HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Staunton, M. Articles by Malone, D. E. Search for Related Content PubMed PubMed Citation Articles by Staunton, M. Articles by Malone, D. E.

Finding Evidence-based Answers to Practical Questions in Radiology: Which Patients with Inoperable Hepatocellular Carcinoma Will Survive Longer after Transarterial Chemoembolization?¹

¹ From the Department of Radiology (M.S., J.D.D., D.E.M.) and the Irish National Liver Unit (P.A.M.), St Vincent's University Hospital, Elm Park, Dublin 4, Ireland; and Department of Radiology, Hamilton Health Sciences Corp, McMaster University Medical Centre, Hamilton, Ontario, Canada (M.S.). Received January 10, 2004; revision requested March 11; revision received September 27; accepted October 20. Address correspondence to D.E.M. (e-mail: D.Malone{at}st-vincents.ie).

	ABSTRACT

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 
To some, evidence-based practice (EBP) means the identification of centers that produce evidence reports and technology assessments to support guideline development. To others, EBP is the best research evidence integrated with clinical expertise and patient values. Inherent in the first approach is the implication that only central academic organizations can produce valid, reliable analyses of existing literature, which will then be distributed to ordinary practitioners. The second approach implies that ordinary practitioners can learn to use a stepwise approach and a preprepared set of rules and tools to effectively find the best current literature, appraise it, and then apply local circumstances to these rules and tools in their hospital. Paul Glasziou, director of the Centre for Evidence-based Practice in Oxford, England, has coined the phrases top-down EBP and bottom-up EBP to describe these approaches. In this article, the authors describe how knowledge gaps in an ordinary radiology practice can be addressed by using stepwise bottom-up EBP techniques. The following clinical scenario is used: Your hospital's recently appointed chief hepatobiliary surgeon questions the use of transarterial chemoembolization for inoperable hepatocellular carcinoma because of his concerns after reading a recent review article suggesting that there is no clear survival benefit to using this procedure. What would you do? Here is how the authors would do it.

© RSNA, 2005

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 
You are an attending radiologist in a busy third-level referral hospital. The hepatobiliary surgical unit in your hospital has recently appointed a new chairman with a strong academic background. He has contacted your chairman in writing concerning the use of interventional radiology techniques in patients with inoperable hepatocellular carcinoma (HCC). The letter reads as follows:

As you know, HCC is the fifth most common neoplasm worldwide and the third major cancer-related cause of death (1). It is an increasing clinical problem in Western countries. Between 1976 and 1995 there was a 41% increase in mortality attributed to HCC in the United States, and the age-specific incidence of HCC progressively shifted toward younger people (2). The three main risk factors for HCC are infection with hepatitis B virus, infection with hepatitis C virus, and alcohol-related liver disease. Despite success in reducing the incidence of new hepatitis B and hepatitis C infections and improvements in antiviral therapy, it is estimated that the prevalence of hepatitis C–related HCC may increase by as much as 88% during the next 2 decades in the United States (3). HCC most frequently occurs in patients who already have cirrhosis and has an annual prevalence of 3%–4% (4). Because patients with early tumors may be cured with resection or transplantation, screening for HCC has been recommended and widely adopted for patients with cirrhosis (1).

In two large prospective screening studies involving Italian patients, 14% (four of 29) and 24% (15 of 61) of the tumors identified were potentially resolvable with surgery or liver transplantation, respectively (5,6). An effective alternative treatment for the majority of tumors detected is clearly needed, hence the interest in the relative roles of interventional treatments such as percutaneous alcohol injection, percutaneous radiofrequency (RF) ablation, and transarterial chemoembolization (TACE). I am keen to build an evidence-based program for the nonsurgical treatment of selected patients with inoperable cancer and would like to work with your department on this. I must express some concern about radiology's continued recommendation to use TACE for many patients. The randomized controlled trials (RCTs) published in the medical literature have produced disappointing results; this has been acknowledged by your own Journal of Vascular and Interventional Radiology experts, who recognize the lack of a clearly proved survival benefit from TACE (7). I hope for a long-term collaborative approach with your department, as TACE can have clinically important side effects and I feel strongly that it is now necessary for those in the radiology field to either justify the continued use of TACE—with clarification of the selection criteria—or cease recommending the use of it in your reports. I enclose the Journal of Vascular and Interventional Radiology article of interest for your consideration and look forward to hearing from you at your earliest convenience.

Your chairman has discussed the letter with several senior colleagues. They suspect that the surgeon is conducting an academic "reconnaissance by fire" as he settles into his new post and are anxious for radiology to win a prominent and respected place in his view of allied specialties. Your chairman has read the article, which concluded that there is no clear survival benefit from TACE (7). He tried a literature search but immediately come across several problems. First, there was a large quantity of published articles on TACE. In PubMed, a search of the term hepatocellular carcinoma yielded 35 564 articles, and a search of the term TACE yielded 240 articles. Even a practitioner with the most generous amount of research time could not hope to cover so much material (8). Second, assuming that the authors of the Journal of Vascular and Interventional Radiology article did an appropriate literature search, he observed that there was no consensus by the experts on the prevalence of HCC in this patient subgroup—that is, patients with HCC and chronic liver disease. These authors had concluded the following:

TACE is the mainstay of treatment for patients with unresectable HCC. The goal of TACE is to cause tumor necrosis and control tumor growth while preserving as much functional liver tissue as possible. The ultimate purpose, however, is to prolong life. Several different TACE protocols have been developed, with no consensus as to the most effective techniques. The effect of TACE on patient survival remains unclear. Several nonrandomized studies have demonstrated a beneficial effect of TACE on survival. This result has not been confirmed with randomized trials (7).

