Results are shown for the tumors in patients who did and did not undergo chemotherapy after portal vein embolization (PVE) (P < .001). Boxes represent upper and lower quartiles of data; horizontal line in the box, median value; limit lines, maximum and minimum values excluding outliers; and solid circles, outliers (more or less than 3/2 times the upper or lower quartiles).
Results are shown for patients after liver resection analyzed according to whether they received or did not receive chemotherapy after portal vein embolization (PVE) (P < .006).
Results are shown for patients’ survival according to whether they received or did not receive chemotherapy after PVE(P < .007).
Fischer C, Melstrom LG, Arnaoutakis D, Jarnagin W, Brown K, D’Angelica M, Covey A, DeMatteo R, Allen P, Kingham TP, Tuorto S, Kemeny N, Fong Y. Chemotherapy After Portal Vein Embolization to Protect Against Tumor Growth During Liver Hypertrophy Before Hepatectomy. JAMA Surg. 2013;148(12):1103–1108. doi:10.1001/jamasurg.2013.2126
Portal vein embolization improves the safety of liver resection by increasing the size of residual liver, but the embolization may increase tumor growth during the waiting period before definitive hepatectomy.
To determine whether the administration of chemotherapy mitigates tumor growth after portal vein embolization (PVE) performed before major hepatectomy for metastatic colorectal cancer.
Design, Setting, and Participants
Review of prospectively collected data at Memorial-Sloan Kettering Cancer Center was conducted. The database included patients subjected to PVE before major hepatectomy for metastatic colorectal cancer.
Main Outcomes and Measures
Lesions in both the embolized and nonembolized lobes of the liver before and 1 month after PVE were measured and Response Evaluation Criteria in Solid Tumors were applied to assess disease status. Assessment of survival was based on receipt of post-PVE chemotherapy and then stratified by subsequent resectability.
Two hundred eight tumors were measured in 64 patients; 53 tumors were in patients undergoing post-PVE chemotherapy. Approximately one-third of the lesions progressed after PVE when no chemotherapy was administered. This did not differ significantly according to whether tumors were ipsilateral or contralateral to the PVE. When chemotherapy was administered, there was a significantly lower rate of progression (18.9%, P = .03). In long-term follow-up, treatment with post-PVE chemotherapy was also independently associated with improved survival (P < .006).
Conclusions and Relevance
Chemotherapy does not retard growth of the liver after PVE and may prevent cancer progression. Thus, the combination of PVE and chemotherapy may enhance both oncologic and operative safety.
Portal vein embolization (PVE) has become a standard technique for increasing residual liver mass before major hepatectomy1- 3 and has been shown to improve clinical outcome.2 This technique was first proposed by Makuuchi et al4 for improving the outcome of patients needing major hepatectomy for hilar cholangiocarcinoma. Portal vein embolization has proven to be a very safe procedure5,6 with few major periprocedural complications. One concern raised regarding PVE is whether local factors stimulating growth of liver parenchyma after PVE will also stimulate growth of local macroscopic and microscopic tumors (Table 1).7- 13
Metastatic colorectal cancer is the tumor most commonly resected by hepatectomy in Western populations. In this disease, great advancements in chemotherapy have been made recently, including the development and approval for use of irinotecan,14 oxaliplatin,15 cetuximab,16 and bevacizumab.17 These agents are often used to produce reduction of tumor bulk in initially unresectable cases to switch patients’ status to candidates for potentially curative resections.18 Many other patients also receive these agents as neoadjuvant therapy before liver resection. We2 previously demonstrated that PVE can be performed while the patient is continuing neoadjuvant or downstaging chemotherapy and results in growth of the liver to enhance subsequent hepatectomy. In the present study, we assessed tumor growth in the liver during the first month after PVE to determine whether concurrent neoadjuvant chemotherapy will influence the pace of tumor progression after PVE.
We examined 64 consecutive PVEs performed in patients with metastatic colorectal cancer between September 1999 and September 2004. An extended liver resection (≥4 segments of liver) was anticipated in all of these patients. Those receiving chemotherapy after PVE (n = 25) were compared with patients who did not receive chemotherapy after PVE (n = 50).
