bs_bs_banner
Journal of Digestive Diseases 2014; 15; 553–560
doi: 10.1111/1751-2980.12177
Original article
Histological features and severity of oxaliplatin-induced liver injury and clinical associations ILKe NALBANTOGLU,* Benjamin R TAN Jr,† David C LINEHAN,‡ Feng GAO§ & Elizabeth M BRUNT* *Departments of Pathology and Immunology, †Internal Medicine, Division of Medical Oncology, ‡General Surgery, Section of Hepato-Pancreato-Biliary Surgery, and §Division of Biostatistics, Siteman Cancer Center, Washington University School of Medicine, Saint Louis, Missouri, USA
OBJECTIVE: Oxaliplatin, a component of chemotherapy for colorectal carcinoma liver metastases, can result in hepatic sinusoidal injury; rarely, the injury is fatal. The manifestations of injury are variable. There are no known predictors of susceptibility and outcome. A semi-quantitative system for assessing histological features in non-tumor liver was designed to compare with clinical short-term and long-term outcomes. METHODS: A review of 47 patients with metastatic colorectal carcinoma who received liver resection utilizing a system for an aggregate liver injury score (0–4) included hepatocellular and sinusoidal features. Immunohistochemistry (IHC) for aberrant capillarization was included. The proliferation of hepatocytes and sinusoidal lining cells was evaluated with Ki-67 stain. RESULTS: In total, 32 (68.1%) cases showed light microscopic lesions of oxaliplatin-induced liver injury, KEY WORDS:
in which 26 were moderate to severe. Elevated preoperative aspartate aminotransferase (AST) and alkaline phosphatase levels were noted with higher injury scores (P = 0.01). Patients with higher injury scores had no significant increase in short-term postoperative complications, with one notable exception, who died of liver failure 10 months postoperatively. Increased CD34 expression was associated with higher injury scores (P = 0.00004), and abnormal AST levels (P = 0.04). Preoperative use of bevacizumab was not associated with lower injury scores. Steatosis was correlated with body mass index (P = 0.052) but not with exposure to oxaliplatin, bevacizumab or irinotecan. CONCLUSIONS: The proposed liver injury scoring system encompasses the spectrum of sinusoidal and hepatocellular lesions in oxaliplatin-induced liver injury and is correlated with serum liver enzyme levels in this group. Most patients recovered without complications during the 93-month follow-up, indicating that these lesions are reversible.
liver sinusoid, oxaliplatin, scoring.
Correspondence to: ILKe NALBANTOGLU, Department of Pathology and Immunology, Washington University School of Medicine, Campus Box 8118, 660 S. Euclid Avenue, St Louis, MO 63110, USA. Email: [email protected] Conflicts of interest: None. Partial results of this study were presented at United States & Canadian Academy of Pathology as an abstract in 2010. © 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
INTRODUCTION Neoadjuvant chemotherapy for colorectal carcinoma is widely used as it may shrink the sizes of the tumors and thus enable the surgical removal of previously unresectable tumors.1–5 The most current neoadjuvant chemotherapy for colorectal cancer is oxaliplatinbased regimen. However, oxaliplatin may result in sinusoidal injury in the liver and it may also be associated with adverse outcomes following surgery.6,7 The
553
554
I Nalbantoglu et al.
current literature suggests that the use of additional agents such as bevacizumab, a humanized monoclonal antibody that inhibits vascular endothelial growth factor A (VEGF-A) and angiogenesis, not only ameliorates the pathological response but also protects potentially against liver injury.8 Aspirin, possibly as an anti-platelet aggregate, may also be protective against oxaliplatin-related liver injury.9 Rubbia-Brandt et al.10 first described the histological lesions of oxaliplatininduced hepatic sinusoidal injury as being primarily centered in acinar zone 3. Additional studies have shown the lesions are not limited to sinusoidal lining cells (SLCs) but also involve the adjacent parenchyma with focal disruption and necrosis of hepatocytes and hemorrhage into the cords. The lesions are not restricted to zone 3, but may be found to be accentuated adjacent to the liver capsule10–14 or around the large hilar structures.11 Additionally, perisinusoidal fibrosis may also occur. In the current study we aimed to focus on an aggregate of histological features of oxaliplatin-induced liver injury, including additional findings of sinusoidal capillarization, which results in the loss of fenestrae of the endothelial cells and the deposition of basement membrane in the space of Disse, regenerative activity of hepatocytes and SLCs in areas of injury. We also analyzed the associations of chemotherapy-related injury with various preoperative and postoperative clinical outcomes.
Journal of Digestive Diseases 2014; 15; 553–560 biochemical parameters including alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), total bilirubin, prothrombin time, partial thromboplastin time (PTT) and calculated international normalized ratio (INR), intra-operative surgical observations of the liver, length of postoperative hospital stay, postoperative complications including infections, and overall survival. The clinical follow-up time ranged from 0.5 to 93 months. Histochemistry and immunohistochemistry (IHC) Non-tumor liver tissues were taken at least 2 cm from the tumor to exclude the mass effect and were free of extensive cautery artifacts. The tissues were cut into 5-μm sections, embedded in paraffin, and stained with hematoxylin and eosin (HE) and Gomori trichrome. Anti-Ki-67 and anti-CD34 IHC stain was also performed by using prediluted rabbit monoclonal Ki-67 (clone 30-9, prediluted; Ventana Medical Systems, Tucson, AZ, USA) and mouse monoclonal anti-CD34 (clone QBend/10, prediluted; Ventana) on the Ventana Benchmark autostainer (Ventana). Ventana CC1, EDTA-Tris, pH 8.0 solution was utilized for antigen retrieval. 3,3′-diaminobenzidine tetrahydrochloride chromogen was used for the detection. Tonsil tissues were used as positive controls for both Ki-67 and CD34. Histological examination and liver injury scoring
PATIENTS AND METHODS Patients This study was approved by the Washington University Institutional Review Board (IRB). A waiver for obtaining patient’s informed consent was given and permission to collect relevant clinical and pathological data was granted by the IRB. A retrospective search for hepatic resection due to metastatic colorectal carcinoma from 2004 to 2011 was performed based on the database of the Department of Pathology and Immunology, Washington University School of Medicine. Exclusion criteria were patients with known chronic liver disease and a lack of detailed clinical history, laboratory and postoperative clinical data. The following clinical parameters were collected: age, gender, ethnicity/race, body mass index (BMI), treatment dose and chemotherapy regimen, cumulative dose of oxaliplatin, use of anticoagulation (aspirin and warfarin), use of additional relevant medications such as bevacizumab, preoperative and postoperative liver
A global scoring system incorporating previously described histological features of oxaliplatin-induced sinusoidal injury10,14 and additional features, which will be described below, had been proposed and presented in abstract form.11 The specimens were examined and scored by two pathologists independently who were blinded to the clinical information of the patients. Any disagreement was resolved by a second assessment of the specimens until a consensus was reached. The spectrum of liver injury and the details of the scoring method are summarized in Table 1 and shown in Fig. 1a–e. Additional findings that were included in the system but not previously described were the presence of hepatocellular anisonucleosis and basophilic discoloration of the hepatocytes. The latter is observable at low-power field (LPF) under microscope with trichrome stain. Finally, nonzonality of hepatic sinusoidal injury rather than preferential zone 3 localization and the predilection of lesions to be found around the large collagenized portal and vascular structures were other features often noted.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560 Table 1.
