- Email: [email protected]
Liver steatosis assessed by preoperative MRI: An independent risk factor for severe complications after major hepatic resection
ARTICLE IN PRESS
Liver steatosis assessed by preoperative MRI: An independent risk factor for severe complications after major hepatic resection Gaspard d’Assignies, MD, PhD,a,b,c Cindy Fayard, MD,a,b,c Helena Leitao, MD, PhD,a,b,c Toni Alfaiate, MD,d,e Florence Tubach, MD, PhD,b,d,e Safi Dokmak, MD, PhD,f Valerie Paradis, MD, PhD,b,c,g Bernard E. Van Beers, MD, PhD,a,b,c Maxime Ronot, MD, PhD,a,b,c and Valerie Vilgrain, MD, PhD,a,b,c Clichy, Hauts-de-Seine and Paris, France
Background. Steatosis assessed by histology is commonly considered to be a significant risk factor for liver surgery. MRI is considered as the new gold standard for noninvasive liver fat quantification. The purpose was to assess whether liver steatosis determined by preoperative MRI is an independent risk factor of complications after major liver resection. Methods. All patients who underwent liver MRI before major liver resection in our institution between January 2001 and December 2011 were included in this retrospective study. The liver fat fraction (LFF) was assessed on in- and opposed-phase T1-weighted dual echo gradient echo MRI and steatosis was defined as a MRI LFF of $5%. The association between steatosis and postoperative complications (Clavien-Dindo classification, ascites >500 mL at day 5, 50-50 criteria, fistula/collection, blood liver test alterations, pulmonary complications, nonpulmonary complications, >1 complication, duration of stay in the intensive care unit, duration of hospital stay, and death) was assessed by multivariate analysis using the appropriate model. Results. A MRI LFF of $5% was associated with severe postoperative complications (Clavien-Dindo score $ IIIa; P = .04), more pulmonary complications (P = .02), and longer duration of hospital stay (P = .02) on the multivariate model adjusted for confounding factors. The postoperative aminotransferase levels were higher in patients with a MRI LFF of $5%, than in other patients (P = .0008). Conclusion. Liver steatosis assessed by routine preoperative MRI is shown to be an independent risk factor of severe postoperative complications after major liver resection. (Surgery 2015;j:j-j.) From the Department of Radiology,a APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine; University Paris Diderot,b Sorbonne Paris Cite , Paris; INSERM U1149,c centre de recherche biome dicale Bichat-Beaujon, CRB3, Paris; De partement d’Epide miologie et Recherche Clinique,d APHP, University Hospitals Paris Nord Val de Seine, Bichat, Paris; INSERM CIC 1425-EC,e UMR 1123 ECEVE, Paris; Department of Surgery,f APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hautsde-Seine; and Department of Pathology,g APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France
NONALCOHOLIC FATTY LIVER DISEASE (NAFLD) includes a wide spectrum of liver damage, ranging from simple steatosis to steatohepatitis and cirrhosis. Although it was unrecognized before the 1980s, the increasing epidemic of obesity has increased awareness of NAFLD, which is currently considered Authors have no conflict of interest. Accepted for publication October 6, 2015. Reprint requests: Maxime Ronot, MD, PhD, H^ opital Beaujon, Department of Radiology, 100 boulevard du, General Leclerc, 92100 Clichy, France. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.10.008
to be the most common cause of chronic liver disease affecting #40% of the general adult population in the United States.1 Thus, the incidence of steatosis during liver surgery has increased significantly. Furthermore, steatosis and steatohepatitis have been associated with irinotecan-based regimens used in patients with colorectal liver metastases whose curative treatment is liver surgery. NAFLD also increases the risk of developing telangiectatic/inflammatory hepatocellular adenoma and hepatocellular carcinoma, both tumors that may require operative resection.2 Steatosis is considered commonly to be a significant risk factor for liver surgery, especially in patients undergoing major resection.3-6 One SURGERY 1
ARTICLE IN PRESS 2 d’Assignies et al
systematic review and metaanalysis has shown a significant association between the degree of steatosis and an increased risk of postoperative complications and mortality.4 All these studies used semiquantitative pathologic examination of the surgical specimen to quantify the degree of steatosis. Indeed, knowing the degree of preoperative steatosis could alter patient management. To our knowledge, only 1 study has shown that patients with moderate to severe steatosis (liver fat content >10% on MRI) have a greater risk of life-threatening complications after major liver resection than patients without steatosis or with mild steatosis.7 Nevertheless, that study did not evaluate the effect of possible confounding factors such as age, American Society of Anesthesiologists (ASA) score, body mass index >25 kg/m2, diabetes, and liver fibrosis. Presently, MRI is the most sensitive imaging modality for detecting liver steatosis and MR spectroscopy (MRS) is the most accurate method.8-10 MRS provides information about the chemical composition in an organ and can quantify small amounts of fat. However, MRS of the liver is usually only used in research protocols. In contrast, the breath-hold dual-echo chemical shift gradient echo (GRE) sequence is technically simple and systematically included in MRI protocols of the liver to estimate liver fat. The purpose of this retrospective study was to assess whether liver steatosis of any grade, determined by preoperative dual-echo chemical shift GRE MRI is an independent risk factor of complications after major liver resection. METHODS Patients. The review board at our institution approved this single-center, retrospective clinical study and informed consent was waived. From January 2001 to December 2011, all consecutive patients >18 years old, scheduled for major liver resection ($3 liver segments), who underwent preoperative liver MRI were identified in the prospective database of the department of hepatobiliary surgery. Based on these criteria, 171 patients were included. Exclusion criteria were multiple liver surgeries and living donors (n = 11) or liver MRI >120 days before liver resection (n = 35). Preoperative volumetric calculations were performed and patients were only operated on if the volume of the future liver remnant was acceptable. Patients underwent major liver resection for colorectal liver metastases in 36 patients (29%), hepatocellular carcinomas in 20 (16%), cholangiocarcinomas in 28 (22%), other malignant
Surgery j 2015
lesions in 27 (22%), and benign liver lesions in 14 (11%). The following preoperative and perioperative data were recorded: patient age, gender, ASA score, body mass index >25 kg/m2, presence of significant liver fibrosis (F0-F1 vs F2-F4), comorbidities such as diabetes mellitus, type of resection (left or right hepatectomy), duration of surgery, vascular clamping, biliary resection–reconstruction, and vascular resection–reconstruction. The right side resection included right hepatectomy extended to segments I and/or IV and/or wedge resection within the left liver. Intraoperative and postoperative variables were recorded as estimated blood loss, transfusion, liver-specific complications, blood liver tests, pulmonary complications, nonpulmonary complications and classified according to the Clavien-Dindo classification.11 Reported blood liver tests were aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels at postoperative day 1, bilirubin levels at postoperative day 5, and maximum peak bilirubin levels. Liver-specific complications were defined as any of the following: biliary fistula/collections requiring percutaneous drainage, liver failure according to 50/50 criteria, defined as the association of prothrombin time <50% and serum bilirubin >50 mL/L on postoperative day 5,12 ascites (>500 mL at day 5). Pulmonary complications included pulmonary embolism, acute respiratory failure, pleural effusion, and pneumonia. Nonpulmonary complications included all other medical major events such as urinary infection, acute renal failure, cardiac arrhythmia, or bowel occlusion. The durations of the intensive care unit (ICU) and hospital stay were recorded and postoperative mortality was defined as any death in the hospital or within 90 days after surgery. Postoperative complications were categorized in 7 grades according to the Clavien-Dindo classification.11 Complications requiring a surgical, endoscopic or radiologic intervention (grade IIIa or higher) were considered severe complications.11,13 MRI. MRI was performed with a 1.5-T MR scanner (Intera, Philips Medical Systems, Best, The Netherlands), with a maximum gradient strength of 40 mT/m and a slew rate of 200 mT/m/ms. A 4-element surface coil was used. Our routine MRI protocol for the liver systematically included a dual echo GRE sequence to quantify liver fat with the following parameters: repetition time/echo time 145/2.3 (in-phase) and 4.6 ms (opposed-phase); matrix 256 3 256 pixels, 213 3 305-mm2 field of view, slice thickness 6 mm, 2 averages, and 808 flip angle. The acquisition time was 25 s.
