Hepatic and Extrahepatic Colorectal Metastases: When Resectable, Their Localization Does Not Matter,...

Hepatic and Extrahepatic Colorectal Metastases: WhenResectable, Their Localization Does Not Matter, But Their

Total Number Has a Prognostic Effect

Dominique Elias, MD, PhD,1 Gabriel Liberale, MD,1 Dewi Vernerey, MSc,2

Marc Pocard, MD, PhD,1 Michel Ducreux, MD, PhD,3 Valerie Boige, MD,3

David Malka, MD, PhD,3 Jean-Pierre Pignon, MD, PhD,2 and Philippe Lasser, MD1

1Department of Surgical Oncology, Institut Gustave Roussy, Comprehensive Cancer Center,39 Rue Camille Desmoulins, 94805 Villejuif Cedex, France

2Department of Medical Statistics, Institut Gustave Roussy, Comprehensive Cancer Center,39 Rue Camille Desmoulins, 94805 Villejuif Cedex, France

3Department of Medical Oncology, Institut Gustave Roussy, Comprehensive Cancer Center,39 Rue Camille Desmoulins, 94805 Villejuif Cedex, France

Background: The presence of extrahepatic disease (EHD) is considered a contraindicationto hepatectomy in patients with colorectal liver metastases. After resection, the prognosis isbased more on the total number of resected metastases (located inside and outside the liver)than on the site of these metastases (only inside the liver or not).Methods: A total of 308 patients with colorectal cancer underwent hepatectomy, and 84

(27%) also underwent resection of miscellaneous EHD. The study was a prospective dataregistration and retrospective analysis. When considering the total number of resectedmetastases, each liver metastasis and each EHD location was counted as one lesion. Uni-variate and multivariate analyses were performed.Results: The median follow-up was 99 months. The overall 5-year survival rate was 32%. In

the multivariate analysis, the total number of metastases (inside or outside the liver) had agreater prognostic value than the criterion ‘‘presence or absence of EHD.’’ Considering thetotal number of resected metastases (whatever their site), 5-year survival rates were38% (SD: 4%) in the group with one to three metastases, 29% (SD: 5%) in patients with fourto six metastases, and 18% (SD: 5%) in patients with more than six metastases (P = .002). Avery simple prognostic score based on sex and the total number of metastases is proposed.Conclusions: EHD, when resectable, is no longer a contraindication to hepatectomy. More

importantly, the total number of the metastases, whatever their location, has a strongerprognostic effect than the site of these metastases.Key Words: Colorectal cancer—Liver metastases—Extrahepatic disease.

Hepatic resection is traditionally the sole treatmentthat produces long-term survival in patients withcolorectal liver metastases (LM), with a 25% to 40%

5-year survival rate.1–10 Extrahepatic disease (EHD)has been considered a contraindication to hepatec-tomy because subgroup analyses of large series pub-lished before 1990 showed a very unfavorableprognosis.1–6 Similarly, lung metastases were rarelyresected even when they were technically resectable,and peritoneal carcinomatosis was never consideredas potentially curable.11,12 However, over the last fewyears, some series reported interesting 5-year survival

Received January 19, 2005; accepted June 28, 2005; publishedonline September 26,2005.Address correspondence and reprint requests to: Dominique

Elias, MD, PhD; E-mail: [email protected]

Published by Springer Science+Business Media, Inc. � 2005 The Society ofSurgical Oncology, Inc.

Annals of Surgical Oncology, 12(11): 900)909DOI: 10.1245/ASO.2005.01.010

900

rates after resection of a high number of LMs,7,13,14

of multiple lung metastases,15,16 of EHD synchro-nous to LM,13,17–19 and of peritoneal carcinomato-sis.20–23 During the same time, the concept that theprognosis of metastasis is better when it is limited toonly one organ is inconsistent with recent studies oncirculating tumor cells and the pathophysiology ofmetastases.24 In contradiction to this concept, werecently reported that the crude overall 5-year sur-vival rate of 75 patients who underwent hepatectomywith resection of EHD was 29%19; these results implythat EHD does not absolutely contraindicate hepa-tectomy for colorectal LM. This article deals with thesame topic but with a new concept: to base theinterpretation of the results of metastasis resectionnot on the sites, as is the case traditionally, but on thetotal number—regardless of the number of invadedsites.