1 2 3
4
555
Histological spectrum and grading of oxaliplatin-induced liver injury
Injury score 0
Oxaliplatin-induced liver injury
HE stain
Trichrome stain
No histological abnormality with or without hepatocyte anisonucleosis Anisonucleosis Extensive anisonucleosis Sinusoidal prominence; ± red blood cells or debris in sinusoids; ± sinusoidal ectasias: parenchymal, subcapsular or near large vascular structures; ± cord atrophy Definite extensive sinusoidal injury with or without sinusoidal disruption, cord atrophy, multifocal hemorrhage
Negative Delicate PSF Delicate PSF and positive clusters of basophilic hepatocytes As in HE, and PSF outlines affected sinusoids
As in HE, multifocal dense PSF and intra-sinusoidal debris
HE, hematoxylin and eosin; PSF, perisinusoidal fibrosis. 11 reproduced with permission
Features that were not associated with oxaliplatininduced liver injury, steatosis and steatohepatitis were also scored per the Kleiner system.15
a
b
c
d
e
f
Figure 1. Histopathological characteristics of oxaliplatininduced liver injury. (a) Extensive anisonucleosis of hepatocytes can be noted (HE stain, ×400). (b) Trichrome stain highlights basophilic discoloration of hepatocytes and non-zonal distribution of injury on low-power examination (×100). (c) Trichrome stain demonstrates perisinusoidal fibrosis and basophilic discoloration of hepatocytes near a large portal tract (×200). (d) Trichrome stain highlights sinusoidal ectasia accompanied by perisinusoidal fibrosis and basophilic discoloration of hepatocytes, an example of injury score 3 (×400). (e) Disrupted sinusoids with peliosis and cord atrophy are noted in this photomicrograph. This is an example of injury score 4 (HE stain, ×200). (f) Aberrant capillarization of the sinusoids is highlighted by the CD34 immunohistochemistry (×200).
Lesions observed in oxaliplatin-exposed liver also included the presence of aberrant capillarization in the foci of the injury. These were suspected by the presence of perisinusoidal fibrosis and confirmed by CD34 stain. CD34 immunoexpression was scored based on the extent of the reactivity within the acini: 0, immediate sinusoidal expression noted in zone 1 (normal expression); 1, more extensive zone 1 or 3 expression, or both; and 2, for extensive multifocal expression within the acini (Fig. 1f). The cases were scored based on the areas with the most intense expression. Additionally, the proliferation of both hepatocytes and SLCs, as indicated by nuclear Ki-67, was noted. For each case, rough visual estimates were made as to the equivalence of hepatocytes and SLCs (H = SLCs) or the predominance of hepatocytes (H > SLCs) or the predominance of SLCs (H < SLCs) (not shown). The final global scoring system was a combination of HE and trichrome findings of the abovementioned features: anisonucleosis, basophilia and sinusoidal alterations, with the goal of application for the ease of use and reproducibility. Statistical analysis Statistical analysis was performed using SAS 9.2 (SAS Institute, Cary, NC, USA). The global injury scores were categorized based on a three-tier system: 0, no injury; 1–2, mild injury; and 3–4, moderate to severe injury. And then a two-tier system identifying the presence of liver injury (scores 1–4) versus no injury (score 0), for
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
556
I Nalbantoglu et al.
Journal of Digestive Diseases 2014; 15; 553–560
Histological findings of oxaliplatin-induced liver injury
lesions of oxaliplatin-induced liver injury that was confirmed by HE stain. Six of the cases showed mild injury (anisonucleosis with or without perisinusoidal fibrosis) and 26 had moderate to severe injury characterized by varying degrees of sinusoidal dilatation and disruption, hemorrhage and peliosis. The lesions were not necessarily centered around the terminal hepatic venule in zone 3 but were scattered throughout the acini. These findings could also be readily identified along the subcapsular parenchyma as well as at the periportal and perivenular spaces of large collagenized structures. The trichrome stain was useful for identifying inconspicuous areas of injury by HE, as hepatocytes in foci of injury appeared basophilic. These foci usually contained perisinusoidal fibrosis (Fig. 1b–e). The extent of liver injury and IHC findings, along with patients’ characteristics, are summarized in Table 2.
Among the 47 cases who had hepatic resection for metastatic colorectal carcinoma, 32 (68.1%) showed
Aberrant capillarization of the sinusoids by CD34 stain (available in all but one of the cases) was found
statistical analyses. For patients’ individual and clinical characteristics, their associations with the status of oxaliplatin-induced liver injury were summarized using contingency tables and assessed by Fisher’s exact test (injury vs no injury) and Jonckheere’s trend test (for three-tier scoring system to preserve the priori ordering). For continuous variables such as BMI and AST were expressed with mean ± standard deviation and were compared by ANOVA or Kruskal-Wallis, a nonparametric rank sum test, when appropriate. All the tests were two-sided and a P value of 0.05 or less was considered as statistical significance. RESULTS
Table 2.
Clinical and pathological characteristics of patients
Characteristics
Gender, n (%) Male Female Ethnicity, n (%) AA/His Caucasian Additional therapy, n (%) Bevacizumab use† No bevacizumab† Anti-coagulation‡ Postoperative complication, n (%) Infection‡ Others†† Preoperative liver function test (U/L, mean) AST ALT ALP CD34 expression, n (%)‡ 0 1 2 Ki-67 stain, n (%) H = SLCs H > SLCs H < SLCs
Total patients
Injury score
P value
0 (no injury)
1–2 (mild)
3–4 (moderate to severe)
28 (59.6) 19 (40.4)
9 6
4 2
15 11
8 (17.0) 39 (83.0)
2 13
0 6
6 20
15 (33.3) 30 (66.7) 14 (30.4) 14 (30) 6 (12.8) 8 (17.0) Range 11–129 11–194 39–533
2 12 4 5 3 2
2 4 4 2 1 1
11 14 6 7 2 5
26.2 ± 10.52 28.6 ± 15.01 84.7 ± 30.39
34.67 ± 17.82 52.3 ± 60.85 99.5 ± 9.79
41.2 ± 22.84 50.0 ± 38.24 127.8 ± 95.90
1 (2.2) 26 (56.5) 19 (41.3)
1 12 1
0 6 0
0 8 18
12 (25.5) 23 (49.0) 12 (25.5)
5 5 5
0 3 3
7 15 4
0.83
0.16
0.06 0.40 0.37 0.21
0.03 0.09 0.03 0.00004
0.19
†Data available in 45 patients. ‡Data available in 46 patients. ††Others include persistent bilirubin and liver enzyme elevations, wound leak, ascites, hematoma, cerebrovascular and cardiac events. Normal range: aspartate aminotransferase (AST) 0–40 U/L, alanine aminotransferase (ALT) 0–41 U/L, alkaline phosphatase (ALP) 25–129 U/L. AA/His, African American or Hispanic; H, hepatocytes; SLCs, sinusoidal lining cells.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560 in 45 cases. Of these, 26 had focal capillarization limited to a small area around the portal tracts or central veins, whereas the other 19 showed extensive capillarization throughout the lobule (Fig. 1f). Aberrant capillarization was associated with the severity of liver injury (P = 0.00004, Table 2). Ki-67 expression was higher in the hepatocellular component in 49.0% of the cases, whereas 25.5% showed equal expression in hepatocytes and SLCs. No correlation between the severity of liver injury and its expression patterns was noted (P = 0.19, Table 2). Steatosis In all, 23 (48.9%) of the liver specimens showed steatosis, among which 16 were grade 1 (5–33%), 5 were grade 2 (34–66%) and 2 were grade 3 (>66%). Two patients were diagnosed as steatohepatitis. The presence of steatosis was correlated with BMI (P = 0.052, not shown) but not with oxaliplatin, bevacizumab or irinotecan treatment, or the preoperative AST and ALT levels. Clinical features and patient characteristics Altogether 47 patients were included in the study; of these patients, 19 (40.4%) were women. And 39 (83.0%) were Caucasians, 1 (2.1%) was Hispanic and the other 7 (14.9%) were African Americans. Their BMI ranged from 19 to 41 kg/m2 (mean 28.5 kg/m2). The mean number of oxaliplatin cycles was 3.8. The minimum interval from the last dose of oxaliplatin to surgery was 19 days (median 63 days). In all, 15 patients additionally received bevacizumab and 14 patients were given anti-coagulation treatment, including aspirin (n = 10) and warfarin (n = 4) preoperatively. Eight patients had prior liver resections and two had preoperative portal vein embolization. In all, 14 (29.7%) patients had postoperative complications, including four vascular events (one myocardial infarction, two deep vein thromboses, and one cerebrovascular accident); six had an infection (one with persistently markedly elevated bilirubin and liver enzyme levels, one with wound leak, one with small bowel obstruction, one with pleural effusion, one with hematoma formation and one had both a vascular event and infection). Postoperative survival time was 0.5–93 months (median 17 months). One patient died of liver failure at 10 months post-surgery. The patient who died of liver failure had been diagnosed with colorectal cancer and multiple liver
Oxaliplatin-induced liver injury
557