Third, several studies cited in the above review were performed with patient populations that were different from your own. Your chairman wonders how best to evaluate this review and remembered a discussion with you about a series of evidence-based radiology workshops that you attended at the 2004 International Congress of Radiology in Montreal, Quebec, Canada. He asks if you will take on the job of formulating a reply by using evidence-based practice (EBP) methods. You agree to do so. He thanks you and suggests that you discuss your findings at next month's journal club meeting. You prepare the following presentation:

	EBP: THE STEPWISE PROCESS

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 
There are five steps to applying the evidence-based approach (9). In this analysis, we will consider the first four steps as they relate to this clinical problem.

Step 1: Ask
Information needs relevant to individual patients are converted into answerable, or focused, questions. The PICO (patient, intervention, comparison intervention, and outcome of interest) format was used for this particular topic. In text form, this question would read as follows: "In which patients with inoperable HCC does the use of TACE alone, compared with the use of conservative therapy, improve survival?"

Step 2: Search
A search strategy is very useful when it combines an EBP theory with a simple, locally applicable approach to informatics. First, identify suitable key words that can be used to perform an up-to-date search. The review article read by the hepatobiliary surgeon (7) served as a useful way of identifying a representative, or index, article on the subject (10). The Medline abstract of this index article was retrieved by using the National Library of Medicine online search engine PubMed (10). Changing the retrieval function from "Display Abstract" to "Display Citation" yielded the medical index subject headings under which this index article had been filed by the National Library of Medicine. Two medical index subject headings—"carcinoma, hepatocellular" and "chemoembolization, therapeutic"—that were linked by "AND" were chosen as the basis for the search. These headings were used to search the Clinical Evidence (from the British Medical Journal Publishing Group [11]) and Cochrane Library (12) databases before they were entered into the Systematic Review search engine in the Clinical Queries link on the PubMed services sidebar. The retrieved abstracts of systematic reviews (SRs) and the abstracts reporting the effects of treatment on patient outcome were identified. Next, the most recent of the relevant SRs was retrieved and appraised. The bibliographies of these SRs were used to identify any other relevant outcome reports. Finally, to double check the search strategy, a librarian was asked to perform a computer search by using our focused clinical question (described in Step 1: Ask) as the starting point.

Results from Computer Search Engines
Search of the Clinical Evidence database (11) of the British Medical Journal Publishing Group yielded no relevant retrievals. The initial search for SRs performed by using the Clinical Queries link on the PubMed services sidebar yielded 16 retrievals. Those reviews that did not address patient outcomes after TACE and narrative reviews were excluded. This left seven SR retrievals, which were published between 1994 and 2003, that were relevant to our question (13–19). The search of the Cochrane Library database, which includes both the Cochrane database of SRs (2935 in all) and the database of abstracts of reviews of effects (4006 in all), yielded one SR (16), which had already been retrieved. When individual studies were considered, the review of SR bibliographies yielded seven RCTs in addition to our index study, for a total of eight relevant RCTs published between 1990 and 2002 (10,20–26). The librarian's search yielded no additional SRs or meta-analyses.

Step 3: Appraise
It was decided that the results for Asian and European patient subgroups would be analyzed separately because it has been shown that there are differences in the cause, clinical behavior, and imaging features of HCC between these groups (27–29). Tumor encapsulation, arterioportal shunting, and fatty metamorphosis all occur with less frequency in the non-Asian population (29). Because the biologic features of HCC tumors differ between these population subgroups, the responses to TACE might also differ.

The retrieved literature was then critically appraised by using EBP methods. Two authors (M.S., D.E.M.) performed the appraisals independently, compared the results, and resolved any differences by consensus. Once the appraisal of validity was completed, the strength of the data was appraised by using EBP indexes of benefit and harm (30–32). Once the evidence was explicitly appraised, the best current evidence was applied to the problem in hand—in a clinical resolution—as a basis for step 4 (discussed later in the text).

EBP appraisal of an SR.—The seven relevant SRs had been published between 1994 and 2003 (13–19). In the 2003 review, which contained the most up-to-date literature, it was concluded that TACE might benefit a subset of patients with unresectable HCC (20). The 2003 SR was considered the best current evidence and was appraised in detail. The inclusion criteria reported in this review excluded all but seven RCTs from the meta-analysis. There were five trials from European centers (1990–2002) (10,20,23,24,26) and two from Asian centers (1988, 2002) (22,25).

Several standard questions are asked when a review article is appraised by using EBP methods (33,34). First, what is the validity of the SR? The following related questions can be asked:

1. Did the review explicitly address a sensible clinical question? Yes. For unresectable HCC, the primary end point was to see if medical interventions (TACE or tamoxifen therapy) have a survival benefit compared with conservative management.

2. Was the search for relevant studies detailed and exhaustive? No. The search was restricted to English-language publications. The authors did not seek out all other reviews, consult experts directly, or search the "gray literature" (ie, internal reports, pharmaceutical industry data, and non–peer-reviewed publications [35]) or unpublished data. Therefore, there may have been publication bias (33) (further addressed later in The Explicit Appraisal Process section).

3. Were the primary studies of high methodologic quality? Yes. The authors selected RCTs only and submitted each selected trial for analysis before including it in the meta-analysis.

4. Were the assessments of studies reproducible? Yes.

What is the strength of the SR? Statistical calculations were used to establish a minimum study size required to produce a strong meta-analysis. For arterial embolization, expecting 2-year survival rates of 55% for the treatment group and 35% for the control group, with a statistical power of 90% and a two-tailed type I error of 5%, the minimal sample size needed to perform a strong meta-analysis should be 256 patients (128 patients per arm). Inclusion criteria included defined methods of TACE. Study investigators had to report death rates at 2 years.

