Perioperative Chemotherapy With or Without Bevacizumab in Patients With Metastatic Colorectal Cancer Undergoing Liver Resection

Transcription

1 Original Study Perioperative With or Without Bevacizumab in Patients With Metastatic Colorectal Cancer Undergoing Liver Resection Anastasia Constantinidou, 1 David Cunningham, 1 Fatima Shurmahi, 1 Uzma Asghar, 1 Yolanda Barbachano, 2 Aamir Khan, 3 Satvinder Mudan, 3 Sheela Rao, 1 Ian Chau 1 Abstract The impact of adding bevacizumab to perioperative chemotherapy in patients with colorectal cancer (CRC) undergoing liver resection is yet to be defined. A retrospective review of our patient records showed that the addition of bevacizumab did not increase morbidity or mortality related to liver resection. Pathologic complete response (CR) is associated with prolonged survival. Background: Patients with colorectal cancer (CRC) and liver metastases benefit from perioperative chemotherapy and liver resection. The potential benefit of adding bevacizumab is yet to be defined. The impact of bevacizumab on liver resection complications has been explored in a small number of retrospective studies. Methods: The records of patients with CRC and liver metastases who underwent liver resection and had received perioperative chemotherapy were reviewed. Complications were reported separately for 2 groups (chemotherapy alone vs chemotherapy and bevacizumab). Survival outcomes (progression-free survival [PFS] and overall survival [OS]) for responders and nonresponders were estimated using the Kaplan-Meier method. Results: Fifty-two patients received chemotherapy alone and 42 patients received chemotherapy and bevacizumab. The median time from the end of systemic treatment to liver resection was 59 days ( days) for the chemotherapy group and 62 days ( days) for the chemotherapy and bevacizumab group. Postoperative complications developed in 54% of the chemotherapy group and in 48% of the chemotherapy and bevacizumab group. Severe complications (grade III-V) occurred in only 13% and 12%, respectively (P.822). Pathologic complete response (CR) was seen in 11/94 patients. Poor performance status (PS) before starting chemotherapy was associated with higher rates of complications (P.002), and severe complications led to prolonged hospital admission (P.001). Patients with pathologic CR had longer OS (P.0275), but there was no difference in OS between responders and nonresponders (P.778). Conclusion: The addition of bevacizumab to chemotherapy does not increase liver resection complication rates. Pathologic CR is associated with prolonged survival. Clinical Colorectal Cancer, Vol. 12, No. 1, Elsevier Inc. All rights reserved. Keywords: Bevacizumab, Colorectal cancer, Liver resection, Perioperative chemotherapy, Pathologic complete response 1 Department of Medicine 3 Department of Academic Surgery 2 Department of Clinical Research and Development Royal Marsden Hospital, London and Surrey, UK Submitted: Jan 30, 2012; Revised: Jun 3, 2012; Accepted: Jul 9, 2012; Epub: Sep 26, 2012 Address for correspondence: David Cunningham, MD, FRCP, Department of Medicine, Royal Marsden Hospital, Sutton SM2 5PT, UK contact: david.cunningham@rmh.nhs.uk Introduction The liver is the most common site of metastatic spread in colorectal cancer (CRC) with approximately 50% of patients experiencing liver metastases during the course of their disease. 1,2 Over the past decade, significant advances in the management of patients with oligometastatic disease have led to improvements in survival. Liver resection has been established as the treatment of choice, and with the appropriate selection of patients 5-year survival rates approach 35% to 40%. 3-5 According to the refined criteria for resectability, patients are eligible for liver resection if all visible metastatic lesions /$ - see frontmatter 2013 Elsevier Inc. All rights reserved. Clinical Colorectal Cancer March

2 Bev and Liver Resection in Colorectal Cancer can be eliminated and the remaining liver remnant exceeds 30% of the original liver volume, with adequate vascular supply and biliary drainage. 6 Extrahepatic disease is not considered an absolute contraindication for liver surgery provided it can also be resected. 7 Despite surgical advances, only 10% to 15% of patients have resectable liver disease at presentation. 8,9 Preoperative chemotherapy has been introduced to increase the number of patients that may be eligible for liver resection by downsizing liver metastases. 10 Other potential benefits of preoperative chemotherapy have been recognized, including the treatment of micrometastatic disease, the early identification of chemoresistant disease, and the avoidance of liver surgery in patients with chemotherapy-resistant progressive disease. In patients with unresectable liver disease that becomes resectable after preoperative chemotherapy, 5-year survival rates may reach 30%. 