metastases and received four cycles of oxaliplatin preoperatively (cumulative dose 340 mg/m2). He then received right hepatectomy, but was not treated with bevacizumab or anti-coagulation treatment or portal vein embolization. The interval between the last dose of chemotherapy and surgery was 35 days. His preoperative liver biochemical results were normal except the AST level was 44 U/L. No gross abnormalities of the liver had been noted during surgery. Histological analysis of the resected liver tissues showed severe liver injury with marked anisonucleosis and basophilia of hepatocytes, multifocal disruption and enlargement of sinusoids, and multiple foci of intraluminal debris within sinusoids (injury score 4). CD34 stain showed extensive multifocal aberrant capillarization of the sinusoids (injury score 2). Ki-67 nuclear reactivity was noted to be greater in the sinusoidal compartment than in the hepatocytes. The levels of peripheral bilirubin (0.672 μmol/L), AST (78 U/L), ALT (79 U/L) and ALP (473 U/L) were all elevated after operation, but in the absence of other immediate postoperative complications. Imaging studies were negative for biliary obstruction, and the patient was discharged on postoperative day 17. He was then re-admitted to the hospital many times over the next 10 months for abnormal liver function test, failure to thrive, jaundice and melena. At 5 months post operation, a follow-up liver biopsy showed severe acute cholestasis, hepatocellular injury, terminal hepatic venule obliteration and dense zone 3 perisinusoidal fibrosis. Despite supportive treatment, his liver function continued to deteriorate (bilirubin 1.27 μmol/L, ALP 209 U/L, AST 4096 U/L, ALT 3737 U/L) and the patient died due to liver failure at 10 months post-resection. In all remaining patients, short-term postoperative complications, peak INR values, bilirubin levels and survival time (up to 93 months [median 17 months]) were not associated with the extent of liver injury. Table 2 summarizes the characteristics of patients (age and BMI not shown) and the extent of liver injury. There were no statistical differences in gender and ethnicity among patients with different degrees of liver injury. Oxaliplatin dose and the interval between the last oxaliplatin dose and surgery were not correlated with presence of liver injury. Additionally, there was no association with the presence of liver injury in the subgroup of patients who received immediate preoperative oxaliplatin (33/47, 70.2%). The administration of bevacizumab in 15 patients and anticoagulation therapy in 14 patients were not associated with a protection from liver injury; of these
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
558
I Nalbantoglu et al.
Journal of Digestive Diseases 2014; 15; 553–560
patients, 13 who received bevacizumab and 10 who received some forms of anticoagulation showed evidence of some degree of sinusoidal injury (Table 2). Interestingly, when serially drawn, rising preoperative AST (P = 0.01), ALT (P = 0.03) and ALP levels (P = 0.01), even if the values remained within normal ranges, were associated with the presence of liver injury (injury score: 0 vs 1–4, Table 3). Higher injury scores were also associated with rising AST and ALP levels (P = 0.03, Table 2), indicating that even mild increases in liver enzyme levels may raise a concern for the presence of liver injury. Abnormal preoperative AST levels noted in 14 of the patients were associated with higher injury scores (P = 0.01, data not shown). The gross appearance of the liver during surgery was noted in operative notes as ‘mottled, blue’ in two cases, both of whom had the highest injury scores (of 4). One of the patients had abnormal preoperative AST of 42 U/L and ALT of 53 U/L, whereas the other had abnormal preoperative ALP of 144 U/L. The length of postoperative hospital stay were 7 and 8 days, respectively, for them. No postoperative complications were noted in either. In four cases with liver resection, parenchymal nodularity was noted microscopically but not grossly. Among them, 3 had abnormal preoperative AST levels, 3 had had preoperative portal vein embolization and one had had a history of hepatic surgery. One was diagnosed with nodular regenerative hyperplasia (NRH) on a pre-resection biopsy, which was done when oxaliplatin injury was suspected. This patient had no liver function abnormalities. A limited resection was performed, showing similar NRH findings in the biopsy; no adverse events occurred postoperatively. In 49.0% of the cases proliferation, as detected by Ki-67 stain, was greater in the hepatocytes than in the sinusoids. There seemed to be a trend between greater Table 3.
Abnormal AST levels were associated with aberrant capillarization, as indicated by higher CD34 scores (P = 0.04). Notably, there was no association of aberrant sinusoidal CD34 expression with short-term or long-term liver function abnormalities (preoperative and postoperative AST, ALT and ALP levels), INR, peak bilirubin and length of hospital stay or preoperative bevacizumab and anti-coagulation treatment. DISCUSSION Our study has confirmed and extended prior histology-based observations of oxaliplatin-induced chemotherapy injury in the liver.1,6,12,14,16,17 Oxaliplatin has been shown to damage the endothelial cells within the hepatic sinusoids.5,18,19 We have additionally documented an associated hepatocellular response, as indicated by basophilia, anisonucleosis and increased proliferation. Since the original description by Rubbia-Brandt et al., additional features of oxaliplatin-induced liver injury, including hepatocyte atrophy and perisinusoidal fibrosis,7 centrilobular necrosis and NRH have been described14,16,20–22 and the need for a semi-quantitative method for scoring liver injury is discussed.14,22 Rubbia-Brandt et al.14 described its histological features as parenchymal, stromal and vascular, whereas Ryan et al.22 combined grading schemes for fatty liver disease and sinusoidal dilatation along with fibrosis and assigned them a combined vascular injury score. Hepatic plate disruption associated with the necrosis of hepatocytes and perisinusoidal fibrosis, hemorrhage into cords and sinusoidal endothelial damage and necrosis are well recognized in oxaliplatin-induced liver injury. The lesions may be nonzonal and subcapsular,12 both of which were confirmed in this study. The predilection of lesions to be found around the large collagenized portal and vascular structures is a new feature. We propose that the sinusoids near these anchored structures are tethered and less
Preoperative liver function test and presence of liver injury
Preoperative liver function test (U/L, mean ± SD) AST ALT ALP
hepatocellular Ki-67 expression and a rising postoperative ALT levels (P = 0.059).
No injury (n = 15)
Injury present (n = 32)
P value
26.3 ± 10.52 28.6 ± 15.01 84.67 ± 30.39
39.9 ± 21.88 50.47 ± 42.16 112.5 ± 86.94
0.01 0.03 0.01
Aspartate aminotransferase (AST) 0–40 U/L, normal range: alanine aminotransferase (ALT) 0–41 U/L, alkaline phosphatase (ALP) 25–129 U/L. SD, standard deviation.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560 flexible, and therefore, more readily disrupted once the lining is damaged in these areas. Another possible explanation could be zonal differences in drug distribution and utilization within the hepatic sinusoids. Hepatocellular anisonucleosis can be observed in several other pathological conditions but in our case series it was easily recognized microscopically under LPF, and may be an initial clue to oxaliplatin-induced liver injury, which can then alert the pathologist to look for other described lesions (Fig. 1a). Even though anisonucleosis may be a sign of senescence, it is also a feature of hepatocellular regeneration. Ectatic sinusoids, some with intraluminal debris, can be noted scattered in a nonzonal fashion in the acini. These lesions may be seen on HE stained specimens when they are large and filled with red cells, but smaller foci are often best appreciated by the concomitant altered basophilic hepatocellular discoloration on trichrome (Fig. 1b–e). Many of the cases also showed features of aberrant capillarization, as indicated not only by collagen deposition but also by re-endothelialization of nonfenestrated endothelium, as detected by sinusoidal reactivity with anti-CD34. Aberrant capillarization related to oxaliplatin use was reported by Narita et al.2 The investigators also found that the number of cycles and abnormal liver function were also associated with CD34 expression.2 However, the dose of oxaliplatin did not correlate with the severity of aberrant capillarization in our data set. A higher percentage of proliferative hepatocytes detected by Ki-67 is indicative of hepatocellular regeneration in the face of sinusoidal damage. Hepatocellular regeneration may, in part, explain the lack of association with histological findings and adverse clinical outcomes. Other investigators have suggested similar results.14 Previous and most current experimental animal18,23 studies showed that steatosis was not associated with oxaliplatin-induced liver injury. Histological analysis of 334 non-neoplastic human livers by Ryan et al.22 demonstrated that oxaliplatin exposure was related to the development of vascular injury but not steatohepatitis. Our findings are in keeping with these studies. The only statistically significant correlation in our study was between BMI and steatosis. In the era of an obesity epidemic, we are more likely to find background steatosis and steatohepatitis. This, however, should not be misinterpreted as oxaliplatin-induced liver injury and needs to be evaluated separately.
