced dynamic imaging and diffusion-weighted imaging for the preoperative evaluation of colorectal liver metastases. J Magn Reson Imaging 2011;34(2):345–353.
5. Nasu K, Kuroki Y, Nawano S, et al. Hepatic metastases: diffusion-weighted sensitivityencoding versus SPIO-enhanced MR imaging. Radiology 2006;239(1):122–130.
14. Muhi A, Ichikawa T, Motosugi U, et al. Diagnosis of colorectal hepatic metastases: comparison of contrast-enhanced CT, contrast-enhanced US, superparamagnetic iron oxide-enhanced MRI, and gadoxetic acid-enhanced MRI. J Magn Reson Imaging 2011;34(2):326–335.
6. Parikh T, Drew SJ, Lee VS, et al. Focal liver lesion detection and characterization with diffusion-weighted MR imaging: comparison with standard breath-hold T2weighted imaging. Radiology 2008;246(3): 812–822.
15. Donati OF, Hany TF, Reiner CS, et al. Value of retrospective fusion of PET and MR images in detection of hepatic metastases: comparison with 18F-FDG PET/CT and Gd-EOB-DTPA-enhanced MRI. J Nucl Med 2010;51(5):692–699.
Radiology: Volume 274: Number 3—March 2015 n radiology.rsna.org
18. Garden OJ, Rees M, Poston GJ, et al. Guidelines for resection of colorectal cancer liver metastases. Gut 2006;55(suppl 3): iii1–iii8. 19. Dodd GD 3rd. An American’s guide to Couinaud’s numbering system. AJR Am J Roentgenol 1993;161(3):574–575. 20. Onishi H, Murakami T, Kim T, et al. He patic metastases: detection with multi-detector row CT, SPIO-enhanced MR imaging, and both techniques combined. Radiology 2006;239(1):131–138. 21. Rappeport ED, Loft A, Berthelsen AK, et al. Contrast-enhanced FDG-PET/CT vs. SPIOenhanced MRI vs. FDG-PET vs. CT in patients with liver metastases from colorectal cancer: a prospective study with intraoperative confirmation. Acta Radiol 2007;48(4): 369–378. 22. Ward J, Robinson PJ, Guthrie JA, et al. Liver metastases in candidates for hepatic resection: comparison of helical CT and gadolinium- and SPIO-enhanced MR imaging. Radiology 2005;237(1):170–180. 23. Goshima S, Kanematsu M, Watanabe H, et al. Hepatic hemangioma and metastasis: differentiation with gadoxetate disodiumenhanced 3-T MRI. AJR Am J Roentgenol 2010;195(4):941–946. 24. Hardie AD, Naik M, Hecht EM, et al. Diagnosis of liver metastases: value of diffusion-weighted MRI compared with gadolinium-enhanced MRI. Eur Radiol 2010;20(6): 1431–1441. 25. Shimada K, Isoda H, Hirokawa Y, Ari zono S, Shibata T, Togashi K. Comparison of gadolinium-EOB-DTPA-enhanced and diffusion-weighted liver MRI for detection of small hepatic metastases. Eur Radiol 2010;20(11):2690–2698. 26. Coenegrachts K, ter Beek L, Haspeslagh M, Bipat S, Stoker J, Rigauts H. Comparison of respiratory-triggered T2-weighted turbo spin-echo imaging versus breath-hold T2weighted turbo spin-echo imaging: distinguishing benign from malignant liver lesions in patients with colorectal cancer. JBR-BTR 2009;92(4):195–201.
721
GASTROINTESTINAL IMAGING: Incremental Liver MR in Curable Colorectal Hepatic Metastasis
27. Low RN, Gurney J. Diffusion-weighted MRI (DWI) in the oncology patient: value of breathhold DWI compared to unenhanced and gadolinium-enhanced MRI. J Magn Reson Imaging 2007;25(4):848–858. 28. Miller FH, Hammond N, Siddiqi AJ, et al. Utility of diffusion-weighted MRI in distinguishing benign and malignant hepatic lesions. J Magn Reson Imaging 2010;32(1):138–147. 29. Taouli B, Koh DM. Diffusion-weighted MR imaging of the liver. Radiology 2010;254(1): 47–66.
722
Kim et al
30. Chakraborty DP. Recent advances in observer performance methodology: jackknife free-response ROC (JAFROC). Radiat Prot Dosimetry 2005;114(1-3):26–31. 31. Chakraborty DP. New developments in observer performance methodology in medical imaging. Semin Nucl Med 2011;41(6): 401–418. 32. Chakraborty DP, Berbaum KS. Observer studies involving detection and localization: modeling, analysis, and validation. Med Phys 2004;31(8):2313–2330.
33. Bartolozzi C, Donati F, Cioni D, et al. Detection of colorectal liver metastases: a prospective multicenter trial comparing unenhanced MRI, MnDPDP-enhanced MRI, and spiral CT. Eur Radiol 2004;14(1):14–20. 34. Scharitzer M, Ba-Ssalamah A, Ringl H, et al. Preoperative evaluation of colorectal liver metastases: comparison between gadoxetic acid-enhanced 3.0-T MRI and contrast-enhanced MDCT with histopathological correlation. Eur Radiol 2013;23(8): 2187–2196.
radiology.rsna.org n Radiology: Volume 274: Number 3—March 2015
