Aortic Aneurysm Wall Stress Analysis
A description and invitation for participation in a multicenter evaluation.
Aortic aneurysm, and in particular abdominal aortic aneurysm (AAA), is a leading cause of death in the US. Predicting rupture risk is critically important to reducing aneurysm-related mortality without unnecessarily increasing the rate of surgical intervention. The current standard for predicting rupture risk is maximum AAA diameter, but this standard is clearly not ideal. Even with a high rate of intervention in patients undergoing frequent and reliable surveillance, the rupture rate for AAAs may still be greater than 2% per year in some patient populations.1,2 Difficulties in managing patients with aortic aneurysms include the fact that small AAAs do rupture,2-5 some patients are at high risk for aneurysm repair, and in thoracic or throacoabdominal aortic aneurysms (TAAs) the rupture risk is not as well defined as it is for AAAs. In older, high-risk patients, more than 50% of patients with aneurysms larger than 5.5 cm will rupture when surgery is deferred due to high operative risk,6,7 many within the first year of observation.6,8
We have previously demonstrated that finite element analysis of AAA wall stress using 3-D CT reconstructions is better than diameter for differentiating AAAs near the time of rupture,9 and that wall stress is superior to AAA diameter for predicting rupture risk in patients under observation.10 A detailed analysis of two-dimensional parameters has yielded additional information that helps somewhat in differentiating patients who are at higher-than-average risk for rupture, but two-dimensional variables do not differentiate high-risk aneurysms as well as stress analysis.9,11 Our center and others have also confirmed the importance of a multivariate analysis that includes demographic variables. For example, women have at least a threefold higher risk of AAA rupture than men, regardless of AAA diameter.2,10,12 Other factors such as current smoking, hypertension, family history, COPD, and AAA expansion rate have also been implicated as potentially independent factors that may affect rupture rate of aortic aneurysms.2,11-13 Aortic and aneurysm morphology is also known to change with age.14 Thus, when evaluating outcomes related to aortic aneurysms, it is important to collect information on adequate numbers of patients to account for confounding factors, such as gender, blood pressure, smoking, and age. Thus, a large series of patients would need to enroll concurrently, preferably from a wide geographic distribution with a diverse patient population. Such a series can only be collected in the context of a large, multicenter experience.
Identification of Potential Participating Physicians and Centers
The goal of the study is to enroll patients from both large and small centers, so as not to compile results from only a few major centers. Potential participating physicians and centers are being identified in one of two ways: 1) by virtue of being known large centers for aortic aneurysm repair (preferably distributed widely in geographic terms); and 2) by expressing interest in participation after learning about the study. This could potentially allow wide participation by physicians and centers that treat smaller numbers of patients individually, but when combined, account for a large percentage of the aneurysms treated in the US. Thus, a mixture of large and small centers is more typical of real world management of aortic aneurysms. Once interest is identified, a study protocol and instructions for obtaining IRB approval and HIPAA compliance will be sent to the interested practitioner. Once these credentials have been established, the physician will be approved for data entry in the study. Multiple physicians or practitioners can be approved during the process of approval for an individual hospital or center. Initially, participation in the study will require access to the internet for Web-based data entry on a secure HIPAA-compliant Web site. The initial goal of the study is enrollment of 500 patients, with potential for larger enrollment depending on rate of accrual and percentage lost to follow-up.
Patient Population and Subgroups
The patient population will be composed of patients being treated or observed in the course of routine care by participating practitioners. Patients enrolled in the study can be conveniently grouped into two separate subgroups: 1) patients who have a 3-D reconstruction from CT obtained during the course of routine care; and 2) patients who are followed by conventional imaging (CT, MR, or ultrasound) only.
Patients who have a 3-D reconstruction from CT obtained during the course of routine care (eg, looking for unusual shapes that may increase rupture risk, to evaluate unusual anatomy, to evaluate potential for endovascular repair, to evaluate occlusive disease in the aorta or branch vessel, to evaluate potential clamp sites and risk for open repair, or to evaluate other nonvascular structures) will undergo aneurysm wall stress analysis that will not incur cost to the patient, the participating practitioner, or the participating center. The costs of performing, analyzing, and reporting on the 3-D aortic aneurysm wall stress map will be borne by the study. The stress map data may be reported to the participating practitioner, or they may remain blinded.
