Available online at www.sciencedirect.com
ScienceDirect EJSO xx (2014) 1e7
www.ejso.com
Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis F. Ratti*,1, F. Cipriani 1, M. Catena 1, M. Paganelli 1, L. Aldrighetti 1 Department of General Surgery, Ospedale San Raffaele, Via Olgettina 60, 20132 Milano, Italy Accepted 26 June 2014 Available online - - -
Abstract Aim: An accurate and noninvasive tool to predict Chemotherapy Associated Liver Injury (CALI) still lacks. Study aimed to evaluate chronic liver disease scores (Aspartate aminotransferase to Platelet Ratio Index, APRI and Fibrosis-4, FIB-4) as Postoperative Liver Failure (PLF) predictors in patients treated with Oxaliplatin for Colorectal Liver Metastases (CLM). Methods: 8 patients who developed PLF after major hepatectomy (Group B) were compared to 24 patients who did not develop PLF (Group A) in a case-matched analysis for patients and disease characteristics. ROC curves analysis was performed to assess score accuracy. Results: In Group A number of CT cycles was lower, (6 vs 9, p NS), interval between treatment and surgery was longer (11 vs 7 weeks, p < 0.05) and bevacizumab was more frequently administered (66.7% vs 37.5%, p < 0.05). In Group B median APRI score was 0.53 (range: 0.86e4.26) whereas in Group A was 0.30 (range: 0.06e2.21, p < 0.05). Median FIB-4 score was 2.46 (range: 0.86e13.65) in Group B and 1.58 (range: 0.27e7.68) in Group A ( p < 0.001). Multivariate analysis showed a significant correlation between APRI and the onset of PLF. A good accuracy of APRI score was evident in ROC curves with an area under the curve of 0.72 ( p 0.003). Conclusions: APRI score is calculated considering both liver damage and platelet count, it is cost effective and easily available. This study demonstrates that there is a good accuracy in PLF prediction and consequently in CT induced liver damage evaluation. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Chemotherapy; Liver surgery; Sinusoidal obstruction; postoperative liver failure; Metastases
Introduction Almost 50% of patients affected by colorectal cancer develop liver metastases (CLM) and the only curative treatment for this subset of patients is represented by surgical removal of lesions1: hence the need for a multidisciplinary management of these patients, together with radiologists, radiotherapists and medical oncologists, to increase resectability rate, therefore improving long term outcome. An optimal timing for systemic ChemoTherapy (CT) administration has not been defined yet: despite this, usefulness * Corresponding author. Tel.: þ39 02 26437811, þ39 348 2411961; fax: þ39 02 26437807. E-mail address: [email protected] (F. Ratti). 1 Hepatobiliary Surgery Unit.
and advantages resulting from NeoAdjuvantChemoTherapy (NACT) are well recognized.2 Indeed, it allows tumor downstaging, expanding both rate of R0 resections and number of candidates to surgery among patients with initially unresectable disease. In addition, preoperative identification of good responders is possible, while patients with disease progression during therapy might be excluded from surgery.3 In spite of its benefits, CT may induce liver injury on nonneoplastic parenchyma leading to increased morbidity and mortality after liver resection in patients who received oxaliplatin or irinotecan based regimens.4e7 Oxaliplatin in particular is associated with Sinusoidal Obstruction Syndrome (SOS), responsible for a significant risk of bleeding and Postoperative Liver Failure (PLF).8 PLF is one of the most severe causes of mortality following liver surgery9: indeed the capability of the liver in tolerating extended resections is based
http://dx.doi.org/10.1016/j.ejso.2014.06.011 0748-7983/Ó 2014 Elsevier Ltd. All rights reserved. Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
2
F. Ratti et al. / EJSO xx (2014) 1e7
upon parenchymal functional reserve, that can be potentially impaired as a result of prolonged chemotherapy.4e8 A non-invasive and accurate tool to predict Chemotherapy Induced Liver Injury (CALI) still lacks, but should be essential to evaluate risk of PLF resulting from parenchymal impairment, therefore allowing better patients selection before major resections. APRI (Aspartate aminotransferase to Platelet Ratio Index)10 and FIB-4 (Fibrosis-4)11 scores have been reported to grade liver fibrosis in patients with hepatitis C virus infection, even though their importance has been recently high lightened in prediction of high grade lesions of SOS.12e14 Aim of the study was to identify clinical and biochemical factors associated to PLF in patients affected by CLM who underwent major liver surgery and previous oxaliplatin based CT in a case-matched analysis. Analysis was focused on chronic liver disease scores. Secondary aim was to assess accuracy of chronic liver disease scores as predictors of PLF. Methods Patients From January 2004 to December 2012, 1413 liver resections were performed at the Hepatobiliary Surgery Unit of San Raffaele Hospital. Data from these patients have been collected in a prospective database and are now retrospectively reviewed. 463 (32.8%) liver resections were performed in patients affected by CLM. Of these, 258 patients were treated with systemic chemotherapy before resection and in 161 (34.7%) of them oxaliplatin-based regimens were used. Out of 101 patients who underwent oxaliplatin based NACT within 4 months before liver resection, without demonstrated chronic liver disease, 54 major or extended hepatectomies were performed. Among this last group, 8 patients developed PLF according to ISGLS definition (in particular all patients had grade B or C PLF, while none of them had grade A PLF)15 and were compared in a case-matched analysis to 24 patients, belonging to the same population, who did not develop PLF in the postoperative period. The two groups (Group A including 24 patients who underwent major hepatectomy without PLF and Group B including 8 patients who underwent major hepatectomy and developed PLF) were compared in a case-matched analysis with a 3:1 ratio according to patients baseline characteristics as shown in Table 1.