11 More recently, perioperative chemotherapy has attracted attention, with evidence from prospective randomized data supporting the advantages of this approach. 12 However, survival data and data on the potential additional benefit of targeted agents are limited. Bevacizumab is a humanized monoclonal antibody targeting vascular endothelial growth factor (VEGF) and thus inhibiting its interaction with the VEGF receptor. Prospective studies have shown that when combined with chemotherapy, bevacizumab may improve survival in metastatic CRC The combination of bevacizumab and chemotherapy has been reported in the preoperative management of patients undergoing liver resection, 16 and concerns about the impact of bevacizumab on surgical complications have been explored in a small number of retrospective studies Despite the fact that different grading systems have been used to assess complications in these studies, it appears that the addition of bevacizumab to preoperative chemotherapy does not increase the morbidity or mortality related to liver resection, but optimal timing of surgery after the last infusion of bevacizumab has yet to be defined. The aim of this large single-institution retrospective study is to report complications after liver resection and the survival outcomes of patients with colorectal liver metastases who underwent liver resection and received perioperative chemotherapy with or without bevacizumab. Patients and Methods Approval was obtained by the Committee for Clinical Research at the Royal Marsden Hospital and the Institute of Cancer Research (CCR 3356 SE). Patient Identification Using the electronic patient records at the Royal Marsden Hospital, we assembled a retrospective liver resection database. Patients with primary CRC who were managed with perioperative systemic therapy and liver resection in our institution until September 1, 2010 were included. Patients with resectable colorectal liver metastases who received any combination of chemotherapy agents with or without bevacizumab during this period were eligible for inclusion. Patients were ineligible if they had been treated with liver resection alone, if they had received postoperative (adjuvant) therapy only, or if they had not proceeded to liver resection despite receiving preoperative therapy. When patients had perioperative therapy and liver resection on more than 1 occasion during the course of their disease, only the first occurrence was included in the analysis. Data on demographics and disease characteristics were obtained from the electronic patient records and included age at presentation of the colorectal primary lesion and liver metastases, sex, primary site (rectum, rectosigmoid junction, sigmoid, descending colon, and ascending colon) and whether the primary tumor had been surgically removed. Specific information on metastatic liver disease included differentiation between synchronous ( 3 months) and metachronous ( 3 months) presentation as well as documentation of the number of liver lesions, their distribution, and the size of the largest lesion as determined by preoperative imaging. Additional information included details on the presence of extrahepatic metastatic disease and whether it was removed along with the liver resection. Carcinoembryonic antigen (CEA) levels were also monitored throughout treatment (start of systemic treatment, preoperatively, postoperatively, and at the end of postoperative systemic therapy). Liver Resection All cases were discussed in the specialist gastrointestinal and liver multidisciplinary meeting. A perioperative strategy was recommended if the metastatic liver disease was deemed potentially operable after review of baseline imaging of the chest, abdomen, and pelvis. Computed tomography, magnetic resonance imaging, and in most cases positron emission tomography were used to determine the status of disease and response to treatment. Response to systemic therapy was assessed by Response Evaluation Criteria in Solid Tumors. The type and extent of curative liver resection (hemihepatectomy, segmentectomy, trisectionectomy, wedge resection) were determined by the hepatobiliary surgeon in the multidisciplinary meeting but were modified accordingly during the operation if necessary. Resections were defined as either complete with negative microscopic margins (R0) or incomplete with microscopically positive margins (R1) or macroscopically positive margins (R2). The additional use of radiofrequency ablation was documented. Postoperative Complications Information on postoperative complications was obtained from the intensive care unit (ICU) and surgical follow-up records. Immediate complications were categorized as follows: wound complications (dehiscence or infection that required antibiotics or drainage or surgical management), hepatobiliary