PATIENTS AND METHODS

Patients

All patients with colorectal LM who underwenthepatectomy, possibly associated with EHD resec-tion, and who had at least 4.5 years of follow-up(except for three foreign patients who were lost tofollow-up) were retrospectively selected from a pro-spective database. The following patient-selectioncriteria were used: the discovery of EHD on preop-erative imaging or at laparotomy was not a contra-indication to hepatectomy if the lesions seemedcompletely resectable, whatever their site. The resec-tion of EHD was performed during the same sessionas hepatectomy, with intestinal anastomosis (if nec-essary) and with immediate intraperitoneal chemo-therapy (with or without hyperthermia) in case ofperitoneal seeding. The only exceptions were lungmetastases, which, if present, were resected 2 monthslater.Hepatic resection was attempted irrespective of the

number of metastases provided that all detected tu-mors could be removed completely, with potentiallynegative surgical margins, and that an acceptableamount of liver parenchyma (more than 33% of thetotal liver volume) could be preserved. In this study, anegative surgical margin was thought to have beensecured when the tumor was not exposed at the cutsurface of the liver either macroscopically or micro-scopically, irrespective of the extent of the margin. Ifthe remaining liver was <33% of the total volume asa result of extensive resection, then preoperative

portal vein embolization was conducted to inducecompensatory hypertrophy of the remaining livervolume and thereby reduce the risk of postoperativehepatic insufficiency.25,26

Resection of all detectable intrahepatic and extra-hepatic disease was considered complete according tothe surgeon�s judgment, even if the final postoperativepathologic or imaging results indicated an incompleteresection (R1 or R2). This definition of the com-pleteness (radicality) of the resection by the surgeonhas the merit of being very close to clinical practiceand being easily reproducible. In other words, thesurgeon always considered a complete (curative)resection to have been performed, even if in a fewcases it was finally incomplete (palliative) afterpathologic examination.

Methods

Positron emission tomographic imaging was notused in this study. Intraoperative ultrasonographywith a high-resolution probe was used systematicallyto detect possible occult tumors that were not foundon preoperative imaging or by inspection and pal-pation. Hepatectomy was performed anatomically ornonanatomically under intermittent clamping of thepedicle27 and, during the last years, under intermit-tent exclusion of the liver without interruption of thevena caval flow.28 EHD was resected with a similarcurative intent. These procedures included lympha-denectomies (hepatic or lateroaortic), resection oflocal recurrences of the primary tumor, and resectionof organs (such as the ovary, adrenal glands, orspleen) or parts of organs (e.g., lungs) with metastaticdisease. For peritoneal carcinomatosis, a completecytoreductive operation was performed before thehepatectomy began.22,23,29,30 It was followed withimmediate intraperitoneal chemotherapy with theaim of curing the remaining microscopic peritonealdisease.23 We used different and successive types ofintraperitoneal chemotherapies, always in the settingof prospective trials with signed consent from thepatients, with a constant improvement of the tech-nique.31–33

Systemic chemotherapy was given in all patientseither before or after hepatectomy. All patients withEHD detected before surgery received preoperativechemotherapy and, in case of objective response, re-ceived it again after hepatectomy unless they hadlimiting toxicity. Regimens of chemotherapy werebased on 5-fluorouracil and leucovorin, and after1997, oxaliplatin or irinotecan were added to thisbasic regimen. Eighty three percent of them received

NUMBER OF RESECTED METASTASES MORE IMPORTANT THAN SITE 901

Ann. Surg. Oncol. Vol. 12, No. 11, 2005

at least two lines of chemotherapy. Because of thenumerous regimens of chemotherapy used during thislong period of time and their miscellaneous combi-nations before or after hepatectomy, we do not reportthem in this article.Follow-up after hepatectomy was as follows: after

discharge, patients were seen every 3 months duringthe first 2 years and every 6 months during the next 3years. Tumor markers were evaluated if they wereincreased before hepatectomy, and the most accuratepersonal imaging studies were performed. Beforesurgery, each patient underwent an ultrasoundexamination and a computed tomographic scan ofthe abdomen, pelvis, and thorax. If LM were noteasily imaged, a magnetic resonance imaging studywas performed. In case of recurrence, surgery wasconsidered first, followed by chemotherapy.Survival curves were constructed by the Kaplan-

