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


     

DOI: 10.1148/radiol.2423060008
This Article
Right arrow Abstract Freely available
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 Maher, M. M.
Right arrow Articles by Kalra, M. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Maher, M. M.
Right arrow Articles by Kalra, M. K.
(Radiology 2007;242:658-670.)
© RSNA, 2007

Evidence-based Radiology Series

Evidence-based Practice in Radiology: Steps 3 and 4—Appraise and Apply Interventional Radiology Literature1

Michael M. Maher, MD, FRCSI, FRCR, FFR (RCSI), Philip A. Hodnett, MB, BCh, MRCPI and Mannudeep K. Kalra, MD, DNB

1 From the Department of Radiology, Cork University Hospital, Wilton, Cork, Ireland (M.M.M., P.A.H.); Department of Radiology, Mercy University Hospital, Grenville Place, Cork, Ireland (M.M.M., P.A.H.); Department of Radiology, University College Cork, Cork, Ireland (M.M.M., P.A.H.); and Department of Radiology, Emory University Hospital, Atlanta, Ga (M.K.K.). Received January 4, 2006; revision requested March 2; revision received May 3; final version accepted June 19. Address correspondence to M.M.M. (e-mail: m.maher{at}ucc.ie).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EBP IN INTERVENTIONAL RADIOLOGY...
 CLINICAL RESOLUTION
 DISCUSSION
 APPENDIX
 References
 
When introducing new interventional radiology techniques or devices, it is important to learn from previous experiences and to remember that there are numerous examples of new techniques that were initially enthusiastically promoted and then subsequently abandoned when early promise was not realized. Appropriateness of new or established interventional radiology techniques to specific clinical conditions must be determined from clinical experience, from communication with experts in the field and/or careful review of available medical literature, and on an individual patient basis by means of review of clinical notes and diagnostic imaging studies. Several paradigms for evidence-based practice (EBP) exist. One model proposes that a central specialized process involving academic centers should primarily construct valid guidelines to direct practice at all levels of medical practice ("top-down" model). An alternative model integrates "the best research evidence with clinical expertise and patient values" ("bottom-up" model). This article will focus on the bottom-up model and describe the use of EBP by individual practitioners or groups of practitioners in optimizing literature review and critical appraisal. EBP is applied to two scenarios as a means of deciding the appropriateness of introducing interventional radiology techniques in a community hospital setting. The authors will also briefly discuss other applications for EBP techniques in interventional radiology, including development of practice guidelines or policy to ensure appropriate and safe practices.

© RSNA, 2007


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EBP IN INTERVENTIONAL RADIOLOGY...
 CLINICAL RESOLUTION
 DISCUSSION
 APPENDIX
 References
 
The clinical scenario: At the monthly meeting of an interventional radiology group of a community hospital, two radiologists informed the group that they had been consulted separately by a general surgeon and a rheumatologist regarding the appropriateness of new interventional techniques for the management of specific clinical problems. The two interventions were radiofrequency ablation (RFA) for treatment of hepatocellular carcinoma and vertebroplasty for management of painful or symptomatic osteoporotic vertebral collapse. Neither of these techniques were available at the institution. The group discussed the issues regarding introduction of both techniques into their practice.

Most of the radiologists agreed that the introduction of new techniques is always exciting but presents new challenges. Several questions were addressed, such as whether the new techniques were feasible at a local level, and, if feasible, whether these new techniques should replace older and more established methods. There was a lengthy discussion regarding RFA for hepatocellular carcinoma, and two of the radiologists suggested that perhaps percutaneous ethanol injection (PEI) could be offered as an alternative service to the general surgeon in place of RFA. There was no consensus following initial discussions; one argument was that both techniques should be introduced, and the other argument was that these patients should be referred to the nearest tertiary referral center.

A senior interventional radiologist stated that the education co-coordinator of the radiology group, who had experience with evidence-based practice (EBP), had convened and acted as a tutor to a "practice-based learning" resident group. The senior radiologist was interested in EBP and was now a regular attendee at these meetings. He suggested that the appropriateness of introducing these two techniques to the hospital was an issue that should be referred to that forum so that these decisions could be made with full knowledge of the best available evidence.

Many of the interventional radiologists had heard of EBP but were unfamiliar with how it could be applied by practicing radiologists to "everyday" clinical problems. The senior radiologist explained that EBP can be used to direct literature search and analysis in difficult or "gray areas" of the radiology literature, and findings of such searches can be applied to individual clinical settings. The senior radiologist explained that EBP has gained importance in ensuring that evidence from the best available literature supports decision-making and policy development (1–7). He continued to explain that EBP provides guidelines to find the most valid "best evidence" from studies, which are graded on the basis of strength of study design (1–7). A valid study is defined as one that is free from bias (2). Likewise, according to EBP criteria, a strong study is one that is free from bias or substantial design flaws (4).

One of the younger interventional radiologists questioned how EBP could be applied to interventional radiology. The senior radiologist explained that EBP had been applied to many clinical scenarios in interventional radiology as an aid to decision-making (3,7). An important question when introducing new techniques is whether a new technique should replace an established surgical technique (3,4). Such replacement of an established technique with a new technique requires a benefit-versus-harm analysis. EBP facilitates an independent statistical assessment of whether the data provided, with the number of patients studied, can support the conclusions made, which is particularly important in deciding the safety and effectiveness of a new technique over an old, established technique (3,4). A short discussion ensued among the group, which led to unanimous agreement that the tools of EBP would be ideally suited to these problems.


   EBP IN INTERVENTIONAL RADIOLOGY PRACTICE: THE STEPWISE APPROACH
TOP
 ABSTRACT
 INTRODUCTION
 EBP IN INTERVENTIONAL RADIOLOGY...
CLINICAL RESOLUTION
 DISCUSSION
 APPENDIX
 References
 
There are five steps in applying the "evidence-based" approach (1–7). These steps are as follows: ask, search, appraise, apply, and evaluate (Appendix). For the purposes of this report, these steps were undertaken between July 2004 and July 2005. The literature was updated further during manuscript submission, and the final literature searches were completed in December 2005.

Step 1: Ask
The first step of the evidence-based approach involves the formulation of a question in text form that asks a focused clinical question (1–4,6–8). This step involves four components by using the PICO format: patient, investigation, comparison, and outcome of interest. The components are most useful if they are used as medical subject headings. A search for suitable medical subject heading terms for any topic can be found by using the Preview/Index tab on the PubMed homepage, as demonstrated earlier in this series (8).

The PICO format can be applied to address the two questions presented in the clinical scenario:

Question A.—To determine the appropriateness of RFA for treatment of hepatocellular carcinoma, the first step is to restructure the question into the PICO format (1–4,6–8): "P" (patient) = hepatocellular carcinoma, "I" (intervention) = RFA, "C" (comparison) = PEI or surgical resection, and "O" (outcome) = survival. Thus, the question may read, "In patients with hepatocellular carcinoma, how does RFA compare with other interventional techniques or resection for survival?"

Question B.—To determine the appropriateness of vertebroplasty for the management of vertebral collapse, again the first step is to restructure the question into the PICO format: "P" (patient) = osteoporotic vertebral collapse, "I" (intervention) = vertebroplasty, "C" (comparison) = conservative treatment, and "O" (outcome) = pain relief, complications, and outcome. Thus, the question may read, "In patients with osteoporotic vertebral collapse, how does vertebroplasty compare with conservative management for pain relief, complications, and outcome?"