complications (hyperbilirubinemia 7 mg/dl or 63 mol/l, 21 bile leakage/biloma), pulmonary complications (pleural effusion requiring drainage, pneumonia requiring antibiotics, invasive or noninvasive ventilation), cardiac complications (arrhythmia, myocardial infarction, cardiovascular instability requiring ICU admission), infection other than wound and pneumonia (sepsis, urinary tract infection, abscess requiring drainage), deep vein thrombosis/pulmonary embolus, blood loss requiring blood transfusion, bowel perforation, and renal impairment requiring dialysis. Death during the inpatient stay or within 30 days of surgery was documented. Complications were graded on a scale of 1 to 5 using the reporting system proposed by Dindo et al. 22 Absence of complications was graded as 0. Grade I complications were defined as complications managed expectantly or by medical means alone, such as electrolyte supplementation, pain medication, and antiemetic medication. Grade II complications included infection managed with antibiotics, arrhythmia treated with intravenous medication, total parenteral 16 Clinical Colorectal Cancer March 2013

3 Anastasia Constantinidou et al feeding, and blood product transfusion. Grade III complications involved operative or interventional procedures, including drainage of an abscess, endoscopy, stent placement, bronchoscopy, or reoperation. Grade IV toxicity comprised severe organ damage requiring admission to the ICU for intubation, ventilation, hemodialysis, hemofiltration, or pacemaker/defibrillator placement, bowel diversion, and organ resection. Death resulting from a complication was graded as V. Systemic Therapy Information regarding preoperative chemotherapy comprised performance status (PS) at the start of treatment when reported, type of chemotherapy, addition of bevacizumab or not, total number of cycles administered, the date of start of therapy, and the day the last dose was administered. Not all patients received postoperative systemic therapy but for those who did, data collection included type of chemotherapy, addition of bevacizumab or not, start and end dates, and total number of cycles administered. Some patients received bevacizumab in the context of the BOXER study, a phase II study assessing the efficacy and safety of bevacizumab in combination with chemotherapy in patients with CRC with liver-only metastases or with synchronous presentation with liver metastases, deemed unsuitable for upfront liver resection. Grade III and/or IV toxicities per the National Cancer Institute Common Toxicity Criteria Adverse Events version 3.0 grading system were documented for all patients. Statistical Analysis Descriptive statistics comprised median and range for continuous variables and proportions (%) for categorical variables. These were reported separately for the 2 groups (chemotherapy only vs. chemotherapy with bevacizumab) regarding the postoperative complications. The 2 test was used to evaluate the association of categorical variables. Progression-free survival (PFS) and overall survival (OS) for the whole group (responders vs. nonresponders) were estimated using the Kaplan-Meier method. Survival outcomes between responders and nonresponders were compared with the log-rank test. The Cox regression model was used to test for possible prognostic factors (including age at presentation, sex, PS at the start of chemotherapy, number of metastatic lesions, size of largest metastasis, synchronous vs. metachronous presentation with metastatic liver disease, negative microscopic margins [R0 vs. R1], and treatment with bevacizumab) and their interaction with treatment. A P value of less than 0.05 was considered statistically significant. Results Demographics and Tumor Characteristics Up until September 1, 2010, 97 patients with CRC metastatic to the liver underwent liver resection in combination with perioperative chemotherapy. Three patients who received treatment in the context of the HORIZON study (oxaliplatin, 5-fluoruracil plus either cediranib/placebo or bevacizumab/placebo) were excluded from this analysis because it was not possible to know whether they had received bevacizumab or not in the context of the study. Fifty-two patients received chemotherapy alone (chemotherapy) and 42 patients received chemotherapy in combination with bevacizumab. Patient and tumor characteristics were similar in the 2 groups apart from the fact that the majority of patients in the chemotherapy and Table 1 Variable Age at Presentation Age at Diagnosis of Liver Metastases Demographics and Tumor Characteristics n 52 and Bevacizumab n 42 P Value 63 (38-79) 62 (29-79).721 a 64 (38-79) 63 (29-79).429 a Sex Female b Male Primary Tumor Rectum b Rectosigmoid junction/ sigmoid 7 7 Descending colon Ascending colon 12 8 Primary Tumor Removed Yes c No 1 1 Liver Metastases Synchronous b Metachronous Distribution