ARTICLE IN PRESS d’Assignies et al 3
Surgery Volume j, Number j
Histologic analysis. Liver resection specimens were evaluated in all patients by an experienced hepatobiliary pathologist (V.P.) who was unaware of the imaging results and surgical outcomes. The degree of liver steatosis was graded according to the Brunt classification14: mild (5–33%), moderate (34–66%), or severe (67–100%). Fibrosis was assessed according to the METAVIR scoring system15 and then classified into 2 groups, with and without advanced fibrosis (F # 2 and F $ 3). Image interpretation. To measure liver fat fraction (LFF), 3 circular regions of interest with a mean area of 5.5 ± 0.4 cm2 were drawn by a radiologist with 5 years experience in abdominal imaging (C.F.; Figs 1 and 2) on the in-phase/ opposed-phase dual echo GRE MRI, in 3 different areas of the liver (the right anterior segment, the right posterior segment, and the left lobe) avoiding major blood vessels, as 3 measurements have been shown highly reliable.16 Small areas of focal fatty sparing or focal fatty deposition near the hilum were avoided. If more diffuse heterogeneous steatosis (such as lobar distribution) was seen on opposed-phase GRE T1-weighted MR sequence, regions of interest were placed in the more fatty segments. Liver signal intensities (in phase [SIiP] and opposed phase [SIoP]) were measured using a commercial software viewer. The LFF was calculated in each voxel with the following equation: LFFð%Þ ¼ ðSIiP SIoP Þ=2 SIiP
½1
The fat fraction for each patient was considered to be the average of the 3 measurements. At MRI, the patients were considered to have liver steatosis when their LFF was $5%. This threshold was chosen in accordance to previous cohort study.17 No receiver operating characteristic curve analysis was performed. Statistical analysis. Patient characteristics were described using frequencies and percentages for categorical variables, and means and standard deviations for continuous variables. Preoperative, intraoperative, and perioperative characteristics associated with the presence of LFF of $5% were identified by univariate analysis (Student t test or Wilcoxon test for continuous variables, and Chisquare or Fisher exact test for categorical variables). The correlation and association between MR LFF and histologic grade of steatosis was assessed using the Spearman correlation coefficient and Kruskall–Wallis test, respectively. The association between LFF on MRI of $5% and each perioperative, intraoperative, and postoperative
variable (blood loss, transfusion, Clavien-Dindo score, ascites >500 mL at day 5, 50/50 criteria, biliary fistula/collection, AST at day 1, ALT at day 1, bilirubin at day 5, peak bilirubin, pulmonary complication, nonpulmonary complication, >1 complication, duration of ICU stay, duration of hospital stay, and death) was assessed using adjusted models, namely analyses of variance when the dependent outcome (postoperative complication) was continuous, logistic regression when the dependent outcome was dichotomous, and multinomial logistic regression when the dependent outcome was categorical and had >2 levels. For each adjusted model, variables to be considered were identified by univariate analysis. For a given postoperative complication, the association between the postoperative complication and each potential adjustment variable (age, ASA score, body mass index >25 kg/m2, diabetes, liver fibrosis >F2, vascular reconstruction, biliary reconstruction, vascular clamping, right or left hepatectomy, duration of surgery) was tested using the Student t test or Wilcoxon test, Chi-square, or Fisher exact test, Pearson or Spearman correlation coefficient. Potential adjustment variables with a P of <.05 were entered into the adjusted model. We also performed this analysis to evaluate the association between the MRI LFF of $5% and each postoperative complication in the subgroup of patients who underwent right side resection (analysis planned a priori). The association between MRI LFF and perioperative and postoperative continuous variables (duration of surgery, perioperative blood loss, AST, ALT, bilirubin at day 5, duration of ICU stay, and duration of postoperative stay) was assessed by the Spearman correlation coefficient. Statistical analyses were performed with SAS software version 9.2 (Institute Inc, Cary, NC). RESULTS The study sample included 125 patients (59 women, and 66 men) with a mean age of 54.6 ± 14.7 years (54.4 ± 13.4 for women; 54.8 ± 15.8 for men). No diffuse heterogeneous steatosis was seen on opposed-phase GRE T1weighted MR sequence. Therefore, the regions of interest were placed in 3 different areas of the liver (the right anterior segment, the right posterior segment, and the left lobe) avoiding major blood vessels. The demographic, preoperative intraoperative, and perioperative data of all patients and patients with an LFF of <5 or $5% on MRI are shown in Table I. No differences were found between
ARTICLE IN PRESS 4 d’Assignies et al
Surgery j 2015
Fig 1. In-phase (left) and opposed-phase (right) dual echo gradient echo MRI of an 18-year-old man with a large telangiectatic adenoma of the right liver. Fat fraction is calculated within the regions of interest (white circle). The preoperative MRI liver fat fraction was 34.6%. Right side resection was performed. Clavien-Dindo IIIb postoperative complications were reported. The duration of intensive care stay was 19 days and hospital stay, 32 days.
Fig 2. In-phase (left) and opposed-phase (right) dual echo gradient echo MRI of a 43-year-old man with a large post– hepatitis B cirrhosis hepatocellular carcinoma of the right liver. The fat fraction is calculated in the regions of interest (white circle). The preoperative MRI liver fat fraction was <5%. Right side resection was performed. No postoperative complications were reported. The duration of intensive care stay was 1 day and the hospital stay, 6 days.
patients with an LFF of <5 or $5% on MRI at baseline. The Spearman correlation coefficient between the MRI LFF and the histologic grade of steatosis was significant (r = 0.558; P < .0001). Table II shows the relationship between MRI LFF and the histologic grade of steatosis. An MRI LFF of $5% adjusted for confounding factors was associated with severe postoperative complications (Clavien-Dindo score $ IIIa; P = .042; Table III; Figs 1 and 2). Furthermore, pulmonary complications were more frequent (P = .021) and the duration of hospital stay was greater (P = .022) for patients with an MRI LFF of $5%. In the subgroup of patients with right hepatectomy, an MRI LFF of $5% was associated with severe postoperative complications (P = .009; Table IV). After right hepatectomy, patients who had an MRI LFF of $5% more frequently had $1 complication (P = .033) and the duration of
hospital stay was greater than for patients without steatosis (P = .025). Significant but modest correlations were found between increased AST and ALT at day 1 and LFF assessed on MRI (r = 0.19 [P = .045] and r = 0.28 [P = .002], respectively). There were stronger correlations in the subgroup with a right side resection (r = 0.38 [P < .001] and r = 0.39 [P < .001] for ALT and AST, respectively). No correlation was found between the MRI LFF and serum bilirubin at Day 5 (P = .381). There was a trend (although not significant) for an increase in perioperative blood loss in patients with an MRI LFF of $5%. DISCUSSION In our study, liver steatosis quantified by preoperative MRI was found to be an independent risk factor for complications after major liver resection. Patients with an MRI LFF of $5% were found to
ARTICLE IN PRESS d’Assignies et al 5
Surgery Volume j, Number j
Table I. Demographic and perioperative data of patients according to MRI liver fat fraction measurements
Characteristic Preoperative Age (y), mean [SD] Sex ratio (M/F) Body mass index > 25, n (%) Diabetes, n (%) ASA grade, n (%) I II III Perioperative Right hepatectomy,* n (%) Biliary reconstruction, n (%) Vascular reconstruction, n (%) Clamping, n (%) Duration of surgery (min), mean [SD] Liver fibrosis >F2, n (%)
All patients (n = 125)
Patients with MRI liver fat fraction <5% (n = 100)
Patients with MRI liver fat fraction $5% (n = 25)
55 [15] 66/59 38 (32) 10 (8)
54 [15] 52/48 28 (30) 8 (8)
56 [14] 14/11 10 (40) 2 (8)
12 (9.6) 107 (85.6) 6 (4.8)
11 (11) 83 (83) 6 (6)
1 (4) 24 (96) 0 (0)
81 (65) 13 (10) 18 (14) 48 (39) 353.0 [103.3] 17 (14)
67 11 15 38 346.7 14
(67) (11) (15) (39) [101.2] (14)
14 2 3 10 377.2 3
(56) (8) (12) (40) [109.9] (12)
P value .581 .722 .347 .630 .319
.303 1.000 1.000 .940 .200 1.00
*Right side resection including 65 right hepatectomy and 26 right hepatectomy extended to segments I and/or IV and/or wedge resections within the left liver. ASA, American Society of Anesthesiologists; SD, standard deviation.