Step 2: Literature Search
The second step in the EBP approach follows a search strategy, which incorporates EBP theory into a simple, locally applicable approach to informatics. This search strategy has been described in detail in an earlier article in the series (8). The use of the EBP approach requires understanding of the "evidence pyramid" (8,9). In EBP terms, evidence-based reviews are classified as the "secondary" literature and should be distinguished from original published journal articles, which are classified as the "primary" literature (2,7). The secondary literature is a rapidly proliferating and evolving resource that can be used as part of an EBP approach to decision-making in health care (2,7).

The primary literature forms the base of the pyramid, and secondary literature is placed higher in the pyramid, because evidence-based reviews follow strict methodologic criteria and are believed to provide more reliable information than traditional "expert" reviews (2,7).

The secondary literature comprises review publications that have been preprocessed, filtered, and explicitly validated with EBP methods (10,11). These structured reviews provide clinicians with an overview of the entire body of evidence addressing a focused clinical question. Systematic reviews and secondary sources of information, including databases, gateways, and some search engines that filter evidence according to quality and relevance, are increasing in number in the medical literature. They differ from traditional (narrative) reviews in the explicit methodology of their construction. The next article in this series will discuss the place of systematic reviews and meta-analyses in EBP (12). At the top of the evidence pyramid are evidence-based guidelines that summarize important and relevant topics in clinical medicine. An example of such a system is Clinical Evidence from the British Medical Journal Publishing Group (10). The next level is made up of evidence-based journals, such as the American College of Physicians Journal Club (11). The next level, which is lower again in the evidence pyramid, includes evidence-based reviews, guidelines, and databases—for example, the Cochrane Collaboration (13).

As discussed in an earlier article in this series (8), a literature search performed by using the EBP approach can follow several strategies, which can vary with the question, the operator, and the available search tools. Prior to the search, appropriate key words must be identified so that they can be used to perform an up-to-date search. These key words must be sensitive to all relevant articles and specific enough to avoid the problem of nonrelevant articles being identified. Finding a frequently cited "index" article on the subject can be enormously helpful in choosing appropriate key words. The MEDLINE abstract of an index article can then be retrieved by using PubMed, the National Library of Medicine online search engine. Changing the retrieval from Display "Abstract" to Display "Citation" (options listed at the top of the PubMed retrieval page) leads to index (medical subject heading) terms under which this index article had been filed by the staff of the National Library of Medicine.

Appropriate terms chosen for the two aforementioned questions included the search performed for question A (carcinoma, hepatocellular, AND catheter ablation) linked by the Boolean operator AND and the search performed for B (vertebroplasty AND osteoporosis) linked by AND. The fact that no medical subject heading term was available for "vertebroplasty" at the time of preparation of this review highlights the recent nature of vertebroplasty. These medical subject heading terms were used to search the Clinical Evidence database and the Cochrane Library before being entered into the "Systematic Review" search engine in the Clinical Queries section of PubMed (the link to Clinical Queries is on the PubMed services sidebar). The retrieved abstracts of systematic review (SR) articles and those articles reporting the effects of treatment on patient outcome were identified. Next, the most recent of the relevant SR articles were retrieved and appraised. Their bibliographies were used to identify any other relevant outcome reports.

If the services of a librarian are available, the search strategy can be double-checked by asking a librarian to perform a computer search with the focused clinical question as his or her starting point. Because the main aim of this article is to illustrate how EBP can be used to direct decision-making in interventional radiology rather than to definitively address the two chosen questions, the services of a librarian were not used.

Inclusion criteria.—Determination of inclusion criteria prior to literature search followed guidelines of the National Health Service Centre for Evidence-Based Medicine, University of Oxford, England. This body has developed a table of "levels of evidence" (freely accessible online) (14), which is used to establish the strength of the available evidence. Following these criteria, levels were assigned to each article retrieved from PubMed. Those articles with the lowest level of evidence did not require further analysis; investigation was confined to those articles with the highest level of evidence (7,8). Performing a search by using these methods substantially reduces workload (8).

Results from assigning a level of evidence for search A.—For search A ("In patients with hepatocellular carcinoma, how does RFA compare with other interventional techniques or resection for survival?"), the review of the primary literature (PubMed limits selected were "All Adult: 19+ years," "5 years" [ie, articles published in the past 5 years], "English," and "Humans") yielded 232 articles. The review of the secondary literature yielded two SRs (published in 2002 and 2004) that were relevant to the PICO question (15,16). The initial search of the Clinical Evidence database yielded no relevant retrievals (10). The search for SRs performed by using the Clinical Queries tab on the PubMed Services sidebar yielded 10 reviews, six of which appeared relevant to the question (16–21). The search of the Cochrane Library, including the Cochrane Database of Systematic Reviews and the Database of Abstracts of Reviews of Effects, yielded one SR published in 2004 (16). The search of the American College of Physicians Journal Club yielded one additional SR published in 2002 (15).

The 2002 SR (15) included four randomized controlled trials (RCTs) that were reported on between 1999 and 2002; in three of these RCTs (22–24), RFA was compared with PEI and in the other RCT (25), RFA was compared with microwave coagulative therapy. The 2004 SR (16) was based on two RCT articles that were published in 2002 and 2003; in one RCT, RFA was compared with PEI (26), and in one RFA was compared with microwave coagulative therapy (25). The RCT performed by Shibata et al (25) was the only study common to both SRs. Overall, on review of the individual studies from the two SRs, the review of the bibliographies of the SRs yielded five RCTs (22–26) with results reported between 1999 and 2003, four RCTs in which RFA was compared with PEI (22–24,26), and one RCT in which RFA was compared with microwave coagulation therapy (25). Lencioni et al reported on two RCTs in 1999 (22) and 2003 (26). There was only one comparative study (27) in which RFA was compared with surgical resection; this study achieved only level 3 evidence, and it was included in the 2002 SR (15). A search of the primary literature (232 articles) showed that no additional SRs were available when compared with the results of the search of the secondary literature.

There were three additional articles on RCTs in which RFA was compared with percutaneous injection of alcohol that were published subsequent to the 2004 SR; these articles therefore were outside the search period covered by both SRs (28–30). These articles included two on RCTs performed by Lin et al (28,29) and another on an RCT performed by Shiina et al (30). In one of the RCTs performed by Lin et al (28), RFA was compared with PEI in 157 patients with hepatocellular carcinoma; in the other RCT performed by Lin et al (29), RFA was compared with PEI and percutaneous acetic acid injection in a different cohort of 187 patients. Shiina et al (30) compared RFA with PEI in 232 patients.

Because the 2004 SR was the most recent review and was based on published RCT results rather than data from RCTs that were reported at scientific meetings, we considered this 2004 SR (16) to represent the current best evidence. A more recent review published by Lencioni and Crocetti (17) in Clinics in Liver Disease was very thorough, but the authors did not employ either EBP techniques or meta-analyses, and the review was considered lower on the evidence pyramid than the 2004 SR found in the Cochrane Database. Similarly, the remaining reviews (18–21) again did not employ either EBP techniques or meta-analysis, and, although the reviews were thorough and extensive in scope, because they belonged to the category of freestyle or narrative reviews they were also considered lower on the evidence pyramid. The three RCT articles published subsequent to the 2004 SR (28–30) addressed the question of RFA versus PEI and were used to supplement the information in the 2004 review. The 2004 SR did not address the issue of RFA versus open surgical resection, but the 2002 SR did address this subject, and findings were based on a single study that had only grade 3 evidence (27).

