Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2333032052
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gazelle, G. S.
Right arrow Articles by Weinstein, M. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gazelle, G. S.
Right arrow Articles by Weinstein, M. C.
(Radiology 2004;233:729-739.)
© RSNA, 2004

Evidence-based Practice

Metastatic Colorectal Carcinoma: Cost-effectiveness of Percutaneous Radiofrequency Ablation versus That of Hepatic Resection1

G. Scott Gazelle, MD, MPH, PhD, Pamela M. McMahon, BS, Molly T. Beinfeld, MPH, Elkan F. Halpern, PhD and Milton C. Weinstein, PhD

1 From the Institute for Technology Assessment and Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Zero Emerson Place, Suite 2H, Boston, MA 02114 (G.S.G., P.M.M., M.T.B., E.F.H.); and Center for Risk Analysis and Department of Health Policy & Management, Harvard School of Public Health, Boston, Mass (G.S.G., P.M.M., M.C.W.). Received December 17, 2003; revision requested February 24, 2004; revision received March 9; accepted April 12. Supported in part by the National Cancer Institute under R01-CA/HS83960 and the U.S. Department of the Army under DAMD 17-99-2-9001. Address correspondence to G.S.G.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
PURPOSE: To evaluate the relative cost-effectiveness of radiofrequency (RF) ablation and hepatic resection in patients with metachronous liver metastases from colorectal carcinoma (CRC) and compare the outcomes, cost, and cost-effectiveness of a variety of treatment and follow-up strategies.

MATERIALS AND METHODS: A state-transition decision model for evaluating the (societal) cost-effectiveness of RF ablation and hepatic resection in patients with CRC liver metastases was developed. The model tracks the presence, number, size, location, growth, detection, and removal of up to 15 individual metastases in each patient. Survival, quality of life, and cost are predicted on the basis of disease extent. Imaging, ablation, and resection affect outcomes through detection and elimination of individual metastases. Several patient care strategies were developed and compared on the basis of cost, effectiveness, and incremental cost-effectiveness (expressed as dollars per quality-adjusted life-year [QALY]). Extensive sensitivity analysis was performed to evaluate the impact of alternative scenarios and assumptions on results.

RESULTS: A strategy permitting ablation of up to five metastases with computed tomographic (CT) follow-up every 4 months resulted in a gain of 0.65 QALYs relative to a no-treat strategy, at an incremental cost of $2400 per QALY. Compared with this ablation strategy, a strategy permitting resection of up to four metastases, one repeat resection, and CT follow-up every 6 months resulted in an additional gain of 0.76 QALYs at an incremental cost of $24 300 per QALY. Across a range of model assumptions, more aggressive treatment strategies (ie, ablation or resection of more metastases, treatment of recurrent metastases, more frequent follow-up imaging) were superior to less aggressive strategies and had incremental cost-effectiveness ratios of less than $35 000 per QALY. Findings were insensitive to changes in most model parameters; however, results were somewhat sensitive to changes in size thresholds for RF ablation, the number of metastases present, and surgery and treatment costs.

CONCLUSION: RF ablation is a cost-effective treatment option for patients with CRC liver metastases. However, in most scenarios, hepatic resection is more effective (in terms of QALYs gained) than RF ablation and has an incremental cost-effectiveness ratio of less than $35 000 per QALY.

© RSNA, 2004

Index terms: Cost-effectiveness • Economics, medical • Liver neoplasms, 761.33 • Liver neoplasms, therapy, 761.1269 • Radiofrequency (RF) ablation, 761.1269


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Colorectal carcinoma (CRC) is the second leading cause of cancer deaths in the United States. For the year 2003, it had been estimated that 147 500 new cases of CRC would be diagnosed and that the disease would be responsible for approximately 57100 deaths (1). Most CRC deaths can be attributed to metastatic disease rather than the primary tumor. In many patients the liver is the only—or principal—site of metastatic disease. Unfortunately, the prognosis for patients with untreated CRC metastases is dismal, with 5-year survival rates reported to be less than 1% and median survival estimated at less than 12 months (2–12).

Several factors have been observed to correlate with survival in patients with metastatic CRC. Principal among these is the extent of liver replaced by tumor (3,7–9,11–16). Other factors, such as serum alkaline phosphatase levels (2,3,7–9,13,15,17) and carcinoembryonic antigen levels (8,15), have also been identified as important predictors, although they may in fact serve as indicators of the extent of liver metastasis (14) rather than as independent predictors of outcome.

The predominant effect of liver tumor involvement suggests that a successful locally targeted therapy could increase life expectancy. Though based only on observational trials, there is now substantial evidence that resection of liver metastases from CRC can result in long-term survival in some patients (10,16,18–38). Less invasive alternatives to surgical resection include percutaneous tumor ablation by using methods such as radiofrequency (RF) ablation (39–68), cryosurgery (69–73), or interstitial laser thermotherapy (74–79). In most cases these procedures can be performed percutaneously on an outpatient basis, although laparotomy is sometimes required to protect adjacent viscera or, for larger lesions, to occlude vascular inflow. Percutaneous tumor ablation is particularly attractive owing to its predictable results, relatively low cost, and low complication rate.

We recently published the results of a cost-effectiveness analysis of hepatic resection in patients with CRC liver metastases (80). The purpose of our current study was to evaluate the relative cost-effectiveness of RF ablation and hepatic resection in patients with metachronous liver metastases from CRC, and more generally, to compare the outcomes, cost, and cost-effectiveness of a variety of treatment and follow-up strategies.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
A "base-case" analysis (ie, one involving our best estimates [consensus of all authors] for all model parameters and event probabilities) was performed from a societal perspective following the consensus recommendations of the U.S. Department of Health and Human Services Panel on Cost-Effectiveness in Health and Medicine (81–84). Sensitivity analyses were performed across a range of assumptions concerning the number of metastases per patient, the rate of tumor growth, surgical mortality, disease-related mortality and quality of life (QOL), the costs of surgery and patient care, the discount rate used for both costs and life-years, and patient age.

Cost-effectiveness Model
For the analysis, we developed a state-transition Monte Carlo decision model (85–87). With Monte Carlo simulation, it is possible to simulate cohorts of hypothetical patients. For each patient, probabilistic events and outcomes are randomly assigned values from population distributions, and the expected outcomes are determined. The process is repeated for each individual in the cohort, and averages are computed for the entire cohort. The model used in this study was based on a model previously used to evaluate the cost-effectiveness of surgical removal of metastases in patients with liver metastases from CRC (80). The model contains three states: alive _treat, alive_notreat, and dead. All patients begin in the alive_treat state. Patients in the alive_treat state are potential candidates for treatment (ie, tumor ablation or resection). Patients transition to the alive_notreat state when they have either been found to be "untreatable" (ie, to have more metastases than the threshold for treatment in the strategy being considered) or when they have had the maximum allowable number of treatments for the strategy under consideration. The dead state is self-explanatory.

At the end of each cycle, patients return to one of the three model states according to event and transition probabilities defined in the model. This process continues until all patients in the initial cohort reach the dead state, at which point the simulation is terminated. The model includes only one generic imaging and treatment strategy, which is defined by using specific model parameters (eg, treatment threshold, image and possibly treat interval, test sensitivity).