Table II. Relation between MRI liver fat fraction and histologic grade of steatosis Histologic grade of steatosis 0 1 2 3
n (%) 70 42 8 4
(56.8) (33.6) (6.4) (3.2)
MRI liver fat fraction mean % (SD) 0.9 3.1 16.2 19
(0.02) (0.03) (0.12) (0.14)
P value* <.001
*Kruskall–Wallis test. SD, Standard deviation.
have higher Clavien-Dindo scores and greater durations of hospital stay. These patients were also more prone to pulmonary complications. Data on the correlation between liver steatosis and perioperative complications are scarce and contradictory.3,5,18-21 However, a recent systematic review and metaanalysis that included 6 studies with a total of 1,000 patients with major liver resection4 has shown an increased risk of postoperative complications in patients with steatosis assessed on pathologic examination of the liver specimen. The relative risk of postoperative complications was 1.53 and 2.01 in patients with <30% and $30% steatosis, respectively. An increased risk of postoperative death was only seen in patients with $30% steatosis. Our results using preoperative MRI to quantify liver fat support the results of the metaanalysis
showing an increased risk of morbidity. Moreover, in our study, patients with an MRI LFF of >5 % were found to have more pulmonary complications. The relationship between an MRI LFF of >5% and complications was even stronger in our subgroup of patients who underwent right side resection, a procedure known to be associated with an increased risk of liver failure.6,22 All published studies that have evaluated the presence and degree of steatosis as a risk factor after major hepatic resection except one, have used a pathologic specimen as the method of reference. This has 2 major drawbacks. First, the pathologic examination is no longer the most accurate method and MR techniques have clearly been shown to be more accurate.9,16,23-25 Although MRS was initially considered to be the best technique to quantify liver fat,17,26,27 its use is limited mainly to academic and research centers. Chemical shift–based MRI techniques that are included in routine liver protocols have been shown to be as accurate as MRS.23-26,28 A recent paper from Leiber et al10 showed that MRI evaluation of steatosis was superior to the semiquantitative pathologic assessment. Authors have compared both techniques using an animal model and a chemical reference technique (the quantification of triglyceride content). Second, a preoperative assessment of liver fat seems more useful to adapt management in higher risk patients. To our knowledge, only 1 study has evaluated liver steatosis on preoperative MRI.7 Those authors
ARTICLE IN PRESS 6 d’Assignies et al
Surgery j 2015
Table III. Relationship between MRI liver fat fraction and postoperative outcomes
Postoperative outcomes
All patients (n = 125)
Clavien score (7 classes), n (%) I 84 (67) II 1 (1) IIIa 12 (10) IIIb 6 (5) IVa 11 (9) IVb 6 (5) V 5 (4) Clavien score (binary variable: 40 (32) $IIIa) n (%) Liver-specific complications, n (%) Ascites >500 mL at day 5 28 (24) 50-50 at day 5 3 (2) Fistula/collection 22 (18) Perioperative blood loss (mL), 651.5 [536.5] mean [SD] Transfusion 37 (30) Blood liver tests, mean [SD] AST at day 1 (U/L) 538 [448] ALT at day 1 (U/L) 488 [368] Bilirubin at day 5 (mmol/mL) 39.0 (25.0) [40.4] Peak bilirubin (mmol/mL) 58.7 (38.0) [56.5] Pulmonary complications, n (%) 49 (39) Nonpulmonary complications, n (%) 19 (15) Complications $ 1, n (%) 63 (52) ICU stay (d), mean [SD] 6 [11] Hospital stay (d), mean [SD] 16 [15] Postoperative death, n (%) 5 (4)
Patients with MR Patients with MR liver fat fraction liver fat fraction <5 % (n = 100) $5 % (n = 25) 71 1 11 3 6 4 4 28
24 2 16 612.9
(71) (1) (11) (3) (6) (4) (4) (28)
(26) (2) (16) [483.0]
13 0 1 3 5 2 1 12
4 1 6 800.0
(52) (0) (4) (12) (20) (8) (4) (48)
(17) (4) (24) [697.5]
29 (29)
8 (32)
470 [327] 425 [318] 38 [41] 57 [56] 35 (35) 17 (17) 46 (48) 6 [11] 15 [13.8] 4 (4)
785 [693] 719 [447] 40 [38] 63 [59] 14 (56) 2 (8) 17 (68) 8 [10] 20 [17] 1 (4)
OR* (95% CI)
P value*
0.41 (0.16–1.04)
.060
2.93 (1.04–8.27)
.042
0.54 (0.15–1.91) 2.39 (0.18–30.95) 1.86 (0.55–6.36)
.336 .506 .321 .090 .810 .002 <.001 .563 .391
*OR and P value adjusted for potential confounding factors in the multivariate statistical analysis. Potential confounding factors used as variables of adjustments were: Age, sex, American Society of Anesthesiologists score, body mass index >25 kg/m2, liver fibrosis >F2, diabetes mellitus, duration of surgery, vascular clamping, biliary reconstruction and vascular reconstruction, blood loss and transfusion. ALT, Alanine aminotransferase; AST, aspartate aminotransferase; ICU, intensive care unit; OR, odds ratio; SD, standard deviation.
showed that an MRI LFF of >10% was associated with more serious postoperative complications, such as liver or multiorgan failure and sepsis after major liver resection.7 Our study has shown that the risk of postoperative complications including IIIa, IIIb, and IV Clavien-Dindo scores was increased with an even lower threshold of 5% LFF on MRI. We feel that this threshold is of particular interest, because it is also the accepted threshold to discriminate patients with from those without liver steatosis on MRI, and has been confirmed by studies with large patient cohorts.17,23,29 However, further study including a subset analysis for the degree of MRI LFF is required to identify the best surgically relevant cutoff, and to better stratify patients before surgery. Almost none of the previous studies including the MRI study analyzed the numerous confounding factors in these patients. Our study, including
125 patients who underwent major hepatic surgery, showed that a preoperative an MRI LFF of >5% is an independent risk factor of severe postoperative complications and greater duration of hospitalization. Increased postoperative serum liver enzymes are considered to be surrogate markers for liver injury.7 Raptis et al7 found a correlation between the MRI LFF and the degree of injury assessed by AST, ALT, and lactate dehydrogenase serum levels in a model of ischemia/reperfusion in ob/ob mice. These experimental results were confirmed by our study, which found a significant positive correlation between preoperative MRI LFF and postoperative transaminase levels. The presence of preoperative an MRI LFF of >5% could, therefore, alter patient management, requiring preoperative diet modification and/or specific treatment.30 In that way, Reeves and al31
ARTICLE IN PRESS d’Assignies et al 7
Surgery Volume j, Number j
Table IV. Relationship between MRI liver fat fraction and post operative outcomes for the subgroup of patients with right side resection*
Postoperative outcomes Clavien score (7 classes), n (%) I IIIa IIIb IVa IVb V Clavien score (binary variable: $IIIa) n (%) Liver-specific complications, n (%) Ascites >500 mL at day 5 50–50 at day 5 Fistula/collection Perioperative blood loss (mL), mean [SD] Transfusion Blood liver tests, mean [SD] AST at day 1 (U/L) ALT at day 1 (U/L) Bilirubin at day 5 (mmol/mL) Peak bilirubin (mmol/mL) Pulmonary complications, n (%) Non pulmonary complications, n (%) Complications $ 1, n (%) ICU stay (d), mean (median) [SD] Hospital stay (d), mean (median) [SD] Postoperative death, n (%)
All patients (n = 81) 53 5 5 8 5 5 53
22 3 10 671.3
Patients with MRI Patients with MR liver fat fraction liver fat fraction <5 % (n = 67) $% 5 (n = 14)
(65) (6) (6) (10) (6) (6) (65)
(30) (4) (12) [516.3]
48 5 2 5 3 4 48
19 2 7 635.1
26 (32) 542 482 41 63 38 15
[478] [397] [37] [53] (47) (18)
(72) (7) (3) (7) (4) (6) (72)
(31) (3) (10) [468.6]
22 (33) 458 407 39 59 28 14
[302] [313] [35] [47] (42) (21)
5 0 3 3 2 1 5
3 1 3 839.3
(35) (0) (21) (21) (14) (7) (36)
(23) (7) (21) [694.0]
ORy (95% CI)
P valuey
0.22 (0.06–0.75)
.016
6.84 (1.61–29.12)
.009
0.67 (0.14–3.11) 2.46 (0.21–29.20) 2.92 (0.57–15.2)
.609 .475 .201 .210
4 (29) 914 815 52 82 10 1
[839] [551] [47] [71] (71) (7)
.748
3.48 (0.91–13.27) 0.27 (0.03–2.56)
.002 .001 .341 .119 .067 .255
46 (59) 8 [11] 17 [15]
34 (53) 7 [11] 15 [13]
12 (86) 11 [12] 25 [20]
10.32 (1.21–87.75)
.033 .213 .025
5 (6)
4 (6)
1 (7)
1.51 (0.07–33.12)
.795
*Right side resection including 65 right hepatectomy and 26 right hepatectomy extended to segments I and/or IV and/or wedge resections within the left liver. yOR and P value adjusted for potential confounding factors in the multivariate statistical analysis. Potential confounding factors used as variables of adjustments were age, sex, American Society of Anesthesiologists score, body mass index >25 kg/m2, liver fibrosis >F2, diabetes mellitus, duration of surgery, vascular clamping, biliary reconstruction and vascular reconstruction, blood loss, and transfusion. ALT, Alanine aminotransferase; AST, aspartate aminotransferase; ICU, intensive care unit; OR, odds ratio; SD, standard deviation.