There were two additional nonrandomized studies in which RFA was compared with surgical resection (31,32); the reports on these studies were published subsequent to the 2002 review and thus were included in the analysis. These, again, were both level 3b studies. However, methods in the study by Vivarelli et al (32) were further weakened by the fact that the patients in the RFA group and the surgical resection group were based at different treatment units in different Italian cities. We were unable to obtain a full-text version of the article by Hong et al (31), and therefore limited appraisal of the abstract was performed. Therefore, the study analyzed in the 2002 SR and the two additional studies described above, which addressed the question of surgical resection versus RFA, were considered to be representative of the best current evidence and were used in this review to address this question (27,31,32).

Results from assigning a level of evidence for search B.—For search B ("In patients with osteoporotic vertebral collapse, how does vertebroplasty compare with conservative management for pain relief, complications, and outcome?"), the search of the Clinical Evidence database (10) yielded no relevant retrievals. The search for SRs performed by using the Clinical Queries tab on the PubMed Services sidebar yielded eight results. There were two relevant results: one systematic review (33), and a set of guidelines from the Society of Interventional Radiology (34). The search of the Cochrane Library, which included the Cochrane Database of Systematic Reviews and the Database of Abstracts of Reviews of Effects, yielded one nonrandomized trial (level 3b) in which the authors compared percutaneous vertebroplasty with conservative management (35). The search of the American College of Physicians Journal Club yielded no SRs. A search of the primary literature was performed by using the medical subject heading terms vertebroplasty AND osteoporosis (limits selected were "All Adults: 19+ years," "10 years," "English," and "Humans"). This search yielded 123 publications and two published nonrandomized comparisons—one of these was the nonrandomized trial (level 3b) identified in the search of the secondary literature, in which percutaneous vertebroplasty was compared with conservative management (35), and the other was a trial by Grohs et al (36), in which vertebroplasty was compared with balloon kyphoplasty. Overall, no RCT was identified in the primary and secondary literature.

The trial by Grohs et al (36), in which vertebroplasty and balloon kyphoplasty were compared, was published in the Journal of Spinal Disorders and Techniques; this journal was not available in any form at our institution. Because the study by Grohs et al was no higher in the evidence pyramid than was the study by Diamond et al (35) and because the authors compared vertebroplasty with balloon kyphoplasty, another technique that is in development, we decided to exclude this study from our review. Thus, the systematic review (33) and the nonrandomized trial (level 3b) (35) were considered as representative of the best current evidence in the medical literature.

Step 3: Critical Appraisal
In this step, critical appraisal of the retrieved literature is performed by using EBP methods. Individual articles from the primary literature are graded according to the levels of evidence described by the National Health Service Centre for Evidence-Based Medicine (14). Once the appraisal of validity is completed, the strength of data are appraised by using EBP indexes of benefit and harm (37–40). Raw data published in the Results sections of the validated articles are used to establish event rates for the control group ("control event rate") and for the group undergoing the intervention ("experimental event rate") (see Appendix). These data are subjected to evidence-based analysis by using a specially designed spreadsheet (37). Validity can be further assessed with EBP methods (41–43). Once the evidence is explicitly appraised, the best current evidence is applied to the problem at hand ("clinical resolution") as a basis for step 4.

EBP appraisal of SRs for search A.—For question A ("In patients with hepatocellular carcinoma, how does RFA compare with other interventional techniques or resection for survival?"), the 2004 review (16) represented best available evidence for comparing RFA with PEI (26) and with microwave coagulative therapy (25).

For the 2004 SR, the inclusion criteria were based on indexes of methodologic quality of the RCT (including generation of allocation sequences, allocation concealment, blinding or masking, and strength of follow-up data). Only two RCTs, one performed by Lencioni et al (RFA vs PEI) (26) and one by Shibata et al (RFA vs microwave thermal coagulation) (25), fulfilled the criteria for entry into the study. The methods used in the RCT by Lencioni et al were adequate, although lack of blinding was highlighted as a methodologic weakness (26). In addition, withdrawal of two patients from the experimental arm of the trial made it difficult to perform an intention-to-treat analysis. The study performed by Shibata et al was described as an RCT; however, the method of randomization sequence was deemed to be unclear (25). This trial was also completed without blinding. These two RCTs were deemed as representative of the best current evidence and underwent detailed appraisal.

Additional RCTs performed by Lin et al (28,29) and Shiina et al (30) had strong methods, and in all three studies, authors reported better clinical outcomes for RFA in comparison with PEI or percutaneous acetic acid injection. In the first RCT by Lin et al (28), RFA was compared with PEI. A weakness of the study was that six patients in the conventional-dose PEI group, three patients in the higher-dose PEI group, and two patients in the RFA group who did not receive a full course of treatment or did not achieve full tumor necrosis underwent transarterial chemoembolization. In the second RCT by Lin et al (29), RFA was compared with PEI and percutaneous acetic acid injection for treatment of hepatocellular carcinomas of 3 cm or smaller in 187 patients. This study was performed to compare RFA (62 patients) with PEI (62 patients) and percutaneous acetic acid injection (63 patients) in a total cohort of 187 patients (29). In the RCT by Shiina et al (30), RFA was compared with PEI injection in 232 patients.

Establishing validity of SRs for search A.—The following questions can establish validity of SRs (2):

1. Did the SR explicitly address a sensible clinical question? The 2004 SR partially addressed question A. RFA was compared, in two RCTs, with PEI and microwave thermal coagulation (16). However, comparison between RFA and surgical resection was not addressed in these studies, because no such RCT existed. The 2002 SR included an appraisal of the level 3 study that addressed this question (27).

2. Was the search for relevant studies detailed and exhaustive? Yes, a detailed and exhaustive search for relevant studies was performed. Such detailed searches should typically include electronic searches of the Cochrane Hepato-Biliary Group Controlled Trials Register, the Cochrane Library, MEDLINE, EMBASE, CancerLit, and Current Contents from initiation of databases to 2003. The reference lists of the identified articles should be checked for further references. Attempts should also be made to search the "gray literature" (internal reports, pharmaceutical industry data, non–peer-reviewed publications, or unpublished data). In the performance of an SR following EBP criteria, the principal authors of the included trials can be contacted and asked about additional trials of which they were aware (4,37). Industrial sources involved with the new technology or technique can be asked for the results of unpublished data or trials. These additional efforts at accessing the gray literature, therefore, can limit the risk of publication bias (44,45).

3. Were the primary studies of high methodologic quality? Yes, authors of the 2004 SR selected RCTs following strict inclusion criteria, which were based on methodologic quality such as generation of allocation sequence, allocation concealment, blinding or masking, and follow-up (16).

4. Were assessments of studies reproducible? Yes.

Determining strength of the SRs for search A.—Authors of the 2004 SR (16) highlighted the fact that only two trials—one with 104 patients (26) and one with 72 patients (25)—fulfilled the criteria for inclusion in the review. Authors of this SR commented that with only two trials, which included a limited number of patients, unequivocal evidence supporting the value of RFA for treatment of hepatocellular carcinoma is not yet in existence (16). The 2004 SR authors used EBP hierarchical methods to determine the best available evidence. The review was completed according to the predetermined, peer-reviewed, published protocol of the group. The study group had intended to perform intention-to-treat analyses. However, two patients were withdrawn from the experimental arm of the RCT by Lencioni et al (RFA vs PEI) (26); therefore, intention-to-treat analysis could not be performed (16). For the study by Shibata et al (25), intention-to-treat analysis could not be performed because only lesion-based data were provided. The lack of patient-related survival data did not allow an adequate survival analysis. Five trials were listed as excluded trials; in four cases this was because they were not controlled trials, and in one case this was because it was not a randomized trial (16).

