SEARCH

SEARCH BY CITATION

Keywords:

  • colorectal cancer;
  • liver metastases;
  • chemotherapy;
  • vascular invasion;
  • vascular attachment

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

BACKGROUND:

Although tumor reduction via present-day prehepatectomy chemotherapy can render initially unresectable disease potentially resectable, little is known about the effects of such chemotherapy on liver metastases with known attachment to or invasion of major intrahepatic vessels. We histologically assessed the relationships of liver tumors to major intrahepatic vessels after chemotherapy.

METHODS:

In 45 patients who underwent chemotherapy and hepatectomy with pretreatment images showing metastases attached to or invading major intrahepatic vessels, 77 metastases showed attachment to or invasion of 96 vessels.

RESULTS:

Using postchemotherapy imaging, 11 of 77 metastases (14.3%) appeared separated from 12 of 96 major hepatic vessels (12.5%). Among 83 vessels later examined pathologically, 29 showed direct invasion (35%) and 10 showed attachment (12%). Tumors involved another 9 vessels (11%) that were separated surgically from the tumor and preserved during hepatectomy. Tumor attachment that exceeded 25% of vessel circumferences via imaging after chemotherapy was a factor associated with pathological vascular invasion or attachment according to multivariate analysis (relative risk, 8.449; 95% confidence interval, 1.961-36.415; P = .0042).

CONCLUSIONS:

Liver metastasis attachment to or invasion of major intrahepatic vessels is difficult to eradicate even with otherwise effective chemotherapy. Cancer 2012;. © 2011 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Recent advances in chemotherapy and surgical improvements including portal vein embolization, staged hepatectomy, and refined vascular resection and reconstruction techniques allow expansion of indications for surgery to treat multiple, massive, unfavorably located tumors invading major vascular structures. Increasingly, preoperative chemotherapy is administered to patients with initially unresectable disease, sometimes reducing tumor bulk to the extent that a patient with initially unresectable disease may become a candidate for potentially curative resection. At present, resectability has become an important endpoint for chemotherapy, focusing on curative potential of treatment as opposed to classic endpoints such as response or progression-free survival.

Liver resection with a microscopically negative margin has been considered crucial for long-term survival. Therefore, when liver tumors appear to invade major hepatic vessels in preoperative images, combined resection of the vessels and areas they supply or drain is performed; alternatively, the vessels are resected and reconstructed. If most of the tumor periphery is eradicated by chemotherapy to result in a smaller tumor, such combined resection of vessels sometimes might be unnecessary. However, according to a previous study of tumors not treated with chemotherapy,1 viable cells predominate in the periphery of the metastases while necrosis predominates in the center. Even after a response to chemotherapy, a concentric pattern of metastasis shrinkage from central fibrosis occurs, showing little change in distribution of viable cells from prechemotherapy.1, 2 Ng et al1 reported viable tumor cells to be more frequent in the periphery of metastases, whether treated with chemotherapy or not. Further, in metastases treated with chemotherapy, the proliferative index, an estimate of the proportion of viable tumor cells staining for Ki67, showed more proliferating tumor cells at the periphery of the tumor than centrally.

Our clinical experience suggested that liver metastases attached to or invading major intrahepatic vessels shrank with chemotherapy in a generally concentric fashion while maintaining vascular attachment or invasion even though images show an overall response to chemotherapy (Figure 1). We therefore hypothesized that liver metastases involving major intrahepatic vessels tend to maintain this involvement, despite effective chemotherapy shown to shrink the tumors. Liver metastasis attachment to and invasion of main hepatic vessels are important findings that directly influence decision-making concerning liver resection procedures. If liver metastases remain attached to or still invade major intrahepatic vessels after chemotherapy, ability to “downstage” patients to stages permitting resection may be limited.

thumbnail image

Figure 1. Abdominal computed tomography (CT) scans during prehepatectomy chemotherapy are shown. (A) CT scan shows a liver tumor before initiation of chemotherapy. The tumor is attached to the right and middle hepatic veins (arrows) and to the hepatic vena cava. (B) CT scan shows the tumor 3 months after chemotherapy initiation. (C) CT scan shows the tumor 5 months after initiation of chemotherapy. The tumor remained attached to each involved intrahepatic vessel throughout chemotherapy.