Oxaliplatin-induced liver injury
559
The pathogenesis of oxaliplatin-induced liver injury is not completely understood. Yeong et al. described a rodent model of pyrrolizidine alkaloids-induced liver injury and noted that a loss of endothelium was associated with extensive red blood cell extravasation and the presence of collagen fibers in the space of Disse in severe sinusoidal injury.18 Recently, a mouse model was developed for oxaliplatin-induced liver injury.23 In this model, the mouse liver specimens showed the development of liver sinusoidal dilatation, the disruption of endothelial cells within sinusoids, and hepatocellular injury, evidenced by elevated ALT and AST levels when exposed to oxaliplatin.23 In addition, early collagen deposition was noted.23 Similar observations have been made14,22 in human liver with oxaliplatin-induced liver injury.14,22 Recent studies have shown that hepatic sinusoidal injury results in up-and-down regulation of genes in several inflammatory, vascular and fibrogenic pathways.23,24 In addition, lower levels of ERCC2, a nucleotide excision repair gene, was found to be associated with sinusoidal injury with oxaliplatin-based chemotherapy.25 To date, however, in spite of wellcharacterized injury patterns, there are no laboratory tests, serum or genetic markers to predict susceptibility to oxaliplatin-induced sinusoidal injury. The limitations of this retrospective study are acknowledged. As with any retrospective study, many variables were not controlled. The sample size was relatively small, while the selection and exclusion of the histological samples were subjected to strict criteria, sampling variability cannot be completely eliminated since oxaliplatin-induced liver injury is a heterogeneous process. However, the strengths of the study included its detailed systematic approach to histopathological analyses and the thorough evaluation for relevant short-term and long-term clinical associations. In conclusion, this study described a method of semiquantitatively evaluating histological features of oxaliplatin-induced liver injury, and combined the histological spectrum of oxaliplatin-induced liver injury into a system that is simple and practical to apply which can be performed with routine histochemical stains. The injury score that we proposed was developed on resected specimens once the injury occurred. The injury scoring system along with CD34 stain deserves to be tested prospectively in biopsy samples in efforts to guide clinical teams for future treatment or surgery.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
560
I Nalbantoglu et al.
ACKNOWLEDGMENT The authors acknowledge intradepartmental research funding from the Department of Pathology and Immunology and the support of Biostatistics Core, Siteman Comprehensive Cancer Center and the National Cancer Institute Cancer Center support grant P30 CA091842. REFERENCES 1 Adam R, Aloia T, Lévi F et al. Hepatic resection after rescue cetuximab treatment for colorectal liver metastases previously refractory to conventional systemic therapy. J Clin Oncol 2007; 25: 4593–602. 2 Narita M, Oussoultzoglou E, Chenard MP et al. Liver injury due to chemotherapy-induced sinusoidal obstruction syndrome is associated with sinusoidal capillarization. Ann Surg Oncol 2012; 19: 2230–7. 3 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 randomised controlled trial. Lancet 2008; 371: 1007–16. 4 Wolf PS, Park JO, Bao F et al. Preoperative chemotherapy and the risk of hepatotoxicity and morbidity after liver resection for metastatic colorectal cancer: a single institution experience. J Am Coll Surg 2013; 216: 41–9. 5 Cleary JM, Tanabe KT, Lauwers GY, Zhu AX. Hepatic toxicities associated with the use of preoperative systemic therapy in patients with metastatic colorectal adenocarcinoma to the liver. Oncologist 2009; 14: 1095–105. 6 Aloia T, Sebagh M, Plasse M et al. Liver histology and surgical outcomes after preoperative chemotherapy with fluorouracil plus oxaliplatin in colorectal cancer liver metastases. J Clin Oncol 2006; 24: 4983–90. 7 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; 247: 118–24. 8 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; 110: 2761–7. 9 Brouquet A, Benoist S, Julie C et al. Risk factors for chemotherapy-associated liver injuries: A multivariate analysis of a group of 146 patients with colorectal metastases. Surgery 2009; 145: 362–71. 10 Rubbia-Brandt L, Audard V, Sartoretti P et al. Severe hepatic sinusoidal obstruction associated with oxaliplatin-based chemotherapy in patients with metastatic colorectal cancer. Ann Oncol 2004; 15: 460–6. 11 Klonowski PW, Tan B, Linehan DC, Porembka MR, Brunt EM. Capillarization of hepatic sinusoids after neoadjuvant therapy. Lab Invest 2010; 90: 360A.
Journal of Digestive Diseases 2014; 15; 553–560 12 Rubbia-Brandt L. Sinusoidal obstruction syndrome. Clin Liver Dis 2010; 14: 651–68. 13 Rubbia-Brandt L. Hepatic lesions induced by systemic chemotherapy for digestive cancer. Ann Pathol 2010; 30: 421–5 (in French). 14 Rubbia-Brandt L, Lauwers GY, Wang H et al. Sinusoidal obstruction syndrome and nodular regenerative hyperplasia are frequent oxaliplatin-associated liver lesions and partially prevented by bevacizumab in patients with hepatic colorectal metastasis. Histopathology 2010; 56: 430–9. 15 Kleiner DE, Brunt EM, Van Natta M et al.; Nonalcoholic Steatohepatitis Clinical Research Network. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005; 41: 1313–21. 16 Arotçarena R, Calès V, Berthelémy P et al. Severe sinusoidal lesions: a serious and overlooked complication of oxaliplatin-containing chemotherapy? Gastroenterol Clin Biol 2006; 30: 1313–6. 17 Kandutsch S, Klinger M, Hacker S, Wrba F, Gruenberger B, Gruenberger T. Patterns of hepatotoxicity after chemotherapy for colorectal cancer liver metastases. Eur J Surg Oncol 2008; 34: 1231–6. 18 Yeong ML, Wakefield SJ, Ford HC. Hepatocyte membrane injury and bleb formation following low dose comfrey toxicity in rats. Int J Exp Pathol 1993; 74: 211–7. 19 Vreuls CP, Van Den Broek MA, Winstanley A et al. Hepatic sinusoidal obstruction syndrome (SOS) reduces the effect of oxaliplatin in colorectal liver metastases. Histopathology 2012; 61: 314–8. 20 Slade JH, Alattar ML, Fogelman DR et al. Portal hypertension associated with oxaliplatin administration: clinical manifestations of hepatic sinusoidal injury. Clin Colorectal Cancer 2009; 8: 225–30. 21 Tamandl D, Klinger M, Eipeldauer S et al. Sinusoidal obstruction syndrome impairs long-term outcome of colorectal liver metastases treated with resection after neoadjuvant chemotherapy. Ann Surg Oncol 2011; 18: 421–30. 22 Ryan P, Nanji S, Pollett A et al. Chemotherapy-induced liver injury in metastatic colorectal cancer: semiquantitative histologic analysis of 334 resected liver specimens shows that vascular injury but not steatohepatitis is associated with preoperative chemotherapy. Am J Surg Pathol 2010; 34: 784–91. 23 Robinson SM, Mann J, Vasilaki A et al. Pathogenesis of FOLFOX induced sinusoidal obstruction syndrome in a murine chemotherapy model. J Hepatol 2013; 59: 318–26. 24 Rubbia-Brandt L, Tauzin S, Brezault C et al. Gene expression profiling provides insights into pathways of oxaliplatin-related sinusoidal obstruction syndrome in humans. Mol Cancer Ther 2011; 10: 687–96. 25 Pilgrim CH, Brettingham-Moore K, Pham A et al. mRNA gene expression correlates with histologically diagnosed chemotherapy-induced hepatic injury. HPB (Oxford) 2011; 13: 811–6.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560
doi: 10.1111/1751-2980.12177
Original article
Histological features and severity of oxaliplatin-induced liver injury and clinical associations ILKe NALBANTOGLU,* Benjamin R TAN Jr,† David C LINEHAN,‡ Feng GAO§ & Elizabeth M BRUNT* *Departments of Pathology and Immunology, †Internal Medicine, Division of Medical Oncology, ‡General Surgery, Section of Hepato-Pancreato-Biliary Surgery, and §Division of Biostatistics, Siteman Cancer Center, Washington University School of Medicine, Saint Louis, Missouri, USA
OBJECTIVE: Oxaliplatin, a component of chemotherapy for colorectal carcinoma liver metastases, can result in hepatic sinusoidal injury; rarely, the injury is fatal. The manifestations of injury are variable. There are no known predictors of susceptibility and outcome. A semi-quantitative system for assessing histological features in non-tumor liver was designed to compare with clinical short-term and long-term outcomes. METHODS: A review of 47 patients with metastatic colorectal carcinoma who received liver resection utilizing a system for an aggregate liver injury score (0–4) included hepatocellular and sinusoidal features. Immunohistochemistry (IHC) for aberrant capillarization was included. The proliferation of hepatocytes and sinusoidal lining cells was evaluated with Ki-67 stain. RESULTS: In total, 32 (68.1%) cases showed light microscopic lesions of oxaliplatin-induced liver injury, KEY WORDS:
in which 26 were moderate to severe. Elevated preoperative aspartate aminotransferase (AST) and alkaline phosphatase levels were noted with higher injury scores (P = 0.01). Patients with higher injury scores had no significant increase in short-term postoperative complications, with one notable exception, who died of liver failure 10 months postoperatively. Increased CD34 expression was associated with higher injury scores (P = 0.00004), and abnormal AST levels (P = 0.04). Preoperative use of bevacizumab was not associated with lower injury scores. Steatosis was correlated with body mass index (P = 0.052) but not with exposure to oxaliplatin, bevacizumab or irinotecan. CONCLUSIONS: The proposed liver injury scoring system encompasses the spectrum of sinusoidal and hepatocellular lesions in oxaliplatin-induced liver injury and is correlated with serum liver enzyme levels in this group. Most patients recovered without complications during the 93-month follow-up, indicating that these lesions are reversible.