Of 328 retrieved trials, 267 were discarded because they were not considered to meet the initial eligibility criteria. This left 61 trials, but 35 of these had an inadequate control arm. Twenty-six trials had an adequate control arm, but 12 of these had too small a sample size for meta-analysis. Of the remaining 14 trials, only seven were focused on TACE. In six of these seven trials, 2-year survival rates were reported, and, thus, these six trials were included in the meta-analysis, the results of which were tabulated (Table 1) (10,20,22,24–26).

View larger version (27K): [in this window] [in a new window] [Download PPT slide]   Meta-analysis of RCTs for analysis of 2-year patient survivals after chemoembolization and/or embolization as compared with 2-year patient survivals after conservative management or suboptimal therapies for unresectable HCC. In the random effects model, the odds ratio was 0.53 (95% CI: 0.32, 0.89; P < .017). (Reprinted, with permission, from reference 18.)

In the European center, the surgical inclusion criteria were a single HCC lesion smaller than 5 cm in diameter or three lesions each smaller than 3 cm in diameter. Transplantation was considered if there was one nodule smaller than 5 cm or three nodules each smaller than 3 cm, if the patient had portal hypertension, or if the patient had an abnormal serum bilirubin level. Resection was considered if the patient had a single tumor smaller than 5 cm, no portal hypertension, and a normal serum bilirubin level (26). The resection criteria used in the Asian center were not specified (25).

Table 2 shows the TACE exclusion criteria used for the European and Asian patients in the RCTs (25,26), and Table 3 shows the EBP indexes of therapeutic effectiveness used in the two patient groups. In the Asian study, the calculations were based on actual reported survivals. In that study, three of the original 80 patients were not included in the final data analysis: One control patient was excluded secondarily when metastases were identified, one control patient was lost to follow-up, and one patient who underwent TACE was lost to follow-up. When the EBP indexes of therapeutic effectiveness were calculated by presuming that all of the patients had poor outcomes, the relative risk ratio, absolute risk ratio, and number needed to treat (32–34) remained significant. When the data were analyzed by presuming that all of the patients had good outcomes, the results were not significant until the 3-year follow-up point was reached. The actual survival values are close to these extreme values and were used to calculate the data shown in Table 3. Table 4 shows the TACE exclusion criteria that have been used for European patients in RCTs to date. The new, more stringent criteria used by Llovet et al (18) also are cited in Table 4.

Clinical Resolution: Formulating a Reply to the Surgeon
Let us return to the clinical scenario described at the beginning of this article, having completed steps 1–3 of the EBP process. Your chairman is pleased with your presentation at the radiology journal club. Together, you and the chairman write back to the surgeon:

EBP techniques were used to establish which patients with inoperable HCC will survive longer when they are treated with TACE alone; the results indicate that only those who meet the inclusion criteria used in two RCTs published in 2002 should receive this treatment. There was no evidence to support the use of TACE alone in the other patients with inoperable HCC. In contrast to the conclusion reached in the 2002 narrative review (7), the results of this evidence-based evaluation suggest that TACE alone can no longer be regarded as the mainstay of therapy for inoperable HCC. Recent radiology literature suggests that we should now be exploring the outcome of ablation combined with TACE in selected patients. Perhaps we could meet to discuss this further.

Step 4: Apply
This step encourages the application of explicitly appraised evidence to local circumstances and individual patients while taking local or individual factors and patient preferences into account (35,38). A week later, your chairman contacts you to tell you that the surgeon has accepted your analysis and a meeting has been arranged. You are invited to attend the meeting and participate in the development of multidisciplinary protocols in which surgery, ablative techniques, and TACE are incorporated for HCC treatment in your hospital.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 
There is increasing emphasis on the identification and regular updating of the evidence base for best patient care in current clinical practices. This is particularly important for radiologists, who interface with all other clinical specialties on a daily basis. Given the ongoing technologic advances in radiology, the increasing complexity of medical and surgical subspecialties, and the explosion of information on the World Wide Web, it has become nearly impossible to balance current best practice with the vast volume of new data that are available from journals, meetings, textbooks, and the Internet.

EBP is a discipline that combines the current best evidence from the literature with clinical expertise and patient values (8,39). It gives physicians who are not epidemiologists the tools to find, competently appraise, and interpret the medical literature. It encourages them to use their clinical experience and judgment with appraisals of relevant current literature when applying conclusions regarding the treatment of an individual patient or a departmental problem.

The problem we addressed in the described evidence-based evaluation was that of identifying criteria that would enable one to determine which patients with inoperable HCC would survive longer if they were treated with TACE alone. Because there is much published literature on the subject but no expert consensus (7), the problem appeared to be ideally suited for the EBP method. This is because two of the core elements of EBP—a hierarchy of evidence and an explicit appraisal process—"decode" the conflicting literature during step 3 of the process (critical appraisal) and allow strong conclusions to be made.

Hierarchy of Evidence
The selection criteria by which the patients who will survive longer after undergoing TACE alone can be identified are a fundamental issue. The narrative review addressing TACE described above (7) represents the kind of conflict of evidence situation that EBP methods are designed to address. The key EBP concept that there exists a hierarchy of evidence (40) should help radiologists to navigate the vast amount of available literature, sift the least biased and strongest studies from the rest, and establish strong resolutions to problems arising in practice.