of Lesions Unilobar b Bilobar Number of Liver Lesions Median (range) Size of Largest Lesion No. of Previous Regimens for Advanced Disease a Mann-Whitney U test. b 2 test. c Fisher exact test. 2(1to 5) 2(1to 10).310 a excluding multiple lesions 2.2 (0.7-17) 2.8 (1-11.2).383 a 0 (0-2) 0 (0-2).488 a bevacizumab group (31/42 [74%]) had presented with synchronous metastatic liver disease. The median time interval between the primary diagnosis and the development of liver metastases was 167 days in the chemotherapy group and 16 days in the chemotherapy and bevacizumab group. Details of patient and tumor characteristics are shown in Table 1. Perioperative Therapy In the chemotherapy group, 34 patients received both pre- and postoperative chemotherapy, whereas the remaining 18 patients received preoperative chemotherapy alone. In the chemotherapy and Clinical Colorectal Cancer March

4 Bev and Liver Resection in Colorectal Cancer bevacizumab group, 27 patients received both pre- and postoperative therapy and 15 patients received preoperative chemotherapy. The PS at the start of systemic therapy was not recorded in approximately half of the patients in each group (chemotherapy 19 patients, chemotherapy and bevacizumab 17 patients), whereas documented PS was between 0 and 1 in the majority of patients (chemotherapy 32 patients, chemotherapy and bevacizumab 25 patients). Capecitabine-oxaliplatin (CAPOX) was the most commonly used regimen both in the chemotherapy and the chemotherapy and bevacizumab groups. The median time between the last administration of chemotherapy and surgery was 59 days and that between chemotherapy and bevacizumab and surgery was 62 days. Bevacizumab was discontinued in 1 patient in whom atrial thrombosis developed while receiving treatment. There were no other significant grade III or IV toxicities associated with bevacizumab. All the chemotherapy regimens used and details on duration of therapy are summarized in Table 2. Liver Resection The type of liver resection was comparable in the 2 groups. In 77% of patients in the chemotherapy group and in 69% of patients in the chemotherapy and bevacizumab group, R0 resection was achieved. Pathologic CR was achieved in 11 patients 6/52 (12%) in the chemotherapy group and 5/42 (12%) in the chemotherapy and bevacizumab group. The estimated blood loss was similar in both groups, as was the hemogloblin drop and the blood transfusion requirements. Details of liver resection in both groups are shown in Table 3. Postoperative Complications Twenty-four patients in the chemotherapy group (46%) and 22 patients in the chemotherapy and bevacizumab group (52%) had no postoperative complications. There was no significant difference between the rates of complications in the 2 groups (P.548). The severity of complications was similar in both groups ( 2 P.691), with the rate of grade III or IV complications being comparable in the 2 groups (13% in the chemotherapy group vs. 12% in the chemotherapy and bevacizumab group; P.822). Two cases of bowel obstruction/ileus were observed postoperatively in the chemotherapy group, but none was seen in the chemotherapy and bevacizumab group. Five cases of bleeding/hematoma were documented in the chemotherapy group, but none was seen in the chemotherapy and bevacizumab group. There were no bowel perforation episodes in either of the groups. There was 1 death during the hospital stay after surgery in the chemotherapy group and no deaths in the chemotherapy and bevacizumab group. The grade of each complication and its management are shown on Table 4. There was no correlation between the type of preoperative therapy or the type of liver resection and postoperative complications. There was a correlation between grade III/IV toxicity and prolonged inpatient stay (Mann-Whitney P.001). Binary logistic regression, which was used to test for predictors of complications in the whole group (94 patients), showed that worse PS at the start of chemotherapy was associated with a higher risk of complications (P.002). Response and Survival Outcomes After preoperative chemotherapy, 68 patients (72%) achieved partial response (PR) in the liver (responders), whereas 26 had stable disease (SD 22) or progressive disease (PD 4) (nonresponders). There was no difference in response rate between oxaliplatin- and irinotecan-containing regimens (P.148). The most common sites of relapse involved the lung and liver. The median follow-up period for this study was 27.4 months. The median PFS for the whole group was 18 months (95% confidence interval [CI], months) and the median OS was 65 months (95% CI, months). There was a difference in OS between patients with pathologic CR and those without (P.0275), but there was no difference in OS between responders and nonresponders (P.778). The Kaplan-Meier curves are shown in Figure 1. No significant