Meier method and were compared by using the log-rank test. The cutoff date for analysis was November15, 2004. Postoperative deaths were included in thesecurves. Only three foreign patients were lost to fol-low-up, and they were excluded from the analysis atthe date of the last follow-up. Statistical calculationwere performed with SAS version 8.02 (SAS Institute,Inc., Cary, NC). Qualitative variables were comparedby v2 test, and quantitative variables were comparedby Student�s t-test. Univariate analysis was per-formed with the log-rank test. All significant vari-ables in the univariate analysis were considered formultivariate analysis with a Cox proportional hazardregression model. Statistical significance was definedas P < .05. All statistical tests were two sided. Allclassic parameters (age, sex, lymph node status of theprimary tumor, operation, resection radicality, pa-thologic characteristics, chemotherapy, and follow-up) were prospectively registered in a specific data-base. However, the carcinoembryonic antigen levelwas not measured initially in 37% of cases and wastherefore excluded from analysis.In this study, the number of resected lesions was

considered from two different perspectives for com-parison. In the classic analysis, survival was consid-ered according to the number of resected LMs, withthree groups (one to three LMs, four to six LMs, andmore than six LMs). In the nonclassic analysis, sur-vival was considered according to the total number ofresected intrahepatic but also extrahepatic metasta-ses, with the same three categories of patients (one tothree metastases, four to six metastases, and morethan six metastases). For the second analysis, wedecided arbitrarily to count one metastasis for eachLM, each lung metastasis, each distant invaded

lymph node, each local recurrence (whatever its size),each ovary or adrenal gland, or each miscellaneousmetastasis. For peritoneal carcinomatosis, we coun-ted one for each peritoneal tumor seeding up to fivebut kept the maximum of five when there was morenumerous tumor seeding. This was decided after afirst analysis of the data to keep a balance betweenthe number of metastases according to the differentsites of disease.

RESULTS

Patient Characteristics

A total of 308 patients underwent operation be-tween January 1, 1987, and January 1, 2000, with 160men and 148 women (mean age, 57 ± 10 years;median, 58 years; range, 18–86 years). The primarytumor was in the colon in 204 cases and in the rectumin 104 cases. One hundred thirty-seven (45%) weresynchronous LMs (i.e., they were discovered at thetime of or within 6 months after discovery of theprimary tumor). The population included in thisstudy is an extension (more patients and longer fol-low-up) of previously published series.17,19

Among these 308 patients who underwent hepa-tectomy for colorectal LM, 84 (27%) also had syn-chronous EHDs, which were resected. These EHDswere known before surgery in 55 and were discoveredat laparotomy in 29.The details of these localizations are listed in

Fig. 1, as are the number of resected LMs. They in-cluded peritoneal carcinomatosis, lung metastases,different sites of distant lymph node metastases (inthe hepatic pedicle or in the left lateroaortic zone),local recurrence of the primary tumor, and ovarianand other miscellaneous metastases. Peritoneal car-cinomatosis was macroscopically present, resected,and histologically proven in 37 patients. Amongthem, 15 patients had 1 or 2 peritoneal implants, 10had 3 to 6 implants, and 12 had more than 10 im-plants.Thirty-four patients underwent a preoperative right

portal embolization. The median number of resectedLMs was 2 (range, 1–14), and the mean diameter ofthe largest LM was 5 cm (SD, 4 cm). The sectionmargin in the liver was invaded in 9 cases and was 0mm but was considered as noninvaded in 71 cases(23%). The section margin was invaded in nine casesat the site of the resected extrahepatic localization.Eighteen patients underwent a palliative resection,even if the surgeon estimated that he had performed a

D. ELIAS ET AL.902


complete operation. After recurrence, 86 patientsunderwent a second operation for metastasis; 33 pa-tients, a third one; and 9, a fourth one.

Mortality and Morbidity

For the 308 patients, mortality until hospitaldischarge was 3% (n = 10). There was no mortalityafter lung resection performed 2 months later. Mor-bidity was 18% (n = 54), including that for lungresections. Mortality and morbidity rates wererespectively 3% and 22% for the 84 patients whounderwent a hepatectomy with resection of 1 or moreEHDs, and they were 3% and 20% for the 224 pa-tients without resection of EHD.The median follow-up was 99 months (range, 8–

205 months). Minimum follow-up was 54 months,except for the three foreign patients who were lost tofollow-up.