The model tracks up to 15 individual hepatic metastases in each patient, specifying and updating tumor location, size, rate of growth, detection, and removal or ablation. The model was analyzed by simulating cohorts of hypothetical patients, one at a time, from initial presentation until death (1 month cycle length) while tracking disease progression, diagnostic tests, procedures, complications, survival, and costs.

For each hypothetical patient, the number of metastases present is drawn at random from a population distribution. Metastases are then distributed throughout the liver. We assume that each metastasis has an independent, equal probability of being located in each of the eight liver segments (88–90). The size of each metastasis is similarly drawn at random from a population distribution. Over time, metastases that are present grow, may be detected with imaging tests, and may be removed during resection. Growth in tumor volume is assumed to be exponential, at rates determined from the literature (91). We assume that no new liver metastases develop over time, since all patients are assumed to have undergone removal of their primary tumors. However, metastases can be missed at initial diagnosis but can later be detected with repeat imaging tests. Such metastases would account for the "new metastases," which were probably present but undetected earlier, that have been reported to occur in clinical series (92–94).

The frequency with which follow-up imaging tests are performed is explicitly modeled. Each time a patient undergoes a diagnostic imaging test, metastases may be (independently) detected or missed. Tumor detection is based on test sensitivity, although it is assumed that below a certain (definable) size threshold, all metastases are missed. We assume that helical contrast material–enhanced computed tomographic (CT) scanning is used in all patients.

On the basis of the CT results, a decision is made regarding the candidacy of each patient for treatment (or for repeat treatment, in the event of newly detected metastases or local treatment failure). We allow the possibility of different criteria for determining treatment candidacy, depending on the number of metastases identified (less than or equal to one, two, three, four, five, or six) and the number of hepatic segments involved. In patients undergoing surgical removal of metastases, additional metastases may be found at the time of surgery, either by palpation of the liver or by using intraoperative ultrasonography (US). The detection of these additional metastases could influence the decision to proceed with surgical removal of metastases and/or modify the plan for resection.

As a result of treatment, patients are either free of metastatic disease or have residual metastases, depending on the number and location of metastases detected, the location of metastases not detected, and the procedure performed. Patients may also suffer surgical morbidity or mortality. Each of these events is explicitly modeled.

For the current study, several modifications to the surgical-removal-of-metastases model were required. These are described below. A schematic diagram of the revised model is illustrated in Figure 1.



View larger version (33K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Structure of the state-transition model. app = apparent; cand = candidate; Clone 2 = clone of subtree, labeled "2"; comp = complication; intraop = intraoperative; lap = laparotomy; mets = metastases; nomets = no metastases; nores = no resection; p = probability; preop = preoperative; proc = procedure; res = resection; Tx = treatment.

 
Structural Changes to Cost-effectiveness Model
Percutaneous RF ablation differs fundamentally from hepatic resection in several respects. To begin with, ablation targets individual metastases rather than segments containing metastases. Thus, unlike segmental hepatic resection, ablation can at best result in the elimination of specific lesions that are detected and targeted, and will not result in the elimination of undetected lesions, even if those lesions are located in the same segments as lesions that are detected. We therefore modified the model to allow treatment of specific lesions (ie, rather than removal of entire liver segments). We also allowed incomplete treatment of individual lesions because ablation may fail to achieve complete tumor necrosis in some cases. In the event that incomplete tumor necrosis was detected (ie, at a follow-up imaging examination), repeat treatment with RF ablation was permitted, up to a maximum (definable) number of treatment sessions.

Parameter Estimation
Model parameters were estimated from the literature where valid data were available and otherwise consisted of our best estimates (consensus of authors). To determine parameter estimates, we performed a comprehensive review of the English-language literature by using PubMed and reviewing the bibliographies of articles thus identified. Base-case estimates and ranges used in sensitivity analyses are summarized in Table 1.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Parameter Estimates Used in Base-Case and Sensitivity Analyses

 
Our approach to estimating model parameters has been previously described (80). For the current study, several new parameters or modifications of existing parameters were required. For example, surgical strategies are specified on the basis of the maximum number of lesions or segments resected; the maximum number of repeat resections performed; the sensitivity for detecting lesions at surgery; the nature, severity, and frequency of complications; surgical mortality; anticipated length of hospital stay; postoperative recovery time; and all associated costs. These parameters were estimated for ablation. Estimates were made on the basis of values recently reported in the literature (39,48–51,58,59,112–116), otherwise consisted of our best estimates, and are described below.

We estimated the morbidity and mortality probabilities associated with liver tumor ablation to be 0.02 and 0.003, respectively (101). We used the same estimates for the probability of morbidity and mortality for all instances of liver tumor ablation (ie, initial or repeat ablation).

We used 0.85 as our base-case estimate for CT sensitivity. This was increased from the estimate used in our analysis of hepatic resection to reflect the ongoing improvements in CT technology (117). Lower estimates were used in sensitivity analyses.

Costs
Costs were estimated from a societal perspective. Costs were converted to 1998 dollars by using the medical care component of the Consumer Price Index (118). For the base-case analysis, costs and quality-adjusted life-years (QALYs) were discounted at a real annual rate of 3%. We also considered discount rates of 0% and 5% in sensitivity analyses. Base-case estimates for all costs are summarized in Table 2. Our approach to estimating costs for the model has been previously described (80). We describe our approach to deriving costs that were added to the revised model below.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Cost Estimates Used in Base-Case and Sensitivity Analyses

 
Costs for all tests and procedures were estimated on the basis of the Medicare payment schedule. We assumed that all ablation procedures would be performed on an outpatient basis, as is typically the case in the United States. Ambulatory Practice Center codes were therefore used to derive estimates for hospital costs, while Current Procedural Terminology codes were used to derive estimates for physician costs. For RF ablation, the relevant Ambulatory Practice Center code was 0980 and the relevant Current Procedural Terminology code was 47382; both codes represented percutaneous RF ablation in the liver. We used year 2001 codes because no code was available for liver tumor RF ablation in 1998. Cost estimates were then derived by multiplying the relative value units by the 1998 Medicare conversion factor.

We estimated the time required for RF ablation as 1 day, which included the time associated with travel, waiting, the procedure, and postprocedural monitoring. As in our previous analysis of hepatic resection (80), we did not include the costs of transportation, care by family and friends, or household modification related to diagnostic testing or treatment.

We assumed that up to three liver tumors could be treated with ablation on 1 day. Therefore, the total estimated cost of ablation was the same for patients undergoing treatment of one to three metastases (on any single day). For patients undergoing treatment of four to six metastases, it was assumed that two treatment sessions would be required. Cost estimates were adjusted accordingly. The cost of repeat RF ablation was assumed to be the same as the cost of initial therapy.

We estimated that, on average, complications from ablation would require 1 day of hospitalization. It was believed that resource use during this day of hospitalization would be comparable to resource use for patients sustaining laparotomy-related complications. We therefore used the estimated per diem routine (ie, non–intensive care unit) care costs among patients having diagnosis-related group, or DRG, codes 191/192 who underwent hepatic resection at our institution in fiscal year 1998 (that were used to estimate laparotomy costs) to estimate complication costs for ablation. These cost estimates were derived by using software from Eclypsis (previously Transition Systems; Boca Raton, Fla) that interfaces with the hospital accounting system and can access patient records according to DRG and/or Current Procedural Terminology codes, admitting and/or discharge diagnosis, or principal procedure performed, among other criteria.