have recently shown in 111 patients that reduction of steatosis by a short-term diet was associated with a decrease in perioperative blood loss during liver resection. Moreover, in our study there was a trend (although not significant) for an increase in perioperative blood loss in patients with an MRI of LFF $5%, which should be anticipated before surgery. This study has certain limitations. First, it is a retrospective clinical study with various indications for liver resection and different operative procedures. However, the patients were included consecutively and representative of everyday clinical practice. We also analyzed a subgroup of patients with right side resections.
Second, we used the classic chemical shift– based MRI sequence, which does not correct for T2* decay. There is greater T2* decay in patients with excess liver iron overload, which could alter the fat quantification. However, the sequence used in this study is used in all MRI liver protocols and has been shown to be reliable in clinical practice.32 Third, few patients receiving chemotherapy (especially an irinotecan-based regimen) between MR imaging and surgery might have developed steatosis, and could be misclassified. Nevertheless, only few patients with colorectal liver metastases received such chemotherapy in our series, and we believe that this possible bias remains limited.
ARTICLE IN PRESS 8 d’Assignies et al
In conclusion, an LFF of >5% assessed by routine clinical MRI is an independent predictor of severe postoperative complications after major hepatectomy. Because MRI is an important tool for tumor staging, assessing the future liver remnant volume and liver anatomy, we suggest that the MRI LFF should be systematically determined as a biomarker of postoperative complications before performing major liver resection. The authors would like to acknowledge Jacques Belghiti. Gaspard d’Assignies and Valerie Vilgrain had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. REFERENCES 1. Williams CD, Stengel J, Asike MI, et al. Prevalence of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis among a largely middle-aged population utilizing ultrasound and liver biopsy: a prospective study. Gastroenterology 2010;140:124-31. 2. Rahman R, Hammoud GM, Almashhrawi AA, et al. Primary hepatocellular carcinoma and metabolic syndrome: An update. World J Gastrointest Oncol 2013;5:186-94. 3. Gomez D, Malik HZ, Bonney GK, et al. Steatosis predicts postoperative morbidity following hepatic resection for colorectal metastasis. Br J Surg 2007;94:1395-402. 4. de Meijer VE, Kalish BT, Puder M, Ijzermans JN. Systematic review and meta-analysis of steatosis as a risk factor in major hepatic resection. Br J Surg 2010;97:1331-9. 5. Behrns KE, Tsiotos GG, DeSouza NF, et al. Hepatic steatosis as a potential risk factor for major hepatic resection. J Gastrointest Surg 1998;2:292-8. 6. Bachellier P, Rosso E, Pessaux P, et al. Risk factors for liver failure and mortality after hepatectomy associated with portal vein resection. Ann Surg 2011;253:173-9. 7. Raptis DA, Fischer MA, Graf R, et al. MRI: the new reference standard in quantifying hepatic steatosis? Gut 2011;61:117-27. 8. Bannas P, Kramer H, Hernando D, et al. Quantitative MR Imaging of Hepatic Steatosis: Validation in Ex Vivo Human Livers. Hepatology 2015;62:1444-55. 9. Ronot M, Van Beers BE, Vilgrain V. Quantification of hepatic fat: Which reference should be the reference? Diagn Interv Imaging 2015;96:841-2. 10. Leiber LM, Boursier J, Michalak S, et al. MRI versus histological methods for time course monitoring of steatosis amount in a murine model of NAFLD. Diagn Interv Imaging 2015;96:915-22. 11. Clavien PA, Barkun J, de Oliveira ML, et al. The ClavienDindo classification of surgical complications: five-year experience. Ann Surg 2009;250:187-96. 12. Balzan S, Belghiti J, Farges O, et al. The ‘‘50-50 criteria’’ on postoperative day 5: an accurate predictor of liver failure and death after hepatectomy. Ann Surg 2005;242:824-8 . discussion 828–9. 13. McKay A, Sutherland FR, Bathe OF, Dixon E. Morbidity and mortality following multivisceral resections in complex hepatic and pancreatic surgery. J Gastrointest Surg 2008;12:86-90.
Surgery j 2015
14. Brunt EM, Tiniakos DG. Histopathology of nonalcoholic fatty liver disease. World J Gastroenterol 2010;16:5286-96. 15. Bedossa P, Poynard T. An algorithm for the grading of activity in chronic hepatitis C. The METAVIR Cooperative Study Group. Hepatology 1996;24:289-93. 16. d’Assignies G, Kauffmann C, Boulanger Y, et al. Simultaneous assessment of liver volume and whole liver fat content: a step towards one-stop shop preoperative MRI protocol. Eur Radiol 2010;21:301-9. 17. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab 2005;288: E462-8. 18. Cho JY, Suh KS, Kwon CH, et al. Mild hepatic steatosis is not a major risk factor for hepatectomy and regenerative power is not impaired. Surgery 2006;139:508-15. 19. Kooby DA, Fong Y, Suriawinata A, et al. Impact of steatosis on perioperative outcome following hepatic resection. J Gastrointest Surg 2003;7:1034-44. 20. McCormack L, Petrowsky H, Jochum W, et al. Hepatic steatosis is a risk factor for postoperative complications after major hepatectomy: a matched case-control study. Ann Surg 2007;245:923-30. 21. Nagai S, Fujimoto Y, Kamei H, et al. Mild hepatic macrovesicular steatosis may be a risk factor for hyperbilirubinaemia in living liver donors following right hepatectomy. Br J Surg 2009;96:437-44. 22. Dokmak S, Raut V, Aussilhou B, et al. Laparoscopic left lateral resection is the gold standard for benign liver lesions: a case-control study. HPB (Oxford) 2013;16:183-7. 23. d’Assignies G, Ruel M, Khiat A, et al. Noninvasive quantitation of human liver steatosis using magnetic resonance and bioassay methods. Eur Radiol 2009;19: 2033-40. 24. El-Badry AM, Breitenstein S, Jochum W, et al. Assessment of hepatic steatosis by expert pathologists: the end of a gold standard. Ann Surg 2009;250:691-7. 25. Ratziu V, Charlotte F, Heurtier A, et al. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology 2005;128:1898-906. 26. Hamilton G, Yokoo T, Bydder M, et al. In vivo characterization of the liver fat (1)H MR spectrum. NMR Biomed 2011; 24:784-90. 27. Longo R, Pollesello P, Ricci C, et al. Proton MR spectroscopy in quantitative in vivo determination of fat content in human liver steatosis. J Magn Reson Imaging 1995;5:281-5. 28. d’Assignies G, Fontes G, Kauffmann C, et al. Early detection of liver steatosis by magnetic resonance imaging in rats infused with glucose and intralipid solutions and correlation to insulin levels. Metabolism 2013;62:1850-7. 29. Marsman H, Matsushita T, Dierkhising R, et al. Assessment of donor liver steatosis: pathologist or automated software? Hum Pathol 2004;35:430-5. 30. Di Minno MN, Russolillo A, Lupoli R, et al. Omega-3 fatty acids for the treatment of non-alcoholic fatty liver disease. World J Gastroenterol 2012;18:5839-47. 31. Reeves JG, Suriawinata AA, Ng DP, et al. Short-term preoperative diet modification reduces steatosis and blood loss in patients undergoing liver resection. Surgery 2013;154:1031-7. 32. Borra RJ, Salo S, Dean K, et al. Nonalcoholic fatty liver disease: rapid evaluation of liver fat content with in-phase and out-of-phase MR imaging. Radiology 2009;250:130-6.