Causes of Death

A total of 236 patients died. Death was due to thecancer in 197 cases (83%), to iatrogenic reasons(including postoperative deaths) in 26 cases, to non-tumor causes in 8 cases, and to unknown causes in 5cases.The 5-year overall survival of the 308 patients was

32% (95% confidence interval 3%), and median sur-vival was 38 months (confidence interval, 36–43months). For patients without EHD, the 5-year sur-vival was 34%—significantly (P= .04) higher than the28% for patients with EHD (Fig. 2). Table 1 showsthe survival rates according to different parametersand analyses, which are detailed afterward. The age,lymph node status of the primary tumor, synchronousor metachronous appearance of the LM, largest sizeof the LM, and free margin had no significant effecton the prognosis, contrary to the sex (P = .01), in

FIG. 1. Distribution of liver metastases (LMs) and extrahepatic disease (EHD) in the 308 patients. Nb, number.



favor of women, and to the completeness of theresection (P < .001).Concerning the number of resected lesions, Fig. 3

shows the overall survival rates of the 308 patientsaccording to the number of LMs (classic analysis) andaccording to the total number of resected metastases,whatever their localization (nonclassical analysis). Thenumber of lesions had a significant prognostic effect inthese two analyses, but it was more powerful in thenonclassic analysis, which considered the total number

of metastases (P = .002), than in the classic analysis,which considered only the number of LMs (P= .04).When comparing overall survival rates according

to the total number of metastases (whatever their site)in patients with and without EHD, analyzed in threegroups (one to three metastases, four to six metas-tases, and more than six metastases; Fig. 4), we didnot find any significant difference within thesegroups. This means that the presence or absence ofEHD has no major effect on prognosis.

Su

rviv

al (

%)

FIG. 2. Overall survival rates ofpatients with and without extrahe-patic disease (EHD).

TABLE 1. Overall survival rates at 3 and 5 years according to the main prognostic factors (univariate analysis)

Survival rate (%)

Variable n 3 y 5 y P value

No. of patients 308 53.2 ± 2.9 31.9 ± 2.7Sex .0119Male 160 51.0 ± 4.0 27.1 ± 3.6Female 148 54.7 ± 4.4 31.7 ± 4.1

Age (y) .783£60 180 52.2 ± 3.8 32.1 ± 3.5>60 128 54.7 ± 4.4 31.7 ± 4.1

Lymph node status of the primary .094N0 95 55.6 ± 5.1 40.5 ± 5.1N+ 213 52.2 ± 3.4 28.1 ± 3.1

Extrahepatic metastases .0431No 224 55.9 ± 3.3 33.5 ± 3.2Yes 84 46.2 ± 5.5 27.6 ± 5.0

Timing of metastases .774Synchronous 137 51.0 ± 4.3 29.0 ± 3.9Metachronous 171 55.0 ± 3.8 34.2 ± 3.7

No. of resected liver metastases .01701–3 217 57.4 ± 3.4 35.9 ± 3.34–6 63 48.0 ± 6.4 24.4 ± 5.6>6 28 32.1 ± 8.8 16.7 ± 7.3

No. of resected hepatic and extrahepatic metastases .002101–3 170 61.0 ± 3.8 37.7 ± 3.74–6 82 50.7 ± 5.6 29.2 ± 5.1>6 56 33.4 ± 6.4 18.1 ± 5.3

Radicality of resection <.0001R0 257 56.7 ± 3.1 35.3 ± 3.0R1 33 43.1 ± 8.9 19.3 ± 7.3R2 18 22.2 ± 9.8 5.6 ± 5.4

D. ELIAS ET AL.904


First, a multivariate analysis was classically per-formed by considering the R0 to R2 status(Table 2). This analysis showed the effect of thisstatus on the prognosis, followed by the totalnumber of metastases, whatever their site. Second,two multivariate analyses were conducted withoutconsidering the R0 to R2 status, which is knownonly after surgery (Tables 3 and 4). The classicstudy (Table 3) confirmed the traditional prognosticeffect of the number of LMs and the presence ofEHD. However, if we considered the total numberof metastases, whatever their site (Table 4), thepresence of EHD was not any more prognostic(P = .3).Finally, Fig. 3B reports the overall survival rate of

all 308 patients according to the total number of re-sected metastases, whatever their site, inside and/oroutside the liver (P = .002). This new prognosticclassification of the patients, based on the totalnumber of metastases, seems more reliable than the

classic one, which is based only on the number ofLMs (P = .01; Fig. 3A).