Treatment Outcomes
Several reports have described the "natural history" of patients with untreated metastases (3,7–9,11,12,14–16). Although these reports have spanned 3 decades and represent the experience at several institutions in different countries, they are remarkably consistent. The single most important determinant of survival appears to be the extent of liver replacement by tumor. Different investigators have reported liver involvement by using different criteria. These include the following: less than 25% versus 25% or more of total LVRT (3,7,15), unilateral versus bilateral metastases (9,11,12,14,15), and fewer than four versus four or more metastases (8,9,16). Each of these criteria is likely to have divided groups of patients along roughly the same lines (eg, patients with unilateral metastases are unlikely to have ≥25% LVRT). Furthermore, estimates of median survival that are based on the different categorizing criteria are similar. We assigned to patients with less than 25% (but more than 0%) LVRT a median survival of 11.5 months, and to patients with LVRT of 25% or more a median survival of 6.3 months on the basis of a weighted average of the studies reporting survival as a function of percentage LVRT (3,7,15). Because we explicitly modeled the number, size, and growth of all metastases, it was possible to calculate percentage LVRT in each patient and to revise it with each cycle of the model. State-to-state transition probabilities for patients with liver metastases were calculated from the corresponding estimated median survivals assuming constant hazard rates (ie, exponential survival curves) (120,121).

For our base-case analysis, we assigned mortality hazard rates to patients without liver metastases (either initially or following successful treatment) that were twice the age- and sex-adjusted rates for the U.S. population (as reported in the 1988 U.S. Life Tables); data from the National Cancer Institute Survival, Epidemiology, and End Results, or SEER, registry were consistent with our estimates of hazard rates for patients with no metastases. To estimate hazard rates for the analyses by using the SEER database, the age- and sex-specific 1-year survival probabilities were used to calculate the hazard rates (by age and sex) for patients with "localized" disease. The hazard rates for patients with localized disease derived from the SEER data are approximately the same as those assigned to patients with no metastases in our base-case analysis. Lower mortality hazard rates were considered in sensitivity analyses; see below for details.

The effectiveness of ablation depends on the size of the lesions treated. We derived the probability of obtaining complete tumor necrosis, according to tumor size, from a study by Solbiati et al (52), in which 117 patients with 179 CRC liver metastases were treated with RF ablation and followed up for up to 5 years. On the basis of the results of this study, we estimated that complete tumor necrosis would be achieved in 78.4%, 47.2%, and 31.6% of tumors 2.5 cm or smaller, larger than 2.5 cm but smaller than or equal to 4 cm, and larger than 4 cm, respectively. For lesions in which only partial necrosis was achieved, the model assigned a postablation tumor volume that was equal to 5% of the pretreatment volume. Imaging and treatment strategies involving in situ ablation were specified to permit multiple treatment sessions. For the base-case analysis, we limited the total number of ablation sessions to three per lesion.

QALYs were used as the primary end point with which the effectiveness of each of the testing strategies was assessed. We assigned years of life without hepatic metastases QALY weights equal to age- and sex-adjusted population values (122). The limited data that are available (123,124) regarding QOL in patients with metastatic CRC suggest that the majority (approximately 95%) of survival is spent with a normal (age-adjusted) QOL. QOL then declines rapidly at the end of life (median time at which QOL declines: last 12 days of life). We therefore modeled QALYs by subtracting a toll in the final cycle (month) before death, estimating that QOL in the month prior to death was 60% of that for age- and sex-matched control subjects. QALY tolls were also assessed to account for decreased QOL following RF ablation, hepatic resection, or laparotomy.

To evaluate the effect of improvements in RF ablation, we created and evaluated two additional scenarios: "improved RF," in which complete tumor necrosis was achieved in 100%, 75%, and 50% of treated tumors 2.5 cm or smaller, larger than 2.5 cm but smaller than or equal to 4 cm, and larger than 4 cm but smaller than or equal to 10 cm, respectively; and "perfect RF," in which complete tumor necrosis was achieved in all treated tumors. For the base-case analysis, patients with one or more tumors larger than 5 cm in diameter were not considered candidates for ablation. Different thresholds for treatment candidacy were evaluated in sensitivity analyses.

Resection strategies were similar to those included in the initial analysis of hepatic resection alone (80), with two exceptions. First, we permitted one repeat resection in all considered strategies. Second, we considered more thresholds for surgical candidacy (less than or equal to one, two, three, four, five, or six metastases seen).

Analyses Performed
The model was analyzed as a first-order Monte Carlo simulation. We simulated cohorts of 100 000 hypothetical patients, one at a time, from initial presentation until death, while tracking disease progression, tests, procedures, complications, survival, and costs. Sample size was determined adaptively to achieve convergence of estimates and separation of strategies in terms of cost, effectiveness, and cost-effectiveness. We modeled several strategies for RF ablation and hepatic resection, including those that have been advocated in the literature or used in practice. Strategies for RF ablation differed in the maximum number of lesions that could be treated, the maximum number of repeat ablations allowed, and the frequency of follow-up imaging. Resection strategies differed in approach (segmental resection vs wedge excision), maximum number of lesions or segments removed, and frequency of follow-up imaging. Ablation and resection strategies were compared in similar cohorts (ie, identically defined target populations), but with allowance for random variation. All comparisons also included costs and outcomes for patients in the entry cohort who were not candidates for the therapy being evaluated.

The base-case analysis was performed by using cohorts of 65-year-old men, a 3% discount rate for both cost and effectiveness, and estimates for costs, treatment effectiveness, and other event probabilities as described above and in Table 1. Included in the comparison is a reference strategy in which neither ablation nor resection is offered and thus follow-up CT is not performed (the "no-treat" strategy). For both RF ablation and hepatic resection, we simulated all possible combinations of follow-up interval (4, 6, or 12 months) and treatment threshold (less than or equal to one, two, three, four, five, or six metastases seen). For each strategy, we calculated total costs and QALYs and compared strategies by using incremental cost-effectiveness ratios (ICERs). Strategies that cost more but yielded lower total benefits—that is, they were "dominated"—and those that had a higher ICER than the next most costly and effective strategy—that is, strategies that were "weakly dominated"—were eliminated.

Extensive sensitivity analysis was performed to investigate the effects of changes in model parameters on estimated costs and effectiveness. The following parameters were varied: patient age and sex; mortality hazard rates (for patients with and those without metastases); number of metastases; tumor size and doubling time; diagnostic test sensitivity; maximum tumor size treatable with RF ablation; success in achieving local control of treated lesions; costs of ablation, resection, and laparotomy; costs of patient care; and the discount rate used for costs and QALYs. Additional analyses in which several of the parameters were simultaneously varied were performed (Tables 1 and 2).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Base-Case Analysis
The results of the base-case analysis are summarized in Figure 2 and Table 3. After dominated and weakly dominated strategies (defined in Materials and Methods) were eliminated, five strategies remained. These are presented in order of increasing ICER. For example, a strategy involving a 12-month follow-up interval and ablation of six or fewer metastases had an ICER of $1300 per QALY, and a strategy involving a 4-month follow-up interval and resection of six or fewer metastases had an ICER of $31 200 per QALY (compared with the next most effective nondominated strategy). Although these data are not included in Table 3, the ICER of this same resection strategy relative to the most effective nondominated ablation strategy (ie, ablation of up to six metastases with 12-month follow-up) was $17 800 per QALY.