Table 1 Baseline characteristics. Variable Age (yr), mean SD Gender (M/F), n (%) Body mass index (kg/m2), mean SD Primary cancer (colon/rectum) Liver metastases at diagnosis Sinchronous/metachronous Number of liver lesions, median (range) Total tumor diameter in cm, mean SD Resectable/unresectable Preoperative portal vein embolization Two stage hepatectomy program ASA > 2
Group A
Group B
24
8
P
62 8 59 12 NS 13/11 (54.2/45.8) 5/3 (62.5/37.5) NS 22.9 3.6 25.5 5 NS 15/9 (62.5/37.5)
6/2 (75/25)
NS
14/10 (58.3/41.7) 5/3 (62.5/37.5) NS 4 (1e10) 2 (1e4) NS 7.7 4.1
5.1 4.5
NS
19/5 (79.1/20.9) 3 (12.5)
6/2 (75/25) 1 (12.5)
NS NS
3 (12.5) 4 (16.7)
1 (12.5) 2 (25)
NS NS
was used in addition to CT as a first line treatment. Response to CT was evaluated every 4e6 weeks using computed tomography, according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria. Preoperative workup Before surgery, all patients were evaluated by thoracoabdominal and pelvic imaging, blood tests including serum tumor markers, colonoscopy and/or rectal echoendoscopic ultrasound (EUS). Selected patients also underwent Positron Emission Tomography (PET), to exclude presence of extrahepatic disease. The following biochemical data were assessed before liver resection: hemoglobin level, serum creatinine level, serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels, gamma glutamyltransferase and alkaline phosphatase levels, serum total bilirubin level, serum albumin level and International Normalized Ratio (INR). APRI score was calculated for every patient as follows: 2 3 U 7 6 L 4AST level ULN 5
9 Platelet count 10L
100:
FIB-4 score was instead calculated as follows: Neoadjuvant systemic chemotherapy Patients with unresectable or marginally resectable disease were candidates to CT to downsize liver metastases and convert them to resectability. All the patients were treated by a combination of fluorouracil, leucovorin and oxaliplatin. In most of them a biologic agent (bevacizumab)
"
age years AST UL sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi# Platelet count
109 L
ðALTÞ
U L :
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
F. Ratti et al. / EJSO xx (2014) 1e7
Surgical procedure Liver resection was performed at least four weeks after the last administration of CT. The technique for liver resection is described elsewhere, both for open and laparoscopic liver resection.16 Resection of 3 or more liver segments according to Couinaud was considered a major hepatectomy.17 Abdominal exploration and intraoperative ultrasound were used to determine the extent of liver disease. Transection of hepatic parenchyma was carried out by a combination of ultrasonic dissector, harmonic scalpel and wet bipolar forceps and was performed under low central venous pressure (5 cm H2O). Intermittent Pringle maneuver was used on demand during liver transection, to control intraoperative blood loss. Intraoperative and postoperative outcome Postoperative complications were reviewed for 90 days following liver resection and were graded retrospectively according to Dindo-Clavien classification of surgical complications.18 PLF was defined according to ISGLS definition, as the persistence of increased International Normalized Ratio (INR) and serum bilirubin on or after POD5. In this classification PLF is stratified into three severity grades: grade A implies abnormality of laboratory parameters without any need for treatments or deviations from normal postoperative course; grade B includes patients with signs of PLF requiring only non-invasive treatments; grade C includes patients requiring even invasive treatments.15 Postoperative mortality was defined as any death during postoperative hospitalization or within 90 days after resection. Statistical analysis Pathologic data, operative details, surgical outcomes, biochemical parameters and chronic liver disease scores were compared using the c2 test or Fisher’s exact test for categorical data and the ManneWhitney U test for ordinal data in a case-matched analysis. All data were expressed as mean plus the standard deviation or median and range when appropriate. Significance was defined as p < 0.05. Univariate analysis was used to examine the relationship between PLF incidence and preoperative and/or postoperative variables. Pearson coefficient was used to analyze correlation between chronic liver disease scores and biochemical data. A ROC curve analysis was performed to assess the accuracy of APRI and FIB-4 scores in prediction of PLF. All variables associated with incidence of PLF with p < 0.05 in univariate analysis were entered into a multivariate logistic regression analysis. All analyses were performed using the statistical package SPSS 18.0 (SPSS, Chicago, IL, USA). Results Clinical and pathological characteristics of patients are summarized in Table 1. Group A included 24 patients
3
undergoing major or extended hepatectomy without signs of PLF in the postoperative period. CLM presentation was synchronous in 14 (58.3%) patients and metachronous in 10 (41.7%); 19 patients (79.1%) had resectable disease at diagnosis, while in 5 patients (20.9%) CLM were unresectable at presentation. Median number of lesions was 4 (range: 1e10), with a mean tumor diameter of 7.7 4.1 cm. Group B consisted of 8 patients who developed PLF after major or extended liver resection: incidence of PLF in patients affected by CLM who underwent oxaliplatinbased NACT within four months before liver resection was significantly higher compared both to PLF incidence in patients undergoing liver surgery for CLM in our institution (4%) and PLF incidence in the whole series of liver resections (5.4%) (p < 0.05). Six (75%) patients had primary colonic cancer, while 2 (25%) had rectal cancer. CLM presentation was synchronous in 5 (62,5%) patients and metachronous in 3 (37.5%); 6 patients (75%) had resectable disease at diagnosis, while in 3 (37.5%) patients CLM were unresectable at presentation. Median number of lesions at presentation was 2 (range: 1e4), with a mean total tumor diameter of 5.1 4.5 cm. Comparison in terms of preoperative characteristics is shown in Table 2. Patients in Group B underwent a median of 9 (range: 6e16) cycles of CT before surgery, while patients in Group An underwent a median of 6 cycles (range: 3e14), without statistical significance. 16 (66.7%) patients belonging to Group A received preoperative administration Table 2 Preoperative characteristics. Variable
Chemotherapy regimen Number of cycles, median (range) Bevacizumab, n (%) Interval between chemotherapy and surgery, median (range) Partial or complete radiological response to CT, n (%) Preoperative biochemical tests, mean SD Serum aspartate aminotransferase, U/L Alkalin phosphatase, U/L Gamma glutamyl transferase, U/L Serum bilirubin, mg/dL International Normalized Ratio (INR) Platelet count , 109/L Median Fib-4a score (range) Median APRIb score (range) a b
Group A
Group B
P
24
8
6 (3e14)
9 (6e16)
NS
16 (66.7) 11 (6e15)
3 (37.5) 7 (3e10)
p < 0.05 p < 0.05
16 (66.7)
5 (62.5)
NS
35 41
62 43
p < 0.05
97 48 112 36
190 56 189 34
NS NS
0.8 0.3 0.95 0.25
0.9 0.3 0.97 0.31
NS NS
180 129 1.58 (0.27e7.68) 0.30 (0.06e2.21)
105 95 2.46 (0.86e13.65) 0.53 (0.14e4.26)
p < 0.05 p < 0.05 p < 0.05
FIB-4 ¼ fibrosis-4. APRI ¼ aspartate aminotransferase to platelet ratio index.
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
4
F. Ratti et al. / EJSO xx (2014) 1e7
of bevacizumab, while a significantly lower number of patients in Group B (3; 37.5%) received it (p < 0.05). Median interval between last administration of chemotherapy and surgery was significantly shorter in Group B (7 weeks, range: 3e10) than in Group A (11 weeks, range: 6e15). Blood tests showed no significant differences in terms of preoperative alkaline phosphatase, gamma glutamyltransferase, total bilirubin and INR. Serum AST level was significantly lower in Group A (35 41 U/L), compared to Group B (62 43 U/ L) (p < 0.05); on the contrary, in Group A platelet count was significantly higher (180 129 10^9/L) compared to Group B (105 95 10^9/L) (p < 0.05). Median FIB-4 score was respectively 1.58 (range: 0.27e7.68) in Group A and 02.46 (range: 0.86e13.65) in Group B (p < 0.05). Patients in Group A had a significantly lower median APRI score compared to Group B (respectively 0.30, range: 0.06e2.21; and 0.53 range: 0.14e4.26) with p < 0.05. Comparison between Groups in terms of intra and postoperative course is shown in Table 3. Mortality was nil in Group A and 62.5% in Group B (p < 0.05): four patients in the latter group died as a consequence of progressive and irreversible PLF and one following septic shock secondary to severe cholangitis. Other causes of morbidity, other than PLF, were significantly more frequent in Group B (100% versus 13.2% in Group A, p < 0.05). One patient (4.2%) in Group A and 2 in Group B developed biliary fistula, classified as Grade II complication according to Clavien. Drainage of pleural effusion was required in 2 patients of Group A and 5 patients in Group B. 4.2% of patients in Group A and 87.5% in Group B developed ascites in the postoperative period. 12.5% of patients in Group A and 100% in Group B developed other complications (wound infection, urinary tract infection, atrial fibrillation) after surgery, as summarized in Table 3. Univariate and multivariate analyses (Table 4) were used to identify independent predictive factors associated with PLF. Univariate analysis showed that length of portal clamping (p ¼ 0.03), number of CT cycles (p ¼ 0.02) and interval CT/surgery (p ¼ 0.0024), platelet count (p ¼ 0.008), preoperative ALT level (p ¼ 0.009), FIB-4 score (p ¼ 0.005) and APRI score (p ¼ 0.0001), total bilirubin (p ¼ 0.0034) and INR (p ¼ 0.049) were correlated to PLF. AST level and platelet count were excluded from multivariate analyses because they were correlated to APRI score. At multivariate analysis APRI score resulted independently associated to PLF (p ¼ 0.004), together with total serum bilirubin level. ROC curve analysis of APRI score is shown in Fig. 1 and demonstrated a good accuracy of APRI score in PLF risk prediction. In addition, Pearson correlation test was used to verify association between APRI score and variables regarding postoperative course. There was a significant correlation between APRI score and total serum bilirubin in POD5 (p ¼ 0.0001) and between APRI score and intraoperative blood loss (p ¼ 0.012).