The determination to be blinded to the stress results may be determined on a case-by-case basis by the practitioner, to simulate real world scenarios in which the practitioner may not obtain stress analysis for all cases. The reported stress map and other results will generate a report, as will be subsequently described. Patients who undergo 3-D reconstructions from MR data of adequate quality may also be candidates for stress mapping, but reports will not allow estimation of rupture risk initially because there are few data on stress maps from MR. One goal of the study is to compile wall stress data on a large, diverse patient population to establish norms related to aneurysm diameter, gender, and other variables. Thus, enrollment is encouraged even for patients who are soon to undergo aortic aneurysm repair because they establish group norms and form an important control function.9 Patients who have already undergone endovascular repair will not be included in this study, but may be included in a separate study.
Patients who are followed by routine CT, MR, or ultrasound may also be entered in the study to enhance analysis of patient variables affecting rupture risk (eg, gender, smoking, etc.) and to compare outcomes for patients who do not undergo stress analysis or 3-D imaging. This limb of the study will also allow practitioners to obtain patient and center-specific reports on all patients even if only a subset of their patients undergo 3-D reconstruction from CT data.
Aortic Aneurysm Wall Stress Analysis
The methods used for aneurysm wall stress analysis using finite element modeling techniques and specialized software have been described previously in detail.9,10 In brief, patients who undergo 3-D reconstruction from CT or MR obtained during the course of routine care will be eligible for 3-D aneurysm wall stress mapping and analysis (Figure 1). This will preferably be performed in the context of the patients’ blood pressure data (systolic, diastolic, and mean), but will also be performed for each patient at standard pressures as a means of comparison and to help establish the importance of blood pressure for the results. The blood pressure data preferably includes the blood pressure on the day of the CT (or other imaging), but the enrollment data will also allow entry of data for a highest and ?representative? blood pressure. Reported stress map results will include stress at multiple blood pressures to aide in patient management, and this will be enhanced by typical patient data. Stress analysis will also include the concept of ?equivalent diameter? in an effort to provide appropriate clinical context for stress map results (eg, a patient may have a maximum AAA diameter of 4.8 cm but a peak AAA wall stress equivalent to a typical 6-cm AAA).
Patient Enrollment Including Demographic Data
Patient enrollment will include entry of a small amount of data into a Web-based datasheet (limited to a size that would fill less than one side of an 8.5 x 11-inch sheet of paper). The datasheet can be printed so the patient can provide information (eg, smoking history) in a manner typical of a normal office visit. This information would be a typical part of routine patient evaluation when considering risk versus benefit of observation versus intervention for an aortic aneurysm. If enough interest exists, a sheet may be developed that can be scanned electronically to enter data rather than requiring Web-based entry. Demographic data will include variables known to affect (or thought to potentially affect) aneurysm rupture risk or patient life expectancy, including age, gender, smoking history (current/former/never), family history of AAA (number of relatives), COPD history, hypertension history, recorded blood pressure from clinic visit, height, weight (for body mass index), cardiac history (MI, CHF, CABG), stroke history, and pertinent medications (eg, beta blockers, steroids, statins, aspirin). Although data collection by the patient’s health care provider will of course not be blinded, the data will be entered on a secure Web site in HIPAA-compliant fashion. When a new patient is enrolled, a unique patient ID is generated. This patient ID can then be used (by the provider only) on subsequent visits so that subsequent information (diameter, date of imaging, blood pressure, smoking cessation, etc.) can be entered with the proper patient association and linked in such a way that reports can be generated for sites and physicians without compromising patient privacy. Public reports will never include patient-identifiable information.
Patient Follow-Up and Updating the Database for Individual Patients
The participating provider will be able to update the Web site database when an enrolled patient has a new imaging study, pertinent demographic data update, or a study-designated endpoint (outcome measure). The participating provider will receive alerts when follow-up duration has exceeded preset limits without a known outcome (eg, patient with ID#112233 with aneurysm diameter 5 cm has not had imaging follow-up in more than 7 months). This will encourage complete follow-up and alert clinicians to patients who may be in danger of being lost to follow-up (see Reports section). Patients who are initially enrolled with a 3-D wall stress map need not have this study for every follow-up (eg, the next follow-up or the next several follow-up imaging studies may be ultrasound). Maintaining an appropriate level of follow-up will be a criterion for remaining a participating center (eg, if more than 10% of patients have lack of data entry within standard follow-up time periods for an aneurysm of a particular size, participation may be in jeopardy).
Outcome Measures for Individual Patients
Outcome measures for the study are elective repair (open or endovascular), emergent repair for acute AAA symptoms or rupture, death from rupture, or death from other causes. If known, the location of rupture (eg, anterior, posterior, right lateral, left lateral, etc.) will be entered. Entry will also be made for patients lost to follow-up or withdrawn from study. This information can be entered via the secure Web site only by the patients own participating practitioner.