The techniques of PVE at our institution have been described.19 Briefly, an ipsilateral puncture of the portal vein is performed to decrease the likelihood of iatrogenic damage to the liver intended to be the remnant after resection. Embolization was performed using polyvinyl alcohol particles (200-300 µm).6 Metallic coils were avoided to prevent interference with a concurrent protocol using magnetic resonance spectroscopy for assessment of liver function.
Volumetric analysis of liver hypertrophy and atrophy was performed by computed tomography (CT) volumetrics comparing pre-embolization scans and scans performed 1 month later. All CT scanning was performed with a helical scanner, and the CT data were transferred to an independent workstation for assessment.
All tumors visible by CT were assessed. Tumors were measured along the longest axis and compared before and 1 month after PVE. Volume of the right and left sides of the liver was determined by semiautomated contouring of the scans. This was performed on serial transverse scans. On each section, the total liver was outlined and the sum of the sections was calculated by integrated software techniques using density threshold. This was repeated solely for the right side of the liver according to the plane of the middle hepatic vein. The difference between total liver volume and the volume of the right side of the liver was considered the volume of the left side. Results were recorded as percentage of volume change of the right and left sides of the liver.
Clinical data, including complications and long-term survival, were collected and evaluated. This review of data was approved by the Memorial Sloan-Kettering Cancer Center institutional review board. All patients were reviewed and classified using a preoperative Clinical Risk Score (CRS) prognostic scoring system defined for colorectal metastasis.20 This 5-point preoperative clinical scale evaluates the condition of patients by (1) disease-free interval of less than 12 months, (2) carcinoembryonic antigen greater than 200 ng/dL, (3) lymph node positive primary, (4) more than 1 hepatic lesion, and (5) hepatic lesion larger than 5 cm.
Comparisons were performed by an unpaired t test and by χ2 test as indicated. Nonparametric factors were compared by the Mann-Whitney test.21
Survival probabilities were estimated using the Kaplan-Meier method.22 Univariate associations between potential risk factors and survival were assessed using the log-rank test. Independent predictors of survival were determined using a proportional hazards regression model.23
The demographics of the study patients are listed in Table 2 and Table 3. There were 38 men and 26 women. The age of the patients was typical for this disease, with a median of 55 years. Most embolizations were right PVE (n = 55) in preparation for right-sided resections. Twenty-five of the patients received chemotherapy in the month after PVE. The regimens of chemotherapy used were varied and included systemic chemotherapy in only 17 cases (68%) and regional plus systemic chemotherapy in 8 cases (32%). The demographics of the patients (age, sex, or CRS) were similar in patients not receiving chemotherapy and those receiving chemotherapy (Table 3).
In these 64 patients, 208 lesions were assessed. Forty-seven of the lesions were contralateral to the side of PVE, and 161 lesions were ipsilateral to the PVE. For patients who did not receive chemotherapy, 34.2% of lesions progressed by Response Evaluation Criteria in Solid Tumors criteria24 in the month after PVE (Table 4) and did not differ significantly whether the lesion was ipsilateral or contralateral to the PVE.
The changes in tumor size in centimeters are shown in Figure 1. Tumors in patients undergoing chemotherapy after PVE had suppressed growth compared with lesions in patients not subjected to chemotherapy (P < .001). For patients not receiving chemotherapy, the tumors had a mean (SD) growth of 8% (3%) in size, and the lesions in patients receiving chemotherapy had a −13% (8%) mean regression in size (P < .001).
Overall, patients subjected to PVE experienced minimal adverse effects; 42% of the patients (n = 27) developed a temperature above 38°C (41% in the no-chemotherapy group and 44% in the chemotherapy group). The fever usually occurred on day 1 after embolization and was self-limiting. Twenty-seven percent (n = 17) of the patients had leukocytosis (white blood cells >11 000/μL [to convert to cells ×109/L, multiply by 0.001]; no chemotherapy, 26%; and chemotherapy, 28%). Neither of these effects was associated with infection. No other complications were noted. Within 30 days, the mean growth of the side of the liver contralateral to the side embolized was 22% (4%) in the no-chemotherapy group and 24% (0.5%) in the chemotherapy group (P > .99).