View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Results of base-case analysis showing average cost (in dollars) and effectiveness (in QALYs) of all strategies.

 

View this table:
[in this window]
[in a new window]

  		TABLE 3. Results of Base-Case Analysis: Resection versus Ablation

 
Sensitivity Analysis
The results of the sensitivity analysis are summarized in Table 4. We list ICERs for two clinically relevant comparisons: (a) RF ablation with a treatment threshold of five or fewer metastases seen and a 4-month follow-up interval (denoted "RF") versus a no-treat strategy, and (b) the same RF strategy (five or fewer metastases, 4-month follow-up) versus resection by using an operative threshold of four or fewer metastases seen and a 6-month follow-up interval (denoted "resection"). Note that this presentation differs from the base-case analysis, which involved comparison of all nondominated strategies.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Results of Sensitivity Analysis

 
In Table 4, we present the ICERs of two comparisons of interest for each strategy, for each of the three scenarios concerning local control rates for RF ablation (ie, base case, improved RF, and perfect RF); average costs and QALYs for the management strategies are not presented. Some general patterns emerge. In all cases, the ICER of RF ablation versus the no-treat strategy was less than $5000 per QALY, and, with very few exceptions, the ICER of resection versus RF ablation was less than $30 000 per QALY.

Across all scenarios, when all possible RF ablation and resection strategies were considered (results not shown), more aggressive surgical strategies were superior (ie, resulted in more QALYs gained) to RF ablation. In most cases, the ICERs of even the most aggressive surgical strategies were less than $35 000 per QALY. An additional finding was that some of the less aggressive RF ablation strategies (eg, treatment of one to two metastases) were both less expensive and more effective than the no-treat strategy.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
REFERENCES
 
The results of our study suggest that both RF ablation and hepatic resection are relatively cost-effective management strategies for patients with limited hepatic metastases from CRC. In some instances, particularly with more restrictive selective criteria for treatment candidacy (ie, those based on the number of lesions detected), ablation was found to be cost saving relative to no treatment. However, throughout the analysis, more aggressive strategies (ie, the treatment of patients with more tumors, more frequent posttreatment follow-up regimens, and surgery rather than ablation) were superior to more conservative strategies. For example, a strategy permitting ablation of up to five metastases with CT follow-up every 4 months resulted in a gain of 0.65 QALY relative to the no-treat strategy, at an incremental cost of $2400 per QALY. Compared with this ablation strategy, a strategy permitting resection of up to four metastases, one repeat resection, and CT follow-up every 6 months resulted in an additional gain of 0.76 QALY at an incremental cost of $24 300 per QALY.

These results suggest that physicians performing ablation and resection should be encouraged to select patients for therapy on the basis of technical feasibility rather than numerical thresholds and to pursue repeat treatment when new lesions are identified after initial therapy. Our results suggest that increasing the threshold for treatment candidacy could result in moderate population-wide QALY gains.

For example, in our base-case scenario, moving from a strategy of treating (with ablation) only one metastasis with 12-month follow-up to a strategy of treating up to six metastases with 4-month follow-up increased the quality-adjusted life expectancy (averaged across the entire patient cohort, including both patients who were and patients who were not candidates for therapy) from 0.8381 to 1.3270. Applying this to a population of 10 000 patients would represent a gain of 489 QALYs. When resection is considered, the resulting gain would be even greater: 2516 QALYs. It should also be noted that there are generally no other potentially curative therapeutic options for these patients. Furthermore, the results suggest that RF ablation of larger lesions should be encouraged, even though local control rates (on an individual-lesion basis) diminish as the size of the target tumor increases. Raising the size threshold for treatment permits more patients to be considered candidates for therapy and results in substantial QALY gains across the population.

The results were relatively insensitive to changes in most model parameters. However, the results were sensitive to changes in the number of metastases present (ie, the assumed population distribution), the size threshold above which lesions were considered untreatable with RF ablation, and the cost of RF ablation. More generally, any changes in model parameters that increased the effectiveness of RF resulted in an increase in the ICER for resection relative to RF ablation. Furthermore, as local control rates were increased (ie, while moving from the base-case assumptions to perfect RF), the ICER of RF ablation versus "no-test/no-treat" decreased, and the ICER of resection versus RF ablation increased. Improved local control results in longer survival and fewer repeat procedures, both of which tended to favor RF ablation in the comparisons. Of note, however, the magnitude of these changes was not always large. For example, in our base-case scenario (CT sensitivity 0.85), moving from base-case to perfect RF only resulted in a gain of 0.0297 QALY and had a correspondingly small effect on the ICER of resection versus ablation.

Increasing the number of metastases present reduced the number of patients in the cohort who were candidates for therapy and thus decreased the average effectiveness, cost, and total gain in life expectancy (that might be expected across a population) of even the most aggressive management strategy. However, given base-case assumptions regarding local control rates, this reduction in the number of treatment candidates had little effect on the ICER of ablation versus the no-treat reference strategy. As local control rates were increased (ie, when moving to improved RF and perfect RF), the effect of reduced candidacy rates was to decrease the ICER of ablation versus the no-treat strategy, as compared with the base-case analysis results. In essence, improved RF effectiveness resulted in both longer life expectancy and lower costs. These effects (ie, the benefits associated with achieving local control) were more pronounced as the number of metastases per patient increased.

In contrast, increasing the number of metastases present substantially increased the ICER of resection versus ablation. Because fewer patients would be candidates for resection, and because postoperative recurrence rates would increase with more metastases per patient, gains in QALYs across the cohort were more expensive to achieve. It is important to note, however, that even given the assumption that the average number of metastases per patient is 10 (as opposed to six in the base-case scenario), the ICER of the reference ablation strategy (relative to the no-treat strategy) was $3500 per QALY, and the ICER of the reference resection strategy (relative to the ablation strategy) was $47 500 per QALY.

Increasing the size threshold for treatment with RF ablation resulted in more patients being eligible for treatment and in greater QALY gains across the cohort of patients treated with RF. It also substantially increased the ICER of resection relative to RF ablation because the cost and effectiveness of surgery were not affected. It is important to note that these results were generally consistent across different scenarios (ie, base case, improved RF, perfect RF).

The sensitivity to the costs of RF ablation is not particularly surprising: Increasing or decreasing these costs would be expected to affect cost-effectiveness ratios. However, even when the cost of RF ablation is increased by 50%, a strategy permitting ablation of five or fewer metastases with 4-month follow-up had an ICER of less than $5000 per QALY relative to the no-treat reference strategy.