Assessing overall results of SRs for search A.—Authors of the 2004 SR concluded that RFA for hepatocellular carcinoma seems to achieve higher recurrence-free survival rates (63% during follow-up at 2 years) in comparison with PEI (43%, P < .01) (16). However, no difference with respect to overall survival was found (P = .14) (16). Authors of the SR also commented that, on the basis of results of one small trial (25), RFA does not appear to show differences with microwave coagulation with regard to safety and efficacy. The authors of the SR concluded that the literature still awaits evidence on the effects of RFA for treatment of hepatocellular carcinoma and that, at present, there is insufficient data to recommend RFA for patients with hepatocellular carcinoma (16).

The 2002 SR had almost identical results (15). In this SR, authors concluded that none of the available trials thoroughly assess the procedure, and they recommended further RCTs of RFA and other ablative therapies (with adequate patient numbers) to assess long-term and overall recurrence in a standardized way. Results of the first RCT by Lin et al (28) showed complete tumor necrosis rates of 88% for the conventional-dose PEI group, 92% for the high-dose PEI group, and 96% for the RFA group; thus, the results showed an advantage of RFA over PEI, but this difference was not statistically significant. Significantly fewer treatment sessions were required to achieve complete tumor necrosis in the RFA group versus the two PEI groups (P < .01) (28). The local tumor progression rate was lowest in the RFA group compared with the conventional-dose PEI group (P = .012) and the high-dose PEI group (P = .037). Also, the overall survival rate was highest in the RFA group compared with the conventional-dose PEI group (P = .014) and the high-dose PEI group (P = .023). Likewise, the cancer-free survival rate was highest in the RFA group when compared with the conventional-dose PEI group (P = .019) and the high-dose PEI group (P = .024).

A multivariate analysis revealed that tumor size (≤3 cm vs >3 cm, P = .011), Edmondton grade of tumor (I or II vs III or IV, P = .026), and the method of treatment (both PEI groups vs RFA group, P = .037) were critical factors affecting cancer-free survival (28). Therefore, the conclusions based on the data from this RCT by Lin et al (28) differed from those from the RCT by Lencioni et al (26) and the 2004 SR (16). Results of this first RCT by Lin et al (28) showed that overall survival was significantly higher with RFA than with PEI; in comparison, results of the study by Lencioni et al (26) showed no difference with respect to overall survival. Findings of a second RCT by Lin at al (29) concurred with those of the previous RCT by Lin et al by showing significant improvement in survival in the RFA group versus the PEI group. Shiina et al (30) reported better survival rates at 4 years after treatment in the RFA group (74%) versus the PEI group (57%) in their RCT, but this difference was not significantly different. RFA showed a 46% smaller risk of death (adjusted relative risk, 0.54 [95% confidence interval: 0.33, 0.89]; P < .05) than did PEI (30).

With regard to complications, severe pain was experienced by one of 52 patients (1.9%) treated with conventional-dose PEI, three of 53 patients (5.7%) treated with higher-dose PEI, and three of 52 patients (5.7%) during RFA (P = .573, RFA vs both PEI groups) (28). Transient pleural effusion was encountered in one of 52 patients (1.9%) after RFA ablation (28). No other severe adverse events were observed (28). In comparing the RFA group with the entire PEI group, there were four adverse events (severe pain or pleural effusion) in the RFA group (experimental event rate, 0.077) and four complications in the PEI group (control event rate, 0.038). The number needed to treat was –25.8 (95% confidence interval: –8.3, 23.6) (the negative value means that adverse effects are more likely with RFA than with PEI). However, the RFA group (experimental event rate, 0.038) had lower rates of failure of complete tumor necrosis compared with the entire PEI group (control event rate, 0.0086). The number needed to treat was 21.8 (95% confidence interval: 8.2, 33.6) (the positive value implies that incomplete tumor necrosis is less likely with RFA than with PEI). Similarly, RFA (experimental event rate, 0.14) was less associated with local tumor recurrence after complete tumor necrosis than was PEI (control event rate, 0.292). The number needed to treat was 6.6 (95% confidence interval: 3.5, 51.8) (the positive value implies that RFA is less likely to be associated with local recurrence after complete tumor necrosis than is PEI).

A limitation of this study (28) was that RFA was performed by using a lower-power generator and smaller-diameter electrode than are typically available on most RFA machines used in current practice. The authors did advise that use of newer generators with greater power and greater coagulation diameter following electrode deployment can lead to better rates of complete tumor necrosis and less risk for local tumor recurrence. The use of more modern technology therefore has potential to further strengthen the advantages of RFA over PEI for complete tumor necrosis following treatment, with less chance for local tumor recurrence. Findings of the second RCT by Lin et al (29) also showed major complications in the RFA group (4.8%; hemothorax [n = 2], gastric perforation [n = 1]) and no complications in the PEI and percutaneous acetic acid injection groups (0%; P = .035), and the authors concluded that RFA caused more major complications when compared with percutaneous acetic acid injection. Shiina et al (30) differed on this subject in comparison with the findings of all the other reports by suggesting that the incidence of adverse events is not different between the two groups.

With regard to RFA versus surgical resection, authors of the 2002 SR (15) concluded that RFA was associated with a higher rate of recurrence and a shorter time interval to recurrence compared with the surgical group. Authors of the 2002 SR did warn that the study they analyzed for the comparison of surgical resection and RFA (27) was limited by (a) lack of randomization and (b) the resulting fact that RFA and resection had been performed in dissimilar groups of patients. This was also one of the major methodologic weaknesses of the study by Vivarelli et al (32) for the comparison of RFA and surgical resection. The patient populations in the two groups were very different, with a statistically larger number of patients in the RFA group having higher Child-Pugh scores (Child B: 46% vs 11%, P < .0001) and multiple tumors (42% vs 13%, P < .01) (32). Procedure-related mortality was higher for resection than for RFA (3.8% vs 0%) and mean hospital stay was longer for resection than for RFA (9 days vs 1 day) (32). The overall survival and disease-free survival rates were significantly higher for those patients treated with resection (P = .002 and .001, respectively); further analysis showed that patients with a Child-Pugh A classification survived longer when treated with surgery (P = .02) (32).

We could not get full-text access to the article by Hong et al (31), and therefore we were unable to complete a full appraisal of this study. However, in this study of 148 patients treated surgically (n = 93) and by means of RFA (n = 55), Hong et al concluded that the rate of local recurrence was significantly higher in the RFA group than in the surgery group but that the incidence of remote recurrence was similar in both groups. The 1- and 3-year cumulative overall survival rates and the 1- and 3-year recurrence-free survival rates were not statistically different between the two groups (31). Overall, the results of this review and of others suggest that the literature examining the question of using RFA versus surgical resection is limited and needs to be assessed in carefully conducted RCTs (19).

EBP appraisal of RCTs for search A.—As outlined above, when performing an SR by using EBP techniques, standard outcome measures of individual RCTs can be individually analyzed by using EBP indexes of therapeutic efficacy, which include relative risk ratio, absolute risk ratio, and number needed to treat. These three analyses were performed in both the 2004 SR (16) and the 2002 SR (15).

EBP appraisal of SRs for search B.—For question B ("In patients with osteoporotic vertebral collapse, how does vertebroplasty compare with conservative management for pain relief, complications, and outcome?"), there was an evidence-based evaluation published in 2000 (33) that was found following a search of the Clinical Queries section of the PubMed Web site. The evidence-based review was based on seven publications (five case series, and two uncontrolled prospective studies) and included one case series with 187 patients. Authors of this evaluation identified studies following searches of databases that included PRE-MEDLINE, MEDLINE, EMBASE, and HealthStar. The authors did not use EBP indexes of therapeutic efficacy, but the lack of RCTs and the paucity of data for comparing vertebroplasty with other treatments would have made this difficult and possibly misleading. In the conclusions of the review, the authors stated that percutaneous vertebroplasty was a new procedure and that evidence was therefore sparse. They suggested that early results were promising—up to 80% of patients with pain unresponsive to medical treatment prior to the percutaneous vertebroplasty procedure experienced a substantial degree of pain relief after the procedure. The SR authors also reported that few serious complications have been reported. They concluded that vertebroplasty is still in the investigational stages but may be appropriate for patients with no other reasonable options for medical treatment.