Thirty-six patients (56%) in this series underwent resection (no chemotherapy, 53% ; chemotherapy, 69% ). The reason for unresectable disease was identification of occult additional disease in all cases. Operative procedures for the 2 groups were similar (no chemotherapy, 8 trisectionectomy plus contralateral minor procedures and 12 lobectomies plus contralateral minor procedures; chemotherapy, 7 trisectionectomies plus and 9 lobectomies plus contralateral minor procedures). The complication rate overall was 24% (no chemotherapy, 30%; chemotherapy, 19%). The complications experienced by the no-chemotherapy group included 1 deep vein thrombosis, 1 intra-abdominal abscess, 3 wound infections, and 1 chemotherapy pump misperfusion. The complications experienced by the chemotherapy group included 1 wound infection and 2 pump misperfusions.
Thirty of the 64 patients died. The median survival of those who died was 15 months, and the median follow-up of survivors was 51 months. The long-term survival of the patients is shown in Figure 2. Even though the 2 groups were equivalent according to age, sex, CRS, and resectability (Table 1), patients subjected to chemotherapy after PVE had a significantly better survival (median: no chemotherapy, 23.7 months vs chemotherapy, 50.1 months; 5-year survival: no chemotherapy, 24% vs chemotherapy, 49%) (P = .006). The survival for patients separated by resectability is shown in Figure 3.
To determine whether chemotherapy after PVE may be an independent predictor of outcome, we used multivariate analysis to examine the influence of chemotherapy and high CRS (Table 5). Use of chemotherapy remained highly correlated with improved outcome.
Portal vein embolization was first proposed 20 years ago as an effective means of increasing the functional residual liver by inducing liver hypertrophy in the unembolized lobe.4 Originally, this procedure was proposed for use in hilar cholangiocarcinoma, in which a very large amount of liver is usually removed for a very small tumor. With the safety of PVE becoming well documented, this procedure is increasingly used for patients with metastatic colorectal cancer because of the increasing extent of resections surgeons are willing to undertake to eradicate disease. It is also becoming important in treatment of metastatic colorectal cancer because the various preoperative chemotherapeutic regimens often produce liver parenchymal damage and hinder posthepatectomy recovery.25 There is no doubt that PVE has improved our ability to safely resect colorectal cancer that has metastasized to the liver. Two studies found that patients who had PVE prior to resection had survival rates equivalent to those in patients who did not have PVE despite the fact that patients who underwent PVE had more severe disease.5,26 The present study examined a series of patients with advanced hepatic metastases (median CRS, 3) and demonstrated the role that PVE may play in complex multimodality treatment strategies.
Recent data have been reassuring for the immediate periprocedural safety of PVE. Although there had been concerns regarding potential detrimental effects of PVE on liver function, clinical data clearly have shown PVE not to have major negative effects on hepatic function.27 Our study2 on the safety of these procedures showed PVE to be associated with only mild postembolization syndrome (fever, thrombocytopenia, and local pain). However, there remains concern regarding the potential for PVE to stimulate tumor growth. It has been hypothesized that by increasing portal blood flow to the unembolized lobe and inducing local growth factors, tumors in this lobe would enlarge. Many small studies3,7- 12 have shown conflicting results, with some studies demonstrating tumor growth by size7 or DNA measurements.8 In the present study, we demonstrated that, after PVE, one-third of the tumors grow sufficiently in the first month to be considered progression. We also showed that simultaneous administration of chemotherapy retarded such tumor progression and was associated with improved long-term survival.