Statistical Evaluation and Authorship
All statistical evaluation will be performed using standard software programs, as in previous studies, for univariate and multivariate analysis with regard to stress analysis and aneurysm rupture risk.9-11 The values are reported as mean ± standard deviation unless otherwise specified. Determination of authors for publication of results will be determined by ideas proposed for study (beyond those already outlined) and by the number of patients entered into the study by the author’s center. A list of participating centers and physicians will be provided as part of the study and for reference in publications (unless a participating provider wishes to have this information withheld). As per the Uniform Requirements for Manuscripts Submitted to Biomedical Journals: The authors will be involved in the study design, have access to the data in this study, take complete responsibility for the integrity of the data, the accuracy of the data, the accuracy of the data analysis and interpretation, and for writing the manuscript and submitting it for publication.15
Three kinds of reports will be available from the Web site database, determined by the level of access appropriate for patient privacy. First is a patient-specific aneurysm risk report for the participating health care provider to use for patient education (eg, while seeing the patient in clinic). This report can be shared with the patient and includes identifiable information because it is only for the patient and the health care providers (a copy can be given to the patient). It will include diameter data plotted versus time. If the stress results are available, a graph of effective diameter versus time can also be made, which takes blood pressure and AAA shape into account, as well as an effective diameter at standard pressure (eg, 120 mm Hg) for comparison of shape changes. Potentially, either this report or the Web site will include a color picture of the patient’s own 3-D aneurysm wall stress map from the last scan (in a standard four views, similar to the two views seen in Figure 1). The risk for rupture based on the maximum diameter or effective diameter is also presented (based on data published to date). In addition, other modifiable risk factors for AAA rupture are elucidated (based on literature to date on rupture risk related to current smoking, gender, hypertension, etc.). This report serves as a vehicle for the health care provider to effectively communicate to the patient the current rupture risk and ways to modify that risk, such as smoking cessation and blood pressure control.
The second report is a patient status and follow-up report that will be used by each participating center (or each participating physician) to see which of their patients is in need of further imaging or in danger of being lost to follow-up. The content will include the patient’s name, patient ID, diameter, peak wall stress and effective diameter, if available, date of last imaging study, and time elapsed since last imaging study.
The last report is for study-wide aggregate data and is designed to be shared among all of the study investigators. This report will provide HIPAA-compliant aggregate statistics for the entire study, including the number of patients enrolled, values for aneurysm diameter and maximum wall stress (mean, median, standard deviation, range), and rupture risk as correlated to wall stress and demographic information (based on last analysis). Data for norms (eg, typical stress for a given diameter) will be calculated and reports updated for each 100 patients enrolled (ie, at 100 patients, 200 patients, and so on). Normative data will be stratified for AAA versus isolated thoracic aneurysms versus thoracoabdominal aortic aneurysms. Reported rupture risk will be based on existing literature until adequate data have been accumulated (eg, studies cited herein). Analysis of rupture risk based on the proposed study will be at timepoints suggesting potential statistical significance based on previous studies or for predetermined milestones (eg, when 100 patients have at least 1 year of follow-up). Once the enrolled patient number reaches adequate size, aneurysm wall stress data can be stratified for key variables (eg, gender). Of course, all patient-identifiable information will be excluded from this or any other public report.
Receiver Operating Characteristic Analysis for Predicting Rupture
Receiver operating characteristic (ROC) analysis will be performed to determine sensitivity, specificity, and accuracy for peak aneurysm wall stress in comparison to maximum aneurysm diameter.10 The optimal thresholds for aneurysm diameter and for peak wall stress will be determined for AAAs and TAAs, including comparison of the area under the ROC curve.
Multivariate analysis will include a proportional hazards analysis of rupture risk over time. A portion of this analysis will be a life-table report of freedom from rupture over time, with patients censored with aneurysm repair, death from other causes, withdrawal from the study, or loss to follow-up. Freedom from other outcome measures will be analyzed and reported similarly.
Maximum AAA diameter has been the standard index of AAA rupture risk for four decades, but it remains far from ideal. Other anatomic indices of rupture risk have been proposed, but none have been validated in other studies.9,14 We have previously shown that a noninvasive in vivo analysis of 3-D AAA wall stress from CT is superior to AAA diameter for predicting rupture risk in patients under observation,10 but this remains to be reproduced on a wide-scale, multicenter fashion. The proposed study should provide a mechanism to evaluate aneurysm wall stress analysis in a “real world” manner. Ideally, this study will form the foundation for a more formal trial of the technology by establishing patient norms and validating large-scale processing of stress analysis data using finite element modeling.