DISCUSSION

For the first time, this study shows that for meta-static colorectal tumors, when they are resectable, thetotal number of metastases, whatever their locations,is a more important prognostic factor than the site ofthe metastases. Therefore, it seems that one LMassociated with one EHD has almost the sameprognostic value as two isolated LMs.Also, the liver as the first site of extension no

longer seems to be a privileged site for resection ofmetastases but is, rather, similar to any other site.This was recently underlined for peritoneal carcino-matosis, for which the prognostic effect of completecytoreductive surgery is great29,34 and for which thecombination of complete resection with intraperito-

A

B

FIG. 3. Overall survival rates ofthe 308 patients according to the nu-mber of metastases in three groups.(A) According to the number of livermetastases (classic analysis); (B)according to the total number of me-tastases (liver metastases + extrahe-patic disease; nonclassic analysis).



neal chemohyperthermia results in curing of 20% to40% of patients.20–23 These results are close to thoseobtained 10 years earlier with hepatectomy for LMonly.35

Therefore, in our study, the site of metastaseswas not a preponderant prognostic factor, providedthat they were completely resectable. We consideredthe completeness of resection according to the

364353699311613915224610131718

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 1 2 3 4 5 67Time after operation (years)

Su

rviv

al (

%)

without EHD (n=152)

with EHD (n=18)

chi2 = 2.695 , p = 0.1007

At risk:

679132439475077111417232832

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 1 2 3 4 5 6 7

Time after operation (years)

Su

rviv

al (

%)

without EHD (n=50)

with EHD (n=32)

chi2 = 0.682 , p = 0.4091

At risk:

A

B

C

33377131822346711142634

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 1 2 3 4 5 6 7

Time after operation (years)

Su

rviv

al (

%) Without EHD (n=22)

With EHD (n=34)

chi2 = 0.416 , p = 0.519

At risk:

FIG. 4. Overall survival rates whenthe total number of metastases (livermetastases + extrahepatic disease [-EHD]) is (A) one to three, (B) four tosix, and (C) greater than six in the g-roup of patients without EHD and inthe group of patients with EHD.

D. ELIAS ET AL.906


surgeon�s macroscopic judgment, which seems to bethe most useful and applicable in a clinical ap-proach. This is why some of our patients who werepostoperatively staged as uncompletely resectedwere kept in our analysis and had very classically alow prognosis.After complete resection of metastases, whatever

their site, it is their number that seems to be the mostimportant prognostic factor. The effect of the numberof metastases was reported earlier for LMs,1–10 lungmetastases,15,16 and peritoneal carcinomatosis.20–23

In this study, the number of resected metastases wasalso preponderant, but any type of localization seemsequivalent if the metastases are adequately treated.It is important to underline that these results con-

cern a selected population for which we estimatedthat all the metastases, whatever their site, but alwaysincluding LMs, could be resected and also for whichthe surgeon considered that a complete resection was

performed. These results are also supported becauseequal or similar mortality and morbidity rates wereobserved in the group of patients who underwenthepatectomy only (3% and 20%, respectively) and inthe group of patients who underwent hepatectomyassociated with EHL resection (3% and 22%,respectively).Times are changing, and so are therapeutic con-

cepts. Traditionally it was considered that only a lo-cal treatment, such as surgery, could cure a cancerwith local extension. Over the last 20 years, we havelearned that micrometastases and circulating cancercells in blood, bone marrow, and lymph nodes arealways present in advanced-stage cancer.35 Moreover,it was shown that most cells derived from the primarytumor are not stopped by the liver and enter thesystemic circulation.37

Thus, traditional radical surgery is, strictlyspeaking, only a cytoreductive surgery (a treatmentthat does not completely eradicate all tumor cells,even if it entirely resects all visible and detectabledisease). However, cure can be achieved with theassistance of the patient�s natural immunologicaldefenses, chemotherapy, or both. Surgical tumorburden reduction may provide an immunologicalbenefit, because tumor cells produce immunosup-pressive cytokines, precipitate immune complexes,and produce peptides that interfere with normaldefenses.38,39 Furthermore, according to the log-killhypothesis, each dose of chemotherapeutic agentkills a constant fraction of cells, rather than a spe-cific number of cells.40 Therefore, reducing the initialtumor volume increases the likelihood that chemo-therapy will reduce the number of viable tumor cellstoward the desired end point (zero). Reducing thetotal tumor volume also substantially diminishes therisk for development of drug resistance, which in-