The principal limitations of our study were related to the necessary simplification of reality that was required in developing a tractable model, as well as to the uncertainty that surrounds some of our parameter estimates. For example, we did not specifically model chemotherapy costs. As in our prior study, we used aggregate estimates for the costs of caring for patients that were based on their disease extent. Therefore, it was not possible to investigate the effects of chemotherapy costs (alone) on cost-effectiveness. However, our sensitivity analysis examined the effect of decreasing overall patient care costs by 50%, and results were similar to those of the base-case scenario. We also did not specifically model extrahepatic metastases, and, although our results appear consistent with those of reported clinical series, one must be careful not to generalize to populations that are dissimilar—for example, one in which there is a much higher incidence of extrahepatic metastases.

With respect to parameter uncertainty, we attempted to provide reasonable estimates for each of the specific model parameters and probability distributions and to justify these estimates with references to the published literature and by validating the results of our simulations against published results from clinical trials. The greatest parameter uncertainty undoubtedly relates to life expectancy (and therefore our transition probabilities) in patients with no metastases (see above), and, to a lesser extent, in patients with limited metastases. Few would dispute that patients with extensive liver metastases have an extremely poor prognosis. Our estimated life expectancy in these patients (ie, those with ≥25% LVRT) of 6.3 months is supported by results of several published studies (3,7,16). However, patients who undergo hepatic resection are a highly selected subgroup of all patients with metastatic CRC, and it is difficult to estimate how long these patients might live without surgery. Our approach, by modeling the actual size and growth of individual metastases and basing life expectancy on percentage LVRT seems reasonable; the only question is whether or not the hazard rate assigned to patients with limited metastases is too high. We addressed this issue through sensitivity analysis, in which the hazard rate in patients with metastases was increased or decreased by 50%, and found little change in results compared with the base-case analysis.

Finally, there is certainly heterogeneity across patients with respect to virtually all model parameters, particularly those relating to outcome and cost. We did not attempt to assess the potential impact of population heterogeneity on our results because of the unavailability of data on which to base range estimates. In addition, through second-order Monte Carlo simulation ("probabilistic sensitivity analysis") it may be possible to assess the global impact of parametric uncertainty on model results. However, superimposing a second-order simulation on our first-order Monte Carlo model was not possible in its current form. Additional data collection and analysis in these areas may be of use in the future.

In conclusion, the results of this study suggest that RF ablation and hepatic resection are relatively cost-effective management strategies for patients with limited hepatic metastases from CRC. Our results further suggest that physicians should be relatively aggressive in their approach to managing these patients and specifically suggest that aggressive surgical approaches are likely to result in substantial QALY gains at a reasonable cost.


   FOOTNOTES
 
Abbreviations: CRC = colorectal carcinoma, ICER = incremental cost-effectiveness ratio, LVRT = liver volume replaced by tumor, QALY = quality-adjusted life-year, RF = radiofrequency, QOL = quality of life

The information does not necessarily reflect the position of the government, and no endorsement should be inferred.

Authors stated no financial relationship to disclose.

Author contributions: Guarantor of integrity of entire study, G.S.G.; study concepts, G.S.G.; study design, G.S.G., M.C.W.; literature research, G.S.G., P.M.M.; data acquisition, G.S.G., P.M.M., M.T.B.; data analysis/interpretation, all authors; statistical analysis, G.S.G., E.F.H.; manuscript preparation, G.S.G.; manuscript definition of intellectual content, editing, revision/review, and final version approval, all authors