Download figure to PowerPoint

The objectives of this study were to use imaging to assess relationships of liver tumors to major intrahepatic vessels before and after chemotherapy, and to then assess vascular invasion and attachment microscopically.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Patients

From 2004 to July 2010, the Department of Gastroenterological Surgery at the Yokohama City University Graduate School of Medicine treated 165 patients in whom colorectal liver metastases were diagnosed at liver resection with curative intent. Among these 165 patients, 62 (38%) underwent preoperative chemotherapy followed by hepatectomy. Of these 62 patients, 45 had metastases attached to or invading major intrahepatic vessels according to preoperative imaging. Patient characteristics are shown in Table 1; the median follow-up duration for these 45 patients was 22 months (range, 1-75 months). This study protocol was approved by the Institutional Review Board of Yokohama City University, and written informed consent was obtained from all patients.

Table 1. Patient Profile
VariablesPatients (n = 45)
  • Data are presented as no. (%) unless noted otherwise.

  • a

    Intra-arterial chemotherapy via the hepatic artery.

Age, y, median (range)66 (37-77)
Sex 
 Men28 (62)
 Women17 (38)
Primary-related 
 Site 
  Colon25 (56)
  Rectum20 (44)
 Histology 
  Well differentiated13 (29)
  Moderately differentiated29 (64)
  Other3 (7)
 Dukes stage 
  B8 (18)
  C37 (82)
Liver-related 
 Timing 
  Synchronous30 (67)
  Metachronous13 (29)
  Recurrence after hepatectomy2 (4)
 Distribution 
  Unilobar8 (18)
  Bilobar37 (82)
 Number, median (range) 
  Prechemotherapy6 (1-23)
  Postchemotherapy5 (1-18)
 Maximum size at resection, mm, median (range)35 (10-140)
 Carcinoembryonic antigen, ng/mL, median (range) 
  Prechemotherapy52 (4-40,610)
  Postchemotherapy11 (2-10,536)
Extrahepatic metastases 
 Present13 (29)
Prehepatectomy chemotherapy 
 Arteriala16 (36)
 Systemic10 (22)
 Arteriala + systemic19 (42)
 Number of cycles, median (range)6 (2-34)
 Number of lines used for administration 
  137 (82)
  23 (7)
  ≥35 (11)

At diagnosis, disease was unresectable in 20 (44%) patients. Chemotherapy was indicated for the other 25 patients because their metastases were defined as marginally resectable. Unresectability was established through multidisciplinary assessment by a team including surgeons and medical oncologists, usually based on insufficient remnant liver (<25%-30%) or excessive risk of surgery considering tumor location, liver function, patient age, and resected volume as defined using a prediction score system introduced by Yamanaka et al.3 Patients with marginally resectable metastases were those with ≥4 metastases, including lesions in both major lobes; tumor ≥80 mm in diameter; or unfavorable tumor location with invasion of major vascular structures.

Preoperative Staging and Imaging Investigation

Preoperative staging included physical examination, measurement of serum carcinoembryonic antigen, and carbohydrate antigen 19-9, colonoscopy, barium enema, abdominal imaging by ultrasonography and computed tomography (CT), and chest imaging via routine radiography or computed tomography (CT). Positron-emission tomography sometimes also was used for preoperative staging. Helical CT with arterioportography usually was performed to define hepatic metastases before introduction of chemotherapy and before hepatectomy. CT was performed with Asteion scanners (Toshiba Medical, Tokyo, Japan). Attachment of liver tumors to major hepatic vessels was evaluated by enhanced helical CT of the liver with 3 image sets (arterial, portal venous, and hepatic venous phases) performed in succession.

Major hepatic vessels were defined as related to the hepatic venous confluence where the right, middle, or left hepatic veins entered the inferior vena cava (IVC) or related to common, right, left, right anterior, or right posterior branches of Glisson's pedicle (or portal vein). When normal liver parenchyma or fat could not be seen between the tumor and major hepatic veins or major branches of Glisson's pedicle (portal branches) on CT, the vessel was defined as attached to or invaded by tumor.

Extent of attachment between tumor and vessels, defined as the maximum percentage of a vessel's circumference in contact with the tumor, also was determined before and after chemotherapy. This extent was graded in steps of 25%: >75%; >50%-75%; >25%-50%; >0%-25%; and 0%.