As discussed previously, subsequent to this SR (33), there was one nonrandomized level 3b trial (35), in which vertebroplasty was compared with conservative management for the management of osteoporotic vertebral fractures, that was considered suitable for appraisal. This study comprised 79 consecutive patients with acute vertebral fractures. Fifty-five patients were treated with percutaneous vertebroplasty and 24 were treated conservatively. The authors concluded that percutaneous vertebroplasty resulted in more prompt pain relief and more rapid rehabilitation when compared with conservative treatment (35). This nonrandomized controlled study (level 3b) was performed in a small number of patients unevenly distributed between the two groups (experimental group, n = 55; control group, n = 24). No power analysis was performed to assess the appropriate patient sample size required to answer the clinical question. However, a strength of the study was that analgesic regimens were similar for vertebroplasty and control groups, and the methods of outcome assessment included objective indexes such as the analogue pain score and Barthel index (quality-of-life analysis).

There were three complications in the experimental group versus none in the conservative management group; this resulted in an experimental event rate of 0.055 versus a control event rate of 0.00. The absolute risk increase was 0.055 (95% confidence interval: –0.005, 0.115). The number needed to harm was 18.3 (95% confidence interval: 8.7, –182.8). Here, the 95% confidence intervals cross zero, which indicates no significant difference between the two groups. The positive number needed to harm implies that vertebroplasty is more likely to be associated with an adverse event than conservative treatment (4,37).

The need for an analgesic at 24 hours, 6 weeks, and 6–12 months were reported. However, on review, the methods were weakened by the fact that no log books were kept; estimates of analgesic intake were recorded as "stopped," "greater than 50% reduction in dosage," or "no change." With regard to the need for an analgesic, at 24 hours, 13 patients had stopped receiving analgesics in the vertebroplasty group versus none in the conservative group (experimental event rate, 0.764; control event rate, 1.0), with relative risk reduction of 0.236 (95% confidence interval: –0.115, 0.341) and absolute risk ratio of 0.236 (95% confidence interval: 0.124, 0.349). A positive number needed to treat of 4.2 suggested that a better outcome was likely with vertebroplasty for this variable than conservative management (95% confidence interval: 2.9, 8.1). With regard to the need for analgesics at 6 weeks, 24 patients in the vertebroplasty group had stopped receiving analgesics versus six patients in the conservative treatment group (experimental event rate, 0.564; control event rate, was 0.750) relative risk ratio of 0.248, absolute risk ratio of 0.186 and number needed to treat of 5.4 (95% confidence interval: –32.4, 2.5). The positive number needed to treat suggested that a better outcome was likely with vertebroplasty for this variable than conservative management. As the 95% confidence interval range crosses zero, data from this trial do not suggest statistically significant differences between the experimental and control groups (4,46).

Establishing validity of SRs for search B: did the study explicitly address a sensible clinical question?—The findings of the nonrandomized controlled trial with weak study methods (35) did appropriately address the question. The SR authors analyzed the role of vertebroplasty for a wide range of indications, including bone metastases, multiple myeloma, and osteoporotic vertebral collapse. The SR, however, was weakened by the small volume of literature (all low in the evidence pyramid) available for analysis.

Determining strength of SRs for search B.—There was one SR and no RCTs retrieved during search B. The SR (limited by scarcity of literature) (33) and the nonrandomized control trial (limited by weakened methodology) (35) represented the best evidence available at the time of the review.

Overall, the assessment of the strength of evidence suggests that the current evidence is weak.

Steps 4 and 5: Apply and Appraise
This step encourages the application of explicitly appraised evidence to local circumstances and individual patients, taking into account patient preferences and local or individual factors. In patients with hepatocellular carcinoma, how does RFA compare with other interventional techniques or resection for survival? Current best evidence, at present suggests insufficient data to recommend RFA for patients with hepatocellular carcinoma.


   CLINICAL RESOLUTION
TOP
 ABSTRACT
 INTRODUCTION
 EBP IN INTERVENTIONAL RADIOLOGY...
 CLINICAL RESOLUTION
DISCUSSION
 APPENDIX
 References
 
Returning to PICO Question for Search A
The senior radiologist consulted the practice-based learning resident group about the scenarios in question, and the group was delighted to accept the challenge. The group met over a period of 4 weeks, and the senior radiologist participated in all meetings. When the EBP appraisal was complete, he returned to the interventional radiology group to present the findings.

The conclusions of the practice-based learning resident group were that current best evidence, based on an EBP review, suggests that RFA for hepatocellular carcinoma seems to result in higher rates of complete tumor necrosis and lower rates of local tumor recurrence when compared with PEI. RFA also appears more likely to be associated with an adverse event than PEI. In regard to comparison of RFA and PEI with reference to overall survival, findings of the early RCTs and the 2004 SR suggested no difference in overall survival between the two groups. However, findings of more recent RCTs (28,29), in which authors compared RFA and PEI, suggested significantly improved survival with RFA in comparison with PEI. The reason for the discrepancy in results between the 2004 SR and the recent RCTs is probably explained by the fact that the follow-up periods in the recent RCTs were longer than those in the RCTs included in the 2004 SR. Current best evidence does not appear to show differences between RFA and microwave coagulation with regard to safety and efficacy. The literature addressing the question of RFA versus surgical resection has methodologic weaknesses, and definitive conclusions cannot be made at this time. Overall, however, at present there is sufficient evidence to recommend RFA over PEI when ablation is indicated, but the indications for RFA relative to transarterial catheter embolization and surgery are as yet not well established.

The conclusion of the senior radiologist was that if a minimally invasive procedure is to be introduced for hepatocellular carcinoma, RFA should be favored over PEI. On the basis of current evidence, RFA of hepatocellular carcinoma should be recommended only after multidisciplinary assessment (for treatments such as resection or transplantation) and should take place as part of a clinical trial (ideally an RCT, but, if not, a nonrandomized controlled trial). As the hospital was a community hospital, patients should at least initially be referred to a tertiary referral center for such multidisciplinary assessment.

The interventional radiology group accepted the recommendation that patients should be referred to a tertiary referral center for multidisciplinary assessment prior to performing RFA. However, with regard to whether RFA could be offered to patients deemed suitable following the multidisciplinary assessment at that hospital, the group decided to consult with referring physicians regarding the desirability and feasibility of establishing an RFA program. Because the hospital was a long distance from the tertiary referral center, the interventional radiology group would consult the general surgeon and request a predicted annual number of patients who would require this procedure and an assessment of the severity of burden for patients to travel to the nearest tertiary referral center for the procedure and follow-up visits. If, after consultations with surgical colleagues, it was clear that a minimally invasive therapy for hepatocellular carcinoma should be introduced at the hospital, the group would decide that RFA was the favored therapy.