Table 3 Intra- and postoperative outcome analysis. Variable
Pringle maneuver, n (%) Lenght of Pringle maneuver, mean SD Blood loss (mL), mean SD Volume of intraoperative blood transfusions (mL), mean SD Total volume of blood transfusions (mL), mean SD Postoperative mortality, n (%) Complications, n (%) Biliary fistula (Grade II) Pleural effusion requiring drainage (Grade IIIa) Ascites (Grade II) Other complications Most severe complication, n (%) Grade I Grade II Grade IIIa Grade Ivb Markers of PLFa, mean SD POD 3 Total bilirubin, mg/dL INRb POD 5 Total bilirubin, mg/dL INR a b
Group A
Group B
24
8
16 (66.7) 28 18
6 (75) 45 25
NS p < 0.05
350 200 380 220
450 320 420 280
NS NS
160 290
590 360
NS
0 4 1 2
5 8 2 5
(62,5) (100) (25) (62.5)
p < 0.05 p < 0.05 NS p < 0.05
2 (8.3) 3 (12.5)
7 (87.5) 8 (100)
p < 0.05 p < 0.05
1 1 2 0
1 2 3 2
(0) (16.7) (4.2) (8.3)
(4.2) (4.2) (8.3) (0)
P
(12.5) (25) (37.5) (25)
1.0 0,3 2.4 2.8 p < 0.05 1.10 0.32 1.45 0.45 p < 0.05 0.95 0.4 4.2 2.5 1.02 0.26 1.46 0.8
p < 0.05 p < 0.05
PLF ¼ postoperative liver failure. INR ¼ international normalized ratio.
Discussion In the last decade, survival of patients with CLM has improved significantly thanks to both an increasing Table 4 Univariate and multivariate analysis. Univariate Multivariate analysis analysis P Hemoglobin, mg/dL Platelet count Creatinine, mg/dL Total bilirubin, mg/dL Preoperative ALTa level, U/L Preoperative ASTb level, U/L Alkaline phosphatase, U/L G-glutamyl transferase, U/L INR, International Normalized Ratio Intraoperative blood loss, mL Length of portal clamping, minutes Interval CTc/surgery Number of CT cycles Fib-1 Score APRI score a b c
NS 0.008 NS 0.0034 0.009 NS NS NS 0.049 0.03 0.03 0.0024 0.02 0.005 0.0001
P
Odds ratio
NS 0.007 2.8 (1.9e11.5) NS
NS NS NS NS NS NS 0.004 3.6 (2.1e29.4)
ALT ¼ alanine aminotransferase. AST ¼ aspartate aminotransferase. CT ¼ chemo therapy.
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
F. Ratti et al. / EJSO xx (2014) 1e7
Figure 1. ROC curve.
diffusion of surgical approach, and major effectiveness of CT regimens leading to better overall survival even in unresected patients.1 In patients with resectable liver metastases at presentation, the correct timing of CT administration has not been clearly defined yet, but frequently preoperative CT with neoadjuvant intent is strongly suggested in these patients, in order to increase the proportion of R0 resection or to perform more conservative resections.19 Among side effects of preoperative CT, risk of liver metastases disappearance at imaging is frequently reported, resulting in a more challenging procedure for the surgeon in terms of tumor recognition and removal.20,21 Secondly, preoperative CT is a factor with demonstrated impact on liver parenchyma, resulting in an impaired function induced by therapy (CALI).4e8 Association of irinotecan based regimens with the presence of steatohepatitis6 has been widely demonstrated in the literature, while oxaliplatin based CT seem to be responsible for vascular endothelial damage, Sinusoidal Obstruction Syndrome (SOS) above all.8 These histopathological abnormalities increase the risk of postoperative complications, impairing both liver function (and therefore functional reserve) and structure: indeed, they seem to be associated to higher risk of bleeding and onset of PLF.5,8 An accurate predictive tool to grade preoperatively CALI and consequently to calculate risk of liver failure presently lacks: availability of a simple and cost effective prognostic marker would instead influence treatment strategies in management of patients with CLM, optimizing outcome of pre-, intra- and postoperative period. This study aimed to identify factors associated with the onset of PLF in patients undergoing hepatic surgery for CLM, focusing on the role of chronic liver disease scores. These were initially introduced for evaluation of grade of liver impairment due to infection by hepatitis C virus,11 but their value could be extended in the field of secondary lesions: indeed, drugs used as first lines chemotherapy treatments, induce a parenchymal damage that was proved to be histologically similar to that detected in patients with cirrhosis.12,13
5
Incidence of PLF in patients undergoing surgery after oxaliplatin based CT was significantly higher compared to PLF incidence in the whole series of liver resections performed in the same institution (including patients with high risk of PLF as those affected by Klatskin tumors) and also in the subset of patients with CLM. The hypothesis that CT has an impact on operative outcome is therefore consistent since, in Group B, a relatively high median number of chemotherapy cycles within limited period of time between CT and surgery was recorded. Nakano et al.4 also reported an association between number of cycles of CT and presence of sinusoidal damage without demonstrating any real importance of dose of CT preoperatively administered. In univariate analysis, Soubrane and his group12 revealed a consistent association between the interval CT/intervention and presence of sinusoidal damage. In present study, patients belonging to Group A were more frequently treated with bevacizumab, in combination with standard CT. This observation may be consistent with a previous study by Ribero et al.22 showing a protective role of bevacizumab towards CALI with a reduction of both its incidence and severity. This is the first study to analyze the relevance of chronic liver disease scores focusing on patients developing PLF and performing a case-matched analysis with patients undergoing major liver resection presenting the same baseline characteristics. In study group both FIB-4 and APRI score were significantly higher compared to Group A, showing a close relationship with biochemical markers currently used to define the onset of the PLF, i.e. INR and total serum bilirubin on POD5.18 APRI score in particular was associated, at multivariate analysis, with the onset of PLF, with a good degree of accuracy (sensitivity 93.8% and specificity of 66.4% for a cutoff value of 0.30) at ROC analysis (AUC 0.72 with p ¼ 0.003). In this experience APRI score represented an accurate index since it considers both existing liver damage (as reflected by transaminases value) and sinusoidal damage (which would have an effect on platelet count). Platelets play a key role in these patients both because thrombocytopenia increases the risk of bleeding (indeed APRI score is correlated with intraoperative blood loss)10 and because platelets seem to be involved in stimulation of hepatocyte regeneration through a mechanism mediated by serotonin.23 CALI seems to play a detrimental role on hepatic functional reserve resulting in increased postoperative morbidity: indeed, it has been demonstrated that risk of PLF is closely dependent from both volume of liver remaining after surgery, and from its functional and regenerative capability. Therefore, liver resection extent is not enough in order to accurately predict postoperative liver function, since it does not provide informations to discriminate an underlying severe liver disease.9 Clinical utility of APRI and FIB-4 scores is based upon use of biological and
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
6
F. Ratti et al. / EJSO xx (2014) 1e7
biochemical blood parameters that are simple, cost effective and easily available preoperatively.12 Furthermore, serum levels of transaminases and platelets count are independently associated with complicated postoperative courses after hepatic resection. Some authors have already analyzed the potential value of preoperative liver biopsy for evaluation of liver impairment resulting from preoperative CT administration: this kind of screening could potentially be performed in a selected cluster of patients, candidates to biopsy according to cutoff value of APRI.24,25 Unfortunately, histological specimen of 15 mm obtained by biopsy is frequently not sufficient to assess CALI since it has widespread but heterogeneous distribution within parenchyma.25 Other possibilities to minimize postoperative morbidity in high risk patients is represented by use of APRI cutoff to screen out potential candidates for surgery. In particular, normalization of biological markers would constitute a warning for correct surgical timing: indeed, hepatectomy is currently performed at least thirty days after the last administration of CT; conversely, APRI evaluation and normalization could shorten this interval or postpone intervention whenever required with a patient-tailored strategy. Secondly, APRI value should be monitored during the period of CT administration so that systemic therapy could be discontinued when the patient has become resectable and/or the value of APRI has become borderline. As stated before, there is no evidence for a precise number of CT cycles conditioning liver damage, so an easy available method to evaluate parenchymal impairing resulting from drug administration might avoid further hepatic functional reserve reduction. This study constitutes a preliminary experience for a possible alternative use of chronic liver disease scores in liver surgery, and may provide basis for further experiences focusing on several aspects of the same topic. First of all, APRI score and cutoff should be validated on a larger scale in a prospectively designed study. Furthermore, score value could be proposed as an indicator of underlying parenchymal impairment to select patients for preoperative portal vein embolization, in association with study of future remnant liver volume: through specific correlation analysis between liver volumes and value of APRI this topic could be deepen. Unfortunately this evaluation could not be performed in present series due to small number of patients undergoing PVE in the setting of two stage hepatectomy program: in these patients APRI score should be evaluated before the first surgical step and then monitored during interval between portal vein occlusion and major hepatectomy to assess liver functional recovery along with time progression from last CT (even though APRI score could be partially influenced by first surgical procedure or PVE, leading to abnormalities in platelet count and transaminases unrelated to CALI). Lastly, APRI variations could be monitored in the period following interruption of chemotherapy in order to evaluate
APRI accuracy as a marker of restored hepatic function along with time progression. In conclusion, APRI score takes into account both liver damage and platelet count: it is reliable in preoperative evaluation of lesions induced by oxaliplatin based CT regimens, which accounts for the high incidence of postoperative liver failure in patients undergoing liver surgery for CLM. Conflict of interest The authors declare that the article is an original work, has not been published before, and is not being considered for publication elsewhere in its final form, in either printed or electronic media. Any republication of the data will not constitute redundant publication, will not breach copyright, and will reference the original publication. The authors declare that they have no financial or any other competing interest in publication. The authors declare that they have no financial or any other competing interest in publication.