association was found between potential predictive factors and greater benefit with the addition of bevacizumab in improving OS. However, regression analysis to test the predictive significance of the same factors in PFS revealed an association between a higher number of liver metastases ( 4) and shorter PFS (P.008 for whole group). The variation in CEA levels for both groups during treatment is shown in Figure 2. Discussion Bevacizumab is increasingly used in addition to perioperative chemotherapy in the management of CRC with liver metastases. In view of the critical role of VEGF in hepatic regeneration and proliferation and the potential toxicity of bevacizumab including arterial and venous thromboembolism, gastrointestinal perforation, bleeding, and poor wound healing concerns have been raised about the safety of liver resection after treatment with this agent. Only a small number of retrospective studies have addressed this issue to date, and their results suggest that the addition of bevacizumab does not increase morbidity or mortality related to liver resection. With this study we add further evidence to support these results, but we also report on the important issues of timing of liver resection after bevacizumab and the impact of chemotherapy with or without bevacizumab to survival outcomes. The severity of postoperative complications in this study was determined using an established grading system, with the exception of mortality, which in our study (and in most other similar studies) was modified to include patient death during inpatient stay or within 30 days of surgery rather than any death associated with a complication. 22,23 Severe complications (grade III-V) were comparable in the 2 groups, occurring in 13% of patients in the chemotherapy group and 12% of patients in the chemotherapy and bevacizumab group, including 1 death in the chemotherapy group. Notably all cases of perforation/ileus (n 2 [2%]) and bleeding/hematoma (n 5 [5%]) occurred in the chemotherapy group and not in the chemotherapy and bevacizumab group. The rates of wound healing complications were similar in both groups, and there were no episodes of bowel perforation in either group. Not surprisingly, there was a statistically significant association between grade III/IV toxicity and prolonged inpatient stay. The chemotherapy alone and the chemotherapy and bevacizumab groups were not only comparable in terms of demographics but also in terms of the type of perioperative chemotherapy administered and 18 Clinical Colorectal Cancer March 2013

5 Anastasia Constantinidou et al Table 2 Preoperative Systemic Therapy Table 3 Liver Resection Variable Performance Status at Start of Treatment n 52 (%) and Bevacizumab n 42 (%) 0 14 (27) 19 (46) 1 18 (35) 6 (14) 2 1 (2) 0 U 19 (36) 17 (40) Type of Preoperative Oxaliplatin/capecitable (CAPOX) Oxaliplatin/5-FU (FOLFOX) Irinotecan/5-FU (FOLFIRI) Irinotecan/capecitabine (CAPIRI) 34 (65) 21 (50) 5 (10) 6 (14) 11 (21) 13 (31) 0 1 (2) Other regimens a 2 (4) 1 (2) Total No of Cycles, Time from Last to Surgery, Median (range) d Time from Last Bev Dose to Surgery, d Time from Surgery to Adjuvant, Median (range) d No. of Patients Receiving Postoperative Type of Postoperative Oxaliplatin/capecitable (CAPOX) Oxaliplatin/5-FU (FOLFOX) Irinotecan/5-FU (FOLFIRI) Irinotecan/capecitable (CAPIRI) P Value (1-12) 4.5 (4-12).035 a 59 (33-181) 62 (44-127).503 a NA 73 (44-141) NA 59 (18-115) 69 (45-108).07 a 37 (71) 30 (71) (55) 14 (47) 4 (12) 1 (3) 4 (12) 10 (33) 0 2 (7) Capecitabine 4 (12) 2 (7) Other regimens a 3 (9) 1 (3) Total No. of Postoperative Cycles 4 (2-8) 1 patient ongoing 4 (3-14) 4 patients ongoing.081 c.056 Abbreviations: Bev bevacizumab; d days; 5-FU 5-fluorouracil; NA not available; U unknown. a Mann-Whitney U test. b 2 test. c Fisher exact test. Variable Type of Liver Resection n 52 (%) Hemihepatectomy and Bevacizumab n 42 (%) P Value Segmentectomy a Other (trisectionectomy, wedge resection) 6 1 Resection Margins 40 (77) 29 (69).732 b R R1 1 1 R2 0 1 U RFA in Addition to Resection (concurrent or within 6 wk) Pathologic CR a Yes 6 (12) 5 (12).636 b No 46 (88) 35 (83) Synchronous Resection Primary tumor b Other metastatic sites 4 6 Estimated Blood loss, ml Blood Transfusion Requirements 650 ( ) 540 ( ).876 c No. of patients b No. of units, median (range) Hemoglobin, g/dl, 2 (1-16) 2 (1-8).400 c Preoperative 12.8 ( ) 13.3 ( ).270 c Postoperative 10.8 (8-16) 10.4 (6-13).124 c Length of Stay in Hospital, d Median (range) Postoperative Complications 10 (4-32) 10 (4-57).223 c No a Yes 28 (54) 20 (48) Grade III/IV Complications No a Yes 7 (13) 5 (12) Abbreviations: RFA radiofrequency ablation; U unknown. a 2 test. b Fisher exact test. c Mann-Whitney U test. the type of liver surgery performed. The majority of patients in both groups had not received chemotherapy for advanced disease previously. PS was 0 or 1 for most patients at the start of chemotherapy, although it was recorded in only half of the patients in each group, Clinical Colorectal Cancer March