Returning to the PICO Question for Search B
The senior radiologist then turned his attention to vertebroplasty for treatment of osteoporotic fractures. He advised that compared with RFA for patients with hepatocellular carcinoma, current best evidence at present is even weaker for percutaneous vertebroplasty in patients with osteoporosis. The only available SR was published in 2000 when the literature on this subject was sparse and even weaker than at the time of this review (33). Authors of this SR concluded that 70%–80% of patients experience improvement in pain after the procedure. The authors discussed the rare occurrence of complications, which included bleeding at the puncture site, transitory worsening of the pain, fever, bone infection or fracture, potential radiculopathy and leakage of material into the paravertebral or epidural spaces, and pulmonary embolism. The other study, which was appraised by means of EBP techniques, was a study with level 3 evidence (35). Not surprisingly, results of this study suggested that percutaneous vertebroplasty is associated with higher complication rates than is conservative management. However, patients require less analgesia at 24 hours and 6 weeks after percutaneous vertebroplasty when compared with conservative management. Overall, however, the quality of evidence available with EBP criteria is weak, and there were insufficient data to recommend percutaneous vertebroplasty for patients with osteoporosis. RCTs are required before any worthwhile recommendation can be made. Therefore, until appropriate evidence becomes available, an interim decision is required.

There were a few different conclusions that could be considered by the interventional radiology group. The first conclusion would be that the results (albeit from weak research that may overestimate benefit) could be interpreted as indicating that the interventional radiology group should begin to perform vertebroplasty in carefully selected patients provided that they continue to observe the literature at regular intervals (the PubMed Web site can be set up to repeat searches at automatic intervals through the "My NCBI" link) and/or they could attempt to enter all patients considered for vertebroplasty at the institution in a multicenter clinical trial. The second conclusion would be that percutaneous vertebroplasty for osteoporosis should be practiced at tertiary referral centers involved in RCTs that compare percutaneous vertebroplasty to other measures in treating osteoporotic fractures and that patients should be referred to such institutions until conclusive evidence became available.

Following discussion, the interventional radiology group agreed on following the second conclusion and decided that the evidence was currently not strong enough to consider introducing the technique at that institution, and patients should be referred to the tertiary referral center until evidence showed clear benefit for vertebroplasty.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EBP IN INTERVENTIONAL RADIOLOGY...
 CLINICAL RESOLUTION
 DISCUSSION
APPENDIX
 References
 
Relevance of EBP in Interventional Radiology Practice
Radiology is a rapidly evolving specialty undergoing constant change with relentless introduction of new technology (2). Currently, new procedures and technologies in interventional radiology frequently diffuse through the radiology community before they have been submitted to thorough technology assessment. RFA and vertebroplasty are examples of two techniques that are currently the subject of such diffusion.

In EBP terms, a study with a control group will always "trump" a case series, so if an RCT cannot be performed, interventional radiology researchers should seek retrospective or nonrandomized controlled data so that comparative studies can replace case series and estimates of equivalence with major therapeutic alternatives can be made. National bodies have recognized these concerns within the specialty. In fact, the British Society of Interventional Radiology concluded in 1999 that only findings of RCTs can provide final proof of procedural appropriateness (4). Increasingly such organizations are recognizing the importance of EBP in developing guidelines and issuing consensus statements—particularly with the paucity of RCTs in the specialty.

Although initial integration of EBP into the radiology literature and everyday practice was much slower than in other specialties, there has been a steady increase in interest in the application of EBP to the practice of radiology (1–4,6,7,37,47).

Top-down EBP
The widespread uptake of clinical governance and the legislation backing it in places such as England requires a rigorous approach to ensure clinical excellence during all stages of patient care (48). This approach requires appropriate design, procurement and commissioning of equipment, quality assurance, and appropriate definition of procedure and practice guidelines for departments (48). The application of the clinical governance model to interventional radiology requires a thorough study of all stages of patient care from initial referral to recovery following a procedure. Examples of direct effects of this model would be requirements that interventional radiologists should have completed standardized training, including advanced cardiovascular life support training, and that the outcome data for each practitioner should be comparable with standardized national data.

EBP is fundamental to issues pertinent to clinical governance and, increasingly, top-down EBP is directing policy in health care (48,49). Basic policy decisions at local and national levels include definition of standard practices or policies in all areas of patient care (for example, patient preparation for interventional radiology procedures). An example of standardized policy would be definition of threshold coagulation parameters (such as prothrombin time and platelet count) below which it is safe to perform interventional radiology procedures. It is now widely accepted that such policy decisions should be made at local and higher levels by following an evidence-based approach. Although these decisions can be very difficult to define and apply with regard to audit and definition of local and national standards, they should be based on thorough literature analysis by using an EBP approach (48). National or research standards may not be routinely applicable locally, and frequent local modification may be necessary on the basis of a combination of best available evidence, respected opinion, and local circumstances.

The Society of Interventional Radiology, or SIR, has published evidence-based standards for clinical excellence with interventional radiology procedures, and these standards are used by the U.S. Food and Drug Administration, hospitals, state regulatory groups, and other medical specialists involved in the practice of interventional procedures (50). These standards outline the criteria for adequate training for specific interventional procedures, as well as the expected success and complication rates. Organizations such as the SIR closely monitor new developments and offer guidelines to aid with decision-making about local introduction of new techniques and also to educate physicians by using information and experience gained at the pioneering institutions (50).

Bottom-up EBP
In contradiction to the top-down EBP approach, the bottom-up EBP approach describes a standardized stepwise approach for individual or nonacademic practitioners to retrieve the best available literature on a given subject, appraise it, and then apply it at the local level (1,37). The approach to resolving the clinical scenarios in this article followed the bottom-up EBP approach.

For 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 below the level of an SR and meta-analysis of RCTs can be disregarded if their results conflict with the SR because their methods are susceptible to bias. If a level 1 study and a level 3b study have disagreeing conclusions, the level 3b study is disregarded, because the potential for flaws in its design introduce a greater potential for bias. Therefore, in practice, one only has to appraise the studies with the highest EBP level of evidence, which makes reviewing the literature more manageable. Fortunately, the secondary literature addressing subjects of interest to radiology is expanding (2). The subject of what to do if no evidence is found, or if the available evidence is weak, will be considered in a later article in this series.

The two clinical questions (ie, RFA for hepatocellular carcinoma and vertebroplasty for osteoporotic vertebral collapse) that form the basis of this review emphasize the differences in approach to EBP-directed literature search, depending on the availability of SRs in the secondary literature. A search of the secondary literature on the question of RFA for hepatocellular carcinoma found two appropriate and valid SRs that addressed the question. For percutaneous vertebroplasty, search of the secondary literature on vertebroplasty identified the highest level studies (an SR [33] and a nonrandomized trial [35]). A review of the primary literature showed no additional studies with a higher level of evidence. The SR was published in 2000, at a time when available literature on the subject was very sparse. Thus, few concrete conclusions could be made. It was therefore necessary to appraise and review the level 3 nonrandomized study in which vertebroplasty was compared with conservative treatment by using the EBP approach. Factors that were considered potential sources of bias in study design were identified, and mathematic analyses to determine the strength, statistical significance, and possible clinical importance of the results were performed. As shown, parameters such as experimental event rate, control event rate, and number needed to treat (and their confidence intervals) are very useful in order to assess how firmly the study conclusions are supported by the reported data (Appendix).

There are limitations to the EBP approach that require discussion. A disadvantage of any review, including an SR, is that it (evidence-based or narrative review) is only a snapshot of the literature at a particular point in time, and, when the results of a review are inconclusive or negative, continued literature surveillance is mandatory for best practice (51). Some authors have suggested that critically appraised topics should acknowledge this limitation and should display an expiry date, which should be determined by the author on the basis of a calculation of the rapidity of research activity on the subject in question (51). In fact, in the search for question A in our article, the recent RCTs reported by Lin et al (28,29) and Shiina et al (30) were not covered in either of the two SRs because their publication dates were outside the time period of the two SRs. Also, the fact that these RCTs were published in Gastroenterology and Gut illustrates the point that it is important to note that, for most questions, single-specialty "journal scanning" is inadvisable, because increasingly articles on interventional radiology techniques are being published in nonradiology journals.