References 1. Kopetz S, Chang GJ, Overman MJ, et al. Improved survival in metastatic colorectal cancer is associated with adoption of hepatic resection and improved chemotherapy. J Clin Oncol 2009 Aug 1;27(22):3677– 83. 2. Nordlinger B, Sorbye H, Glimelius B, et al. Perioperative chemotherapy with FOLFOX4 and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC Intergroup trial 40983): a randomized controlled trial. Lancet 2008 Mar 22; 371(9617):1007–16. 3. Blazer 3rd DG, Kishi Y, Maru DM, et al. Pathologic response to preoperative chemotherapy: a new outcome end point after resection of hepatic colorectal metastases. J Clin Oncol 2008 Nov 20;26(33): 5344–51. 4. Nakano H, Oussoultzoglou E, Rosso E, et al. Sinusoidal injury increases morbidity after major hepatectomy in patients with colorectal liver metastases receiving preoperative chemotherapy. Ann Surg 2008 Jan;247(1):118–24. 5. Karoui M, Penna C, Amin-Hashem M, et al. Influence of preoperative chemotherapy on the risk of major hepatectomy for colorectal liver metastases. Ann Surg 2006 Jan;243(1):1–7. 6. Vauthey JN, Pawlik TM, Ribero D, et al. Chemotherapy regimen predicts steatohepatitis and an increase in 90-day mortality after surgery for hepatic colorectal metastases. J Clin Oncol 2006 May 1;24(13): 2065–72. 7. Pulitano C, Aldrighetti L, Arru M, et al. Influence of preoperative chemotherapy on the risk of major hepatectomy for colorectal liver metastases. Ann Surg 2006 Nov;244(5):833–5. 8. Rubbia-Brandt L, Audard V, Sartoretti P, et al. severe hepatic sinusoidal obstruction associated with oxaliplatin-bades chemotherapy in patients with metastatic colorectal cancer. Ann Oncol 2004 Mar;15(3): 460–6. 9. Hammond JS, Guha IN, Backingham IJ, Lobo DN. Prediction, prevention and management of postresection liver failure. Br J Surg 2011 Sep;98(9):1188–200. 10. Aloia T, Sebagh M, Plasse M, et al. Liver histology and surgical outcomes after preoperative chomotherapy with fluorouracil plus
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
F. Ratti et al. / EJSO xx (2014) 1e7
11.
12.
13.
14.
15.
16.
17.
oxaliplatin in colorectal cancer liver metastases. J Clin Oncol 2006 Nov 1;24(31):4983–90. Wai CT, Greenson JK, Fontana RJ, et al. A simple and noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology 2003 Aug;38(2):518–26. Soubrane O, Brouquet A, Zalinski S, et al. Predicting high grade lesions of sinusoidal obstruction syndrome related to oxaliplatin-based chemotherapy for colorectal liver metastases: correlation with posthepatectomy outcome. Ann Surg 2010 Mar;251(3):454–60. Farges O, Malassagne B, Flejou JF, Balzan S, Sauvanet A, Belghiti J. Risk of major liver resection in patients with underlying chronic liver disease: a reappraisal. Ann Surg 1999 Feb;229(2):210–5. Taketomi A, Kitagawa D, Itoh S, et al. Trends in morbidity and mortality after hepatic resection for hepatocellular carcinoma: an institute’s experience with 625 patients. J Am Coll Surg 2007 Apr; 204(4):580–7. Rahbari NN, Garden OJ, Padbury R, et al. Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS). Surgery 2011;149:713–24. Aldrighetti L, Pulitano C, Arru M, Catena M, Finazzi R, Ferla G. “Technological” approach versus clamp crushing technique for hepatic parenchymal transection: a comparative study. J Gastrointest Surg 2006 Jul-Aug;10(7):974–9. Terminology Committee of the International Hepato-Pancreato-Biliary Association. The IHPBA Brisbane 2000 terminology of liver anatomy and resections. HPB Surg 2000;2:333–9.
7
18. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004 Aug;240(2):205–13. 19. Nordlinger B, Van Cutsem E, Gruenberg T, et al. Combination of surgery and chemotherapy and the role of targeted agents in the treatment of patients with colorectal liver metastases: recommendations from an expert panel. Ann Oncol 2009 Jun;20(6):985–92. 20. Van Vledder MG, de Jong MC, Pawlik TM, Schulik RD, Diaz LA, Choti MA. Disappearing colorectal liver metastases after chemotherapy: should web e concerned? J Gastrointest Surg 2010 Nov; 14(11):1691–700. 21. Auer RC, White RR, Kemeny NE, et al. Predictors of a true complete response among disappearing liver metastases from colorectal cancer after chemotherapy. Cancer 2010 Mar 15;116(6). 22. Ribero D, Wang H, Donadon M, et al. Bevacizumab improves pathologic response and protects against hepatic injury in patients treated with oxaliplatin-based chemotherapy for colorectal liver metastases. Cancer 2007 Dec 15;110(12):2761–7. 23. Lesurtel M, Graf R, Aleil B, et al. Platelet-derived serotonin mediates liver regeneration. Science 2006 Apr 7;312(5770):104–7. 24. McHutchison JG, Blatt LM, de Medina M, et al. Measurement of serum hyaluronic acid in patients with chronic hepatitis C and its relationship to liver histology. Consensus Interferon Study Group. J Gastroenterol Hepatol 2000 Aug;15(8):945–51. 25. Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology 2003 Dec;38(6):1449–57.