Prehepatectomy Chemotherapy

Sixteen patients received final prehepatectomy chemotherapy with a combination of 5-fluorouracil (5-FU), l-folinic acid (FA), and cisplatin (CDDP); in 13 of the 16 patients, a chronomodulated regimen was used. Treatment consisted of a 5-day course of infusion via the hepatic artery through an implanted arterial access port (Vital-Port, Cook Vascular, Leechburg, PA). On each of 5 days, 5-FU (500-600 mg/m2/d), FA (100 mg/m2/d), and CDDP (10 mg/m2/d) were delivered. In principle, this 5-day course was repeated 3 or more times at 9-day intervals. In patients who received chronomodulated infusions, the same drugs were administered using a Graseby Model 3000 infusion pump (Graseby Medical, Watford, UK). Peak delivery was scheduled for 4 AM for 5-FU and FA and 4 PM for CDDP, essentially as described in previous reports.4,5 Ten patients received systemic chemotherapy consisting of 5-FU and FA alone (n = 1) or combined with oxaliplatin (n = 6), irinotecan (n = 2) or both oxaliplatin and irinotecan (n = 1). Bevacizumab was added in 6 patients with 5-FU, FA, and oxaliplatin, and 1 with 5-FU, FA, oxaliplatin, and irinotecan. Cetuximab was added in 1 patient with 5-FU, FA, and irinotecan. Nineteen patients received both 3 cycles of chronomodulated hepatic arterial infusion chemotherapy as outlined above and 3 cycles of systemic infusion of 5-FU and FA with oxaliplatin (FOLFOX-4). CT of the thorax and pelvis and triple-phase helical abdominal CT were performed after every 3 courses of treatment. The response to chemotherapy was evaluated after every 3 cycles of treatment using CT of the liver, according to the Response Evaluation Criteria in Solid Tumors (RECIST) criteria.6

Hepatectomy Procedure

Hepatectomy was not necessarily performed according to anatomical principles of resection; the guiding principle was assurance of tumor-free margins. Intraoperative ultrasonography was used to identify any occult tumors not detected preoperatively, and to confirm relationships between tumors and vasculobiliary structures. Parenchymal dissection was performed using ultrasonic dissectors. When attachment or invasion of a liver tumor to the major hepatic vessels was suspected, either en bloc resection of the area supplied or drained by the resected vessels or combined resection/reconstruction of vessels was performed. When vessel reconstruction was expected to be difficult or resection volume was expected to be excessive, the liver tumor was separated surgically from the major vessels unless the tumor invaded the vessels according to intraoperative findings.

Pathological Examination

Resected specimens were embedded in paraffin, sectioned at a thickness of 3 μm, stained with hematoxylin and eosin, and examined microscopically. Presence or absence of tumor cells in the plane of liver resection was examined when a liver tumor was separated from a major vessel defined preoperatively as attached to a tumor. When a major vessel was resected simultaneously, presence or absence of tumor invasion of the vessel was determined (Figure 2). When judgment of vascular invasion by a tumor was difficult, examination using elastica van Gieson staining was added.

thumbnail image

Figure 2. Microscopic images show relationships between liver metastases and major intrahepatic vessels (hematoxylin and eosin stain; original magnification, ×40). (A) Liver metastasis invading Glisson's pedicle is shown. (B) Liver metastasis attached to Glisson's pedicle is shown. (C) Liver metastasis separated from Glisson's pedicle is shown. (D) Liver metastasis invading a hepatic vein is shown. (E) Liver metastasis attached to a hepatic vein is shown. (F) Liver metastasis separated from a hepatic vein is shown.

Download figure to PowerPoint

Patient Follow-up

Preoperative chemotherapy was continued postoperatively as adjuvant therapy.

Patients underwent follow-up evaluation monthly at our outpatient clinic. No patients were lost to follow-up. Serum carcinoembryonic antigen and carbohydrate antigen 19-9 were measured every month, and CT was performed every 3 months for 5 years after the most recent operation.

Statistical Analysis

Statistical comparisons of baseline data were performed using the Mann-Whitney U test, chi-square test, or Fisher exact test. Independent predictors of tumor invasion or attachment were identified via multivariate analysis using a multiple logistic regression analysis. Univariate and multivariate analysis for continuous variables was performed using a receiver-operating characteristic curve for analysis. Sensitivity versus specificity in predicting tumor invasion or attachment was plotted according to various cutoff points. The most significant values for sensitivity and specificity were chosen as cutoff values for continuous variables. Survival rates were calculated using the Kaplan-Meier method. A difference was considered significant when the 2-sided P value was <.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Imaging before treatment showed 303 liver metastases in the 45 patients. Among these metastases, 77 showing attachment to or invasion of major intrahepatic vessels were analyzed. Attachment or invasion involved 1 intrahepatic vessel in 61 metastases, 2 vessels in 14 metastases, 3 vessels in 1 metastases, and 4 vessels in 1 metastasis. Thus, a total of 96 major intrahepatic vessels showed attachment or invasion by metastases. Specific major vessels involved were the hepatic IVC in 9 metastases, hepatic vein in 57 metastases (left, 8; middle, 25; right, 24), and portal vein or Glisson's pedicle in 30 metastases (main trunk, 2; left branch, 11; right branch, 5; anterior branch, 8; posterior branch, 4).