The lowest and largest level in the hierarchy of evidence is original literature, which is also referred to as primary evidence. The aim of EBP preprocessing is to select the best current evidence. The better the study design and the data analysis, the less likely it is that the conclusions will be wrong. We use the National Health Service–Oxford University Centre for Evidence-based Medicine hierarchy system to rank studies according to their potential for bias that is inherent to the methods used (41). If the conclusions of a level 1 study and level 3b study disagree, the level 3b study is disregarded because the flaws in its design will introduce a greater potential for bias. Therefore, in practice one has only to appraise the highest-level studies retrieved from the literature search of the topic in question. With use of this approach, the results of literature searches will suddenly become more manageable.

However, would it not be even better if you found that someone had already done this for you? That is what is meant by the higher hierarchical levels of secondary literature. These are published reviews that have been preprocessed, filtered, and explicitly validated by using EBP methods. The lowest level of secondary literature is the SR. These structured reviews (for example, the Cochrane Reviews) provide clinicians with an overview of all of the evidence addressing a focused clinical question (12,33). The explicit methodologic features of the construction of these reviews differ from those of traditional (narrative) reviews. The McMaster/Centre for Evidence-based Medicine EBP paradigm does not accept narrative reviews for consideration owing to the historically high potential for bias yielded by the "freestyle" methodology of these reviews (42).

Synopses are the next highest level in the hierarchy of secondary literature. These published materials encapsulate the key methodologic details and results required to apply the evidence to individual patient care. Synopses are published in summary journals such as Evidence-Based Medicine and the American College of Physicians Journal Club.

Finally, systems are the highest level in the hierarchy. These are online EBP guidelines, clinical pathways, or textbook summaries that provide much of the information needed to guide the care of individual patients. Clinical Evidence, which is produced by the British Medical Journal Publishing Group, is one such system (11).

We found that by using the concept of a hierarchy of evidence, we avoided the difficulties encountered by traditional narrative reviewers (7). EBP is a developing field, and within the accepted concept of the hierarchy of evidence there are several schools of thought about how studies should be ranked (43). We used the National Health Service–Oxford University Centre for Evidence-based Medicine hierarchy system because it is internationally established, regularly updated conceptually, and available online without a subscription fee (41).

For the effective practice of bottom-up EBP, it is crucial to stress that in practice, the concept of a hierarchy of evidence means that all studies ranked lower than an SR and a meta-analysis of RCTs should be disregarded if their results conflict with the SR results, because their methodologic features are more subject to bias. We did not have to go through the hundreds of published primary works that our figurative chairman retrieved during his first PubMed search! On the contrary, our search retrievals at the SR level led us to a recently published, methodologically sound overview and meta-analysis, at which we found consistent evidence that TACE may benefit a subset of patients with unresectable HCC (18). The authors did not consider that they had identified the relevant subset of patients during the SR (18).

The Explicit Appraisal Process
The explicit EBP appraisal process revealed only minor flaws in the SR (18). First, random effects meta-analysis enables the pooling of data from small study groups. However, the small size of studies can affect the results, and pooled estimates are susceptible to publication bias (44). Publication bias occurs because some studies with inconclusive or negative results are less likely to be accepted for publication. Therefore, it may be important to contact authors directly and to review unpublished work (eg, directories and databases of research and trial registers, thesis databases, circulars, health service reports, and pharmaceutical industry data), as is done in many SRs, to ensure a thorough literature review (33,45).

The pooled studies were tested for statistical heterogeneity; doing this eliminates the major differences that are due to chance. However, this testing does not eliminate the differences that are due to nonrandom factors (eg, populations, different interventions) (46). A sensitivity analysis was performed; however, it was somewhat limited (47). Therefore, the conclusions were valid when they were judged according to EBP criteria.

Deriving Clinical Resolutions
In the EBP paradigm, it is necessary to reach a clinical resolution by using the best available evidence, even if this evidence is not perfect. (See step 4.) The process of reaching a clinical resolution forms the basis for a planned literature surveillance and an integration of literature on differing techniques into departmental protocols and research studies. The positive results of this SR were largely dependent on two strong RCTs (Figure) (25,26). To reach a clinical resolution during the EBP appraisal of the topic, the relative importance of SRs and meta-analyses, as compared with the importance of RCTs, was considered. Of all the TACE RCTs that were pooled for meta-analysis (10,20–26), only one European and one Asian trial, both of which were published in 2002, revealed that TACE facilitated an outcome benefit in selected patients (25,26). This result is clearly illustrated in the forest plot from the SR (Figure) (18). On the plot, the overall result lies between the negative and indeterminate results of earlier trials and the positive results of the two most recent RCTs.

In practice, meta-analysis involves the pooling of the results of trials performed at different times, with different methods, under different circumstances, and (sometimes) with different populations (47). It is a statistical method that uses pooling of data to increase the number of patients examined in circumstances where published RCTs have been underpowered and/or suboptimal primary research is available. It has been recognized that differing study inclusion and exclusion criteria may lead to SRs and meta-analyses of the same subject in which different conclusions are reached (48). If recent, well-designed, suitably powered RCTs are identified in an updated literature search, pooling the data from these trials with earlier research in a meta-analysis may be superfluous. In this context, considering some of the data from the two RCTs that revealed a survival benefit after TACE proved to be useful.