liver sinusoid, oxaliplatin, scoring.
Correspondence to: ILKe NALBANTOGLU, Department of Pathology and Immunology, Washington University School of Medicine, Campus Box 8118, 660 S. Euclid Avenue, St Louis, MO 63110, USA. Email: [email protected] Conflicts of interest: None. Partial results of this study were presented at United States & Canadian Academy of Pathology as an abstract in 2010. © 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
INTRODUCTION Neoadjuvant chemotherapy for colorectal carcinoma is widely used as it may shrink the sizes of the tumors and thus enable the surgical removal of previously unresectable tumors.1–5 The most current neoadjuvant chemotherapy for colorectal cancer is oxaliplatinbased regimen. However, oxaliplatin may result in sinusoidal injury in the liver and it may also be associated with adverse outcomes following surgery.6,7 The
553
554
I Nalbantoglu et al.
current literature suggests that the use of additional agents such as bevacizumab, a humanized monoclonal antibody that inhibits vascular endothelial growth factor A (VEGF-A) and angiogenesis, not only ameliorates the pathological response but also protects potentially against liver injury.8 Aspirin, possibly as an anti-platelet aggregate, may also be protective against oxaliplatin-related liver injury.9 Rubbia-Brandt et al.10 first described the histological lesions of oxaliplatininduced hepatic sinusoidal injury as being primarily centered in acinar zone 3. Additional studies have shown the lesions are not limited to sinusoidal lining cells (SLCs) but also involve the adjacent parenchyma with focal disruption and necrosis of hepatocytes and hemorrhage into the cords. The lesions are not restricted to zone 3, but may be found to be accentuated adjacent to the liver capsule10–14 or around the large hilar structures.11 Additionally, perisinusoidal fibrosis may also occur. In the current study we aimed to focus on an aggregate of histological features of oxaliplatin-induced liver injury, including additional findings of sinusoidal capillarization, which results in the loss of fenestrae of the endothelial cells and the deposition of basement membrane in the space of Disse, regenerative activity of hepatocytes and SLCs in areas of injury. We also analyzed the associations of chemotherapy-related injury with various preoperative and postoperative clinical outcomes.
Journal of Digestive Diseases 2014; 15; 553–560 biochemical parameters including alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), total bilirubin, prothrombin time, partial thromboplastin time (PTT) and calculated international normalized ratio (INR), intra-operative surgical observations of the liver, length of postoperative hospital stay, postoperative complications including infections, and overall survival. The clinical follow-up time ranged from 0.5 to 93 months. Histochemistry and immunohistochemistry (IHC) Non-tumor liver tissues were taken at least 2 cm from the tumor to exclude the mass effect and were free of extensive cautery artifacts. The tissues were cut into 5-μm sections, embedded in paraffin, and stained with hematoxylin and eosin (HE) and Gomori trichrome. Anti-Ki-67 and anti-CD34 IHC stain was also performed by using prediluted rabbit monoclonal Ki-67 (clone 30-9, prediluted; Ventana Medical Systems, Tucson, AZ, USA) and mouse monoclonal anti-CD34 (clone QBend/10, prediluted; Ventana) on the Ventana Benchmark autostainer (Ventana). Ventana CC1, EDTA-Tris, pH 8.0 solution was utilized for antigen retrieval. 3,3′-diaminobenzidine tetrahydrochloride chromogen was used for the detection. Tonsil tissues were used as positive controls for both Ki-67 and CD34. Histological examination and liver injury scoring
PATIENTS AND METHODS Patients This study was approved by the Washington University Institutional Review Board (IRB). A waiver for obtaining patient’s informed consent was given and permission to collect relevant clinical and pathological data was granted by the IRB. A retrospective search for hepatic resection due to metastatic colorectal carcinoma from 2004 to 2011 was performed based on the database of the Department of Pathology and Immunology, Washington University School of Medicine. Exclusion criteria were patients with known chronic liver disease and a lack of detailed clinical history, laboratory and postoperative clinical data. The following clinical parameters were collected: age, gender, ethnicity/race, body mass index (BMI), treatment dose and chemotherapy regimen, cumulative dose of oxaliplatin, use of anticoagulation (aspirin and warfarin), use of additional relevant medications such as bevacizumab, preoperative and postoperative liver
A global scoring system incorporating previously described histological features of oxaliplatin-induced sinusoidal injury10,14 and additional features, which will be described below, had been proposed and presented in abstract form.11 The specimens were examined and scored by two pathologists independently who were blinded to the clinical information of the patients. Any disagreement was resolved by a second assessment of the specimens until a consensus was reached. The spectrum of liver injury and the details of the scoring method are summarized in Table 1 and shown in Fig. 1a–e. Additional findings that were included in the system but not previously described were the presence of hepatocellular anisonucleosis and basophilic discoloration of the hepatocytes. The latter is observable at low-power field (LPF) under microscope with trichrome stain. Finally, nonzonality of hepatic sinusoidal injury rather than preferential zone 3 localization and the predilection of lesions to be found around the large collagenized portal and vascular structures were other features often noted.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560 Table 1.