Overall survival rates at 1, 3, and 5 years after hepatectomy for the 45 patients were 91.8%, 71.4%, and 48.3%, respectively; disease-free rates at these time points were 39.6%, 20.0%, and 10.0%.

Imaging Status of Metastases After Chemotherapy

Chemotherapy resulted in partial response (PR) in 27 (60%) patients, stable disease (SD) in 11 (24%) patients, and progressive disease (PD) in 7 (16%) patients. Among the 77 metastases involving major hepatic vessels, response to chemotherapy was complete response (CR) in 1 nodule, PR in 48 nodules, SD in 17 nodules, and PD in 11 nodules. Among involved intrahepatic vessels, response to chemotherapy was CR in 1 vessel, PR in 61 vessels, SD in 21 vessels, and PD in 13 vessels.

After chemotherapy, 11 of 77 (14.3%) metastases no longer appeared attached to 12 major hepatic vessels (12/96 [12.5%]) according to preoperative imaging. Among these, 1 metastasis showed complete disappearance from images.

Pathological Examination

Sixty-six major intrahepatic vessels were resected simultaneously with the metastatic nodules attached to or invading the vessels. The 30 remaining vessels were separated from the metastatic nodules at hepatectomy and left in the remnant liver. In 12 simultaneously resected vessels, pathological evaluation was not possible because sections precisely corresponding to areas where imaging showed attachment or invasion by tumor could not be obtained. The main reason was that those metastatic tumors involved multiple vessels, requiring differing planes of sectioning, obscuring precise relationships between the tumor and the vessel. One of 30 portions of liver metastases separated surgically from intrahepatic vessels could not be examined microscopically for the same reason.

Among 54 simultaneously resected vessels, 29 were invaded directly by liver metastatic tumors. Ten vessels were not invaded, but were attached directly with no liver parenchyma between tumor and vessels. No tumor invasion or attachment was observed in the other 15 vessels. Among 29 portions of resected metastases separated from attached hepatic vessels, tumors were evident at the resection plane at the point of separation from the vessels in 9; the other 20 portions of metastases showed no tumor cells at the point of separation (Table 2). Overall, 29 of 83 (35%) vessels were invaded directly by liver metastases, and 19 vessels (23%) were attached to tumors even after chemotherapy.

Table 2. Histological Relationship Between Liver Metastases and Vessels According to Imaging Diagnosis
Microscopic FindingsAll Vessels (n = 96)Imaging Diagnosis After ChemotherapyChemotherapy
Separated from Vessels (n = 12)Attached to Vessels (n = 84)PR (n = 62)SD (n = 21)PD (n = 13)
  1. PD indicates progressive disease; PR, partial response; SD, stable disease.

Liver resection including vessels      
 Tumor invasion of vessels292271766
 Tumor attachment to vessels10010730
 Tumor separation from vessels154111041
 Not evaluable12012444
Liver resection without vessels      
 Tumor exposure at the plane of resection918900
 No tumor exposure at the resection margin205151532
 Not evaluable101010

When vessels were divided into 2 groups according to preoperative findings, invasion/attachment was still seen in 3 of the 12 (25%) vessels that showed separation from tumors in images after chemotherapy, and was seen in 45 of 71 (63%) vessels that were positive for invasion or attachment by tumors in images after chemotherapy (Table 2).

Cancer invasion was seen in 17 of 58 (29%) vessels adjoining tumors responding to chemotherapy, and cancer attachment or presence at the point of surgical separation was seen in 16 (28%). Invasion or attachment was seen in 9 of 16 (56%) vessels adjoining SD tumors and in 6 of 9 (67%) vessels adjoining PD tumors (Table 2).