6 Bev and Liver Resection in Colorectal Cancer Table 4 Postoperative Complications in 28 Patients in the Group and 20 Patients in the and Bevacizumab Group Variable and Bevacizumab Type of Complication No 44 Management Severity (Grade) No 31 Management Wound Dehiscence 4 3 settled conservatively; 1 required debridement I III 4 Settled conservatively I Infection 2 Required oral antibiotics II 2 1 required oral antibiotics; 1 required debridement II III Hepatobiliary Hyperbilirubinemia 63 mol/l 7 Managed expectantly; I 4 Managed expectantly; I Bile leak 3 all 3 required stent insertion III 3 all 3 required stent insertion III Pulmonary Pleural effusion 4 3 managed expectantly; 1 required drainage I III 2 Managed expectantly I Pneumonia 1 Required IV antibiotics II 0 Cardiovascular Arrhythmia 6 Hypertension 2 3 settled conservatively; 3 required cardiovascular medication Required cardiovascular medication III 3 II 1 Hypotension required PM insertion; 1 required an IMD Required cardiovascular medication All required cardiovascular medication Myocardial infarction 1 ICU/organ support IV Pericarditis 1 Managed with electrolytes only I Other Infection Urinary tract 1 Required oral antibiotics II 1 Required IV antibiotics II Sepsis 4 Required IV antibiotics II 2 Required IV antibiotics II Liver abscess 0 1 Required drainage III DVT/PE 0 2 LMWH given II Hematoma/Bleeding 5 4 required drainage; 1 required blood transfusion III II 0 Bowel Obstruction/Ileus 2 1 required reoperation; 1 required TPN III II 0 Other Coagulopathy (INR 1.5) 1 Required FFP II 1 Managed with vitamin K I Anastomotic leak 0 1 Managed expectantly I Death during IP stay/ within 30 d 1 V 0 Severity (Grade) Abbreviations: DVT/PE deep vein thrombosis/pulmonary embolism; FFP fresh frozen plasma; ICU intensive care unit; IMD implantable defibrillator; INR international normalized ratio; IP inpatient; IV intravenous; LMWH low-molecular-weight heparin; PM pacemaker; TPN total parenteral nutrition. IV IV II II and worse PS at the start of chemotherapy was associated with a higher risk of postoperative complications. The most commonly used preoperative chemotherapy regimen was oxaliplatin in combination with capecitabine or 5-FU (75% in the chemotherapy group vs. 64% in the chemotherapy and bevacizumab group). More than 70% of patients in both groups went on to have postoperative chemotherapy. The addition of bevacizumab to chemotherapy was well tolerated, although the time between liver resection and the start of postoperative chemotherapy in the chemotherapy and bevacizumab group was longer, although not statistically significantly, compared with the chemotherapy group (69 vs. 59 days; P.07). No association was found between the type of chemotherapy or the type of liver resection and an increased rate of postoperative complications. The issue of optimal timing of liver resection after neoadjuvant chemotherapy remains unresolved. In the study by Kesmodel et al, 17 no association was found between time from discontinuation of bevacizumab to surgery and postoperative morbidity. Even when liver resection was performed soon after the discontinuation of bevaci- 20 Clinical Colorectal Cancer March 2013

7 Anastasia Constantinidou et al Figure 1 A 100 Probability of PFS (%) (A) Progression-Free Survival (PFS) for the Whole Group (responders and nonresponders). (B) Overall Survival (OS) for the Whole Group (responders and nonresponders) PFS Years from start of chemotherapy B OS 100 Probability of survival (%) Years from start of chemotherapy explain the reason that in this study the number of patients with synchronous liver metastases was significantly higher in the chemotherapy and bevacizumab group compared with the chemotherapy group. The significant proportion of BOXER patients (more advanced liver disease) in the chemotherapy and bevacizumab group could account for the absence of a noticeable impact from the addition of bevacizumab on survival outcomes (results not shown). Data to support the potential survival benefit of targeted agents such as bevacizumab in the perioperative management of CRC are limited. The retrospective nature of this study does not allow a headto-head comparison between the chemotherapy group and the chemotherapy and bevacizumab group in terms of survival outcomes. Instead a comparison between responders and nonresponders showed that there was no statistically significant difference in either PFS or OS for the follow-up period of this study. Consistent with published literature, our results showed that patients with pathologic CR had a prolonged survival. 