Liver steatosis assessed by preoperative MRI: An independent risk factor for severe complications after major hepatic resection Gaspard d’Assignies, MD, PhD,a,b,c Cindy Fayard, MD,a,b,c Helena Leitao, MD, PhD,a,b,c Toni Alfaiate, MD,d,e Florence Tubach, MD, PhD,b,d,e Safi Dokmak, MD, PhD,f Valerie Paradis, MD, PhD,b,c,g Bernard E. Van Beers, MD, PhD,a,b,c Maxime Ronot, MD, PhD,a,b,c and Valerie Vilgrain, MD, PhD,a,b,c Clichy, Hauts-de-Seine and Paris, France
Background. Steatosis assessed by histology is commonly considered to be a significant risk factor for liver surgery. MRI is considered as the new gold standard for noninvasive liver fat quantification. The purpose was to assess whether liver steatosis determined by preoperative MRI is an independent risk factor of complications after major liver resection. Methods. All patients who underwent liver MRI before major liver resection in our institution between January 2001 and December 2011 were included in this retrospective study. The liver fat fraction (LFF) was assessed on in- and opposed-phase T1-weighted dual echo gradient echo MRI and steatosis was defined as a MRI LFF of $5%. The association between steatosis and postoperative complications (Clavien-Dindo classification, ascites >500 mL at day 5, 50-50 criteria, fistula/collection, blood liver test alterations, pulmonary complications, nonpulmonary complications, >1 complication, duration of stay in the intensive care unit, duration of hospital stay, and death) was assessed by multivariate analysis using the appropriate model. Results. A MRI LFF of $5% was associated with severe postoperative complications (Clavien-Dindo score $ IIIa; P = .04), more pulmonary complications (P = .02), and longer duration of hospital stay (P = .02) on the multivariate model adjusted for confounding factors. The postoperative aminotransferase levels were higher in patients with a MRI LFF of $5%, than in other patients (P = .0008). Conclusion. Liver steatosis assessed by routine preoperative MRI is shown to be an independent risk factor of severe postoperative complications after major liver resection. (Surgery 2015;j:j-j.) From the Department of Radiology,a APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine; University Paris Diderot,b Sorbonne Paris Cite , Paris; INSERM U1149,c centre de recherche biome dicale Bichat-Beaujon, CRB3, Paris; De partement d’Epide miologie et Recherche Clinique,d APHP, University Hospitals Paris Nord Val de Seine, Bichat, Paris; INSERM CIC 1425-EC,e UMR 1123 ECEVE, Paris; Department of Surgery,f APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hautsde-Seine; and Department of Pathology,g APHP, University Hospitals Paris Nord Val de Seine, Beaujon, Clichy, Hauts-de-Seine, France
NONALCOHOLIC FATTY LIVER DISEASE (NAFLD) includes a wide spectrum of liver damage, ranging from simple steatosis to steatohepatitis and cirrhosis. Although it was unrecognized before the 1980s, the increasing epidemic of obesity has increased awareness of NAFLD, which is currently considered Authors have no conflict of interest. Accepted for publication October 6, 2015. Reprint requests: Maxime Ronot, MD, PhD, H^ opital Beaujon, Department of Radiology, 100 boulevard du, General Leclerc, 92100 Clichy, France. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.10.008
to be the most common cause of chronic liver disease affecting #40% of the general adult population in the United States.1 Thus, the incidence of steatosis during liver surgery has increased significantly. Furthermore, steatosis and steatohepatitis have been associated with irinotecan-based regimens used in patients with colorectal liver metastases whose curative treatment is liver surgery. NAFLD also increases the risk of developing telangiectatic/inflammatory hepatocellular adenoma and hepatocellular carcinoma, both tumors that may require operative resection.2 Steatosis is considered commonly to be a significant risk factor for liver surgery, especially in patients undergoing major resection.3-6 One SURGERY 1
ARTICLE IN PRESS 2 d’Assignies et al
systematic review and metaanalysis has shown a significant association between the degree of steatosis and an increased risk of postoperative complications and mortality.4 All these studies used semiquantitative pathologic examination of the surgical specimen to quantify the degree of steatosis. Indeed, knowing the degree of preoperative steatosis could alter patient management. To our knowledge, only 1 study has shown that patients with moderate to severe steatosis (liver fat content >10% on MRI) have a greater risk of life-threatening complications after major liver resection than patients without steatosis or with mild steatosis.7 Nevertheless, that study did not evaluate the effect of possible confounding factors such as age, American Society of Anesthesiologists (ASA) score, body mass index >25 kg/m2, diabetes, and liver fibrosis. Presently, MRI is the most sensitive imaging modality for detecting liver steatosis and MR spectroscopy (MRS) is the most accurate method.8-10 MRS provides information about the chemical composition in an organ and can quantify small amounts of fat. However, MRS of the liver is usually only used in research protocols. In contrast, the breath-hold dual-echo chemical shift gradient echo (GRE) sequence is technically simple and systematically included in MRI protocols of the liver to estimate liver fat. The purpose of this retrospective study was to assess whether liver steatosis of any grade, determined by preoperative dual-echo chemical shift GRE MRI is an independent risk factor of complications after major liver resection. METHODS Patients. The review board at our institution approved this single-center, retrospective clinical study and informed consent was waived. From January 2001 to December 2011, all consecutive patients >18 years old, scheduled for major liver resection ($3 liver segments), who underwent preoperative liver MRI were identified in the prospective database of the department of hepatobiliary surgery. Based on these criteria, 171 patients were included. Exclusion criteria were multiple liver surgeries and living donors (n = 11) or liver MRI >120 days before liver resection (n = 35). Preoperative volumetric calculations were performed and patients were only operated on if the volume of the future liver remnant was acceptable. Patients underwent major liver resection for colorectal liver metastases in 36 patients (29%), hepatocellular carcinomas in 20 (16%), cholangiocarcinomas in 28 (22%), other malignant
Surgery j 2015
lesions in 27 (22%), and benign liver lesions in 14 (11%). The following preoperative and perioperative data were recorded: patient age, gender, ASA score, body mass index >25 kg/m2, presence of significant liver fibrosis (F0-F1 vs F2-F4), comorbidities such as diabetes mellitus, type of resection (left or right hepatectomy), duration of surgery, vascular clamping, biliary resection–reconstruction, and vascular resection–reconstruction. The right side resection included right hepatectomy extended to segments I and/or IV and/or wedge resection within the left liver. Intraoperative and postoperative variables were recorded as estimated blood loss, transfusion, liver-specific complications, blood liver tests, pulmonary complications, nonpulmonary complications and classified according to the Clavien-Dindo classification.11 Reported blood liver tests were aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels at postoperative day 1, bilirubin levels at postoperative day 5, and maximum peak bilirubin levels. Liver-specific complications were defined as any of the following: biliary fistula/collections requiring percutaneous drainage, liver failure according to 50/50 criteria, defined as the association of prothrombin time <50% and serum bilirubin >50 mL/L on postoperative day 5,12 ascites (>500 mL at day 5). Pulmonary complications included pulmonary embolism, acute respiratory failure, pleural effusion, and pneumonia. Nonpulmonary complications included all other medical major events such as urinary infection, acute renal failure, cardiac arrhythmia, or bowel occlusion. The durations of the intensive care unit (ICU) and hospital stay were recorded and postoperative mortality was defined as any death in the hospital or within 90 days after surgery. Postoperative complications were categorized in 7 grades according to the Clavien-Dindo classification.11 Complications requiring a surgical, endoscopic or radiologic intervention (grade IIIa or higher) were considered severe complications.11,13 MRI. MRI was performed with a 1.5-T MR scanner (Intera, Philips Medical Systems, Best, The Netherlands), with a maximum gradient strength of 40 mT/m and a slew rate of 200 mT/m/ms. A 4-element surface coil was used. Our routine MRI protocol for the liver systematically included a dual echo GRE sequence to quantify liver fat with the following parameters: repetition time/echo time 145/2.3 (in-phase) and 4.6 ms (opposed-phase); matrix 256 3 256 pixels, 213 3 305-mm2 field of view, slice thickness 6 mm, 2 averages, and 808 flip angle. The acquisition time was 25 s.