Predictive Factors for Microscopic Vascular Invasion or Attachment by Tumors

After excluding 13 vessels that could not be assessed histologically, predictive factors for microscopic tumor invasion or attachment to major hepatic vessels were analyzed. Univariate analysis identified changes in tumor size related to chemotherapy (calculated using the following formula: tumor diameter after chemotherapy / tumor diameter before chemotherapy × 100% [P = 0.038]); response to chemotherapy by tumors showing vascular invasion or attachment (P = .011); extent-of-attachment grade for vessels before (P = .001) and after (P<.001) chemotherapy; and changes in attachment grade related to chemotherapy (P = .041) (Table 3). Multivariate analysis using the factors identified via univariate analysis selected a single factor independently associated with tumor invasion or attachment to vessels: >25% of vessel circumference attached to tumor in images after chemotherapy (relative risk [RR], 8.449; 95% confidence interval [CI], 1.961-36.415; P = .0042) (Table 4). Considering factors predicting direct tumor invasion of vessels, changes of tumor size related to chemotherapy (P = .001), individual tumor response to chemotherapy (P = .025), and extent-of-attachment grade for vessels before (P = .002) and after (P<.001) chemotherapy were identified via univariate analysis (Table 5). Tumor size reduction after chemotherapy (at least 60%; RR, 23.099; 95% CI, 1.839-290.136; P = .015) and attachment to over 25% of the vessel circumference after chemotherapy (RR, 11.973; 95% CI, 2.083 to 68.826; P = .0054) were identified by multivariate analysis (Table 6).

Table 3. Predictive Factors for Microscopically Invaded or Attached Vessels by Tumors
VariablesInvaded or Attached Vessels (n = 48)Separated Vessels (n = 35)P
  • Data are presented as no. (%) unless noted otherwise.

  • a

    Intra-arterial chemotherapy via the hepatic artery.

  • b

    Extent of attachment of tumor and vessels was defined as the percentage of the maximum length of the vessels circumference in contact with the tumor out of length of the whole vessels circumference.

  • c

    Defined according to the changes of grading. When a tumor was shrank after chemotherapy, this was considered a downgrade.

Histology  .201
 Well differentiated12 (25)7 (20) 
 Moderately differentiated35 (73)24 (69) 
 Other1 (2)4 (11) 
Serum carcinoembryonic antigen concentration, ng/mL, median (range)   
 Prechemotherapy67.1 (4.4-40,609)51.8 (3.9-40 609).483
 Postchemotherapy12.1 (1.4-10,536)10.1 (1.4-10 536).754
Size of tumors, mm, median (range)   
 Prechemotherapy42.4 (14.8-147.3)41.8 (6-188.8).772
 Postchemotherapy27.0 (9.1-144.7)25.9 (0-82.0).218
Changes in tumor size with chemotherapy (size after/size before) ×100%, median (range)75.6 (33.0-127.7)55.2 (0-112.8).038
Vessel  .645
 Portal vein18 (38)11 (31) 
 Hepatic vein30 (63)24 (69) 
Prehepatectomy chemotherapy   
 Mode  .355
  Systemic13 (27)5 (14) 
  Arteriala16 (33)15 (43) 
  Arteriala + systemic19 (40)15 (43) 
 Number of cycles, median (range)6 (2-23)6 (2-34).803
 Number of lines used for administration  .745
  135 (73)28 (80) 
  25 (10)3 (9) 
  ≥38 (17)4 (11) 
Response of individual tumors to chemotherapy  .011
 Complete response03 (9) 
 Partial response22 (46)23 (66) 
 Stable disease21 (44)9 (26) 
 Progressive disease5 (10)0 
Extent of attachment between tumors and vessels (% of circumference)b   
 Prechemotherapy  .001
  >7513 (27)3 (9) 
  >50-7514 (29)5 (14) 
  >25-5018 (38)14 (40) 
  >0-253 (6)13 (37) 
Postchemotherapy  <.001
  >754 (8)0 
  >50-7521 (44)2 (6) 
  >25-5013 (27)4 (11) 
  >0-258 (17)22 (63) 
  02 (4)7 (20) 
Changes in attachment grade (before vs after chemotherapy)c  .041
 Upgrade8 (17)1 (3) 
 No change20 (42)11 (31) 
 Downgrade20 (42)23 (66) 
Table 4. Univariate and Multivariate Analyses of Predictive Factors for Microscopic Vascular Invasion or Attachment by Tumors
 Invaded or Attached Vessels (n = 48)Separated Vessels (n = 35)PRR (95% CI)
UnivariateMultivariate
  1. CI, confidence interval; CR, complete response; PD, progressive disease; PR, partial response; RR, relative risk; SD, stable disease.