TABLE 3. Multivariate analysis of the prognostic factorswithout resection completeness considering the no. of resected

liver metastases (classic study)

Variable n HR P value

SexMale 160 1.4 (1.0–1.8)

.0205Female 148 1

Extrahepatic metastasesNo 224 1

.0145Yes 84 1.4 (1.1–1.9)

No. of resected liver metastases1–3 170 14–6 82 1.3 (1.0–1.9) .0240>6 56 1.7 (1.1–2.6)

HR, hazard ratio.

oo�

TABLE 2. Multivariate analysis of prognostic factors,including the completeness of resection


SexMale 160 1.4 (1.1–1.8) .0112Female 148 1

Radicality of resectionR0 257 1 .0002R1 33 1.2 (.8–1.9)R2 18 2.9 (1.7–4.8)

Extrahepatic metastasesNo 224 1 .999Yes 84 1.0 (.7–1.4)

No. of resected intrahepaticand extrahepatic metastases1–3 170 1 .01804–6 82 1.2 (.9–1.7)>6 56 1.7 (1.2–2.5)

HR, hazard ratio.

TABLE 4. Multivariate analysis of the prognostic factorswithout resection completeness considering the no. of resectedintrahepatic and extrahepatic metastases (nonclassic study)


SexMale 160 1.4 (1.1–1.8)

.0112Female 148 1

Extrahepatic metastasesNo 224 1

.3493Yes 84 1.2 (.8–1.6)

No. of resected intrahepaticand extrahepatic metastases1–3 170 14–6 82 1.2 (.8–1.6) .0216>6 56 1.7 (1.2–2.4)

HR, hazard ratio.

oo�



creases directly with the number of cancer cells andthe treatment duration.40

Recently, it was proven that systemic chemother-apy could cure microscopic disease in colorectalcancer, mainly as adjuvant treatment after operationof the primary tumor.41,42 In fact, with the lastgeneration of drugs, this principle has increasedapplication for metastatic stages. For example, it isnow clear that a complete cytoreductive operationtreating macroscopic peritoneal carcinomatosis, fol-lowed by immediate intraperitoneal chemotherapytreating the remaining microscopic peritoneal dis-ease, is able to cure some patients20–23 and results insurvival rates similar to those obtained with hepa-tectomy.35 In the same way, we showed in this studythat the complete resection of metastases, whatevertheir sites, is the first step to a curative intent, inassociation with more and more efficient chemo-therapy. The principle of complete resection ofmetastases, whatever their localization, becomesmore important than the localization of the metas-tases in only one organ. Also, it was reported thatpatients whose metastatic disease did not progresswhile they were receiving neoadjuvant chemotherapyexperienced improved survival after liver resectioncompared with patients who did not receivechemotherapy.18,43

The logical application of this principle is, in thefuture, to rapidly test the chemosensitivity of the pa-tients first, when tumor targets are in place, then tocompletely resect the metastases (when feasible), andthen to continue efficient chemotherapyafter resection.In this view, surgery and chemotherapy have the sameimportance for curing. It is their association that isessential: their combination is efficient and potentiallyable to cure, but onewithout the other is unable to cure.In conclusion, to treat macroscopic and micro-

scopic tumor compounds in a different way (but in acombined approach), complete resection of colorectalmetastases is more important than the sites of thesemetastases. Therefore, extrahepatic metastases are nolonger a contraindication to hepatectomy. The totalnumber of resected metastases, whatever their site,retains a strong prognostic effect.

REFERENCES

1. Hughes KS, Simon R, Songhorabodi S, et al. Resection of theliver for colorectal carcinoma metastases: a multi-institutionalstudy of indications for resection. Surgery 1988;103: 278–88.

2. Nordlinger B, Jaeck D, Guignet M, et al. Resection Chirurgi-cale des Metastases Hepatiques, Enquete de l�AssociationFrancaise de Chirurgie. Paris: Springer-Verlag, 1992:141–75.

3. Eckberg H, Transberg KG, Andersson R, et al. Determinantsof survival in liver resection for colorectal secondaries. Br JSurg 1986;73:727–31.