Survival Benefit after TACE
Llovet et al (26) attributed their success in performing TACE alone in European patients to their more stringent patient selection process compared with the patient selection processes used in previous RCTs (Table 4). They reported that treatment allocation was the sole baseline variable that was independently related to survival (odds ratio, 0.45 [95% CI: 0.25, 0.81]; P = .02) in the Cox regression model. For patients who had sustained objective responses to TACE for at least 6 months, the probabilities of survival at 1, 2, and 3 years were 96%, 77%, and 47% (P = .002 for comparison with patients in whom treatment failed, P = .006 for comparison with the control group). Use of TACE significantly lowered the probability of portal venous invasion (17% in patients with sustained responses vs 58% in control patients at 2 years, P = .005). The survival difference was also significant at sequential triangular testing. Nevertheless, EBP analysis revealed that the 95% CIs of the relative risk ratio, absolute risk ratio, and number-needed-to-treat values for TACE performed in European patients were higher than zero (Table 3). This means that there was either an inadequate sample size or a true lack of a difference in mortality between the patients treated with TACE and the control subjects (49). This suggests that the effectiveness of TACE performed alone in European patients has not been proved beyond a doubt. Because the traditional statistical analysis yielded significant results, we have accepted it as the best current evidence for our clinical resolution while recognizing the caveat.

Lo et al (25) reported that the relative risk of death in the TACE group was 0.50 (95% CI: 0.31, 0.81; P = .005). They attributed the success that they achieved in performing TACE in Asian patients to the chemoembolization technique and regimen used and their patient population: Asian subjects with predominantly hepatitis B virus–related cirrhosis. These patients are more likely to tolerate TACE well than are patients with alcohol-related cirrhosis (10,25). During TACE, they selectively injected a cisplatin-lipiodol emulsion into the artery feeding the tumor, when possible, and used a variable dose of the emulsion based on the tumor size. This resulted in the use of a median cisplatin dose of 10 mg compared with a median dose of 70 mg that was used in another study that involved a fixed-dose regimen (10). The patients may have tolerated the lower dose of cisplatin better (25). The use of this TACE regimen yielded a survival benefit that was significant both at multivariate analysis and at our EBP analysis of the raw data (Table 3).

In summary, Llovet et al (26) and Lo et al (25) suggest that stringent selection criteria and improved TACE methods accounted for their improved outcomes compared with the outcomes in earlier RCTs of TACE for treatment of HCC. For our purposes, the selection criteria and methods reported in these two RCTs can be considered the best current evidence supporting the clinical use of TACE in European and Asian patients.

Comparison of the survival rates between the European and Asian patients revealed that although Asian patients respond to TACE better than European patients (numbers needed to treat for one extra survivor at 1, 2, and 3 years were 4.0 vs 5.5, 4.9 vs 6.7, and 4.3 vs 70.0 [Table 3]), the absolute prognosis for Asian patients with HCC is worse than that for European patients with HCC (1-, 2-, and 3-year survival rates of 57% vs 96%, 31% vs 77%, and 26% vs 47%). This finding is consistent with the other known differences in cause, imaging findings, and outcome with other treatment methods between HCC in Asian patients and HCC in European patients (27–29).

Step 5: Evaluate
To complete the EBP process, we must evaluate, by using prospective research or auditing, our performance in applying the principles derived by using EBP methods. In practice, in our own department, the EBP process has also been used to identify the best current evidence with respect to other radiologic options for treating inoperable HCC. The TACE data must be related to other relevant studies. For example, results of the only published RCT on interventional therapies other than TACE suggest that the recurrence-free survival after percutaneous RF ablation of HCC is better than that achieved with percutaneous alcohol injection (50). Percutaneous RF ablation, where available, should probably replace percutaneous alcohol injection. There is preliminary evidence regarding other treatment options. For example, TACE and percutaneous RF ablation have been combined clinically (51,52), and larger percutaneous RF ablation lesions have been shown to result when percutaneous RF ablation is performed after TACE (53). One of the major reported complications of percutaneous RF ablation of HCC is intraperitoneal hemorrhage (54). Theoretically, performing percutaneous RF ablation after embolization or TACE may reduce the incidence of this complication. Finally, like the outcomes after TACE, the outcomes after percutaneous RF ablation of HCC may vary according to the cause of cirrhosis (55).

The hypothesis that TACE and percutaneous RF ablation may improve outcomes when they are used in combination is level 5 evidence (41) and needs to be tested further. In the future, the strength of scientific arguments (regarding interventional therapy vs surgery) will depend on the quality of the evidence rather than on the number of cases treated (56). This means that in terms of inoperable HCC, controlled trials that yield high-impact data about the survival after combined therapy will be needed. These should ideally be RCTs; however, the use of any control group—prospective or retrospective—will yield a higher level of evidence than a descriptive uncontrolled case series (41,57). Many factors have been identified as important to the study design and the data collection during percutaneous RF ablation trials (58).

The EBP evaluation described in this article revealed the additional possibility that, with regard to HCC, the Asian or European ethnicity of patients and the cause of cirrhosis may have an effect on the clinical prognosis. It also appears reasonable to suggest that the effects of percutaneous RF ablation and TACE on the quality of life of patients with inoperable HCC need to be studied, because the findings of such an investigation might represent a justification for performing interventional therapy with TACE with or without percutaneous RF ablation, even if improved survivals cannot be achieved.

There are still many questions regarding the optimal technical approach for performing TACE and the optimal treatment strategies for patients who undergo TACE. Further studies of TACE performed alone are needed to answer these questions. Trends identified in the recent literature summarized herein suggest that further RCTs of TACE alone versus symptomatic therapy are not likely to be performed.

The key RCTs (25,26) identified in the most recent review had been published since the last time this topic was addressed in the Evidence-based Practice section of Radiology (16). Our report illustrates that any review—whether evidence based or narrative—is only a snapshot of the literature that is available at any particular time and that when the results of a review are inconclusive or negative, continued literature surveillance is mandatory "to help determine the future best practice" (59). The fact that neither key RCT (25,26) was published in the radiology literature demonstrates the limitations of single-specialty journal scanning for continuing professional development.