Changes of tumor size before and after chemotherapy  .014  
 <60%1924   
 ≥60%2911   
Response to chemotherapy, tumor-by-tumor analysis  .013  
 SD, PD269   
 PR, CR2226   
Extent of attachment between tumors and vessels before chemotherapy  <.001  
 >25% of circumference4522   
 ≤25% of circumference313   
Extent of attachment between tumors and vessels after chemotherapy  <.001.00428.449 (1.961-36.415)
 >25% of circumference386   
 ≤25% of circumference1029   
Changes of attachment grade before and after chemotherapy     
 No change or upgrade2812045 .
 Downgrade2023   
Table 5. Predictive Factors for Microscopically Invaded Vessels by Tumors
VariablesInvaded Vessels (n = 24)Attached or Separated Vessels (n = 59)P
  • Data are presented as no. (%) unless noted otherwise.

  • a

    Intra-arterial chemotherapy via the hepatic artery.

  • b

    Extent of attachment of tumor and vessels was defined as the percentage of the maximum length of the vessels circumference in contact with the tumor out of length of the whole vessels circumference.

  • c

    Defined according to the changes of grading. When a tumor was shrank after chemotherapy, this was considered a downgrade.

Histology   
 Well differentiated6 (25)13 (22).338
 Moderately differentiated18 (75)41 (69) 
 Other0 (0)5 (8) 
Serum CEA concentrations, ng/mL, median (range)   
 Prechemotherapy77.3 (4.5-40 609)51.8 (3.9-40 609).250
 Postchemotherapy17.5 (1.4-10 536)11.0 (1.4-10 536).439
Size of tumors, mm, median (range)   
 Prechemotherapy40.1 (14.9-147.3)42.4 (6-188.8).865
 Postchemotherapy27.0 (9.1-144.7)25.9 (0-82.0).131
Changes in tumor size with chemotherapy, (size after/size before) ×100%, median (range)81.7 (45.1-120.9)54.9 (0-127.7).001
Vessel   
 Portal vein9 (38)20 (34).803
 Hepatic vein15 (63)39 (66) 
Prehepatectomy chemotherapy   
 Mode   
  Systemic7 (29)11 (19).572
  Arteriala8 (33)23 (39) 
  Arteriala + systemic9 (38)25 (42) 
 Number of cycles, median (range)6 (2-20)6 (2-34).957
 Number of lines used for administration   
  117 (71)46 (78).380
  24 (17)4 (7) 
  ≥33 (13)9 (15) 
Response of individual tumors to chemotherapy   
 Complete response03 (5).025
 Partial response8 (33)37 (63) 
 Stable disease13 (54)17 (29) 
 Progressive disease3 (13)2 (3) 
Extent of attachment between tumors and vessels (% of circumference)b   
 Prechemotherapy   
  >759 (38)7 (12).002
  >50-759 (38)10 (17) 
  >25-504 (17)28 (47) 
  >0-252 (8)14 (24) 
 Postchemotherapy   
  >753 (13)1 (2)<.001
  >50-7512 (50)11 (19) 
  >25-506 (25)11 (19) 
  >0-252 (8)28 (47) 
  01 (4)8 (14) 
Changes in attachment grade (before vs after chemotherapy)c   
 Upgrade4 (17)5 (8).388
 No change10 (42)21 (36) 
 Downgrade10 (42)33 (56) 
Table 6. Univariate and Multivariate Analyses of Predictive Factors for Microscopic Vascular Invasion by Tumors
 Invaded Vessels (n = 24)Attached or Separated Vessels (n = 59)PRR (95% CI)
UnivariateMultivariate
  1. CI, confidence interval; CR, complete response; PD, progressive disease; PR, partial response; RR, relative risk; SD, stable disease.