1 2 3
4
555
Histological spectrum and grading of oxaliplatin-induced liver injury
Injury score 0
Oxaliplatin-induced liver injury
HE stain
Trichrome stain
No histological abnormality with or without hepatocyte anisonucleosis Anisonucleosis Extensive anisonucleosis Sinusoidal prominence; ± red blood cells or debris in sinusoids; ± sinusoidal ectasias: parenchymal, subcapsular or near large vascular structures; ± cord atrophy Definite extensive sinusoidal injury with or without sinusoidal disruption, cord atrophy, multifocal hemorrhage
Negative Delicate PSF Delicate PSF and positive clusters of basophilic hepatocytes As in HE, and PSF outlines affected sinusoids
As in HE, multifocal dense PSF and intra-sinusoidal debris
HE, hematoxylin and eosin; PSF, perisinusoidal fibrosis. 11 reproduced with permission
Features that were not associated with oxaliplatininduced liver injury, steatosis and steatohepatitis were also scored per the Kleiner system.15
a
b
c
d
e
f
Figure 1. Histopathological characteristics of oxaliplatininduced liver injury. (a) Extensive anisonucleosis of hepatocytes can be noted (HE stain, ×400). (b) Trichrome stain highlights basophilic discoloration of hepatocytes and non-zonal distribution of injury on low-power examination (×100). (c) Trichrome stain demonstrates perisinusoidal fibrosis and basophilic discoloration of hepatocytes near a large portal tract (×200). (d) Trichrome stain highlights sinusoidal ectasia accompanied by perisinusoidal fibrosis and basophilic discoloration of hepatocytes, an example of injury score 3 (×400). (e) Disrupted sinusoids with peliosis and cord atrophy are noted in this photomicrograph. This is an example of injury score 4 (HE stain, ×200). (f) Aberrant capillarization of the sinusoids is highlighted by the CD34 immunohistochemistry (×200).
Lesions observed in oxaliplatin-exposed liver also included the presence of aberrant capillarization in the foci of the injury. These were suspected by the presence of perisinusoidal fibrosis and confirmed by CD34 stain. CD34 immunoexpression was scored based on the extent of the reactivity within the acini: 0, immediate sinusoidal expression noted in zone 1 (normal expression); 1, more extensive zone 1 or 3 expression, or both; and 2, for extensive multifocal expression within the acini (Fig. 1f). The cases were scored based on the areas with the most intense expression. Additionally, the proliferation of both hepatocytes and SLCs, as indicated by nuclear Ki-67, was noted. For each case, rough visual estimates were made as to the equivalence of hepatocytes and SLCs (H = SLCs) or the predominance of hepatocytes (H > SLCs) or the predominance of SLCs (H < SLCs) (not shown). The final global scoring system was a combination of HE and trichrome findings of the abovementioned features: anisonucleosis, basophilia and sinusoidal alterations, with the goal of application for the ease of use and reproducibility. Statistical analysis Statistical analysis was performed using SAS 9.2 (SAS Institute, Cary, NC, USA). The global injury scores were categorized based on a three-tier system: 0, no injury; 1–2, mild injury; and 3–4, moderate to severe injury. And then a two-tier system identifying the presence of liver injury (scores 1–4) versus no injury (score 0), for
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
556
I Nalbantoglu et al.
Journal of Digestive Diseases 2014; 15; 553–560
Histological findings of oxaliplatin-induced liver injury
lesions of oxaliplatin-induced liver injury that was confirmed by HE stain. Six of the cases showed mild injury (anisonucleosis with or without perisinusoidal fibrosis) and 26 had moderate to severe injury characterized by varying degrees of sinusoidal dilatation and disruption, hemorrhage and peliosis. The lesions were not necessarily centered around the terminal hepatic venule in zone 3 but were scattered throughout the acini. These findings could also be readily identified along the subcapsular parenchyma as well as at the periportal and perivenular spaces of large collagenized structures. The trichrome stain was useful for identifying inconspicuous areas of injury by HE, as hepatocytes in foci of injury appeared basophilic. These foci usually contained perisinusoidal fibrosis (Fig. 1b–e). The extent of liver injury and IHC findings, along with patients’ characteristics, are summarized in Table 2.
Among the 47 cases who had hepatic resection for metastatic colorectal carcinoma, 32 (68.1%) showed
Aberrant capillarization of the sinusoids by CD34 stain (available in all but one of the cases) was found
statistical analyses. For patients’ individual and clinical characteristics, their associations with the status of oxaliplatin-induced liver injury were summarized using contingency tables and assessed by Fisher’s exact test (injury vs no injury) and Jonckheere’s trend test (for three-tier scoring system to preserve the priori ordering). For continuous variables such as BMI and AST were expressed with mean ± standard deviation and were compared by ANOVA or Kruskal-Wallis, a nonparametric rank sum test, when appropriate. All the tests were two-sided and a P value of 0.05 or less was considered as statistical significance. RESULTS
Table 2.
Clinical and pathological characteristics of patients
Characteristics
Gender, n (%) Male Female Ethnicity, n (%) AA/His Caucasian Additional therapy, n (%) Bevacizumab use† No bevacizumab† Anti-coagulation‡ Postoperative complication, n (%) Infection‡ Others†† Preoperative liver function test (U/L, mean) AST ALT ALP CD34 expression, n (%)‡ 0 1 2 Ki-67 stain, n (%) H = SLCs H > SLCs H < SLCs
Total patients
Injury score
P value
0 (no injury)
1–2 (mild)
3–4 (moderate to severe)
28 (59.6) 19 (40.4)
9 6
4 2
15 11
8 (17.0) 39 (83.0)
2 13
0 6
6 20
15 (33.3) 30 (66.7) 14 (30.4) 14 (30) 6 (12.8) 8 (17.0) Range 11–129 11–194 39–533
2 12 4 5 3 2
2 4 4 2 1 1
11 14 6 7 2 5
26.2 ± 10.52 28.6 ± 15.01 84.7 ± 30.39
34.67 ± 17.82 52.3 ± 60.85 99.5 ± 9.79
41.2 ± 22.84 50.0 ± 38.24 127.8 ± 95.90
1 (2.2) 26 (56.5) 19 (41.3)
1 12 1
0 6 0
0 8 18
12 (25.5) 23 (49.0) 12 (25.5)
5 5 5
0 3 3
7 15 4
0.83
0.16
0.06 0.40 0.37 0.21
0.03 0.09 0.03 0.00004
0.19
†Data available in 45 patients. ‡Data available in 46 patients. ††Others include persistent bilirubin and liver enzyme elevations, wound leak, ascites, hematoma, cerebrovascular and cardiac events. Normal range: aspartate aminotransferase (AST) 0–40 U/L, alanine aminotransferase (ALT) 0–41 U/L, alkaline phosphatase (ALP) 25–129 U/L. AA/His, African American or Hispanic; H, hepatocytes; SLCs, sinusoidal lining cells.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560 in 45 cases. Of these, 26 had focal capillarization limited to a small area around the portal tracts or central veins, whereas the other 19 showed extensive capillarization throughout the lobule (Fig. 1f). Aberrant capillarization was associated with the severity of liver injury (P = 0.00004, Table 2). Ki-67 expression was higher in the hepatocellular component in 49.0% of the cases, whereas 25.5% showed equal expression in hepatocytes and SLCs. No correlation between the severity of liver injury and its expression patterns was noted (P = 0.19, Table 2). Steatosis In all, 23 (48.9%) of the liver specimens showed steatosis, among which 16 were grade 1 (5–33%), 5 were grade 2 (34–66%) and 2 were grade 3 (>66%). Two patients were diagnosed as steatohepatitis. The presence of steatosis was correlated with BMI (P = 0.052, not shown) but not with oxaliplatin, bevacizumab or irinotecan treatment, or the preoperative AST and ALT levels. Clinical features and patient characteristics Altogether 47 patients were included in the study; of these patients, 19 (40.4%) were women. And 39 (83.0%) were Caucasians, 1 (2.1%) was Hispanic and the other 7 (14.9%) were African Americans. Their BMI ranged from 19 to 41 kg/m2 (mean 28.5 kg/m2). The mean number of oxaliplatin cycles was 3.8. The minimum interval from the last dose of oxaliplatin to surgery was 19 days (median 63 days). In all, 15 patients additionally received bevacizumab and 14 patients were given anti-coagulation treatment, including aspirin (n = 10) and warfarin (n = 4) preoperatively. Eight patients had prior liver resections and two had preoperative portal vein embolization. In all, 14 (29.7%) patients had postoperative complications, including four vascular events (one myocardial infarction, two deep vein thromboses, and one cerebrovascular accident); six had an infection (one with persistently markedly elevated bilirubin and liver enzyme levels, one with wound leak, one with small bowel obstruction, one with pleural effusion, one with hematoma formation and one had both a vascular event and infection). Postoperative survival time was 0.5–93 months (median 17 months). One patient died of liver failure at 10 months post-surgery. The patient who died of liver failure had been diagnosed with colorectal cancer and multiple liver
Oxaliplatin-induced liver injury
557