We believe that this approach will answer the many requests from physicians for access to stress analysis technology, but retain appropriate controls on the technology until it can be validated further on a broad scale. The proposed study also provides a unique opportunity to study other key aspects of aneurysms, such as wall stress in thoracic and thoracoabdominal aortic aneurysms, for which we have much fewer data. Data with regard to rupture risk based on gender, smoking, and other factors is far better known now than in the past, but further data in this area will nonetheless be helpful for placing rupture risk in the proper context, especially for factors that may be independent of wall stress. Lastly, the decision to repair an aneurysm is based on the risk of rupture versus the risk of repair, but within the context of the patients overall life expectancy. Hopefully, this study will also provide further insight into placing rupture risk in the context of life expectancy, adding to the data from other studies to date.
How to Participate
Contact Mimi Dion at Dartmouth Hitchcock Medical Center (Mary.Ellen.Dion@Hitchcock.org), who is compiling a contact list. E-mail is strongly preferred to facilitate group communications. We will then provide a protocol and guidelines for establishing IRB and HIPPA compliance at your institution, including more details. Phone contact and potential group meetings will be determined once the initial group size can be estimated. Estimated start-up time for patient entry is September 1, 2003.
The author acknowledge support for portions of this study from NHLBI NIH R01 HL64351-01: The Role Of Wall Stress Distribution In Abdominal Aortic Aneurysms, and from Medical Media Systems, Inc., West Lebanon, NH.
Mark F. Fillinger, MD, is from the Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire.
1. Brown PM, Pattenden R, Vernooy C, et al. Selective management of abdominal aortic aneurysms in a prospective measurement program. J Vasc Surg. 1996;23:213-220.
2. Brown LC, Powell JT. Risk factors for aneurysm rupture in patients kept under ultrasound surveillance. UK Small Aneurysm Trial Participants. Ann Surg. 1999;230:289-296.
3. Darling RC, Messina CR, Brewster DC, et al. Autopsy study of unoperated abdominal aortic aneurysms: the case for early resection. Circulation. 1977;56(Suppl):II161-164.
4. Cronenwett JL, Murphy TF, Zelenock GB, et al. Actuarial analysis of variables associated with rupture of small abdominal aortic aneurysms. Surgery. 1985;98:472-483.
5. Nicholls SC, Gardner JB, Meissner MH, et al. Rupture in small abdominal aortic aneurysms. J Vasc Surg. 1998;28:884-888.
6. Conway KP, Byrne J, Townsend M, et al. Prognosis of patients turned down for conventional abdominal aortic aneurysm repair in the endovascular and sonographic era: Szilagyi revisited? J Vasc Surg. 2001;33:752-757.
7. Heikkinen M, Salenius J, Zeitlin R, et al. The fate of AAA patients referred electively to vascular surgical unit. Scand J Surg. 2002;91:345-352.
8. Lederle FA, Johnson GR, Wilson SE, et al. Rupture rate of large abdominal aortic aneurysms in patients refusing or unfit for elective repair. JAMA. 2002;287:2968-2972.
9. Fillinger MF, Raghavan ML, Marra SP, et al. In vivo analysis of mechanical wall stress and abdominal aortic aneurysm rupture risk. J Vasc Surg. 2002;36:589-597.
10. Fillinger MF, Marra SP, Raghavan ML, et al. Prediction of rupture risk in abdominal aortic aneurysms during observation: wall stress versus diameter. J Vasc Surg. 2003;37:724-732.
11. Fillinger MF, Racusin J, Baker RK, et al. Anatomic characteristics of ruptured abdominal aortic aneurysms: implications for rupture risk and potential for endovascular repair. J Vasc Surg. 2003; Submitted for publication.
12. Powell JT, Brown LC. The natural history of abdominal aortic aneurysms and their risk of rupture. Adv Surg. 2001;35:173-185.
13. Brewster DC, Cronenwett JL, Hallett JW, Jr., et al. Guidelines for the treatment of abdominal aortic aneurysms. J Vasc Surg. 2003;37:1106-1117.
14. Ouriel K, Green RM, Donayre C, et al. An evaluation of new methods of expressing aortic aneurysm size: relationship to rupture. J Vasc Surg. 1992;15:12-20.
15. Uniform requirements for a manuscript submitted to biomedical journals. JAMA. 1997;277:927-934.