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 

  1. American Cancer Society. Cancer facts and figures Washington, DC: American Cancer Society, 2000.
  2. Bengmark S, Hafstrom L. The natural history of primary and secondary malignant tumors of the liver. Cancer 1969; 23:198-202.[CrossRef][Medline]
  3. Bengtsson G, Carlsson G, Hafstrom L, Jonsson PE. Natural history of patients with untreated liver metastases from colorectal cancer. Am J Surg 1981; 141:586-589.[CrossRef][Medline]
  4. Abrams MS, Lerner HJ. Survival of patients at Pennsylvania Hospital with hepatic metastases from carcinoma of the colon and rectum. Dis Colon Rectum 1971; 14:431-434.[Medline]
  5. Cady B, Monson DO, Swinton NW. Survival of patients after colonic resection for carcinoma with simultaneous liver metastases. Surg Gynecol Obstet 1970; 131:697-700.[Medline]
  6. Baden H, Andersen B. Survival of patients with untreated liver metastatases from colorectal cancer. Scand J Gastroenterol 1975; 10:221-223.[Medline]
  7. Finan PJ, Marshall RJ, Cooper EH, Giles GR. Factors affecting survival in patients presenting with synchronous hepatic metastases from colorectal cancer: a clinical and computer analysis. Br J Surg 1985; 72:373-377.[Medline]
  8. Goslin R, Steele G, Zamcheck N, Mayer R, MacIntyre J. Factors influencing survival in patients with hepatic metastases from adenocarcinoma of the colon or rectum. Dis Colon Rectum 1982; 25:749-754.[Medline]
  9. Lahr CJ, Soong SJ, Cloud G, Smith JW, Urist MM, Balch CM. A multifactorial analysis of prognostic factors in patients with liver metastases from colorectal carcinoma. J Clin Oncol 1983; 1:720-726.[Abstract]
  10. Registry of Hepatic Metastases. Resection of the liver for colorectal carcinoma metastases: a multi-institutional study of indications for resection. Surgery 1988; 103:278-288.[Medline]
  11. Wagner JS, Adson MA, Van Heerden JA, Adson MH, Ilstrup DM. The natural history of hepatic metastases from colorectal cancer. Ann Surg 1984; 199:502-507.[Medline]
  12. Wood CB, Gillis CR, Blumgart LH. A retrospective study of the natural history of patients with liver metastases from colorectal cancer. Clin Oncol 1976; 2:285-288.[Medline]
  13. Jaffe BM, Donegan WL, Watson F, Spratt JS. Factors influencing survival in patients with untreated hepatic metastases. Surg Gynecol Obstet 1968; 127:1-11.[Medline]
  14. Boey J, Choi TK, Wong J, Ong GB. Carcinoma of the colon and rectum with liver involvement. Surg Gynecol Obstet 1981; 153:864-868.[Medline]
  15. Stangl R, Altendorf-Hofmann A, Charnely RM, Scheele J. Factors influencing the natural history of colorectal liver metastases. Lancet 1994; 343:1405-1410.[CrossRef][Medline]
  16. Scheele J, Stangl R, Altendorf-Hofmann A. Hepatic metastases from colorectal carcinoma: impact of surgical resection on the natural history. Br J Surg 1990; 77:1241-1246.[Medline]
  17. Oxley EM, Ellis H. Prognosis of carcinoma of the large bowel in the presence of liver metastases. Br J Surg 1969; 56:149-152.[Medline]
  18. Saenz NC, Cady B, McDermott WV, Steele GD. Experience with colorectal carcinoma metastatic to the liver. Surg Clin North Am 1989; 69:361-370.[Medline]
  19. Fegiz G, Ramacciato G, D’Angelo F, et al. Patient selection and factors affecting results following resection for hepatic metastases from colorectal carcinoma. Int Surg 1991; 76:58-63.[Medline]
  20. Adson MA, VanHeerden JA, Adson MH, Wayner JS, Ilstrup DM. Resection of hepatic metastases from colorectal cancer. Arch Surg 1984; 119:647-651.[Abstract]
  21. Foster JH. Survival after liver resection for secondary tumors. Am J Surg 1978; 135:389-394.[CrossRef][Medline]
  22. Thompson HH, Tompkins RK, Longmire WP. Major hepatic resection. Ann Surg 1983; 197:375-388.[Medline]
  23. Butler J, Attiyeh FF, Daly JM. Hepatic resection for metastases of the colon and rectum. Surg Gynecol Obstet 1986; 162:109-113.[Medline]
  24. Fortner JG, Silva JS, Golbey RB, Cox EB, Maclean BJ. Multivariate analysis of a personal series of 247 consecutive patients with liver mestastases from colorectal cancer. Ann Surg 1984; 199:306-316.[Medline]
  25. Hughes KS, Rosenstein RB, Songhorabodi S, et al. Resection of the liver for colorectal carcinoma metastases: a multi-institutional study of long-term survivors. Dis Colon Rectum 1988; 31:1-4.[Medline]
  26. Cobourn CS, Makowka L, Langer B, Taylor BR, Falk RE. Examination of patient selection and outcome for hepatic resection for metastatic disease. Surg Gynecol Obstet 1987; 165:239-246.[Medline]
  27. Nordlinger B, Parc R, Delva E, Quilichini MA, Hannoun L, Huguet C. Hepatic resection for colorectal liver metastases. Ann Surg 1987; 205:256-263.[Medline]
  28. Schlag P, Hohenberger P, Herfarth C. Resection of liver metastases in colorectal cancer: competitive analysis of treatment results in synchronous versus metachronous metastases. Eur J Surg Oncol 1990; 16:360-365.[Medline]
  29. Younes RN, Rogatko A, Brennan MF. The influence of intraoperative hypotension and perioperative blood transfusion on disease-free survival in patients with complete resection of colorectal liver metastases. Ann Surg 1991; 214:107-113.[Medline]
  30. Doci R, Gennari L, Bignami P, Montalto F, Morabito A, Bozzetti F. One hundred patients with hepatic metastases from colorectal cancer treated by resection: an anlysis of prognostic determinants. Br J Surg 1991; 78:797-801.[Medline]
  31. Scheele J, Stangl R, Altendorf-Hofmann A, Gall FP. Indicators of prognosis after hepatic resection for colorectal secondaries. Surgery 1991; 110:13-29.[Medline]
  32. Scheele J, Stangl R, Altendorf-Hofmann A, Paul M. Resection of colorectal liver metastases. World J Surg 1995; 19:59-71.[CrossRef][Medline]
  33. Rosen CB, Nagorney DM, Taswell HF, et al. Perioperative blood transfusion and determinants of survival after liver resection for metastatic colorectal carcinoma. Ann Surg 1992; 216:493-505.[Medline]
  34. Fong Y, Kemeny N, Paty P, Blumgart LH, Cohen A. Treatment of colorectal cancer: hepatic metastasis. Semin Surg Oncol 1996; 12:219-252.[CrossRef][Medline]
  35. Fong Y, Cohen AM, Fortner JG, et al. Liver resection for colorectal metastases. J Clin Oncol 1997; 15:938-946.[Abstract/Free Full Text]
  36. Lind DS, Parker GA, Horsley JS, et al. Formal hepatic resection of colorectal liver metastases: ploidy and prognosis. Ann Surg 1992; 215:677-684.[Medline]
  37. Petrelli N, Gupta B, Piedmonte M, Herrera L. Morbidity and survival of liver resection for colorectal adenocarcinoma. Dis Colon Rectum 1991; 34:899-904.[CrossRef][Medline]
  38. Petrelli NJ, Nambisan RN, Herrera L, Mittelman A. Hepatic resection for isolated metastasis from colorectal carcinoma. Am J Surg 1985; 149:205-209.[CrossRef][Medline]
  39. Livraghi T, Goldberg SN, Monti F, et al. Saline-enhanced radiofrequency tissue ablation in the treatment of liver metastases. Radiology 1997; 202:205-210.[Abstract/Free Full Text]
  40. Goldberg SN, Gazelle GS, Dawson SL, Rittman WJ, Mueller PR, Rosenthal DI. Tissue ablation with radiofrequency: effect of probe size, gauge, duration, and temperature on lesion volume. Acad Radiol 1995; 2:399-404.[CrossRef][Medline]
  41. Goldberg SN, Gazelle GS, Dawson SL, Rittman WJ, Mueller PR, Rosenthal DI. Tissue ablation with radiofrequency using multiprobe arrays. Acad Radiol 1995; 2:670-674.[Medline]