Another limitation of the EBP hierarchical approach is that the methods are statistically weaker than are the methods used in formal meta-analyses (51,52). However, the EBP approach allows an individual practitioner or a small group with limited resources to address gray areas in the literature pertaining to management of an individual patient or clinical condition in the absence of expertise required to complete a meta-analysis (52). The aspect of the EBP approach that focuses on a specific patient or clinical scenario may limit transferability of the results to another patient or clinical scenario (51). As described above, there is always difficulty in applying results of population studies to individual patients being treated with local expertise and technology (51). There is reluctance among many practitioners to learn EBP techniques. Findings of studies have shown that many practitioners use evidence-based summaries produced by others (72%) and EBP guidelines or protocols (84%), but the majority (95%) believe that learning the skills of EBP is not the most appropriate means of moving to EBP (51).

In conclusion, a stepwise EBP approach can be used in many clinical scenarios in interventional radiology to aid the decision-making process by resolving questions about the introduction of a new technique.


   APPENDIX
TOP
 ABSTRACT
 INTRODUCTION
 EBP IN INTERVENTIONAL RADIOLOGY...
 CLINICAL RESOLUTION
 DISCUSSION
 APPENDIX
References
 
There are five steps in the application of the EBP approach (43,53): (a) Information needs relevant to individual patients are converted into "answerable" or "focused" questions (8). (b) A comprehensive literature search is performed to find the best evidence to help answer these questions (8). (c) The evidence retrieved is critically appraised in an explicit and structured manner to establish its validity, strength, and usefulness in practice. (d) The results of this critical appraisal are then applied to the care of individuals or groups of patients. (e) The clinical performance of the clinicians involved, assessed by using the principles derived from the first four steps, is subjected to evaluation (clinical audit).

To effectively carry out step c above, literature is classified as belonging to one of several "domains." Examples of domains are diagnosis, therapeutic benefit, therapeutic harm, reviews, clinical guidelines, prognosis, economic analysis, and qualitative research. The difference between evidence-based and "traditional" literature analysis is that the evidence-based researchers have described, in the peer-reviewed literature, an explicit process of appraisal for literature in each domain (14). Factors that will act as sources of bias in study design have been identified, and mathematic analyses that will give the reader a clear idea of the strength, statistical significance, and possible clinical importance of the results are described.

Analyses of procedural benefit used in our review are the absolute risk reduction and number needed to treat. Analyses of procedural harm are the absolute risk increase and the number needed to harm. These are calculated from control event rates and experimental event rates. The absolute risk reduction is the arithmetic difference in risk between the two groups (control event rate minus experimental event rate). If the result is a positive number, the control group has a higher risk of this complication. If the result is negative, the experimental group has a higher risk (absolute risk increase). Absolute risk data may be difficult for us to remember, especially when the numbers are less than 1. On the other hand, the inverse of absolute risk reduction (1 divided by absolute risk reduction) is a whole number and has the useful property of telling us the numbers of patients that we need to treat with the experimental therapy for the duration of the trial to prevent one additional bad outcome (number needed to treat). Similarly, the inverse of absolute risk increase (1 divided by absolute risk increase) is also a whole number and tells us the number of patients that we need to treat with experimental therapy for the duration of the trial to encounter one additional negative outcome (number needed to harm) (37,39,53).

In addition to being published in the original literature (54), these principles are now widely available in textbook form (42,53) and on the Internet (14,55). The applicability of evidence-based principles to radiology was discussed in detail in a special review in Radiology in 2001 and in European Radiology in 2004 (2,56).


   FOOTNOTES
 

Abbreviations: EBP = evidence-based practice • PEI = percutaneous ethanol injection • RCT = randomized controlled trial • RFA = radiofrequency ablation • SR = systematic review

Authors stated no financial relationship to disclose.


   References
TOP
 ABSTRACT
 INTRODUCTION
 EBP IN INTERVENTIONAL RADIOLOGY...
 CLINICAL RESOLUTION
 DISCUSSION
 APPENDIX
 References
 