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
ScienceDirect EJSO xx (2014) 1e7
www.ejso.com
Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis F. Ratti*,1, F. Cipriani 1, M. Catena 1, M. Paganelli 1, L. Aldrighetti 1 Department of General Surgery, Ospedale San Raffaele, Via Olgettina 60, 20132 Milano, Italy Accepted 26 June 2014 Available online - - -
Abstract Aim: An accurate and noninvasive tool to predict Chemotherapy Associated Liver Injury (CALI) still lacks. Study aimed to evaluate chronic liver disease scores (Aspartate aminotransferase to Platelet Ratio Index, APRI and Fibrosis-4, FIB-4) as Postoperative Liver Failure (PLF) predictors in patients treated with Oxaliplatin for Colorectal Liver Metastases (CLM). Methods: 8 patients who developed PLF after major hepatectomy (Group B) were compared to 24 patients who did not develop PLF (Group A) in a case-matched analysis for patients and disease characteristics. ROC curves analysis was performed to assess score accuracy. Results: In Group A number of CT cycles was lower, (6 vs 9, p NS), interval between treatment and surgery was longer (11 vs 7 weeks, p < 0.05) and bevacizumab was more frequently administered (66.7% vs 37.5%, p < 0.05). In Group B median APRI score was 0.53 (range: 0.86e4.26) whereas in Group A was 0.30 (range: 0.06e2.21, p < 0.05). Median FIB-4 score was 2.46 (range: 0.86e13.65) in Group B and 1.58 (range: 0.27e7.68) in Group A ( p < 0.001). Multivariate analysis showed a significant correlation between APRI and the onset of PLF. A good accuracy of APRI score was evident in ROC curves with an area under the curve of 0.72 ( p 0.003). Conclusions: APRI score is calculated considering both liver damage and platelet count, it is cost effective and easily available. This study demonstrates that there is a good accuracy in PLF prediction and consequently in CT induced liver damage evaluation. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Chemotherapy; Liver surgery; Sinusoidal obstruction; postoperative liver failure; Metastases
Introduction Almost 50% of patients affected by colorectal cancer develop liver metastases (CLM) and the only curative treatment for this subset of patients is represented by surgical removal of lesions1: hence the need for a multidisciplinary management of these patients, together with radiologists, radiotherapists and medical oncologists, to increase resectability rate, therefore improving long term outcome. An optimal timing for systemic ChemoTherapy (CT) administration has not been defined yet: despite this, usefulness * Corresponding author. Tel.: þ39 02 26437811, þ39 348 2411961; fax: þ39 02 26437807. E-mail address: [email protected] (F. Ratti). 1 Hepatobiliary Surgery Unit.
and advantages resulting from NeoAdjuvantChemoTherapy (NACT) are well recognized.2 Indeed, it allows tumor downstaging, expanding both rate of R0 resections and number of candidates to surgery among patients with initially unresectable disease. In addition, preoperative identification of good responders is possible, while patients with disease progression during therapy might be excluded from surgery.3 In spite of its benefits, CT may induce liver injury on nonneoplastic parenchyma leading to increased morbidity and mortality after liver resection in patients who received oxaliplatin or irinotecan based regimens.4e7 Oxaliplatin in particular is associated with Sinusoidal Obstruction Syndrome (SOS), responsible for a significant risk of bleeding and Postoperative Liver Failure (PLF).8 PLF is one of the most severe causes of mortality following liver surgery9: indeed the capability of the liver in tolerating extended resections is based
http://dx.doi.org/10.1016/j.ejso.2014.06.011 0748-7983/Ó 2014 Elsevier Ltd. All rights reserved. Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
2
F. Ratti et al. / EJSO xx (2014) 1e7
upon parenchymal functional reserve, that can be potentially impaired as a result of prolonged chemotherapy.4e8 A non-invasive and accurate tool to predict Chemotherapy Induced Liver Injury (CALI) still lacks, but should be essential to evaluate risk of PLF resulting from parenchymal impairment, therefore allowing better patients selection before major resections. APRI (Aspartate aminotransferase to Platelet Ratio Index)10 and FIB-4 (Fibrosis-4)11 scores have been reported to grade liver fibrosis in patients with hepatitis C virus infection, even though their importance has been recently high lightened in prediction of high grade lesions of SOS.12e14 Aim of the study was to identify clinical and biochemical factors associated to PLF in patients affected by CLM who underwent major liver surgery and previous oxaliplatin based CT in a case-matched analysis. Analysis was focused on chronic liver disease scores. Secondary aim was to assess accuracy of chronic liver disease scores as predictors of PLF. Methods Patients From January 2004 to December 2012, 1413 liver resections were performed at the Hepatobiliary Surgery Unit of San Raffaele Hospital. Data from these patients have been collected in a prospective database and are now retrospectively reviewed. 463 (32.8%) liver resections were performed in patients affected by CLM. Of these, 258 patients were treated with systemic chemotherapy before resection and in 161 (34.7%) of them oxaliplatin-based regimens were used. Out of 101 patients who underwent oxaliplatin based NACT within 4 months before liver resection, without demonstrated chronic liver disease, 54 major or extended hepatectomies were performed. Among this last group, 8 patients developed PLF according to ISGLS definition (in particular all patients had grade B or C PLF, while none of them had grade A PLF)15 and were compared in a case-matched analysis to 24 patients, belonging to the same population, who did not develop PLF in the postoperative period. The two groups (Group A including 24 patients who underwent major hepatectomy without PLF and Group B including 8 patients who underwent major hepatectomy and developed PLF) were compared in a case-matched analysis with a 3:1 ratio according to patients baseline characteristics as shown in Table 1.