25,26 The limitations associated with a retrospective, single-institution study are acknowledged. First, there is the risk of selection bias for patients with better PS and less advanced disease, which is difficult to avoid in the context of a retrospective analysis. Second, although the total number of patients undergoing perioperative chemotherapy and liver resection steadily increased over the 4-year period of this study (reflecting increasing evidence to support this approach), the addition of bevacizumab in particular increased in the last 2 years primarily because of drug availability through the BOXER study. As a result, during this time there was a preference toward the option of perioperative chemotherapy and bevacizumab rather than chemotherapy alone. Third, increased emphasis on accurate recording of postoperative complications may have influenced not only the number but also the type of complications documented in the recent years of this study as zumab, there was no significant difference in the rates of postoperative complications compared with the group who underwent surgery 60 days after bevacizumab discontinuation (P.21). Although the results implied that surgery might be safely performed at 4 to 6 weeks rather than 8 weeks, the number of patients in each group was small and did not allow definitive conclusions to be drawn. In our study, the median time from the end of chemotherapy to surgery was similar in both groups (59 days for the chemotherapy group and 62 for the chemotherapy and bevacizumab group). In the absence of a gold standard, this interval appears sufficient to allow a safe surgical procedure, and from our data there is no evidence to support that it had an adverse effect on the clinical outcome. On the contrary, monitoring of the CEA levels at 4 key time points (prechemotherapy bevacizumab, preoperatively, postoperatively, and at the end of systemic treatment) showed that the response trend was maintained throughout this period (Figure 2). Because it was not approved by the National Institute of Clinical Excellence, bevacizumab was not routinely available in the National Health System in the United Kingdom during the period of this study for patients with metastatic CRC, but some patients could have access to bevacizumab through clinical trials. About one third of the patients who had bevacizumab in this study (12/42 [28%]) received the drug in the context of the BOXER study. 24 This may Figure 2 Mean/Median CEA Carcinoembryonic (CEA) Levels at 4 Time Points: Prechemotherapy Bevacizumab (BEV), Preoperatively, Postoperatively, and at the End of Systemic Treatment 0 Start of systemic tx pre op post op Timing No Bev-Mean CEA No Bev-Median CEA Bev-Mean CEA Bev-Median CEA end of systemic tx Clinical Colorectal Cancer March

8 Bev and Liver Resection in Colorectal Cancer opposed to the earlier years. Finally, we acknowledge that our statement on prolonged survival in patients with CR, although consistent with the published literature, can only be reported as a descriptive observation in the context of a small retrospective series. Conclusion This study contributes evidence regarding the safety of bevacizumab administration in combination with chemotherapy as perioperative therapy in patients with metastatic CRC undergoing liver resection. Patients achieving pathologic CR enjoy prolonged survival. Clinical Practice Points This study addresses the issue of safety of liver resection after treatment with bevacizumab in CRC and the overall outcome of patients undergoing perioperative chemotherapy and liver resection. Further to evidence provided by a small number of other retrospective studies, the addition of bevacizumab does not increase the rate or severity of postoperative complications related to liver resection. The only factor predicting postoperative complications is the PS at the start of preoperative chemotherapy, and severe complications result in an increased inpatient stay after liver resection. An interval of 8 weeks between bevacizumab and liver resection appears sufficient to allow a safe surgical procedure and does not have an adverse effect on the clinical outcome. With regard to patient outcome, there is no difference in response rate between oxaliplatin-based and irinotecan based regimens. There is no difference in OS between responders (PR) and nonresponders (SD PD). OS in patients with pathologic CR is longer compared with those without (P.0275). The only predictive factor (of PFS) from the univariate analysis was the number of liver lesions ( 4). This study contributes evidence about the safety of perioperative chemotherapy with bevacizumab in patients undergoing liver resection and provides information on the timing of surgery. There appears to be no difference in OS between responders and nonresponders, but pathologic CR is associated with prolonged OS. Acknowledgments We are grateful to clinical nurse specialists Toni Cole and Ramani Sitamvaram and research