4. Cady B, McDermott WV Jr. Major hepatic resection formetachronous metastases from colon cancer. Ann Surg1985;210:204–9.

5. Foster JH. Survival after liver resection for secondary tumors.Am J Surg 1978;135:389–94.

6. Adson MA. The resection of hepatic metastases. Another view.Arch Surg 1989;124:1023–4.

7. Fong Y, Fortner J, Sun RL, Brennan MF, Blumgart LH.Clinical score for predicting recurrence after hepatic resectionfor metastatic colorectal cancer. Analysis of 1001 consecutivecases. Ann Surg 1999;230:309–21.

8. Jamison RL, Donohue JH, Nagorney DM, Rosen CB,Harmsen S, Ilstrup DM. Hepatic resection for metastaticcolorectal cancer results in cure for some patients. Arch Surg1997;132:505–11.

9. Scheele J, Stang R, Altendorf-Hofmann A, Paul M. Resectionof colorectal metastases. World J Surg 1995;19:59–71.

10. Iwatsuki S, Dvorchick I, Madariaga J, et al. Hepatic resectionfor metastatic colorectal adenocarcinoma: a proposal of aprognostic scoring system. J Am Coll Surg 1999;189:291–9.

11. Chu DZ, Lang NP, Thompson C, Osteen PK, Westbrook KC.Peritoneal carcinomatosis in non-gynecologic malignancy.Cancer 1989;63:364–7.

12. Sadeghi B, Arvieux C, Glehen O, et al. Peritoneal carcinoma-tosis from non-gynecologic malignancies. Multicentric pro-spective study. Cancer 2000;88:358–63.

13. Minagawa M, Makuuchi M, Torzilli G, et al. Extension of thefrontiers of surgical indications in the treatment of livermetastases from colorectal cancer. Ann Surg 2000;231:487–99.

14. Imamura H, Seyama Y, Kokudo N, et al. Single and multipleresections of multiple hepatic metastases of colorectal origin.Surgery 2004;135:508–17.

15. Ike H, Shimada H, Ohki S, Togo S, Yamaguchi S, Ichikawa Y.Results of aggressive resection of lung metastases from colo-rectal carcinoma detected by intensive follow-up. Dis ColonRectum 2002;45:468–75.

16. Regnard JF, Grunenwald D, Spaggiari L, et al. Surgicaltreatment of hepatic and pulmonary metastases from colorec-tal cancers. Ann Thorac Surg 1998;66:214–9.

17. Elias D, Ouellet JF, Bellon N, Pignon JP, Pocard M, Lasser P.Extrahepatic disease does not contraindicate hepatectomy forcolorectal liver metastases. Br J Surg 2003;90:567–74.

18. Adam R, Pascal G, Castaing D, et al. Tumor progression whileon chemotherapy. A contraindication to liver resection formultiple colorectal metastases? Ann Surg 2004;240:1052–64.

19. Elias D, Sideris L, Pocard M, et al. Results of R0 resection forcolorectal metastases associated with extrahepatic disease. AnnSurg Oncol 2004;11:274–80.

20. Sugarbaker PH, Jablonski KH. Prognostic features of 51colorectal and 130 appendiceal cancer patients with peritonealcarcinomatosis treated by cytoreductive surgery and intra-peritoneal chemotherapy. Ann Surg 1995;221:124–32.

21. Verwaal VC, van Ruth S, de Bree E, et al. Randomized trial ofcytoreduction and hyperthermic intraperitoneal chemotherapyversus systemic chemotherapy and palliative surgery in patientswith peritoneal carcinomatosis from colorectal cancer. J ClinOncol 2003;21:3737–43.

22. Elias D, Sideris L, Pocard M, et al. Efficacy of intraperitonealchemohyperthermia with oxaliplatin in colorectal peritonealcarcinomatosis. Preliminary results in 24 patients. Ann Oncol2004;15:781–5.

23. Elias D, Ouellet JF. Intraperitoneal chemohyperthermia:rationale, technique, indications, and results. Surg Oncol ClinNorth Am 2001;10:915–33.

24. Filder IJ. Seed and soil revisited. Contribution of the organmicroenvironment to cancer metastases. . Surg Oncol ClinNorth Am 2001;10:257–69.