Changes of tumor size before and after chemotherapy  <.001.015023.099 (1.839-290.136)
 <60%538   
 ≥60%1921   
Response to chemotherapy, tumor-by-tumor analysis  .007  
 SD, PD1619   
 PR, CR840   
Extent of attachment between tumors and vessels before chemotherapy  .134  
 >25% of circumference2245   
 ≤25% of circumference214   
Extent of attachment between tumors and vessels after chemotherapy  <.001.005411.973 (2.083-68.826)
 >25% of circumference2123   
 ≤25% of circumference336   

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

Effective prehepatectomy chemotherapy can reduce tumor size sufficiently to convert unresectable liver metastases to resectable metastases. However, preoperative assessment of the presence or absence of tumor invasion or attachment involving major vessels is crucial for planning surgical strategy. When invasion or attachment of a liver tumor involving major hepatic vessels is suspected preoperatively, en bloc resection of affected vessels and areas perfused or drained should be considered; alternatively, resection and reconstruction of the vessels should be performed. The former strategy usually causes excessive liver parenchymal loss that can lead to clinical manifestations of decompensation including hepatic insufficiency. The latter is a less effective technique, with increased risk of occlusion of reconstructed vessels and of other morbidities. Thus, preoperative decisions concerning extent of liver resection based on imaging of major vessel invasion is likely to strongly influence both short- and long-term outcomes.

In this study, only 12% of major hepatic vessels initially attached to liver metastases showed separation by imaging after chemotherapy. Further, on pathological examination, approximately 60% of hepatic vessels were invaded by or attached to tumors even after chemotherapy. No previous reports have described frequency of pathological tumor invasion of major intrahepatic vessels appearing attached to tumors by preoperative imaging. More generally, in reports of patients suspected to have invasion of the hepatic IVC, cancer invasion of the IVC wall was confirmed histologically in approximately 45%-60% of patients who underwent concomitant liver and IVC resection.7, 8 The frequency of vessel invasion or attachment after chemotherapy in our study differed little from these previous results for patients without chemotherapy. Thus, difficulty of achieving separation of liver metastases and major vessels by prehepatectomy chemotherapy was demonstrated. These vessels should be resected simultaneously with areas they perfuse or drain; otherwise, resection and reconstruction of the vessels is needed. If a predicted positive surgical margin after resection is no longer an absolute contraindication to surgery for treating liver metastases,9 these vessels could be preserved even when they show attachment to liver tumors. More relaxed consideration of the margin may permit extension of surgical indications.

In this study, only 7 patients were treated with chemotherapy regimens including bevacizumab, and 1 patient was treated with a regimen including cetuximab. These few cases could offer little new information about morphological changes associated with monoclonal antibodies. Further, RECIST criteria may have limited ability to assess response to such biological agents.10 After regimens including bevacizumab, optimal response according to imaging was defined as a change in metastases from lesions with heterogeneous attenuation and thick, irregular borders to bland, homogeneously hypodense masses with a sharp interface between tumors and adjacent liver parenchyma. Homogeneous attenuation shown by metastases responding to bevacizumab in combination regimens likely reflects replacement of treated tumor by fibrous connective tissue. Thus, viable tumor cells predominate in the periphery of metastases even after such chemotherapy. The same result was obtained when treating hepatocellular carcinoma with sorafenib.11 In such situations, liver metastases attached to or invading major intrahepatic vessels seemed not easily separable by treatment, even with a regimen including monoclonal antibodies. To assure more uniformity among regimens, we conducted an additional analysis excluding the 8 patients treated with regimens including a monoclonal antibody. In the 37 patients receiving only cytotoxic regimens, 67 of 77 vessels attached to or invaded by 62 metastases could be evaluated pathologically. Of these 67 vessels, 37 (55%) vessels proved pathologically to be invaded by metastases (n = 24) or attached to metastases (n = 6), or were associated with exposed tumor at the resection plane at the point metastases where the specimen was separated from the vessels (n = 7, data not shown). The above analysis of data obtained using regimens of cytotoxic drug combinations alone should serve as a useful baseline for future studies of possible additional effects of monoclonal antibodies.

Extent of tumor attachment to the vessels and deformity of the vessels on CT were reported to be useful indications for concomitant liver and vessel resection, focusing on hepatic caval invasion of the liver tumors.7 In this study, tumor attachment involving >25% of circumference of vessels on imaging after chemotherapy was identified as a predictive factor for microscopic tumor invasion or attachment to vessels. This result is consistent with previous findings that >25% extent of IVC tumor attachment predicted the need for concomitant liver and IVC resection.7 As for factors predicting direct tumor invasion of vessels, tumor size reduction was identified by multivariate analysis, in addition to extent of vessel circumference showing attachment after chemotherapy. Recent reports of a favorable result in R1 resection, a designation denoting histopathologically evident cancer cells at the line of resection, likely reflect various advances—especially increasingly effective chemotherapy regimens.12, 13 According to our present results, if a metastasis was attached to >25% of the extent of a vessel by imaging but tumor size reduction of ≥60% was accomplished with chemotherapy, the metastasis was likely to be attached to the vessel without invasion at the time of pathological examination. If a predicted positive surgical margin after resection is no longer an absolute contraindication to surgery for treating advanced and aggressive liver metastases,12, 13 tumor size reduction by ≥60% could permit resection that preserves the vessel showing attachment, without vascular resection or reconstruction. Such a vessel-sparing procedure can reduce liver resection volume, maintaining sufficient future liver remnant volume. Therefore, information about vascular attachment or invasion by tumor is important for expanding eligibility for liver resection for advanced or aggressive liver metastases that otherwise might require massive liver resection with high risk of postoperative liver failure. The information also is important in making detailed preoperative decisions concerning extent and mode of liver resection.