ARTICLE IN PRESS d’Assignies et al 3
Surgery Volume j, Number j
Histologic analysis. Liver resection specimens were evaluated in all patients by an experienced hepatobiliary pathologist (V.P.) who was unaware of the imaging results and surgical outcomes. The degree of liver steatosis was graded according to the Brunt classification14: mild (5–33%), moderate (34–66%), or severe (67–100%). Fibrosis was assessed according to the METAVIR scoring system15 and then classified into 2 groups, with and without advanced fibrosis (F # 2 and F $ 3). Image interpretation. To measure liver fat fraction (LFF), 3 circular regions of interest with a mean area of 5.5 ± 0.4 cm2 were drawn by a radiologist with 5 years experience in abdominal imaging (C.F.; Figs 1 and 2) on the in-phase/ opposed-phase dual echo GRE MRI, in 3 different areas of the liver (the right anterior segment, the right posterior segment, and the left lobe) avoiding major blood vessels, as 3 measurements have been shown highly reliable.16 Small areas of focal fatty sparing or focal fatty deposition near the hilum were avoided. If more diffuse heterogeneous steatosis (such as lobar distribution) was seen on opposed-phase GRE T1-weighted MR sequence, regions of interest were placed in the more fatty segments. Liver signal intensities (in phase [SIiP] and opposed phase [SIoP]) were measured using a commercial software viewer. The LFF was calculated in each voxel with the following equation: LFFð%Þ ¼ ðSIiP SIoP Þ=2 SIiP
½1
The fat fraction for each patient was considered to be the average of the 3 measurements. At MRI, the patients were considered to have liver steatosis when their LFF was $5%. This threshold was chosen in accordance to previous cohort study.17 No receiver operating characteristic curve analysis was performed. Statistical analysis. Patient characteristics were described using frequencies and percentages for categorical variables, and means and standard deviations for continuous variables. Preoperative, intraoperative, and perioperative characteristics associated with the presence of LFF of $5% were identified by univariate analysis (Student t test or Wilcoxon test for continuous variables, and Chisquare or Fisher exact test for categorical variables). The correlation and association between MR LFF and histologic grade of steatosis was assessed using the Spearman correlation coefficient and Kruskall–Wallis test, respectively. The association between LFF on MRI of $5% and each perioperative, intraoperative, and postoperative
variable (blood loss, transfusion, Clavien-Dindo score, ascites >500 mL at day 5, 50/50 criteria, biliary fistula/collection, AST at day 1, ALT at day 1, bilirubin at day 5, peak bilirubin, pulmonary complication, nonpulmonary complication, >1 complication, duration of ICU stay, duration of hospital stay, and death) was assessed using adjusted models, namely analyses of variance when the dependent outcome (postoperative complication) was continuous, logistic regression when the dependent outcome was dichotomous, and multinomial logistic regression when the dependent outcome was categorical and had >2 levels. For each adjusted model, variables to be considered were identified by univariate analysis. For a given postoperative complication, the association between the postoperative complication and each potential adjustment variable (age, ASA score, body mass index >25 kg/m2, diabetes, liver fibrosis >F2, vascular reconstruction, biliary reconstruction, vascular clamping, right or left hepatectomy, duration of surgery) was tested using the Student t test or Wilcoxon test, Chi-square, or Fisher exact test, Pearson or Spearman correlation coefficient. Potential adjustment variables with a P of <.05 were entered into the adjusted model. We also performed this analysis to evaluate the association between the MRI LFF of $5% and each postoperative complication in the subgroup of patients who underwent right side resection (analysis planned a priori). The association between MRI LFF and perioperative and postoperative continuous variables (duration of surgery, perioperative blood loss, AST, ALT, bilirubin at day 5, duration of ICU stay, and duration of postoperative stay) was assessed by the Spearman correlation coefficient. Statistical analyses were performed with SAS software version 9.2 (Institute Inc, Cary, NC). RESULTS The study sample included 125 patients (59 women, and 66 men) with a mean age of 54.6 ± 14.7 years (54.4 ± 13.4 for women; 54.8 ± 15.8 for men). No diffuse heterogeneous steatosis was seen on opposed-phase GRE T1weighted MR sequence. Therefore, the regions of interest were placed in 3 different areas of the liver (the right anterior segment, the right posterior segment, and the left lobe) avoiding major blood vessels. The demographic, preoperative intraoperative, and perioperative data of all patients and patients with an LFF of <5 or $5% on MRI are shown in Table I. No differences were found between
ARTICLE IN PRESS 4 d’Assignies et al
Surgery j 2015
Fig 1. In-phase (left) and opposed-phase (right) dual echo gradient echo MRI of an 18-year-old man with a large telangiectatic adenoma of the right liver. Fat fraction is calculated within the regions of interest (white circle). The preoperative MRI liver fat fraction was 34.6%. Right side resection was performed. Clavien-Dindo IIIb postoperative complications were reported. The duration of intensive care stay was 19 days and hospital stay, 32 days.
Fig 2. In-phase (left) and opposed-phase (right) dual echo gradient echo MRI of a 43-year-old man with a large post– hepatitis B cirrhosis hepatocellular carcinoma of the right liver. The fat fraction is calculated in the regions of interest (white circle). The preoperative MRI liver fat fraction was <5%. Right side resection was performed. No postoperative complications were reported. The duration of intensive care stay was 1 day and the hospital stay, 6 days.
patients with an LFF of <5 or $5% on MRI at baseline. The Spearman correlation coefficient between the MRI LFF and the histologic grade of steatosis was significant (r = 0.558; P < .0001). Table II shows the relationship between MRI LFF and the histologic grade of steatosis. An MRI LFF of $5% adjusted for confounding factors was associated with severe postoperative complications (Clavien-Dindo score $ IIIa; P = .042; Table III; Figs 1 and 2). Furthermore, pulmonary complications were more frequent (P = .021) and the duration of hospital stay was greater (P = .022) for patients with an MRI LFF of $5%. In the subgroup of patients with right hepatectomy, an MRI LFF of $5% was associated with severe postoperative complications (P = .009; Table IV). After right hepatectomy, patients who had an MRI LFF of $5% more frequently had $1 complication (P = .033) and the duration of
hospital stay was greater than for patients without steatosis (P = .025). Significant but modest correlations were found between increased AST and ALT at day 1 and LFF assessed on MRI (r = 0.19 [P = .045] and r = 0.28 [P = .002], respectively). There were stronger correlations in the subgroup with a right side resection (r = 0.38 [P < .001] and r = 0.39 [P < .001] for ALT and AST, respectively). No correlation was found between the MRI LFF and serum bilirubin at Day 5 (P = .381). There was a trend (although not significant) for an increase in perioperative blood loss in patients with an MRI LFF of $5%. DISCUSSION In our study, liver steatosis quantified by preoperative MRI was found to be an independent risk factor for complications after major liver resection. Patients with an MRI LFF of $5% were found to
ARTICLE IN PRESS d’Assignies et al 5
Surgery Volume j, Number j
Table I. Demographic and perioperative data of patients according to MRI liver fat fraction measurements
Characteristic Preoperative Age (y), mean [SD] Sex ratio (M/F) Body mass index > 25, n (%) Diabetes, n (%) ASA grade, n (%) I II III Perioperative Right hepatectomy,* n (%) Biliary reconstruction, n (%) Vascular reconstruction, n (%) Clamping, n (%) Duration of surgery (min), mean [SD] Liver fibrosis >F2, n (%)
All patients (n = 125)
Patients with MRI liver fat fraction <5% (n = 100)
Patients with MRI liver fat fraction $5% (n = 25)
55 [15] 66/59 38 (32) 10 (8)
54 [15] 52/48 28 (30) 8 (8)
56 [14] 14/11 10 (40) 2 (8)
12 (9.6) 107 (85.6) 6 (4.8)
11 (11) 83 (83) 6 (6)
1 (4) 24 (96) 0 (0)
81 (65) 13 (10) 18 (14) 48 (39) 353.0 [103.3] 17 (14)
67 11 15 38 346.7 14
(67) (11) (15) (39) [101.2] (14)
14 2 3 10 377.2 3
(56) (8) (12) (40) [109.9] (12)
P value .581 .722 .347 .630 .319
.303 1.000 1.000 .940 .200 1.00
*Right side resection including 65 right hepatectomy and 26 right hepatectomy extended to segments I and/or IV and/or wedge resections within the left liver. ASA, American Society of Anesthesiologists; SD, standard deviation.