Table 1 Baseline characteristics. Variable Age (yr), mean SD Gender (M/F), n (%) Body mass index (kg/m2), mean SD Primary cancer (colon/rectum) Liver metastases at diagnosis Sinchronous/metachronous Number of liver lesions, median (range) Total tumor diameter in cm, mean SD Resectable/unresectable Preoperative portal vein embolization Two stage hepatectomy program ASA > 2
Group A
Group B
24
8
P
62 8 59 12 NS 13/11 (54.2/45.8) 5/3 (62.5/37.5) NS 22.9 3.6 25.5 5 NS 15/9 (62.5/37.5)
6/2 (75/25)
NS
14/10 (58.3/41.7) 5/3 (62.5/37.5) NS 4 (1e10) 2 (1e4) NS 7.7 4.1
5.1 4.5
NS
19/5 (79.1/20.9) 3 (12.5)
6/2 (75/25) 1 (12.5)
NS NS
3 (12.5) 4 (16.7)
1 (12.5) 2 (25)
NS NS
was used in addition to CT as a first line treatment. Response to CT was evaluated every 4e6 weeks using computed tomography, according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria. Preoperative workup Before surgery, all patients were evaluated by thoracoabdominal and pelvic imaging, blood tests including serum tumor markers, colonoscopy and/or rectal echoendoscopic ultrasound (EUS). Selected patients also underwent Positron Emission Tomography (PET), to exclude presence of extrahepatic disease. The following biochemical data were assessed before liver resection: hemoglobin level, serum creatinine level, serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels, gamma glutamyltransferase and alkaline phosphatase levels, serum total bilirubin level, serum albumin level and International Normalized Ratio (INR). APRI score was calculated for every patient as follows: 2 3 U 7 6 L 4AST level ULN 5
9 Platelet count 10L
100:
FIB-4 score was instead calculated as follows: Neoadjuvant systemic chemotherapy Patients with unresectable or marginally resectable disease were candidates to CT to downsize liver metastases and convert them to resectability. All the patients were treated by a combination of fluorouracil, leucovorin and oxaliplatin. In most of them a biologic agent (bevacizumab)
"
age years AST UL sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi# Platelet count
109 L
ðALTÞ
U L :
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
F. Ratti et al. / EJSO xx (2014) 1e7
Surgical procedure Liver resection was performed at least four weeks after the last administration of CT. The technique for liver resection is described elsewhere, both for open and laparoscopic liver resection.16 Resection of 3 or more liver segments according to Couinaud was considered a major hepatectomy.17 Abdominal exploration and intraoperative ultrasound were used to determine the extent of liver disease. Transection of hepatic parenchyma was carried out by a combination of ultrasonic dissector, harmonic scalpel and wet bipolar forceps and was performed under low central venous pressure (5 cm H2O). Intermittent Pringle maneuver was used on demand during liver transection, to control intraoperative blood loss. Intraoperative and postoperative outcome Postoperative complications were reviewed for 90 days following liver resection and were graded retrospectively according to Dindo-Clavien classification of surgical complications.18 PLF was defined according to ISGLS definition, as the persistence of increased International Normalized Ratio (INR) and serum bilirubin on or after POD5. In this classification PLF is stratified into three severity grades: grade A implies abnormality of laboratory parameters without any need for treatments or deviations from normal postoperative course; grade B includes patients with signs of PLF requiring only non-invasive treatments; grade C includes patients requiring even invasive treatments.15 Postoperative mortality was defined as any death during postoperative hospitalization or within 90 days after resection. Statistical analysis Pathologic data, operative details, surgical outcomes, biochemical parameters and chronic liver disease scores were compared using the c2 test or Fisher’s exact test for categorical data and the ManneWhitney U test for ordinal data in a case-matched analysis. All data were expressed as mean plus the standard deviation or median and range when appropriate. Significance was defined as p < 0.05. Univariate analysis was used to examine the relationship between PLF incidence and preoperative and/or postoperative variables. Pearson coefficient was used to analyze correlation between chronic liver disease scores and biochemical data. A ROC curve analysis was performed to assess the accuracy of APRI and FIB-4 scores in prediction of PLF. All variables associated with incidence of PLF with p < 0.05 in univariate analysis were entered into a multivariate logistic regression analysis. All analyses were performed using the statistical package SPSS 18.0 (SPSS, Chicago, IL, USA). Results Clinical and pathological characteristics of patients are summarized in Table 1. Group A included 24 patients
3
undergoing major or extended hepatectomy without signs of PLF in the postoperative period. CLM presentation was synchronous in 14 (58.3%) patients and metachronous in 10 (41.7%); 19 patients (79.1%) had resectable disease at diagnosis, while in 5 patients (20.9%) CLM were unresectable at presentation. Median number of lesions was 4 (range: 1e10), with a mean tumor diameter of 7.7 4.1 cm. Group B consisted of 8 patients who developed PLF after major or extended liver resection: incidence of PLF in patients affected by CLM who underwent oxaliplatinbased NACT within four months before liver resection was significantly higher compared both to PLF incidence in patients undergoing liver surgery for CLM in our institution (4%) and PLF incidence in the whole series of liver resections (5.4%) (p < 0.05). Six (75%) patients had primary colonic cancer, while 2 (25%) had rectal cancer. CLM presentation was synchronous in 5 (62,5%) patients and metachronous in 3 (37.5%); 6 patients (75%) had resectable disease at diagnosis, while in 3 (37.5%) patients CLM were unresectable at presentation. Median number of lesions at presentation was 2 (range: 1e4), with a mean total tumor diameter of 5.1 4.5 cm. Comparison in terms of preoperative characteristics is shown in Table 2. Patients in Group B underwent a median of 9 (range: 6e16) cycles of CT before surgery, while patients in Group An underwent a median of 6 cycles (range: 3e14), without statistical significance. 16 (66.7%) patients belonging to Group A received preoperative administration Table 2 Preoperative characteristics. Variable
Chemotherapy regimen Number of cycles, median (range) Bevacizumab, n (%) Interval between chemotherapy and surgery, median (range) Partial or complete radiological response to CT, n (%) Preoperative biochemical tests, mean SD Serum aspartate aminotransferase, U/L Alkalin phosphatase, U/L Gamma glutamyl transferase, U/L Serum bilirubin, mg/dL International Normalized Ratio (INR) Platelet count , 109/L Median Fib-4a score (range) Median APRIb score (range) a b
Group A
Group B
P
24
8
6 (3e14)
9 (6e16)
NS
16 (66.7) 11 (6e15)
3 (37.5) 7 (3e10)
p < 0.05 p < 0.05
16 (66.7)
5 (62.5)
NS
35 41
62 43
p < 0.05
97 48 112 36
190 56 189 34
NS NS
0.8 0.3 0.95 0.25
0.9 0.3 0.97 0.31
NS NS
180 129 1.58 (0.27e7.68) 0.30 (0.06e2.21)
105 95 2.46 (0.86e13.65) 0.53 (0.14e4.26)
p < 0.05 p < 0.05 p < 0.05
FIB-4 ¼ fibrosis-4. APRI ¼ aspartate aminotransferase to platelet ratio index.
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
4
F. Ratti et al. / EJSO xx (2014) 1e7
of bevacizumab, while a significantly lower number of patients in Group B (3; 37.5%) received it (p < 0.05). Median interval between last administration of chemotherapy and surgery was significantly shorter in Group B (7 weeks, range: 3e10) than in Group A (11 weeks, range: 6e15). Blood tests showed no significant differences in terms of preoperative alkaline phosphatase, gamma glutamyltransferase, total bilirubin and INR. Serum AST level was significantly lower in Group A (35 41 U/L), compared to Group B (62 43 U/ L) (p < 0.05); on the contrary, in Group A platelet count was significantly higher (180 129 10^9/L) compared to Group B (105 95 10^9/L) (p < 0.05). Median FIB-4 score was respectively 1.58 (range: 0.27e7.68) in Group A and 02.46 (range: 0.86e13.65) in Group B (p < 0.05). Patients in Group A had a significantly lower median APRI score compared to Group B (respectively 0.30, range: 0.06e2.21; and 0.53 range: 0.14e4.26) with p < 0.05. Comparison between Groups in terms of intra and postoperative course is shown in Table 3. Mortality was nil in Group A and 62.5% in Group B (p < 0.05): four patients in the latter group died as a consequence of progressive and irreversible PLF and one following septic shock secondary to severe cholangitis. Other causes of morbidity, other than PLF, were significantly more frequent in Group B (100% versus 13.2% in Group A, p < 0.05). One patient (4.2%) in Group A and 2 in Group B developed biliary fistula, classified as Grade II complication according to Clavien. Drainage of pleural effusion was required in 2 patients of Group A and 5 patients in Group B. 4.2% of patients in Group A and 87.5% in Group B developed ascites in the postoperative period. 12.5% of patients in Group A and 100% in Group B developed other complications (wound infection, urinary tract infection, atrial fibrillation) after surgery, as summarized in Table 3. Univariate and multivariate analyses (Table 4) were used to identify independent predictive factors associated with PLF. Univariate analysis showed that length of portal clamping (p ¼ 0.03), number of CT cycles (p ¼ 0.02) and interval CT/surgery (p ¼ 0.0024), platelet count (p ¼ 0.008), preoperative ALT level (p ¼ 0.009), FIB-4 score (p ¼ 0.005) and APRI score (p ¼ 0.0001), total bilirubin (p ¼ 0.0034) and INR (p ¼ 0.049) were correlated to PLF. AST level and platelet count were excluded from multivariate analyses because they were correlated to APRI score. At multivariate analysis APRI score resulted independently associated to PLF (p ¼ 0.004), together with total serum bilirubin level. ROC curve analysis of APRI score is shown in Fig. 1 and demonstrated a good accuracy of APRI score in PLF risk prediction. In addition, Pearson correlation test was used to verify association between APRI score and variables regarding postoperative course. There was a significant correlation between APRI score and total serum bilirubin in POD5 (p ¼ 0.0001) and between APRI score and intraoperative blood loss (p ¼ 0.012).