  42. Goldberg SN, Gazelle GS, Halpern EF, Rittman WJ, Mueller PR, Rosenthal DI. Radiofrequency tissue ablation: importance of local temperature along the electrode tip exposure in determining lesion size and shape. Acad Radiol 1996; 3:212-218.[CrossRef][Medline]
  43. Goldberg SN, Gazelle GS, Solbiati L, Rittman WJ, Mueller PR. Radiofrequency tissue ablation: increased lesion diameter with a perfusion electrode. Acad Radiol 1996; 3:636-644.[CrossRef][Medline]
  44. Goldberg SN, Hahn PF, Tanabe KK, et al. Percutaneous radiofrequency tissue ablation: does perfusion-mediated tissue cooling limit coagulation necrosis? J Vasc Interv Radiol 1998; 9:101-111.[Medline]
  45. Goldberg SN, Gazelle GS, Solbiati L, Livraghi T, Tanabe KK, Mueller PR. Ablation of liver tumors using percutaneous RF therapy. AJR Am J Roentgenol 1998; 170:1023-1028.[Free Full Text]
  46. Goldberg SN, Solbiati L, Hahn PF, et al. Radiofrequency tumor ablation using a clustered electrode technique: results in animals and patients with liver metastases. Radiology 1998; 209:371-379.[Abstract/Free Full Text]
  47. Goldberg SN, Hahn PF, Halpern EF, Fogle R, Gazelle GS. Radiofrequency tissue ablation: effect of pharmacologic modulation of blood flow on coagulation diameter. Radiology 1998; 209:761-767.[Abstract/Free Full Text]
  48. Solbiati L, Ierace T, Goldberg SN, et al. Percutaneous US-guided RF tissue ablation of liver metastases: results of treatment and follow-up in 16 patients. Radiology 1997; 202:195-203.[Abstract/Free Full Text]
  49. Solbiati L, Goldberg SN, Ierace T, et al. Hepatic metastases: percutaneous radio-frequency ablation with cooled-tip electrodes. Radiology 1997; 205:367-373.[Abstract/Free Full Text]
  50. Solbiati L, Goldberg SN, Ierace T, Dellanoce M, Livraghi T, Gazelle GS. Long-term follow-up of liver metastases treated with percutaneous US-guided radiofrequency (RF) ablation using internally-cooled electrodes (abstr). Radiology 1998; 209(P):449.[Abstract/Free Full Text]
  51. Solbiati L, Goldberg SN, Ierace T, Dellanoce M, Livraghi T, Gazelle GS. Radiofrequency ablation of hepatic metastases: postprocedural assessment with a microbubble ultrasound contrast agent—early experience. Radiology 1999; 211:643-649.[Abstract/Free Full Text]
  52. Solbiati L, Livraghi T, Goldberg SN, et al. Percutaneous radio-frequency ablation of hepatic metastases from colorectal cancer: long-term results in 117 patients. Radiology 2001; 221:159-166.[Abstract/Free Full Text]
  53. Livraghi T, Goldberg SN, Lazzaroni S, Meloni F, Solbiati L, Gazelle GS. Small hepatocellular carcinoma: treatment with radio-frequency ablation versus ethanol injection. Radiology 1999; 210:655-661.[Abstract/Free Full Text]
  54. Livraghi T, Goldberg SN, Lazzaroni S, et al. Hepatocellular carcinoma: radio-frequency ablation of medium and large lesions. Radiology 2000; 214:761-768.[Abstract/Free Full Text]
  55. McGahan JP, Wei-Zhong G, Brock JM, Tesluk H, Jones CD. Hepatic ablation using bipolar radiofrequency electrocautery. Acad Radiol 1996; 3:418-422.[CrossRef][Medline]
  56. Lorentzen T, Christensen NE, Nolsle CP, Torp-Pedersen ST. Radiofrequency tissue ablation with a cooled needle in vitro: ultrasonography, dose response, and lesion temperature. Acad Radiol 1997; 4:292-297.[CrossRef][Medline]
  57. Leveen RF. Laser hyperthermia and radiofrequency ablation of hepatic lesions. Semin Interv Radiol 1997; 14:313-324.
  58. Jiao LR, Hansen PD, Havlik R, Mitry RR, Pignatelli M, Habib N. Clinical short-term results of radiofrequency ablation in primary and secondary liver tumors. Am J Surg 1999; 177:303-306.[CrossRef][Medline]
  59. Francica G, Marone G. Ultrasound-guided percutaneous treatment of hepatocellular carcinoma by radiofrequency hyperthermia with a "cooled-tip needle": a preliminary clinical experience. Eur J Ultrasound 1999; 9:145-153.[CrossRef][Medline]
  60. McGahan JP, Browning PD, Brock JM, Tesluk H. Hepatic ablation using radiofrequency electrocautery. Invest Radiol 1990; 25:267-270.[Medline]
  61. Buscarini L, Buscarini E, DiStasi M, Quaretti P, Zangrandi A. Percutaneous radiofrequency thermal ablation combined with transcatheter arterial embolization in the treatment of large hepatocellular carcinoma. Ultraschall Med 1999; 20:47-53.[CrossRef][Medline]
  62. Curley SA, Izzo F, Delria P, et al. Radiofrequency ablation of unresectable primary and metastic hepatic malignancies: results in 123 patients. Ann Surg 1999; 230:1-8.[CrossRef][Medline]
  63. Bilchik AJ, Rose DM, Allegra DP, Bostick PJ, Hsueh E, Morton DL. Radiofrequency ablation: a minimally invasive technique with multiple applications. Cancer J Sci Am 1999; 5:356-361.[Medline]
  64. Goldberg SN, Saldinger PR, Gazelle GS, et al. Percutaneous tumor ablation: increased coagulation necrosis with combined radiofrequency ablation and intratumoral doxorubicin injection in a rat breast tumor model. Radiology 2001; 220:420-427.[Abstract/Free Full Text]
  65. McGhana JP, Dodd GD, 3rd. Radiofrequency ablation of the liver: current status. AJR Am J Roentgenol 2001; 176:3-16.[Free Full Text]
  66. Ahmed M, Lobo SM, Weinstein J, et al. Improved coagulation with saline solution pretreatment during radiofrequency tumor ablation in a canine model. J Vasc Interv Radiol 2002; 13:717-724.[Medline]
  67. Monsky WL, Kruskal JB, Lukyanov AN, et al. Radio-frequency ablation increases intratumoral liposomal doxorubicin accumulation in a rat breast tumor model. Radiology 2002; 224:823-829.[Abstract/Free Full Text]
  68. Livraghi T, Solbiati L, Meloni F, Ierace T, Goldberg SN, Gazelle GS. Percutaneous radiofrequency ablation of liver metastases in potential candidates for resection: the "test of time" approach. Cancer 2003; 97:3027-3035.[CrossRef][Medline]
  69. Ravikumar TS, Steele G, Kane R, King V. Experimental and clinical observations on hepatic cryosurgery for colorectal metastases. Cancer Res 1991; 51:6323-6327.[Medline]
  70. Ravikumar TS, Steele GD. Hepatic cryosurgery. Surg Clin North Am 1989; 69:433-440.[Medline]
  71. Silverman SG, Tuncali K, Adams DF, et al. MR imaging-guided percutaneous cryotherapy of liver tumors: initial experience. Radiology 2000; 217:657-664.[Abstract/Free Full Text]
  72. Lee FT, Jr, Chosy SG, Littrup PJ, Warner TF, Kuhlman JE, Mahvi DM. CT-monitored percutaneous cryoablation in a pig liver model: pilot study. Radiology 1999; 211:687-692.[Abstract/Free Full Text]
  73. Kuszyk BS, Boitnott JK, Choti MA, et al. Local tumor recurrence following hepatic cryoablation: radiologic-histopathologic correlation in a rabbit model. Radiology 2000; 217:477-486.[Abstract/Free Full Text]
  74. Amin Z, Lees WR. Local treatment of colorectal liver metastases: a comparison of interstitial laser photocoagulation (ILP) and percutaneous alcohol injection (PAI). Clin Radiol 1993; 48:166-171.[CrossRef][Medline]
  75. Bosman S, Phoa SSK, van Gemert MJC. Effect of percutaneous interstitial thermal laser on normal liver of pigs: sonographic and histopathological correlations. Br J Surg 1991; 78:572-575.[Medline]
  76. Godlewski G, Rouy S, Pignodel C, et al. Deep localized neodymium (Nd)-YAG laser photocoagulation in liver using a new water cooled and echoguided handpiece. Lasers Surg Med 1988; 8:501-509.[Medline]
  77. Vogl TJ, Mack MG, Roggan A, et al. Internally cooled power laser for MR-guided interstitial laser-induced thermotherapy of liver lesions: initial clinical results. Radiology 1998; 209:381-385.[Abstract/Free Full Text]