D. ELIAS ET AL.908


25. Imamura H, Shimada R, Kubota M, et al. Preoperative portalvein embolization: an audit of 84 patients. Hepatology 1999;29:1099–105.

26. Roche A, Soyer P, Elias D, Breittmayer F, Gad M, RougierPH. Pre-operative portal vein embolisation for hepatic metas-tases. J Intervent Radiol 1991;6:63–6.

27. Elias D, Desruennes E, Lasser PH. Prolonged intermittentclamping of the portal triad during hepatectomy. Br J Surg1991;78:42–4.

28. Elias D, Lasser P, Debaene B, et al. Intermittent exclusion ofthe liver without vena cava clamping during major hepatec-tomy. Br J Surg 1995;82:1535–9.

29. Elias D, Delperro JR, Sideris L, et al. Treatment of peritonealcarcinomatosis from colorectal cancer: impact of completecytoreductive surgery and difficulties in conducting rando-mised trials. Ann Surg Oncol 2004;11:518–21.

30. Elias D, Dube P, Bonvalot S, et al. Treatment of livermetastases with moderate peritoneal carcinomatosis byhepatectomy and cytoreductive surgery followed by imme-diate postoperative intraperitoneal chemotherapy. Feasibilityand preliminary results. Hepatogastroenterology 1999;46:360–3.

31. Elias D, Bonnay M, Puizillou JM, et al. Heated intra-operativeintraperitoneal oxaliplatin after complete resection of perito-neal carcinomatosis: pharmacokinetics and tissue distribution.Ann Oncol 2002;13:267–72.

32. Elias D, El Otmany A, Bonnay M, et al. Human pharmaco-kinetic study of heated intraperitoneal oxaliplatin in increas-ingly hypotonic solutions after complete resection of peritonealcarcinomatosis. Oncology 2002;63:346–52.

33. Elias D, Matsuhisa T, Sideris L, et al. Heated intra-opera-tive intraperitoneal oxaliplatin plus irinotecan after completeresection of peritoneal carcinomatosis: pharmacokinetics,

tissue distribution and tolerance. Ann Oncol 2004;15:1558–65.

34. Glehen O, Kwiatkowski F, Sugarbaker PH, et al. Cytoreductivesurgery combined with intraperitoneal chemotherapy for man-agement of peritoneal carcinomatosis from colorectal cancer: amulti-institutional study. J Clin Oncol 2004;22:3284–92.

35. Elias D. Peritoneal carcinomatosis or liver metastases fromcolorectal cancer: similar standards for a curative surgery? AnnSurg Oncol 2004;11:122–3.

36. Tsavellas G, Patel H, Aleen-Mersh TG. Detection and clinicalsignificance of occult tumour cells in colorectal cancer. Review.Br J Surg 2001;88:1307–20.

37. Sadahiro S, Suzuki T, Tokunaga N, et al. Detection of tumorcells in the portal and peripheral blood of patients with colo-rectal carcinoma using competitive reverse transcriptase-poly-merase chain reaction. Cancer 2001;92:1251–8.

38. Pollock RE, Roth JA. Cancer-induced immunosuppression:implications for therapy? Semin Surg Oncol 1989;5:414–9.

39. Kavanaugh DY, Carbone DP. Immunologic dysfunction incancer. Hematol Oncol Clin North Am 1996;10:927–51.

40. Muss HB. Chemotherapy of Gynecologic Cancer. 2nd ed. NewYork: Wiley-Liss, 1990.

41. IMPACT (International Multicentre Pooled Analysis of ColonCancer Trials)Efficacy of adjuvant fluorouracil and folinic acidin colon cancer. Lancet 1995;345:939–44.

42. Gill S, Loprinzi CL, Sargent DJ, et al. Pooled analysis of flu-orouracil-based adjuvant therapy for stage II and III coloncancer: who benefits and by how much? J Clin Oncol2004;22:1797–806.

43. Allen PJ, Kemeny N, Jarnagin W, et al. Importance of re-sponse to neoadjuvant chemotherapy in patients undergoingresection of synchronous colorectal liver metastases. J Gas-trointest Surg 2003;7:109–17.



Hepatic and Extrahepatic Colorectal Metastases: When Resectable, Their Localization Does Not Matter,...

Documents

Transcript of Hepatic and Extrahepatic Colorectal Metastases: When Resectable, Their Localization Does Not Matter,...