Table 3 Intra- and postoperative outcome analysis. Variable
Pringle maneuver, n (%) Lenght of Pringle maneuver, mean SD Blood loss (mL), mean SD Volume of intraoperative blood transfusions (mL), mean SD Total volume of blood transfusions (mL), mean SD Postoperative mortality, n (%) Complications, n (%) Biliary fistula (Grade II) Pleural effusion requiring drainage (Grade IIIa) Ascites (Grade II) Other complications Most severe complication, n (%) Grade I Grade II Grade IIIa Grade Ivb Markers of PLFa, mean SD POD 3 Total bilirubin, mg/dL INRb POD 5 Total bilirubin, mg/dL INR a b
Group A
Group B
24
8
16 (66.7) 28 18
6 (75) 45 25
NS p < 0.05
350 200 380 220
450 320 420 280
NS NS
160 290
590 360
NS
0 4 1 2
5 8 2 5
(62,5) (100) (25) (62.5)
p < 0.05 p < 0.05 NS p < 0.05
2 (8.3) 3 (12.5)
7 (87.5) 8 (100)
p < 0.05 p < 0.05
1 1 2 0
1 2 3 2
(0) (16.7) (4.2) (8.3)
(4.2) (4.2) (8.3) (0)
P
(12.5) (25) (37.5) (25)
1.0 0,3 2.4 2.8 p < 0.05 1.10 0.32 1.45 0.45 p < 0.05 0.95 0.4 4.2 2.5 1.02 0.26 1.46 0.8
p < 0.05 p < 0.05
PLF ¼ postoperative liver failure. INR ¼ international normalized ratio.
Discussion In the last decade, survival of patients with CLM has improved significantly thanks to both an increasing Table 4 Univariate and multivariate analysis. Univariate Multivariate analysis analysis P Hemoglobin, mg/dL Platelet count Creatinine, mg/dL Total bilirubin, mg/dL Preoperative ALTa level, U/L Preoperative ASTb level, U/L Alkaline phosphatase, U/L G-glutamyl transferase, U/L INR, International Normalized Ratio Intraoperative blood loss, mL Length of portal clamping, minutes Interval CTc/surgery Number of CT cycles Fib-1 Score APRI score a b c
NS 0.008 NS 0.0034 0.009 NS NS NS 0.049 0.03 0.03 0.0024 0.02 0.005 0.0001
P
Odds ratio
NS 0.007 2.8 (1.9e11.5) NS
NS NS NS NS NS NS 0.004 3.6 (2.1e29.4)
ALT ¼ alanine aminotransferase. AST ¼ aspartate aminotransferase. CT ¼ chemo therapy.
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
F. Ratti et al. / EJSO xx (2014) 1e7
Figure 1. ROC curve.
diffusion of surgical approach, and major effectiveness of CT regimens leading to better overall survival even in unresected patients.1 In patients with resectable liver metastases at presentation, the correct timing of CT administration has not been clearly defined yet, but frequently preoperative CT with neoadjuvant intent is strongly suggested in these patients, in order to increase the proportion of R0 resection or to perform more conservative resections.19 Among side effects of preoperative CT, risk of liver metastases disappearance at imaging is frequently reported, resulting in a more challenging procedure for the surgeon in terms of tumor recognition and removal.20,21 Secondly, preoperative CT is a factor with demonstrated impact on liver parenchyma, resulting in an impaired function induced by therapy (CALI).4e8 Association of irinotecan based regimens with the presence of steatohepatitis6 has been widely demonstrated in the literature, while oxaliplatin based CT seem to be responsible for vascular endothelial damage, Sinusoidal Obstruction Syndrome (SOS) above all.8 These histopathological abnormalities increase the risk of postoperative complications, impairing both liver function (and therefore functional reserve) and structure: indeed, they seem to be associated to higher risk of bleeding and onset of PLF.5,8 An accurate predictive tool to grade preoperatively CALI and consequently to calculate risk of liver failure presently lacks: availability of a simple and cost effective prognostic marker would instead influence treatment strategies in management of patients with CLM, optimizing outcome of pre-, intra- and postoperative period. This study aimed to identify factors associated with the onset of PLF in patients undergoing hepatic surgery for CLM, focusing on the role of chronic liver disease scores. These were initially introduced for evaluation of grade of liver impairment due to infection by hepatitis C virus,11 but their value could be extended in the field of secondary lesions: indeed, drugs used as first lines chemotherapy treatments, induce a parenchymal damage that was proved to be histologically similar to that detected in patients with cirrhosis.12,13
5
Incidence of PLF in patients undergoing surgery after oxaliplatin based CT was significantly higher compared to PLF incidence in the whole series of liver resections performed in the same institution (including patients with high risk of PLF as those affected by Klatskin tumors) and also in the subset of patients with CLM. The hypothesis that CT has an impact on operative outcome is therefore consistent since, in Group B, a relatively high median number of chemotherapy cycles within limited period of time between CT and surgery was recorded. Nakano et al.4 also reported an association between number of cycles of CT and presence of sinusoidal damage without demonstrating any real importance of dose of CT preoperatively administered. In univariate analysis, Soubrane and his group12 revealed a consistent association between the interval CT/intervention and presence of sinusoidal damage. In present study, patients belonging to Group A were more frequently treated with bevacizumab, in combination with standard CT. This observation may be consistent with a previous study by Ribero et al.22 showing a protective role of bevacizumab towards CALI with a reduction of both its incidence and severity. This is the first study to analyze the relevance of chronic liver disease scores focusing on patients developing PLF and performing a case-matched analysis with patients undergoing major liver resection presenting the same baseline characteristics. In study group both FIB-4 and APRI score were significantly higher compared to Group A, showing a close relationship with biochemical markers currently used to define the onset of the PLF, i.e. INR and total serum bilirubin on POD5.18 APRI score in particular was associated, at multivariate analysis, with the onset of PLF, with a good degree of accuracy (sensitivity 93.8% and specificity of 66.4% for a cutoff value of 0.30) at ROC analysis (AUC 0.72 with p ¼ 0.003). In this experience APRI score represented an accurate index since it considers both existing liver damage (as reflected by transaminases value) and sinusoidal damage (which would have an effect on platelet count). Platelets play a key role in these patients both because thrombocytopenia increases the risk of bleeding (indeed APRI score is correlated with intraoperative blood loss)10 and because platelets seem to be involved in stimulation of hepatocyte regeneration through a mechanism mediated by serotonin.23 CALI seems to play a detrimental role on hepatic functional reserve resulting in increased postoperative morbidity: indeed, it has been demonstrated that risk of PLF is closely dependent from both volume of liver remaining after surgery, and from its functional and regenerative capability. Therefore, liver resection extent is not enough in order to accurately predict postoperative liver function, since it does not provide informations to discriminate an underlying severe liver disease.9 Clinical utility of APRI and FIB-4 scores is based upon use of biological and
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
6
F. Ratti et al. / EJSO xx (2014) 1e7
biochemical blood parameters that are simple, cost effective and easily available preoperatively.12 Furthermore, serum levels of transaminases and platelets count are independently associated with complicated postoperative courses after hepatic resection. Some authors have already analyzed the potential value of preoperative liver biopsy for evaluation of liver impairment resulting from preoperative CT administration: this kind of screening could potentially be performed in a selected cluster of patients, candidates to biopsy according to cutoff value of APRI.24,25 Unfortunately, histological specimen of 15 mm obtained by biopsy is frequently not sufficient to assess CALI since it has widespread but heterogeneous distribution within parenchyma.25 Other possibilities to minimize postoperative morbidity in high risk patients is represented by use of APRI cutoff to screen out potential candidates for surgery. In particular, normalization of biological markers would constitute a warning for correct surgical timing: indeed, hepatectomy is currently performed at least thirty days after the last administration of CT; conversely, APRI evaluation and normalization could shorten this interval or postpone intervention whenever required with a patient-tailored strategy. Secondly, APRI value should be monitored during the period of CT administration so that systemic therapy could be discontinued when the patient has become resectable and/or the value of APRI has become borderline. As stated before, there is no evidence for a precise number of CT cycles conditioning liver damage, so an easy available method to evaluate parenchymal impairing resulting from drug administration might avoid further hepatic functional reserve reduction. This study constitutes a preliminary experience for a possible alternative use of chronic liver disease scores in liver surgery, and may provide basis for further experiences focusing on several aspects of the same topic. First of all, APRI score and cutoff should be validated on a larger scale in a prospectively designed study. Furthermore, score value could be proposed as an indicator of underlying parenchymal impairment to select patients for preoperative portal vein embolization, in association with study of future remnant liver volume: through specific correlation analysis between liver volumes and value of APRI this topic could be deepen. Unfortunately this evaluation could not be performed in present series due to small number of patients undergoing PVE in the setting of two stage hepatectomy program: in these patients APRI score should be evaluated before the first surgical step and then monitored during interval between portal vein occlusion and major hepatectomy to assess liver functional recovery along with time progression from last CT (even though APRI score could be partially influenced by first surgical procedure or PVE, leading to abnormalities in platelet count and transaminases unrelated to CALI). Lastly, APRI variations could be monitored in the period following interruption of chemotherapy in order to evaluate
APRI accuracy as a marker of restored hepatic function along with time progression. In conclusion, APRI score takes into account both liver damage and platelet count: it is reliable in preoperative evaluation of lesions induced by oxaliplatin based CT regimens, which accounts for the high incidence of postoperative liver failure in patients undergoing liver surgery for CLM. Conflict of interest The authors declare that the article is an original work, has not been published before, and is not being considered for publication elsewhere in its final form, in either printed or electronic media. Any republication of the data will not constitute redundant publication, will not breach copyright, and will reference the original publication. The authors declare that they have no financial or any other competing interest in publication. The authors declare that they have no financial or any other competing interest in publication.