	EVIDENCE-BASED PRACTICE RESOURCES

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 
The principles outlined herein, as well as the original literature (60), are now widely available in textbook form (9,34,37,39) and on the Internet (41,61,62). The applicability of EBP principles to radiology has been discussed in detail in a special review in this journal (63). Interested readers are referred to these resources. For this study, the retrieved literature was appraised, as described earlier, by using a previously described interventional spreadsheet (31), which can be downloaded from the "Appraise" section on the Evidence-based Radiology Group Web site (62) to perform the necessary calculations. This Web site also provides a tutorial on EBP critical appraisal and calculations for beginners. In addition, an overview of this Web site can be found at the 2003 EPOS (online poster exhibition) Web site of the European Congress of Radiology (64).

	ESSENTIALS

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 

• The McMaster/Centre for Evidence-based Medicine EBP techniques involve the use of a structured approach to focusing on a question arising from day-to-day clinical practice, searching the literature, evaluating the retrieved published material, reaching and applying conclusions, and evaluating performance.
  
• EBP search methods are designed to minimize wasted time and errors and to ensure that no relevant published material is missed by the reviewer. Interested readers can find more information at http://www.evidencebasedradiology.net.
  
• EBP review of TACE literature revealed selection criteria and treatment protocols that substantially improved the survivals of European and Asian patients with inoperable HCC in RCTs.
  
• More clinical trials are needed to establish the best combination of the multiple therapeutic options, including TACE, that are now available for the treatment of HCC.
  

	ACKNOWLEDGMENTS

 
The authors thank Niamh Lucy, the librarian at St Vincent's University Hospital Education and Research Centre, for her literature search.

	FOOTNOTES

 

Abbreviations: CI = confidence interval • EBP = evidence-based practice • HCC = hepatocellular carcinoma • RCT = randomized controlled trial • RF = radiofrequency • SR = systematic review • TACE = transarterial chemoembolization

	References

TOP ABSTRACT INTRODUCTION EBP: THE STEPWISE PROCESS DISCUSSION EVIDENCE-BASED PRACTICE... ESSENTIALS References

 