Even without these new data, the use of chemotherapy with PVE has practical value. Many patients arrive for evaluation by the hepatic surgeon while receiving neoadjuvant chemotherapy. Furthermore, of patients with unresectable hepatic colorectal metastases, approximately 15% are now routinely converted to resectable cases by chemotherapy. If stopping chemotherapy for 3 to 6 weeks is necessary for PVE and waiting 4 to 6 weeks afterward is necessary to allow adequate hypertrophy, then patients with metastatic disease will need to be without cancer treatment for at least 7 to 12 weeks prior to hepatectomy. That is why Covey et al2 and others28 had previously examined the possibility of performing PVE while the patient is receiving chemotherapy. Initially, it was thought that neoadjuvant chemotherapy would impede liver hypertrophy. Thus, it was stopped several weeks before embolization, allowing for a period of time when potential for tumor growth was uninhibited. Beal et al29 conducted a small retrospective study looking at 15 patients and showed that hypertrophy does occur in those who received PVE and chemotherapy. However, this occurred to a lesser extent than in patients who underwent only PVE. In a slightly larger study, Goéré et al28 demonstrated no significant difference in hypertrophy response when chemotherapy was continued after PVE. Interestingly, they also showed no increase in postoperative complications among patients who continued chemotherapy. Most recently, Covey et al,2 in one of the largest studies to date with a series of 100 patients, concluded that chemotherapy had no negative effects on hypertrophy. They further showed that patients who underwent PVE before resection had improved outcomes measured by reduced need for transfusion and reduced hospital stay after resection. From these and similar studies, it is evident that chemotherapy after PVE does not have a negative effect on liver hypertrophy. The present study shows that not only are the combination of PVE and chemotherapy safe and effective from a liver hypertrophy standpoint, it is associated with retarded tumor growth and improved long-term survival.
Although recent advances in chemotherapy have been a major improvement in treatment for metastatic colorectal cancer, chemotherapy performed before hepatectomy is also producing new challenges for the liver surgeon. With this increasing experience with chemotherapy, a clinically apparent syndrome of liver injury has been noted—chemotherapy-associated steatohepatitis.30 This is not only a pathologic entity but one associated with adverse clinical outcomes.25 Portal vein embolization was first introduced for cases in which very small liver remnants were anticipated31,32 and for patients with liver parenchymal injury from hepatitis33; it is now increasingly used in resections involving parenchymal damage from chemotherapy.2 With the knowledge that the liver can still hypertrophy in a toxic environment, we can safely continue a modality that will shrink tumor burden while also using a strategy to produce more functional residual liver. This dual strategy of neoadjuvant chemotherapy and PVE allows physicians to make use of different approaches simultaneously, providing maximal benefit to our patients. If chemotherapy does not have to be halted after PVE, it eliminates the concern of tumor progression during the standard 3- to 4-week time until resection; moreover, as we have shown, chemotherapy can slow progression of disease.
Accepted for Publication: March 4, 2013.
Corresponding Author: Yuman Fong, MD, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10065 (FongY@mskcc.org).
Published Online: October 30, 2013. doi:10.1001/jamasurg.2013.2126.
Author Contributions: Drs Fischer and Fong had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Arnaoutakis, Allen, Kingham, Tuorto, Kemeny, Fong.
Acquisition of data: Fischer, Arnaoutakis, Jarnagin, Brown, Covey, Tuorto, Fong.
Analysis and interpretation of data: Fischer, Melstrom, D’Angelica, DeMatteo, Kingham, Tuorto, Fong.
Drafting of the manuscript: Fischer, Arnaoutakis, D’Angelica, Kingham, Tuorto, Kemeny, Fong.
Critical revision of the manuscript for important intellectual content: Fischer, Melstrom, Jarnagin, Brown, Covey, DeMatteo, Allen, Kingham, Tuorto, Fong.
Statistical analysis: Fischer, Tuorto, Fong.
Obtained funding: Fong.
Administrative, technical, and material support: Fischer, Arnaoutakis, Jarnagin, D’Angelica, Tuorto, Fong.
Study supervision: Allen, Kingham, Kemeny, Fong.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by US Public Health Service grants RO1CA75416, RO1CA72632, and RO1CA61524 (Dr Fong) from the National Institutes of Health.
Role of the Sponsor: The National Institutes of Health had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.