Grouping metastases according to response to preoperative chemotherapy, cancer invasion was seen in 29% (17/58) of vessels attached to responding metastases, 38% (6/16) of vessels attached to SD tumors, and 67% (6/9) of vessels attached to PD tumors. Involved vessels are not easily separated from tumors by recent accounts in devised regimens, but our results suggest that more effective regimens might increase the likelihood of separation between tumor and vessels by greater reduction of tumor size and extent of vessel attachment. In the future, new chemotherapeutic agents with novel tumor regression mechanisms may accomplish better separation of vessels from tumors.

In conclusion, liver metastases attached to or invading major intrahepatic vessels are not easily separated from them, even by recently developed effective chemotherapy regimens. These vessels should be resected together with their perfusion or drainage area, resected and reconstructed, or sometimes preserved when showing attachment without invasion of vessels. More effective chemotherapy regimens may accomplish more reliable separation between tumor and vessels.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES

No specific funding was disclosed.

CONFLICT OF INTEREST

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES
  • 1
    Ng JK, Urbanski SJ, Mangat N, et al. Colorectal liver metastases contract centripetally with a response to chemotherapy: a histomorphologic study. Cancer. 2008; 112: 362-371.
  • 2
    Rubbia-Brandt L, Giostra E, Brezault C, et al. Importance of histological tumor response assessment in predicting the outcome in patients with colorectal liver metastases treated with neo-adjuvant chemotherapy followed by liver surgery. Ann Oncol. 2007; 18: 299-304.
  • 3
    Yamanaka N, Okamoto E, Oriyama T, et al. A prediction scoring system to select the surgical treatment of liver cancer. Further refinement based on 10 years of use. Ann Surg. 1994; 219: 342-346.
  • 4
    Levi F, Zidani R, Misset JL. Randomised multicentre trial of chronotherapy with oxaliplatin, fluorouracil, and folinic acid in metastatic colorectal cancer. Lancet. 1997; 350: 681-686.
  • 5
    Levi F. Circadian chronotherapy for human cancers. Lancet Oncol. 2001; 2: 307-15.
  • 6
    Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst. 2000; 92: 205-216.
  • 7
    Hashimoto T, Minagawa M, Aoki T, et al. Caval invasion by liver tumor is limited. J Am Coll Surg. 2008; 207: 383-392.
  • 8
    Maeba T, Okano K, Mori S, et al. Extent of pathologic invasion of the inferior vena cava in resected liver cancer compared with possible caval invasion diagnosed by preoperative images. J Hepatobiliary Pancreat Surg. 2000; 7: 299-305.
  • 9
    de Haas RJ, Wicherts DA, Flores E, Azoulay D, Castaing D, Adam R. R1 resection by necessity for colorectal liver metastases: is it still a contraindication to surgery? Ann Surg. 2008; 248: 626-637.
  • 10
    Chun YS, Vauthey JN, Boonsirikamchai P, et al. Association of computed tomography morphologic criteria with pathologic response and survival in patients treated with bevacizumab for colorectal liver metastases. JAMA. 2009; 302: 2338-2344.
  • 11
    Abou-Alfa GK, Schwartz L, Ricci S, et al. Phase II study of sorafenib in patients with advanced hepatocellular carcinoma. J Clin Oncol. 2006; 24: 4293-4300.
  • 12
    de Haas RJ, Wicherts DA, Flores E, et al. R1 resection by necessity for colorectal liver metastases: is it still a contraindication to surgery? Ann Surg. 2008; 248: 626-637.
  • 13
    Tanaka K, Nojiri K, Kumamoto T, et al. R1 resection for aggressive or advanced colorectal liver metastases is justified in combination with effective prehepatectomy chemotherapy. Eur J Surg Oncol. 2011; 37: 336-343.