1. Bruix J, Llovet JM. HCC surveillance: who is the target population? Hepatology 2003;37:507–509.[CrossRef][Medline]
2. El-Serag HB, Mason AC. Rising incidence of hepatocellular carcinoma in the United States. N Engl J Med 1999;340:745–750.[Abstract/Free Full Text]
3. Davis GL, Albright JE, Cook SF, Rosenberg DM. Projecting future complications of chronic hepatitis C in the United States. Liver Transpl 2003;9:331–338.[CrossRef][Medline]
4. Velazquez RF, Rodriguez M, Navascues CA, et al. Prospective analysis of risk factors for hepatocellular carcinoma in patients with liver cirrhosis. Hepatology 2003;37:520–527.[CrossRef][Medline]
5. Colombo M, de Franchis R, Del Ninno E, et al. Hepatocellular carcinoma in Italian patients with cirrhosis. N Engl J Med 1991;325:675–680.[Abstract]
6. Bolondi L, Sofia S, Siringo S, et al. Surveillance program of cirrhotic patients for early diagnosis and treatment of hepatocellular carcinoma: a cost effectiveness analysis. Gut 2001;48:251–259.[Abstract/Free Full Text]
7. Ramsey DE, Kernagis LY, Soulen MC, Geschwind JF. Chemoembolization of hepatocellular carcinoma. J Vasc Interv Radiol 2002;13(9 pt 2):S211–S221.
8. Sackett DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn't. BMJ 1996;312:71–72.[Free Full Text]
9. Sackett DL, Straus SE, Richardson WS, Rosenberg W, Haynes RB, eds. Introduction. In: Evidence-based medicine: how to practice and teach EBM. 2nd ed. Edinburgh, Scotland: Churchill Livingstone, 2000;3–4.
10. Groupe d'Etude et de Traitment du Carcinome Hepatocellulaire. A comparison of lipiodol chemoembolization and conservative treatment for unresectable hepatocellular carcinoma. N Engl J Med 1995;332:1256–1261.[Abstract/Free Full Text]
11. Clinical evidence. London, England: BMJ Publishing Group, BMA House. http://www.clinicalevidence.com. Accessed September 27, 2004.
12. The Cochrane Collaboration. http://www.cochrane.org/index0.htm. Accessed September 27, 2004.
13. Malaguarnera M, Trovato G, Restuccia S, et al. Treatment of nonresectable hepatocellular carcinoma: review of the literature and meta-analysis. Adv Ther 1994;11:303–319.[Medline]
14. Simonetti RG, Liberati A, Angiolini C, Pagliaro L. Treatment of hepatocellular carcinoma: a systematic review of randomized controlled trials. Ann Oncol 1997;8:117–136.[Abstract/Free Full Text]
15. Trevisani F, De Notariis S, Rossi C, Bernardi M. Randomized control trials on chemoembolization for hepatocellular carcinoma: is there room for new studies? J Clin Gastroenterol 2001;32:383–389.[CrossRef][Medline]
16. Camma C, Schepis F, Orlando A, et al. Transarterial chemoembolization for unresectable hepatocellular carcinoma: meta-analysis of randomized controlled trials. Radiology 2002;224:47–54.[Abstract/Free Full Text]
17. Ramsey DE, Geschwind JF. Chemoembolization of hepatocellular carcinoma: what to tell the skeptics—review and meta-analysis. Tech Vasc Interv Radiol 2002;5:122–126.[CrossRef][Medline]
18. Llovet JM, Bruix J. Systematic review of randomized trials for unresectable hepatocellular carcinoma: chemoembolization improves survival. Hepatology 2003;37:429–442.[CrossRef][Medline]
19. Geschwind JF, Ramsey DE, Choti MA, Thuluvath PJ, Huncharek MS. Chemoembolization of hepatocellular carcinoma: results of a metaanalysis. Am J Clin Oncol 2003;26(4):344–349.[CrossRef][Medline]
20. Pelletier G, Roche A, Ink O, et al. A randomized trial of hepatic arterial chemoembolization in patients with unresectable hepatocellular carcinoma. J Hepatol 1990;11:181–184.[CrossRef][Medline]
21. Madden MV, Krige JE, Bailey S, et al. Randomised trial of targeted chemotherapy with lipiodol and 5-epidoxorubicin compared with symptomatic treatment for hepatoma. Gut 1993;34:1598–1600.[Abstract/Free Full Text]
22. Lin DY, Liaw YF, Lee TY, Lai CM. Hepatic arterial embolization in patients with unresectable hepatocellular carcinoma: a randomized controlled trial. Gastroenterology 1988;94:453–456.[Medline]
23. Pelletier G, Ducreux M, Gay F, et al. Treatment of unresectable hepatocellular carcinoma with lipiodol chemoembolization: a multicenter randomized trial. Groupe CHC. J Hepatol 1998;29:129–134.
24. Bruix J, Llovet JM, Castells A, et al. Transarterial embolization versus symptomatic treatment in patients with advanced hepatocellular carcinoma: results of a randomized, controlled trial in a single institution. Hepatology 1998;27:1578–1583.[CrossRef][Medline]
25. Lo CM, Ngan H, Tso WK, et al. Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma. Hepatology 2002;35:1164–1171.[CrossRef][Medline]
26. Llovet JM, Real MI, Montana X, et al. Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial. Lancet 2002;359:1734–1739.[CrossRef][Medline]
27. Beasley RP, Hwang LY, Lin CC, Chien CS. Hepatocellular carcinoma and hepatitis B virus: a prospective study of 22 707 men in Taiwan. Lancet 1981;2:1129–1133.[CrossRef][Medline]
28. Chin PL, Chu DZ, Clarke KG, Odom-Maryon T, Yen Y, Wagman LD. Ethnic differences in the behavior of hepatocellular carcinoma. Cancer 1999;85:1931–1936.[Medline]
29. Freeny PC, Baron RL, Teefey SA. Hepatocellular carcinoma: reduced frequency of typical findings with dynamic contrast-enhanced CT in a non-Asian population. Radiology 1992;182:143–148.[Abstract/Free Full Text]
30. Cook RJ, Sackett DL. The number needed to treat: a clinically useful measure of treatment effect. BMJ 1995;310:452–454.[Free Full Text]
31. MacEneaney PM, Malone DE. Applying "evidence-based medicine" theory to interventional radiology. II. A spreadsheet for swift assessment of procedural benefit and harm. Clin Radiol 2000;55:938–945.
32. Sackett DL, Strauss SE, Richardson WS, Rosenberg W, Haynes RB. Therapy and harm. In: Evidence-based medicine: how to practice and teach EBM. 2nd ed. Edinburgh, Scotland: Churchill Livingstone, 2000;105–167.
33. McGovern DPB. Systematic reviews. In: Levi M, ed. Key topics in evidence-based medicine. Oxford, England: Scion, 2001;17–19.
34. Oxman A, Guyatt G, Cook D, Montori V. Summarizing the evidence. In: Guyatt G, Rennie D, eds. User's guides to the medical literature: a manual for evidence-based practice. Chicago, Ill: American Medical Association, 2002;155–173.
35. Guyatt GH, Sackett DL, Cook DJ. Users' guides to the medical literature. II. How to use an article about therapy or prevention. B. What were the results and will they help me in caring for my patients? Evidence-based Medicine Working Group. JAMA 1994;271:59–63.
36. Guyatt GH, Sackett DL, Cook DJ. Users' guides to the medical literature. II. How to use an article about therapy or prevention. A. Are the results of the study valid? Evidence-based medicine working group. JAMA 1993;270:2598–2601.
37. Guyatt G, Cook D, Devereux PJ, Meade M, Straus S. Part 1, the basics: therapy. In: Guyatt G, Rennie D, eds. User's guides to the medical literature: a manual for evidence-based practice. Chicago, Ill: American Medical Association, 2002;55–79.
38. Straus SE, Sackett DL. Applying evidence to the individual patient. Ann Oncol 1999;10:29–32.[Free Full Text]
39. Sackett DL, Straus SE, Richardson WS, Rosenberg W, Haynes RB, eds. Introduction. In: Evidence-based medicine: how to practice and teach EBM. 2nd ed. Edinburgh, Scotland: Churchill Livingstone, 2000;1.
40. Haynes RB. Of studies, syntheses, synopses, and systems: the "4S" evolution of services for finding current best evidence. ACP J Club 2001;134:A11–A13.[Medline]
41. The NHS Centre for Evidence-Based Medicine, Oxford University (England). Levels of evidence and grades of recommendations. http://www.cebm.net/levels_of_evidence.asp. Accessed September 27, 2004.
42. Oxman AD, Guyatt GH. The science of reviewing research. Ann N Y Acad Sci 1993;703:125–133.[Medline]
43. Upshur RE. Are all evidence-based practices alike? problems in the ranking of evidence. CMAJ 2003;169:672–673.[Free Full Text]
44. Montori V, Guyatt G. Part 2, beyond the basics: summarizing the evidence—publication bias. In: Guyatt G, Rennie D, eds. User's guides to the medical literature: a manual for evidence-based practice. Chicago, Ill: American Medical Association, 2002;529–538.
45. Snowball R. Finding the evidence: an information skills approach. In: Snowball R, ed. Evidence-based practice: a primer for health care professionals. Edinburgh, Scotland: Churchill Livingstone, 1999; 20.
46. Montori V, Hatala R, Guyatt G. Part 2, beyond the basics: summarizing the evidence—evaluating differences in study results. In: Guyatt G, Rennie D, eds. User's guides to the medical literature: a manual for evidence-based practice. Chicago, Ill: American Medical Association, 2002;547–552.
47. McGovern DPB. Meta-analyses. In: Levi M, ed. Key topics in evidence-based medicine. Oxford, England: Scion, 2001;20–25.
48. Linde K, Willich SN. How objective are systematic reviews? differences between reviews on complementary medicine. J R Soc Med 2003;96:17–22.[Abstract/Free Full Text]
49. Guyatt G, Jaeschke R, Heddle N, Cook D, Shannon H, Walter S. Basic statistics for clinicians. II. Interpreting study results: confidence intervals. CMAJ 1995;152:169–173.
50. Lencioni RA, Allgaier HP, Cioni D, et al. Small hepatocellular carcinoma in cirrhosis: randomized comparison of radio-frequency thermal ablation versus percutaneous ethanol injection. Radiology 2003;228:235–240.[Abstract/Free Full Text]
51. Yamakado K, Nakatsuka A, Ohmori S, et al. Radiofrequency ablation combined with chemoembolization in hepatocellular carcinoma: treatment response based on tumor size and morphology. J Vasc Interv Radiol 2002;13:1225–1232.[Medline]
52. Cui SZ, Liu HY, Tang YQ, et al. Radiofrequency thermal ablation combined with arterial embolization for primary liver cancer [in Chinese]. Di Yi Jun Yi Da Xue Xue Bao 2003;23:1093–1094, 1098.[Medline]
53. Kitamoto M, Imagawa M, Yamada H, et al. Radiofrequency ablation in the treatment of small hepatocellular carcinomas: comparison of the radiofrequency effect with and without chemoembolization. AJR Am J Roentgenol 2003;181:997–1003.[Abstract/Free Full Text]
54. 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 study. Radiology 2003;226:441–451.[Abstract/Free Full Text]
55. Giovannini M, Moutardier V, Danisi C, Bories E, Pesenti C, Delpero JR. Treatment of hepatocellular carcinoma using percutaneous radiofrequency thermoablation: results and outcomes in 56 patients. J Gastrointest Surg 2003;7:791–796.[CrossRef][Medline]
56. Reekers JA. Uterine artery embolization: what the others say. Cardiovasc Intervent Radiol 2004;27:305–306.[Medline]
57. Malone DE, MacEneaney PM. Applying ‘technology assessment’ and ‘evidence based medicine’ theory to interventional radiology. I. Suggestions for the phased evaluation of new procedures. Clin Radiol 2000;55:929–937.
58. Goldberg SN, Charboneau JW, Dodd GD 3rd, et al. Image-guided tumor ablation: proposal for standardization of terms and reporting criteria. Radiology 2003;228:335–345.[Abstract/Free Full Text]
59. Stolberg HO. Treatment of abdominal aortic aneurysms: review with evidence-based methods. Radiology 2003;228:614–616.[Free Full Text]
60. Oxman AD, Sackett DL, Guyatt GH; and the Evidence-Based Medicine Working Group. Users' guides to the medical literature. I. How to get started. JAMA 1993;270:2093–2105.
61. McMaster University (Hamilton, Ontario, Canada), Health Information Research Unit. Evidence-based health informatics. http://hiru.mcmaster.ca. Accessed September 27, 2004.
62. The Evidence-based Radiology Group (St. Vincent's University Hospital, Dublin, Ireland). http://www.evidencebasedradiology.net. Accessed September 27, 2004.
63. The Evidence-based Radiology Working Group. Evidence-based radiology: a new approach to the practice of radiology. Radiology 2001;220:566–575.[Abstract/Free Full Text]
64. Malone DE, Staunton M, Skehan SJ, MacEneaney PM, Schranz M. Evidence-based radiology on the internet: a website for education and practice. Eur Radiol 2003;13(suppl 1):435. http://epos.myecr.org/ (ECR password required for access). Accessed September 27, 2004.