nurses Sarah Kerr, Diego Bottero, and Janine Webb for their invaluable contribution to the management of the patients included in this study. We acknowledge National Health Service funding to the National Institute for Health Research Biomedical Research Centre (London, UK). Disclosure DC has conducted advisory boards for Roche and has received research funding from Roche. IC has conducted advisory boards and has received research funding and honoraria from Roche. All other authors state that they have no conflicts of interest. References 1. Faivre J, Manfredi S, Bouvier AM. Epidemiology of colorectal cancer liver metastases [in French]. Bull Acad Natl Med 2003; 187:815-22; discussion Geoghegan JG, Scheele J. Treatment of colorectal liver metastases. Br J Surg 1999; 86: Choti MA, Sitzmann JV, Tiburi MF, et al. Trends in long-term survival following liver resection for hepatic colorectal metastases. Ann Surg 2002; 235: Fong Y, Cohen AM, Fortner JG, et al. Liver resection for colorectal metastases. J Clin Oncol 1997; 15: Rees M, Tekkis PP, Welsh FK, et al. Evaluation of long-term survival after hepatic resection for metastatic colorectal cancer: a multifactorial model of 929 patients. Ann Surg 2008; 247: Van Cutsem E, Nordlinger B, Adam R, et al. Towards a pan-european consensus on the treatment of patients with colorectal liver metastases. Eur J Cancer 2006; 42: Elias D, Sideris L, Pocard M, et al. Results of R0 resection for colorectal liver metastases associated with extrahepatic disease. Ann Surg Oncol 2004; 11: Adam R. and surgery: new perspectives on the treatment of unresectable liver metastases. Ann Oncol 2003; 14 Suppl 2:ii Vibert E, Canedo L, Adam R. Strategies to treat primary unresectable colorectal liver metastases. Semin Oncol 2005; 32(6 suppl 8): Nordlinger B, Van Cutsem E, Rougier P, et al. Does chemotherapy prior to liver resection increase the potential for cure in patients with metastatic colorectal cancer? A report from the European Colorectal Metastases Treatment Group. Eur J Cancer 2007; 43: Adam R, Delvart V, Pascal G, et al. Rescue surgery for unresectable colorectal liver metastases downstaged by chemotherapy: a model to predict long-term survival. Ann Surg 2004; 240:644-57; discussion Nordlinger B, Sorbye H, Glimelius B, et al. Perioperative chemotherapy with FOL- FOX4 and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC Intergroup trial 40983): a randomised controlled trial. Lancet 2008; 371: Giantonio BJ, Catalano PJ, Meropol NJ, et al. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200. J Clin Oncol 2007; 25: Saltz LB, Clarke S, Díaz-Rubio E, et al. Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line therapy in metastatic colorectal cancer: a randomized phase III study. J Clin Oncol 2008; 26: Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350: Gruenberger B, Tamandl D, Schueller J, et al. Bevacizumab, capecitabine, and oxaliplatin as neoadjuvant therapy for patients with potentially curable metastatic colorectal cancer. J Clin Oncol 2008; 26: Kesmodel SB, Ellis LM, Lin E, et al. Preoperative bevacizumab does not significantly increase postoperative complication rates in patients undergoing hepatic surgery for colorectal cancer liver metastases. J Clin Oncol 2008; 26: Reddy SK, Morse MA, Hurwitz HI, et al. Addition of bevacizumab to irinotecan- and oxaliplatin-based preoperative chemotherapy regimens does not increase morbidity after resection of colorectal liver metastases. J Am Coll Surg 2008; 206: D Angelica M, Kornprat P, Gonen M, et al. Lack of evidence for increased operative morbidity after hepatectomy with perioperative use of bevacizumab: a matched case-control study. Ann Surg Oncol 2007; 14: Tamandl D, Gruenberger B, Klinger M, et al. Liver resection remains a safe procedure after neoadjuvant chemotherapy including bevacizumab: a case-controlled study. Ann Surg 2010; 252: Mullen JT, Ribero D, Reddy SK, et al. Hepatic insufficiency and mortality in 1,059 noncirrhotic patients undergoing major hepatectomy. J Am Coll Surg 2007; 204: ; discussion 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; 240: DeOliveira ML, Winter JM, Schafer M, et al. Assessment of complications after pancreatic surgery: a novel grading system applied to 633 patients undergoing pancreaticoduodenectomy. Ann Surg 2006; 244:931-7; discussion Wong R, Cunningham D, Barbachano Y, et al. A multicentre study of capecitabine, oxaliplatin plus bevacizumab as perioperative treatment of patients with poor-risk colorectal liver-only metastases not selected for upfront resection. Ann Oncol 2011; 22: Blazer DG 3rd, 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; 26: Adam R, Wicherts DA, de Haas RJ, et al. Complete pathologic response after preoperative chemotherapy for colorectal liver metastases: myth or reality? J Clin Oncol 2008; 26: Clinical Colorectal Cancer March 2013