Table II. Relation between MRI liver fat fraction and histologic grade of steatosis Histologic grade of steatosis 0 1 2 3
n (%) 70 42 8 4
(56.8) (33.6) (6.4) (3.2)
MRI liver fat fraction mean % (SD) 0.9 3.1 16.2 19
(0.02) (0.03) (0.12) (0.14)
P value* <.001
*Kruskall–Wallis test. SD, Standard deviation.
have higher Clavien-Dindo scores and greater durations of hospital stay. These patients were also more prone to pulmonary complications. Data on the correlation between liver steatosis and perioperative complications are scarce and contradictory.3,5,18-21 However, a recent systematic review and metaanalysis that included 6 studies with a total of 1,000 patients with major liver resection4 has shown an increased risk of postoperative complications in patients with steatosis assessed on pathologic examination of the liver specimen. The relative risk of postoperative complications was 1.53 and 2.01 in patients with <30% and $30% steatosis, respectively. An increased risk of postoperative death was only seen in patients with $30% steatosis. Our results using preoperative MRI to quantify liver fat support the results of the metaanalysis
showing an increased risk of morbidity. Moreover, in our study, patients with an MRI LFF of >5 % were found to have more pulmonary complications. The relationship between an MRI LFF of >5% and complications was even stronger in our subgroup of patients who underwent right side resection, a procedure known to be associated with an increased risk of liver failure.6,22 All published studies that have evaluated the presence and degree of steatosis as a risk factor after major hepatic resection except one, have used a pathologic specimen as the method of reference. This has 2 major drawbacks. First, the pathologic examination is no longer the most accurate method and MR techniques have clearly been shown to be more accurate.9,16,23-25 Although MRS was initially considered to be the best technique to quantify liver fat,17,26,27 its use is limited mainly to academic and research centers. Chemical shift–based MRI techniques that are included in routine liver protocols have been shown to be as accurate as MRS.23-26,28 A recent paper from Leiber et al10 showed that MRI evaluation of steatosis was superior to the semiquantitative pathologic assessment. Authors have compared both techniques using an animal model and a chemical reference technique (the quantification of triglyceride content). Second, a preoperative assessment of liver fat seems more useful to adapt management in higher risk patients. To our knowledge, only 1 study has evaluated liver steatosis on preoperative MRI.7 Those authors
ARTICLE IN PRESS 6 d’Assignies et al
Surgery j 2015
Table III. Relationship between MRI liver fat fraction and postoperative outcomes
Postoperative outcomes
All patients (n = 125)
Clavien score (7 classes), n (%) I 84 (67) II 1 (1) IIIa 12 (10) IIIb 6 (5) IVa 11 (9) IVb 6 (5) V 5 (4) Clavien score (binary variable: 40 (32) $IIIa) n (%) Liver-specific complications, n (%) Ascites >500 mL at day 5 28 (24) 50-50 at day 5 3 (2) Fistula/collection 22 (18) Perioperative blood loss (mL), 651.5 [536.5] mean [SD] Transfusion 37 (30) Blood liver tests, mean [SD] AST at day 1 (U/L) 538 [448] ALT at day 1 (U/L) 488 [368] Bilirubin at day 5 (mmol/mL) 39.0 (25.0) [40.4] Peak bilirubin (mmol/mL) 58.7 (38.0) [56.5] Pulmonary complications, n (%) 49 (39) Nonpulmonary complications, n (%) 19 (15) Complications $ 1, n (%) 63 (52) ICU stay (d), mean [SD] 6 [11] Hospital stay (d), mean [SD] 16 [15] Postoperative death, n (%) 5 (4)
Patients with MR Patients with MR liver fat fraction liver fat fraction <5 % (n = 100) $5 % (n = 25) 71 1 11 3 6 4 4 28
24 2 16 612.9
(71) (1) (11) (3) (6) (4) (4) (28)
(26) (2) (16) [483.0]
13 0 1 3 5 2 1 12
4 1 6 800.0
(52) (0) (4) (12) (20) (8) (4) (48)
(17) (4) (24) [697.5]
29 (29)
8 (32)
470 [327] 425 [318] 38 [41] 57 [56] 35 (35) 17 (17) 46 (48) 6 [11] 15 [13.8] 4 (4)
785 [693] 719 [447] 40 [38] 63 [59] 14 (56) 2 (8) 17 (68) 8 [10] 20 [17] 1 (4)
OR* (95% CI)
P value*
0.41 (0.16–1.04)
.060
2.93 (1.04–8.27)
.042
0.54 (0.15–1.91) 2.39 (0.18–30.95) 1.86 (0.55–6.36)
.336 .506 .321 .090 .810 .002 <.001 .563 .391
*OR and P value adjusted for potential confounding factors in the multivariate statistical analysis. Potential confounding factors used as variables of adjustments were: Age, sex, American Society of Anesthesiologists score, body mass index >25 kg/m2, liver fibrosis >F2, diabetes mellitus, duration of surgery, vascular clamping, biliary reconstruction and vascular reconstruction, blood loss and transfusion. ALT, Alanine aminotransferase; AST, aspartate aminotransferase; ICU, intensive care unit; OR, odds ratio; SD, standard deviation.
showed that an MRI LFF of >10% was associated with more serious postoperative complications, such as liver or multiorgan failure and sepsis after major liver resection.7 Our study has shown that the risk of postoperative complications including IIIa, IIIb, and IV Clavien-Dindo scores was increased with an even lower threshold of 5% LFF on MRI. We feel that this threshold is of particular interest, because it is also the accepted threshold to discriminate patients with from those without liver steatosis on MRI, and has been confirmed by studies with large patient cohorts.17,23,29 However, further study including a subset analysis for the degree of MRI LFF is required to identify the best surgically relevant cutoff, and to better stratify patients before surgery. Almost none of the previous studies including the MRI study analyzed the numerous confounding factors in these patients. Our study, including
125 patients who underwent major hepatic surgery, showed that a preoperative an MRI LFF of >5% is an independent risk factor of severe postoperative complications and greater duration of hospitalization. Increased postoperative serum liver enzymes are considered to be surrogate markers for liver injury.7 Raptis et al7 found a correlation between the MRI LFF and the degree of injury assessed by AST, ALT, and lactate dehydrogenase serum levels in a model of ischemia/reperfusion in ob/ob mice. These experimental results were confirmed by our study, which found a significant positive correlation between preoperative MRI LFF and postoperative transaminase levels. The presence of preoperative an MRI LFF of >5% could, therefore, alter patient management, requiring preoperative diet modification and/or specific treatment.30 In that way, Reeves and al31
ARTICLE IN PRESS d’Assignies et al 7
Surgery Volume j, Number j
Table IV. Relationship between MRI liver fat fraction and post operative outcomes for the subgroup of patients with right side resection*
Postoperative outcomes Clavien score (7 classes), n (%) I IIIa IIIb IVa IVb V Clavien score (binary variable: $IIIa) n (%) Liver-specific complications, n (%) Ascites >500 mL at day 5 50–50 at day 5 Fistula/collection Perioperative blood loss (mL), mean [SD] Transfusion Blood liver tests, mean [SD] AST at day 1 (U/L) ALT at day 1 (U/L) Bilirubin at day 5 (mmol/mL) Peak bilirubin (mmol/mL) Pulmonary complications, n (%) Non pulmonary complications, n (%) Complications $ 1, n (%) ICU stay (d), mean (median) [SD] Hospital stay (d), mean (median) [SD] Postoperative death, n (%)
All patients (n = 81) 53 5 5 8 5 5 53
22 3 10 671.3
Patients with MRI Patients with MR liver fat fraction liver fat fraction <5 % (n = 67) $% 5 (n = 14)
(65) (6) (6) (10) (6) (6) (65)
(30) (4) (12) [516.3]
48 5 2 5 3 4 48
19 2 7 635.1
26 (32) 542 482 41 63 38 15
[478] [397] [37] [53] (47) (18)
(72) (7) (3) (7) (4) (6) (72)
(31) (3) (10) [468.6]
22 (33) 458 407 39 59 28 14
[302] [313] [35] [47] (42) (21)
5 0 3 3 2 1 5
3 1 3 839.3
(35) (0) (21) (21) (14) (7) (36)
(23) (7) (21) [694.0]
ORy (95% CI)
P valuey
0.22 (0.06–0.75)
.016
6.84 (1.61–29.12)
.009
0.67 (0.14–3.11) 2.46 (0.21–29.20) 2.92 (0.57–15.2)
.609 .475 .201 .210
4 (29) 914 815 52 82 10 1
[839] [551] [47] [71] (71) (7)
.748
3.48 (0.91–13.27) 0.27 (0.03–2.56)
.002 .001 .341 .119 .067 .255
46 (59) 8 [11] 17 [15]
34 (53) 7 [11] 15 [13]
12 (86) 11 [12] 25 [20]
10.32 (1.21–87.75)
.033 .213 .025
5 (6)
4 (6)
1 (7)
1.51 (0.07–33.12)
.795
*Right side resection including 65 right hepatectomy and 26 right hepatectomy extended to segments I and/or IV and/or wedge resections within the left liver. yOR and P value adjusted for potential confounding factors in the multivariate statistical analysis. Potential confounding factors used as variables of adjustments were age, sex, American Society of Anesthesiologists score, body mass index >25 kg/m2, liver fibrosis >F2, diabetes mellitus, duration of surgery, vascular clamping, biliary reconstruction and vascular reconstruction, blood loss, and transfusion. ALT, Alanine aminotransferase; AST, aspartate aminotransferase; ICU, intensive care unit; OR, odds ratio; SD, standard deviation.