This article has been cited by other articles:

				E. J. R. van Beek and D. E. Malone Evidence-based Practice in Radiology Education: Why and How Should We Teach It? Radiology, June 1, 2007; 243(3): 633 - 640. [Abstract] [Full Text] [PDF]

				D. E. Malone and M. Staunton Evidence-based Practice in Radiology: Step 5 (Evaluate)--Caveats and Common Questions Radiology, May 1, 2007; 243(2): 319 - 328. [Abstract] [Full Text] [PDF]

				M. M. Maher, P. A. Hodnett, and M. K. Kalra Evidence-based Practice in Radiology: Steps 3 and 4--Appraise and Apply Interventional Radiology Literature Radiology, March 1, 2007; 242(3): 658 - 670. [Abstract] [Full Text] [PDF]

				J. D. Dodd Evidence-based Practice in Radiology: Steps 3 and 4--Appraise and Apply Diagnostic Radiology Literature Radiology, February 1, 2007; 242(2): 342 - 354. [Abstract] [Full Text] [PDF]

				D. E. Malone Evidence-based Practice in Radiology: An Introduction to the Series Radiology, January 1, 2007; 242(1): 12 - 14. [Full Text] [PDF]

				M. Staunton Evidence-based Radiology: Steps 1 and 2--Asking Answerable Questions and Searching for Evidence Radiology, January 1, 2007; 242(1): 23 - 31. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Staunton, M. Articles by Malone, D. E. Search for Related Content PubMed PubMed Citation Articles by Staunton, M. Articles by Malone, D. E.