metastases and received four cycles of oxaliplatin preoperatively (cumulative dose 340 mg/m2). He then received right hepatectomy, but was not treated with bevacizumab or anti-coagulation treatment or portal vein embolization. The interval between the last dose of chemotherapy and surgery was 35 days. His preoperative liver biochemical results were normal except the AST level was 44 U/L. No gross abnormalities of the liver had been noted during surgery. Histological analysis of the resected liver tissues showed severe liver injury with marked anisonucleosis and basophilia of hepatocytes, multifocal disruption and enlargement of sinusoids, and multiple foci of intraluminal debris within sinusoids (injury score 4). CD34 stain showed extensive multifocal aberrant capillarization of the sinusoids (injury score 2). Ki-67 nuclear reactivity was noted to be greater in the sinusoidal compartment than in the hepatocytes. The levels of peripheral bilirubin (0.672 μmol/L), AST (78 U/L), ALT (79 U/L) and ALP (473 U/L) were all elevated after operation, but in the absence of other immediate postoperative complications. Imaging studies were negative for biliary obstruction, and the patient was discharged on postoperative day 17. He was then re-admitted to the hospital many times over the next 10 months for abnormal liver function test, failure to thrive, jaundice and melena. At 5 months post operation, a follow-up liver biopsy showed severe acute cholestasis, hepatocellular injury, terminal hepatic venule obliteration and dense zone 3 perisinusoidal fibrosis. Despite supportive treatment, his liver function continued to deteriorate (bilirubin 1.27 μmol/L, ALP 209 U/L, AST 4096 U/L, ALT 3737 U/L) and the patient died due to liver failure at 10 months post-resection. In all remaining patients, short-term postoperative complications, peak INR values, bilirubin levels and survival time (up to 93 months [median 17 months]) were not associated with the extent of liver injury. Table 2 summarizes the characteristics of patients (age and BMI not shown) and the extent of liver injury. There were no statistical differences in gender and ethnicity among patients with different degrees of liver injury. Oxaliplatin dose and the interval between the last oxaliplatin dose and surgery were not correlated with presence of liver injury. Additionally, there was no association with the presence of liver injury in the subgroup of patients who received immediate preoperative oxaliplatin (33/47, 70.2%). The administration of bevacizumab in 15 patients and anticoagulation therapy in 14 patients were not associated with a protection from liver injury; of these
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
558
I Nalbantoglu et al.
Journal of Digestive Diseases 2014; 15; 553–560
patients, 13 who received bevacizumab and 10 who received some forms of anticoagulation showed evidence of some degree of sinusoidal injury (Table 2). Interestingly, when serially drawn, rising preoperative AST (P = 0.01), ALT (P = 0.03) and ALP levels (P = 0.01), even if the values remained within normal ranges, were associated with the presence of liver injury (injury score: 0 vs 1–4, Table 3). Higher injury scores were also associated with rising AST and ALP levels (P = 0.03, Table 2), indicating that even mild increases in liver enzyme levels may raise a concern for the presence of liver injury. Abnormal preoperative AST levels noted in 14 of the patients were associated with higher injury scores (P = 0.01, data not shown). The gross appearance of the liver during surgery was noted in operative notes as ‘mottled, blue’ in two cases, both of whom had the highest injury scores (of 4). One of the patients had abnormal preoperative AST of 42 U/L and ALT of 53 U/L, whereas the other had abnormal preoperative ALP of 144 U/L. The length of postoperative hospital stay were 7 and 8 days, respectively, for them. No postoperative complications were noted in either. In four cases with liver resection, parenchymal nodularity was noted microscopically but not grossly. Among them, 3 had abnormal preoperative AST levels, 3 had had preoperative portal vein embolization and one had had a history of hepatic surgery. One was diagnosed with nodular regenerative hyperplasia (NRH) on a pre-resection biopsy, which was done when oxaliplatin injury was suspected. This patient had no liver function abnormalities. A limited resection was performed, showing similar NRH findings in the biopsy; no adverse events occurred postoperatively. In 49.0% of the cases proliferation, as detected by Ki-67 stain, was greater in the hepatocytes than in the sinusoids. There seemed to be a trend between greater Table 3.
Abnormal AST levels were associated with aberrant capillarization, as indicated by higher CD34 scores (P = 0.04). Notably, there was no association of aberrant sinusoidal CD34 expression with short-term or long-term liver function abnormalities (preoperative and postoperative AST, ALT and ALP levels), INR, peak bilirubin and length of hospital stay or preoperative bevacizumab and anti-coagulation treatment. DISCUSSION Our study has confirmed and extended prior histology-based observations of oxaliplatin-induced chemotherapy injury in the liver.1,6,12,14,16,17 Oxaliplatin has been shown to damage the endothelial cells within the hepatic sinusoids.5,18,19 We have additionally documented an associated hepatocellular response, as indicated by basophilia, anisonucleosis and increased proliferation. Since the original description by Rubbia-Brandt et al., additional features of oxaliplatin-induced liver injury, including hepatocyte atrophy and perisinusoidal fibrosis,7 centrilobular necrosis and NRH have been described14,16,20–22 and the need for a semi-quantitative method for scoring liver injury is discussed.14,22 Rubbia-Brandt et al.14 described its histological features as parenchymal, stromal and vascular, whereas Ryan et al.22 combined grading schemes for fatty liver disease and sinusoidal dilatation along with fibrosis and assigned them a combined vascular injury score. Hepatic plate disruption associated with the necrosis of hepatocytes and perisinusoidal fibrosis, hemorrhage into cords and sinusoidal endothelial damage and necrosis are well recognized in oxaliplatin-induced liver injury. The lesions may be nonzonal and subcapsular,12 both of which were confirmed in this study. The predilection of lesions to be found around the large collagenized portal and vascular structures is a new feature. We propose that the sinusoids near these anchored structures are tethered and less
Preoperative liver function test and presence of liver injury
Preoperative liver function test (U/L, mean ± SD) AST ALT ALP
hepatocellular Ki-67 expression and a rising postoperative ALT levels (P = 0.059).
No injury (n = 15)
Injury present (n = 32)
P value
26.3 ± 10.52 28.6 ± 15.01 84.67 ± 30.39
39.9 ± 21.88 50.47 ± 42.16 112.5 ± 86.94
0.01 0.03 0.01
Aspartate aminotransferase (AST) 0–40 U/L, normal range: alanine aminotransferase (ALT) 0–41 U/L, alkaline phosphatase (ALP) 25–129 U/L. SD, standard deviation.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 553–560 flexible, and therefore, more readily disrupted once the lining is damaged in these areas. Another possible explanation could be zonal differences in drug distribution and utilization within the hepatic sinusoids. Hepatocellular anisonucleosis can be observed in several other pathological conditions but in our case series it was easily recognized microscopically under LPF, and may be an initial clue to oxaliplatin-induced liver injury, which can then alert the pathologist to look for other described lesions (Fig. 1a). Even though anisonucleosis may be a sign of senescence, it is also a feature of hepatocellular regeneration. Ectatic sinusoids, some with intraluminal debris, can be noted scattered in a nonzonal fashion in the acini. These lesions may be seen on HE stained specimens when they are large and filled with red cells, but smaller foci are often best appreciated by the concomitant altered basophilic hepatocellular discoloration on trichrome (Fig. 1b–e). Many of the cases also showed features of aberrant capillarization, as indicated not only by collagen deposition but also by re-endothelialization of nonfenestrated endothelium, as detected by sinusoidal reactivity with anti-CD34. Aberrant capillarization related to oxaliplatin use was reported by Narita et al.2 The investigators also found that the number of cycles and abnormal liver function were also associated with CD34 expression.2 However, the dose of oxaliplatin did not correlate with the severity of aberrant capillarization in our data set. A higher percentage of proliferative hepatocytes detected by Ki-67 is indicative of hepatocellular regeneration in the face of sinusoidal damage. Hepatocellular regeneration may, in part, explain the lack of association with histological findings and adverse clinical outcomes. Other investigators have suggested similar results.14 Previous and most current experimental animal18,23 studies showed that steatosis was not associated with oxaliplatin-induced liver injury. Histological analysis of 334 non-neoplastic human livers by Ryan et al.22 demonstrated that oxaliplatin exposure was related to the development of vascular injury but not steatohepatitis. Our findings are in keeping with these studies. The only statistically significant correlation in our study was between BMI and steatosis. In the era of an obesity epidemic, we are more likely to find background steatosis and steatohepatitis. This, however, should not be misinterpreted as oxaliplatin-induced liver injury and needs to be evaluated separately.