  78. Vogl TJ, Straub R, Eichler K, Woitaschek D, Mack MG. Malignant liver tumors treated with MR imaging-guided laser-induced thermotherapy: experience with complications in 899 patients (2,520 lesions). Radiology 2002; 225:367-377.[Abstract/Free Full Text]
  79. Pacella CM, Bizzarri G, Magnolfi F, et al. Laser thermal ablation in the treatment of small hepatocellular carcinoma: results in 74 patients. Radiology 2001; 221:712-720.[Abstract/Free Full Text]
  80. Gazelle GS, Hunink MG, Kuntz K, et al. Cost-effectiveness of hepatic metastasectomy in patients with metastatic colorectal carcinoma: a state-transition Monte Carlo decision analysis. Ann Surg 2003; 237:544-555.[CrossRef][Medline]
  81. Russell LB, Gold MB, Siegel JE, Daniels N, Weinstein MC. The role of cost-effectiveness analysis in health and medicine. JAMA 1996; 276:1172-1177.[Abstract]
  82. Siegel JE, Weinstein MC, Russell LB, Gold MR. Recommendations for reporting cost-effectiveness analyses. JAMA 1996; 276:1339-1341.[Abstract]
  83. Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the panel on cost-effectiveness in health and medicine. JAMA 1996; 276:1253-1258.[Abstract]
  84. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health and medicine New York, NY: Oxford University Press, 1996.
  85. Beck JR, Pauker SG. The Markov process in medical prognosis. Med Decis Making 1983; 3:419-458.
  86. Sonnenberg FA, Beck JR. Markov models in medical decision making: a practical guide. Med Decis Making 1993; 13:322-338.
  87. Briggs A, Sculpher M. An introduction to Markov modelling for economic evaluation. Pharmacoeconomics 1998; 13:397-409.[CrossRef][Medline]
  88. Dionne L. The pattern of blood-borne metastasis from carcinoma of rectum. Cancer 1965; 18:775-781.[CrossRef][Medline]
  89. Schulz W, Hagen C, Hort W. The distribution of liver metastases from colonic cancer: a quantitative postmortem study. Virchows Arch A Pathol Anat Histopathol 1985; 406:279-284.[CrossRef][Medline]
  90. Strohmeyer T, Schultz W. The distribution of metastases of different primary tumors in the liver. Liver 1986; 6:184-187.[Medline]
  91. Finlay IG, Meek D, Brunton F, McArdle CS. Growth rate of hepatic metastases in colorectal carcinoma. Br J Surg 1988; 75:641-644.[Medline]
  92. Nordlinger B, Vaillant JC, Guiguet M, et al. Survival benefit of repeat liver resections for recurrent colorectal metastases: 143 cases. J Clin Oncol 1994; 12:1491-1496.[Abstract]
  93. Hughes KS, Simon R, Songhorabodi S, et al. Resection of the liver for colorectal carcinoma metastases: a multi-institutional study of patterns of recurrence. Surgery 1986; 100:278-284.[Medline]
  94. Ekberg H, Tranberg KG, Andersson R, et al. Pattern of recurrence in liver resection for colorectal secondaries. World J Surg 1987; 11:541-547.[CrossRef][Medline]
  95. Urata K, Kawasaki S, Matsunami H, et al. Calculation of child and adult standard liver volume for liver transplantation. Hepatology 1995; 21:1317-1321.[CrossRef][Medline]
  96. Steele G, Bleday R, Mayer RJ, Lindblad A, Petrelli N, Weaver D. A prospective evaluation of hepatic resection for colorectal carcinoma metastases to the liver: gastrointestinal tumor study group protocol 6584. J Clin Oncol 1991; 9:1105-1112.[Abstract]
  97. Iwatsuki S, Esquivel C, Gordon RD, Starzl TE. Liver resection for metastatic colorectal cancer. Surgery 1986; 100:804-810.[Medline]
  98. Holm A, Bradley E, Aldrete JS. Hepatic resection of metastasis from colorectal carcinoma: morbidity, mortality, and pattern of recurrence. Ann Surg 1989; 209:428-434.[Medline]
  99. Nakamura S, Suzuki S, Baba S. Resection of liver metastases of colorectal carcinoma. World J Surg 1997; 21:741-747.[CrossRef][Medline]
  100. Fegiz G, Ramacciato G, Gennari L, et al. Hepatic resections for colorectal metastases: the Italian multicenter experience. J Surg Oncol Suppl 1991; 2:144-154.[Medline]
  101. Livraghi T, Solbiati L, Meloni MF, Gazelle GS, Halpern EF, Goldberg SN. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications encountered in a multicenter trial. Radiology 2003; 226:441-451.[Abstract/Free Full Text]
  102. Wernecke K, Rummeny E, Bongartz G, et al. Detection of hepatic masses in patients with carcinoma: comparative sensitivities of sonography, CT, and MR imaging. AJR Am J Roentgenol 1991; 157:731-739.[Abstract/Free Full Text]
  103. Knol JA, Marn CS, Francis IR, Rubin JM, Bromberg J, Chang AE. Comparisons of dynamic infusion and delayed computed tomography, intraoperative ultrasound, and palpation in the diagnosis of liver metastases. Am J Surg 1993; 165:81-88.[CrossRef][Medline]
  104. Valls C, Lopez E, Gumà A, et al. Helical CT versus CT arterial portography in the detection of hepatic metastasis of colorectal carcinoma. AJR Am J Roentgenol 1998; 170:1341-1347.[Abstract/Free Full Text]
  105. Matsui O, Takashima T, Kadoya M. Liver metastases from colorectal cancers: detection with CT during arterial portography. Radiology 1987; 165:65-69.[Abstract/Free Full Text]
  106. Clarke MP, Kane RA, Steele G, et al. Prospective comparison of preoperative imaging and intraoperative ultrasonography in the detection of liver tumors. Surgery 1989; 106:849-855.[Medline]
  107. Heiken JP, Weyman PJ, Lee JK, et al. Detection of focal hepatic masses: prospective evaluation with CT, delayed CT, CT during arterial portography, and MR imaging. Radiology 1989; 171:47-51.[Abstract/Free Full Text]
  108. Zerhouni EA, Rutter C, Hamilton SR, et al. CT and MR imaging in the staging of colorectal carcinoma: report of the Radiology Diagnostic Oncology Group II. Radiology 1996; 200:443-451.[Abstract/Free Full Text]
  109. Parker GA, Lawrence WJ, Horsley S, et al. Intraoperative ultrasound of the liver affects operative decision making. Ann Surg 1989; 209:569-577.[Medline]
  110. Machi J, Isomoto H, Yamashita Y, Kurohiji T, Shirouzu K, Kakegawa T. Intraoperative ultrasonography in screening for liver metastases from colorectal cancer: comparative accuracy with traditional procedures. Surgery 1987; 101:678-684.[Medline]
  111. Onik G, Kane R, Steele GJ, et al. Society of Gastrointestinal Radiologists Roscoe E. Miller Award: monitoring hepatic cryosurgery with sonography. AJR Am J Roentgenol 1986; 147:665-669.
  112. Lewin JS, Connell CF, Duerk JL, et al. Interactive MR-guided radiofrequency interstitial thermal ablation of abdominal tumors: clinical trial for evaluation of safety and feasibility. J Magn Reson Imaging 1998; 8:40-47.[Medline]
  113. Siperstein AE, Rogers SJ, Hansen PD, Gitomirsky A. Laparoscopic thermal ablation of hepatic neuroendocrine tumor metastases. Surgery 1997; 122:1147-1155.[CrossRef][Medline]
  114. Rossi S, DiStasi M, Buscarini E, et al. Percutaneous RF interstitial thermal ablation in the treatment of hepatic cancer. AJR Am J Roentgenol 1996; 167:759-768.[Abstract/Free Full Text]
  115. Rossi S, Buscarini E, Garbagnati F, et al. Percutaneous treatment of small hepatic tumors by an expandable RF needle electrode. AJR Am J Roentgenol 1998; 170:1015-1022.[Abstract/Free Full Text]
  116. Elias D, Debaere T, Muttillo I, Cavalcanti A, Coyle C, Roche A. Intraoperative use of radiofrequency treatment allows an increase in the rate of curative liver resection. J Surg Oncol 1998; 67:190-191.[CrossRef][Medline]