  1. Staunton M, Malone DE. Can acute mesenteric ischemia be ruled out using computed tomography? critically appraised topic. Can Assoc Radiol J 2005;56:9–12.[Medline]
  2. Dodd JD, MacEneaney PM, Malone DE. Evidence-based radiology: how to quickly assess the validity and strength of publications in the diagnostic radiology literature. Eur Radiol 2004;14:915–922.[CrossRef][Medline]
  3. Maher MM, McNamara AM, MacEneaney PM, Sheehan SJ, Malone DE. Abdominal aortic aneurysms: elective endovascular repair versus conventional surgery—evaluation with evidence-based medicine techniques. Radiology 2003;228:647–658.[Abstract/Free Full Text]
  4. 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.
  5. Baxter R, Baxter H. Clinical governance. I. Evidence-based practice. J Wound Care 2002;11:7–9.
  6. Staunton M, Malone DE. Ultrasonography or computed tomography for diagnosis in haemodynamically stable patients with recent blunt abdominal trauma? critically Appraised Topic. Can Assoc Radiol J 2003;54:279–280.[Medline]
  7. Staunton M, Dodd JD, McCormick PA, Malone DE. Finding evidence-based answers to practical questions in radiology: which patients with inoperable hepatocellular carcinoma will survive longer following transcatheter arterial chemoembolization? Radiology 2005;237:404–413.[Abstract/Free Full Text]
  8. Staunton M. Evidence-based radiology: steps 1 and 2—asking answerable questions and searching for evidence. Radiology 2006;242:23–31.
  9. Haynes RB. Of studies, summaries, synopses and systems: the "4S" evolution of services for finding current best evidence. Evid Based Ment Health 2001;4:37–39.[Free Full Text]
  10. Clinical Evidence Web site. BMJ Publishing Group. http://www.clinicalevidence.com. Accessed December 22, 2005.
  11. American College of Physicians Journal Club Web site. http://www.acpjc.org. Accessed December 22, 2005.
  12. Halligan S, Altman D. Evidence-based practice in radiology: steps 3 and 4—appraise and apply systematic reviews and meta-analyses. Radiology (in press).
  13. The Cochrane Collaboration Web site. http://www.cochrane.org/index0.htm. Accessed December 22, 2005.
  14. Levels of evidence. Oxford Centre for Evidence-Based Medicine Web site. http://www.cebm.net/levels_of_evidence.asp. Accessed December 22, 2005.
  15. Sutherland LM, Williams JA, Padbury RT, Gotley DG, Stokes B, Maddern GJ. A systematic review of radiofrequency ablation for the treatment of liver tumors. ASERNIP-S report no. 28. Adelaide, South Australia: ASERNIP-S, October 2002. http://www.surgeons.org/AM/Template.cfm?Section=ASERNIP_S_Publications&Template=/CM/ContentDisplay.cfm&ContentFileID=1106. Accessed December 22, 2005.
  16. Galandi D, Antes G. Radiofrequency thermal ablation versus other ablations for hepatocellular carcinoma. The Cochrane Database of Systematic Reviews 2004;2. doi:10.1002/14651858.CD003046.pub2. Published February 25, 2004. Accessed December 22, 2005.
  17. Lencioni R, Crocetti L. A critical appraisal of the literature on local ablative therapies for hepatocellular carcinoma. Clin Liver Dis 2005;9:301–314.[CrossRef][Medline]
  18. Gaiani S, Celli N, Cecilioni L, Piscaglia F, Bolondi L. Percutaneous treatment of hepatocellular carcinoma. Aliment Pharmacol Ther 2003;17(suppl 2):103–110.
  19. Ng KK, Lam CM, Poon RT, Ai V, Tso WK, Fan ST. Thermal ablative therapy for malignant liver tumors: a critical appraisal. J Gastroenterol Hepatol 2003;18:616–629.[CrossRef][Medline]
  20. Poon RT, Fan ST, Tsang FH, Wong J. Locoregional therapies for hepatocellular carcinoma: a critical review from the surgeon's perspective. Ann Surg 2002;235:466–486.[CrossRef][Medline]
  21. Ryder SD; British Society of Gastroenterology. Guidelines for the diagnosis and treatment of hepatocellular carcinoma (HCC) in adults. Gut 2003;52(suppl 3):iii1–iii8.[Free Full Text]
  22. Lencioni RA, Cioni D, Donati F, Crocetti L, Granai G, Bartolozzi C. Percutaneous treatment of small hepatocellular carcinoma: radiofrequency ablation versus percutaneous ethanol injection—a prospective randomized controlled trial (final report) [abstr]. Radiology 1999;213(P):123.
  23. Olshewski M, Lencioni R, Allgaier HP, et al. A randomized comparison of radiofrequency thermal ablation and percutaneous ethanol injection for treatment of small hepatocellular carcinoma. Alexandria, Va: American Society of Clinical Oncology, 2001; 126A.
  24. Shiina S, Teratani T, Obi S. Prospective randomised controlled trial comparing percutaneous radiofrequency ablation and percutaneous ethanol injection therapy for small hepatocellular carcinoma. Gastroenterology 2000;118:959A.
  25. Shibata T, Iimuro Y, Yamamoto Y, et al. Small hepatocellular carcinoma: comparison of radiofrequency ablation and percutaneous microwave coagulation therapy. Radiology 2002;223:331–337.[Abstract/Free Full Text]
  26. Lencioni RA, Allgaier HP, Cioni D, et al. Small hepatocellular carcinoma in cirrhosis: randomized comparison of radiofrequency thermal ablation versus percutaneous ethanol injection. Radiology 2003;228:235–240.[Abstract/Free Full Text]
  27. Yu HC, Le JM, Kim DG, Cho BH. Recurrence pattern of the radiofrequency ablation for hepatocellular carcinoma. J Hepatobiliary Pancreat Surg 2002;125(9 suppl 1):F-231.
  28. Lin SM, Lin CJ, Lin CC, Hsu CW, Chen YC. Radiofrequency ablation improves prognosis compared with ethanol injection for hepatocellular carcinoma < or = to 4 cm. Gastroenterology 2004;127:1714–1723.[CrossRef]
  29. Lin SM, Lin CJ, Lin CC, Hsu CW, Chen YC. Randomised controlled trial comparing percutaneous radiofrequency thermal ablation, percutaneous ethanol injection, and percutaneous acetic acid injection to treat hepatocellular carcinoma of 3 cm or less. Gut 2005;54:1151–1156.[Abstract/Free Full Text]
  30. Shiina S, Teratani T, Obi S, et al. A randomized controlled trial of radiofrequency ablation with ethanol injection for small hepatocellular carcinoma. Gastroenterology 2005;129:122–130.[CrossRef][Medline]
  31. Hong SN, Lee SY, Choi MS, et al. Comparing the outcomes of radiofrequency ablation and surgery in patients with a single small hepatocellular carcinoma and well-preserved hepatic function. J Clin Gastroenterol 2005;39:247–252.[CrossRef][Medline]
  32. Vivarelli M, Guglielmi A, Ruzzenente A, et al. Surgical resection versus percutaneous radiofrequency ablation in the treatment of hepatocellular carcinoma on cirrhotic liver. Ann Surg 2004;240:102–107.[CrossRef][Medline]
  33. Levine SA, Perin LA, Hayes D, Hayes WS. An evidence-based evaluation of percutaneous vertebroplasty. Manag Care 2000;9:56–60, 63.[Medline]
  34. McGraw JK, Cardella J, Barr JD, et al. Society of Interventional Radiology quality improvement guidelines for percutaneous vertebroplasty. J Vasc Interv Radiol 2003;14:827–831.[Medline]
  35. Diamond TH, Champion B, Clark WA. Management of acute osteoporotic vertebral fractures: a nonrandomized trial comparing percutaneous vertebroplasty with conservative therapy. Am J Med 2003;114:257–265.[CrossRef][Medline]
  36. Grohs JG, Matzner M, Trieb K, Krepler P. Minimal invasive stabilization of osteoporotic vertebral fractures: a prospective nonrandomized comparison of vertebroplasty and balloon kyphoplasty. J Spinal Disord Tech 2005;18:238–242.[Medline]
  37. 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.
  38. Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 1988;318:1728–1733.[Medline]
  39. Cook RJ, Sackett DL. The number needed to treat: a clinically useful measure of treatment effect. BMJ 1995;310:452–454.[Free Full Text]
  40. Chatellier G, Zapletal E, Lemaitre D, Menard J, Degoulet P. The number needed to treat: a clinically useful nomogram in its proper context. BMJ 1996;312:426–429. [Published correction appears in BMJ 1996;312:563.][Free Full Text]
  41. 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.
  42. 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.
  43. Guyatt G, Cook D, Devereux PJ, Meade M, Straus S. I. The basics: therapy. In: Guyatt G, Rennie D, eds. User's guides to the medical literature: a manual for evidence-based medicine. Chicago, Ill: American Medical Association, 2002; 55–79.
  44. McGovern DPB. Systematic reviews. In: Levi M, ed. Key topics in evidence-based medicine. Oxford, England: Scion, 2001; 17–19.
  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. 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.
  47. Medina LS, Blackmore CC, eds. Evidence-based imaging: optimizing imaging in patient care. New York, NY: Springer Science & Business Media, 2006.
  48. Ayers AB. Clinical governance: is it coming of age? Br J Radiol 2001;74:1083–1087.[Free Full Text]
  49. Hunink MG. Decision making in the face of uncertainty and resource constraints: examples from trauma imaging. Radiology 2005;235:375–383.[Abstract/Free Full Text]
  50. Clinical practice guidelines Society of Interventional Radiology Web site. http://www.sirweb.org/clinical/clinicalPracticeGuidelines.shtml. Accessed December 22, 2005.
  51. Hawkins RC. The evidence based medicine approach to diagnostic testing: practicalities and limitations. Clin Biochem Rev 2005;26:7–18.[Medline]
  52. van Heek NT, Ubbink DT, Legemate DA. Systematic review of studies on elective endovascular repair versus conventional surgery for abdominal aortic aneurysms [letter]. Radiology 2004;232:933–936.[Free Full Text]
  53. Sackett DL, Strauss SE, Richardson WS, Rosenberg W, Haynes RB. Evidence-based medicine: how to practice and teach EBM. 2nd ed. Philadelphia, Pa: Churchill Livingstone, 2000.
  54. Oxman AD, Scakett DL, Guyatt GH. Users guide to the medical literature. I. How to get started. The Evidence-Based Medicine Group. JAMA 1993;270:2093–2095.
  55. Health Information Research Unit (HIRU): evidence-based health informatics. McMaster University Web site. http://hiru.mcmaster.ca/. Accessed December 22, 2005.
  56. Evidence-Based Radiology Working Group. Evidence-based radiology: a new approach to the practice of radiology. Radi