Oxaliplatin-induced liver injury
559
The pathogenesis of oxaliplatin-induced liver injury is not completely understood. Yeong et al. described a rodent model of pyrrolizidine alkaloids-induced liver injury and noted that a loss of endothelium was associated with extensive red blood cell extravasation and the presence of collagen fibers in the space of Disse in severe sinusoidal injury.18 Recently, a mouse model was developed for oxaliplatin-induced liver injury.23 In this model, the mouse liver specimens showed the development of liver sinusoidal dilatation, the disruption of endothelial cells within sinusoids, and hepatocellular injury, evidenced by elevated ALT and AST levels when exposed to oxaliplatin.23 In addition, early collagen deposition was noted.23 Similar observations have been made14,22 in human liver with oxaliplatin-induced liver injury.14,22 Recent studies have shown that hepatic sinusoidal injury results in up-and-down regulation of genes in several inflammatory, vascular and fibrogenic pathways.23,24 In addition, lower levels of ERCC2, a nucleotide excision repair gene, was found to be associated with sinusoidal injury with oxaliplatin-based chemotherapy.25 To date, however, in spite of wellcharacterized injury patterns, there are no laboratory tests, serum or genetic markers to predict susceptibility to oxaliplatin-induced sinusoidal injury. The limitations of this retrospective study are acknowledged. As with any retrospective study, many variables were not controlled. The sample size was relatively small, while the selection and exclusion of the histological samples were subjected to strict criteria, sampling variability cannot be completely eliminated since oxaliplatin-induced liver injury is a heterogeneous process. However, the strengths of the study included its detailed systematic approach to histopathological analyses and the thorough evaluation for relevant short-term and long-term clinical associations. In conclusion, this study described a method of semiquantitatively evaluating histological features of oxaliplatin-induced liver injury, and combined the histological spectrum of oxaliplatin-induced liver injury into a system that is simple and practical to apply which can be performed with routine histochemical stains. The injury score that we proposed was developed on resected specimens once the injury occurred. The injury scoring system along with CD34 stain deserves to be tested prospectively in biopsy samples in efforts to guide clinical teams for future treatment or surgery.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
560
I Nalbantoglu et al.
ACKNOWLEDGMENT The authors acknowledge intradepartmental research funding from the Department of Pathology and Immunology and the support of Biostatistics Core, Siteman Comprehensive Cancer Center and the National Cancer Institute Cancer Center support grant P30 CA091842. REFERENCES 1 Adam R, Aloia T, Lévi F et al. Hepatic resection after rescue cetuximab treatment for colorectal liver metastases previously refractory to conventional systemic therapy. J Clin Oncol 2007; 25: 4593–602. 2 Narita M, Oussoultzoglou E, Chenard MP et al. Liver injury due to chemotherapy-induced sinusoidal obstruction syndrome is associated with sinusoidal capillarization. Ann Surg Oncol 2012; 19: 2230–7. 3 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 randomised controlled trial. Lancet 2008; 371: 1007–16. 4 Wolf PS, Park JO, Bao F et al. Preoperative chemotherapy and the risk of hepatotoxicity and morbidity after liver resection for metastatic colorectal cancer: a single institution experience. J Am Coll Surg 2013; 216: 41–9. 5 Cleary JM, Tanabe KT, Lauwers GY, Zhu AX. Hepatic toxicities associated with the use of preoperative systemic therapy in patients with metastatic colorectal adenocarcinoma to the liver. Oncologist 2009; 14: 1095–105. 6 Aloia T, Sebagh M, Plasse M et al. Liver histology and surgical outcomes after preoperative chemotherapy with fluorouracil plus oxaliplatin in colorectal cancer liver metastases. J Clin Oncol 2006; 24: 4983–90. 7 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; 247: 118–24. 8 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; 110: 2761–7. 9 Brouquet A, Benoist S, Julie C et al. Risk factors for chemotherapy-associated liver injuries: A multivariate analysis of a group of 146 patients with colorectal metastases. Surgery 2009; 145: 362–71. 10 Rubbia-Brandt L, Audard V, Sartoretti P et al. Severe hepatic sinusoidal obstruction associated with oxaliplatin-based chemotherapy in patients with metastatic colorectal cancer. Ann Oncol 2004; 15: 460–6. 11 Klonowski PW, Tan B, Linehan DC, Porembka MR, Brunt EM. Capillarization of hepatic sinusoids after neoadjuvant therapy. Lab Invest 2010; 90: 360A.
Journal of Digestive Diseases 2014; 15; 553–560 12 Rubbia-Brandt L. Sinusoidal obstruction syndrome. Clin Liver Dis 2010; 14: 651–68. 13 Rubbia-Brandt L. Hepatic lesions induced by systemic chemotherapy for digestive cancer. Ann Pathol 2010; 30: 421–5 (in French). 14 Rubbia-Brandt L, Lauwers GY, Wang H et al. Sinusoidal obstruction syndrome and nodular regenerative hyperplasia are frequent oxaliplatin-associated liver lesions and partially prevented by bevacizumab in patients with hepatic colorectal metastasis. Histopathology 2010; 56: 430–9. 15 Kleiner DE, Brunt EM, Van Natta M et al.; Nonalcoholic Steatohepatitis Clinical Research Network. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005; 41: 1313–21. 16 Arotçarena R, Calès V, Berthelémy P et al. Severe sinusoidal lesions: a serious and overlooked complication of oxaliplatin-containing chemotherapy? Gastroenterol Clin Biol 2006; 30: 1313–6. 17 Kandutsch S, Klinger M, Hacker S, Wrba F, Gruenberger B, Gruenberger T. Patterns of hepatotoxicity after chemotherapy for colorectal cancer liver metastases. Eur J Surg Oncol 2008; 34: 1231–6. 18 Yeong ML, Wakefield SJ, Ford HC. Hepatocyte membrane injury and bleb formation following low dose comfrey toxicity in rats. Int J Exp Pathol 1993; 74: 211–7. 19 Vreuls CP, Van Den Broek MA, Winstanley A et al. Hepatic sinusoidal obstruction syndrome (SOS) reduces the effect of oxaliplatin in colorectal liver metastases. Histopathology 2012; 61: 314–8. 20 Slade JH, Alattar ML, Fogelman DR et al. Portal hypertension associated with oxaliplatin administration: clinical manifestations of hepatic sinusoidal injury. Clin Colorectal Cancer 2009; 8: 225–30. 21 Tamandl D, Klinger M, Eipeldauer S et al. Sinusoidal obstruction syndrome impairs long-term outcome of colorectal liver metastases treated with resection after neoadjuvant chemotherapy. Ann Surg Oncol 2011; 18: 421–30. 22 Ryan P, Nanji S, Pollett A et al. Chemotherapy-induced liver injury in metastatic colorectal cancer: semiquantitative histologic analysis of 334 resected liver specimens shows that vascular injury but not steatohepatitis is associated with preoperative chemotherapy. Am J Surg Pathol 2010; 34: 784–91. 23 Robinson SM, Mann J, Vasilaki A et al. Pathogenesis of FOLFOX induced sinusoidal obstruction syndrome in a murine chemotherapy model. J Hepatol 2013; 59: 318–26. 24 Rubbia-Brandt L, Tauzin S, Brezault C et al. Gene expression profiling provides insights into pathways of oxaliplatin-related sinusoidal obstruction syndrome in humans. Mol Cancer Ther 2011; 10: 687–96. 25 Pilgrim CH, Brettingham-Moore K, Pham A et al. mRNA gene expression correlates with histologically diagnosed chemotherapy-induced hepatic injury. HPB (Oxford) 2011; 13: 811–6.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