have recently shown in 111 patients that reduction of steatosis by a short-term diet was associated with a decrease in perioperative blood loss during liver resection. Moreover, in our study there was a trend (although not significant) for an increase in perioperative blood loss in patients with an MRI of LFF $5%, which should be anticipated before surgery. This study has certain limitations. First, it is a retrospective clinical study with various indications for liver resection and different operative procedures. However, the patients were included consecutively and representative of everyday clinical practice. We also analyzed a subgroup of patients with right side resections.
Second, we used the classic chemical shift– based MRI sequence, which does not correct for T2* decay. There is greater T2* decay in patients with excess liver iron overload, which could alter the fat quantification. However, the sequence used in this study is used in all MRI liver protocols and has been shown to be reliable in clinical practice.32 Third, few patients receiving chemotherapy (especially an irinotecan-based regimen) between MR imaging and surgery might have developed steatosis, and could be misclassified. Nevertheless, only few patients with colorectal liver metastases received such chemotherapy in our series, and we believe that this possible bias remains limited.
ARTICLE IN PRESS 8 d’Assignies et al
In conclusion, an LFF of >5% assessed by routine clinical MRI is an independent predictor of severe postoperative complications after major hepatectomy. Because MRI is an important tool for tumor staging, assessing the future liver remnant volume and liver anatomy, we suggest that the MRI LFF should be systematically determined as a biomarker of postoperative complications before performing major liver resection. The authors would like to acknowledge Jacques Belghiti. Gaspard d’Assignies and Valerie Vilgrain had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. REFERENCES 1. Williams CD, Stengel J, Asike MI, et al. Prevalence of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis among a largely middle-aged population utilizing ultrasound and liver biopsy: a prospective study. Gastroenterology 2010;140:124-31. 2. Rahman R, Hammoud GM, Almashhrawi AA, et al. Primary hepatocellular carcinoma and metabolic syndrome: An update. World J Gastrointest Oncol 2013;5:186-94. 3. Gomez D, Malik HZ, Bonney GK, et al. Steatosis predicts postoperative morbidity following hepatic resection for colorectal metastasis. Br J Surg 2007;94:1395-402. 4. de Meijer VE, Kalish BT, Puder M, Ijzermans JN. Systematic review and meta-analysis of steatosis as a risk factor in major hepatic resection. Br J Surg 2010;97:1331-9. 5. Behrns KE, Tsiotos GG, DeSouza NF, et al. Hepatic steatosis as a potential risk factor for major hepatic resection. J Gastrointest Surg 1998;2:292-8. 6. Bachellier P, Rosso E, Pessaux P, et al. Risk factors for liver failure and mortality after hepatectomy associated with portal vein resection. Ann Surg 2011;253:173-9. 7. Raptis DA, Fischer MA, Graf R, et al. MRI: the new reference standard in quantifying hepatic steatosis? Gut 2011;61:117-27. 8. Bannas P, Kramer H, Hernando D, et al. Quantitative MR Imaging of Hepatic Steatosis: Validation in Ex Vivo Human Livers. Hepatology 2015;62:1444-55. 9. Ronot M, Van Beers BE, Vilgrain V. Quantification of hepatic fat: Which reference should be the reference? Diagn Interv Imaging 2015;96:841-2. 10. Leiber LM, Boursier J, Michalak S, et al. MRI versus histological methods for time course monitoring of steatosis amount in a murine model of NAFLD. Diagn Interv Imaging 2015;96:915-22. 11. Clavien PA, Barkun J, de Oliveira ML, et al. The ClavienDindo classification of surgical complications: five-year experience. Ann Surg 2009;250:187-96. 12. Balzan S, Belghiti J, Farges O, et al. The ‘‘50-50 criteria’’ on postoperative day 5: an accurate predictor of liver failure and death after hepatectomy. Ann Surg 2005;242:824-8 . discussion 828–9. 13. McKay A, Sutherland FR, Bathe OF, Dixon E. Morbidity and mortality following multivisceral resections in complex hepatic and pancreatic surgery. J Gastrointest Surg 2008;12:86-90.
Surgery j 2015
14. Brunt EM, Tiniakos DG. Histopathology of nonalcoholic fatty liver disease. World J Gastroenterol 2010;16:5286-96. 15. Bedossa P, Poynard T. An algorithm for the grading of activity in chronic hepatitis C. The METAVIR Cooperative Study Group. Hepatology 1996;24:289-93. 16. d’Assignies G, Kauffmann C, Boulanger Y, et al. Simultaneous assessment of liver volume and whole liver fat content: a step towards one-stop shop preoperative MRI protocol. Eur Radiol 2010;21:301-9. 17. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab 2005;288: E462-8. 18. Cho JY, Suh KS, Kwon CH, et al. Mild hepatic steatosis is not a major risk factor for hepatectomy and regenerative power is not impaired. Surgery 2006;139:508-15. 19. Kooby DA, Fong Y, Suriawinata A, et al. Impact of steatosis on perioperative outcome following hepatic resection. J Gastrointest Surg 2003;7:1034-44. 20. McCormack L, Petrowsky H, Jochum W, et al. Hepatic steatosis is a risk factor for postoperative complications after major hepatectomy: a matched case-control study. Ann Surg 2007;245:923-30. 21. Nagai S, Fujimoto Y, Kamei H, et al. Mild hepatic macrovesicular steatosis may be a risk factor for hyperbilirubinaemia in living liver donors following right hepatectomy. Br J Surg 2009;96:437-44. 22. Dokmak S, Raut V, Aussilhou B, et al. Laparoscopic left lateral resection is the gold standard for benign liver lesions: a case-control study. HPB (Oxford) 2013;16:183-7. 23. d’Assignies G, Ruel M, Khiat A, et al. Noninvasive quantitation of human liver steatosis using magnetic resonance and bioassay methods. Eur Radiol 2009;19: 2033-40. 24. El-Badry AM, Breitenstein S, Jochum W, et al. Assessment of hepatic steatosis by expert pathologists: the end of a gold standard. Ann Surg 2009;250:691-7. 25. Ratziu V, Charlotte F, Heurtier A, et al. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology 2005;128:1898-906. 26. Hamilton G, Yokoo T, Bydder M, et al. In vivo characterization of the liver fat (1)H MR spectrum. NMR Biomed 2011; 24:784-90. 27. Longo R, Pollesello P, Ricci C, et al. Proton MR spectroscopy in quantitative in vivo determination of fat content in human liver steatosis. J Magn Reson Imaging 1995;5:281-5. 28. d’Assignies G, Fontes G, Kauffmann C, et al. Early detection of liver steatosis by magnetic resonance imaging in rats infused with glucose and intralipid solutions and correlation to insulin levels. Metabolism 2013;62:1850-7. 29. Marsman H, Matsushita T, Dierkhising R, et al. Assessment of donor liver steatosis: pathologist or automated software? Hum Pathol 2004;35:430-5. 30. Di Minno MN, Russolillo A, Lupoli R, et al. Omega-3 fatty acids for the treatment of non-alcoholic fatty liver disease. World J Gastroenterol 2012;18:5839-47. 31. Reeves JG, Suriawinata AA, Ng DP, et al. Short-term preoperative diet modification reduces steatosis and blood loss in patients undergoing liver resection. Surgery 2013;154:1031-7. 32. Borra RJ, Salo S, Dean K, et al. Nonalcoholic fatty liver disease: rapid evaluation of liver fat content with in-phase and out-of-phase MR imaging. Radiology 2009;250:130-6.