References 1. Kopetz S, Chang GJ, Overman MJ, et al. Improved survival in metastatic colorectal cancer is associated with adoption of hepatic resection and improved chemotherapy. J Clin Oncol 2009 Aug 1;27(22):3677– 83. 2. Nordlinger B, Sorbye H, Glimelius B, et al. Perioperative chemotherapy with FOLFOX4 and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC Intergroup trial 40983): a randomized controlled trial. Lancet 2008 Mar 22; 371(9617):1007–16. 3. Blazer 3rd DG, Kishi Y, Maru DM, et al. Pathologic response to preoperative chemotherapy: a new outcome end point after resection of hepatic colorectal metastases. J Clin Oncol 2008 Nov 20;26(33): 5344–51. 4. Nakano H, Oussoultzoglou E, Rosso E, et al. Sinusoidal injury increases morbidity after major hepatectomy in patients with colorectal liver metastases receiving preoperative chemotherapy. Ann Surg 2008 Jan;247(1):118–24. 5. Karoui M, Penna C, Amin-Hashem M, et al. Influence of preoperative chemotherapy on the risk of major hepatectomy for colorectal liver metastases. Ann Surg 2006 Jan;243(1):1–7. 6. Vauthey JN, Pawlik TM, Ribero D, et al. Chemotherapy regimen predicts steatohepatitis and an increase in 90-day mortality after surgery for hepatic colorectal metastases. J Clin Oncol 2006 May 1;24(13): 2065–72. 7. Pulitano C, Aldrighetti L, Arru M, et al. Influence of preoperative chemotherapy on the risk of major hepatectomy for colorectal liver metastases. Ann Surg 2006 Nov;244(5):833–5. 8. Rubbia-Brandt L, Audard V, Sartoretti P, et al. severe hepatic sinusoidal obstruction associated with oxaliplatin-bades chemotherapy in patients with metastatic colorectal cancer. Ann Oncol 2004 Mar;15(3): 460–6. 9. Hammond JS, Guha IN, Backingham IJ, Lobo DN. Prediction, prevention and management of postresection liver failure. Br J Surg 2011 Sep;98(9):1188–200. 10. Aloia T, Sebagh M, Plasse M, et al. Liver histology and surgical outcomes after preoperative chomotherapy with fluorouracil plus
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
F. Ratti et al. / EJSO xx (2014) 1e7
11.
12.
13.
14.
15.
16.
17.
oxaliplatin in colorectal cancer liver metastases. J Clin Oncol 2006 Nov 1;24(31):4983–90. Wai CT, Greenson JK, Fontana RJ, et al. A simple and noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology 2003 Aug;38(2):518–26. Soubrane O, Brouquet A, Zalinski S, et al. Predicting high grade lesions of sinusoidal obstruction syndrome related to oxaliplatin-based chemotherapy for colorectal liver metastases: correlation with posthepatectomy outcome. Ann Surg 2010 Mar;251(3):454–60. Farges O, Malassagne B, Flejou JF, Balzan S, Sauvanet A, Belghiti J. Risk of major liver resection in patients with underlying chronic liver disease: a reappraisal. Ann Surg 1999 Feb;229(2):210–5. Taketomi A, Kitagawa D, Itoh S, et al. Trends in morbidity and mortality after hepatic resection for hepatocellular carcinoma: an institute’s experience with 625 patients. J Am Coll Surg 2007 Apr; 204(4):580–7. Rahbari NN, Garden OJ, Padbury R, et al. Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS). Surgery 2011;149:713–24. Aldrighetti L, Pulitano C, Arru M, Catena M, Finazzi R, Ferla G. “Technological” approach versus clamp crushing technique for hepatic parenchymal transection: a comparative study. J Gastrointest Surg 2006 Jul-Aug;10(7):974–9. Terminology Committee of the International Hepato-Pancreato-Biliary Association. The IHPBA Brisbane 2000 terminology of liver anatomy and resections. HPB Surg 2000;2:333–9.
7
18. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004 Aug;240(2):205–13. 19. Nordlinger B, Van Cutsem E, Gruenberg T, et al. Combination of surgery and chemotherapy and the role of targeted agents in the treatment of patients with colorectal liver metastases: recommendations from an expert panel. Ann Oncol 2009 Jun;20(6):985–92. 20. Van Vledder MG, de Jong MC, Pawlik TM, Schulik RD, Diaz LA, Choti MA. Disappearing colorectal liver metastases after chemotherapy: should web e concerned? J Gastrointest Surg 2010 Nov; 14(11):1691–700. 21. Auer RC, White RR, Kemeny NE, et al. Predictors of a true complete response among disappearing liver metastases from colorectal cancer after chemotherapy. Cancer 2010 Mar 15;116(6). 22. Ribero D, Wang H, Donadon M, et al. Bevacizumab improves pathologic response and protects against hepatic injury in patients treated with oxaliplatin-based chemotherapy for colorectal liver metastases. Cancer 2007 Dec 15;110(12):2761–7. 23. Lesurtel M, Graf R, Aleil B, et al. Platelet-derived serotonin mediates liver regeneration. Science 2006 Apr 7;312(5770):104–7. 24. McHutchison JG, Blatt LM, de Medina M, et al. Measurement of serum hyaluronic acid in patients with chronic hepatitis C and its relationship to liver histology. Consensus Interferon Study Group. J Gastroenterol Hepatol 2000 Aug;15(8):945–51. 25. Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology 2003 Dec;38(6):1449–57.
Please cite this article in press as: Ratti F, et al., Liver failure in patients treated with chemotherapy for colorectal liver metastases: Role of chronic disease scores in patients undergoing major liver surgery. A case-matched analysis, Eur J Surg Oncol (2014), http://dx.doi.org/10.1016/j.ejso.2014.06.011
