Published online before print October 23, 2008, doi: 10.1161/STROKEAHA.108.525709
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2009;40:156.)
© 2009 American Heart Association, Inc.
Original Contributions |
Embolic Potential of Cardiac Tumors and Outcome After Resection
A Case–Control Study
From the Department of Surgery (A.W.E.), Brigham and Women’s Hospital, Harvard Medical School, Boston, Mass; and the Division of Cardiovascular Surgery (A.W.E., J.A.D., R.C.D., C.J.M., T.A.O., F.J.P., H.V.S.), Mayo Clinic and Foundation, Rochester, Minn.
Correspondence to Andrew W. ElBardissi, MD, MPH, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115. E-mail aelbardissi{at}partners.org
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Background and Purpose— Embolic events have long been thought to occur in patients with cardiac tumors secondary to embolization of tumor fragments; however, there are no large studies examining the epidemiology and occurrence of embolism in this group of patients.
Methods— From 1957 to 2006, 323 consecutive patients with primary cardiac tumors were treated surgically at our institution. Of these, patients who experienced an embolic event included 80 (cerebrovascular accident 31 [9.7%], transient ischemic attack 30 [9.3%], and other 19 [6%]). Those with no history of an embolic event (n=243 [75%]) were defined as control subjects.
Results— Age was similar between the case and control groups (mean 54.5 versus 53.9 years, P=0.8). A multivariate logistic regression model including tumor location, tumor burden, tumor histology, and cerebrovascular risk factors, indicated that left atrial tumors (OR, 1.95; P=0.04), aortic valve tumors (OR, 4.17; P=0.002), and smaller tumor burden (OR, 2.20; P=0.01) were the most significant factors in the occurrence of embolism (P<0.001). The presence of mitral regurgitation (OR, 0.12; P=0.006) and decreased functional status (New York Heart Association III/IV; OR, 0.31; P<0.001) were protective against the occurrence of embolism. Follow-up was obtained in 82% at a mean follow-up time of 6.17±6.9 years. There were no recurrent embolic events at follow-up. A Kaplan-Meier survival curve demonstrated no difference in survival between both groups (P=0.78).
Conclusion— Aortic valve and left atrial tumors have the greatest anatomic risk for embolism. Furthermore, patients with smaller tumors, minimal symptomatology, and no evidence of mitral regurgitation have a high risk of embolism. Cardiac tumors can be resected with low early mortality, and late survival after operation in the context of an embolic event is similar to patients with cardiac tumors who undergo resection for other indications.
Key Words: cardiac tumor • embolism • stroke • TIA
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Tumors of the heart have the potential to cause obstruction of intracardiac flow, arrhythmias secondary to myocardial invasion, constitutional symptoms, or embolization of tumor fragments.1–6 The scientific literature contains many case reports and small case-series highlighting patients who experience adverse outcomes; however, there have been no studies examining epidemiological and tumor characteristics that may predispose to the occurrence of these clinical events. Arguably, the most feared consequence of cardiac tumors is systemic embolization with some studies estimating this risk to be as high as 30% to 40% in patients with myxomas.4,5,7 The risk of embolization is not limited to myxomas; papillary fibroelastomas and other tumor types have been reported to have high rates of embolization as well.8,9 Due to the low incidence of cardiac tumors, there have been no studies appropriately designed and powered to develop a risk prediction model of tumor and clinical factors that may predispose to embolization. Furthermore, there is uncertainty regarding the timing of surgery when patients experience a systemic embolic event and are subsequently found to have a cardiac tumor.
The purpose of this study was to compare the characteristics of patients with primary cardiac tumors who presented to our institution in the context of a recent embolic event to a cohort of patients with primary cardiac tumors who had no history of an embolic event. Additionally, we sought to compare survival characteristics between patients in our case and control groups to ascertain if immediate surgical resection in the context of an embolic event results in acceptable short- and long-term outcomes.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
From December 11, 1957, to April 19, 2005, 521 consecutive patients presented to our institution and were diagnosed with a cardiac mass suggestive of a cardiac tumor. Of these, 329 patients were diagnosed with a cardiac tumor that originated from the epicardium, myocardium, or endocardium; 323 (98%) patients were treated surgically and 6 (2%) were treated conservatively. The remaining 192 (37%) patients had metastatic cancer to the heart. Case group inclusion criteria included all patients who presented to our institution with evidence of stroke, transient ischemic attack, or other systemic/pulmonary embolic event according to commonly accepted diagnostic criteria.10 On investigation, these patients were found to have a cardiac tumor and underwent surgical resection with subsequent histological confirmation of a cardiac tumor originating from the heart. Case group exclusion criteria included patients who presented to our institution with a history of an embolic event but found to have a cardiac tumor of secondary origin (tumor metastasis to the heart), patients who experienced an embolic event but were not offered surgical resection, or patients in whom a suspected embolic event did not meet predefined and generally accepted criteria.10 The control group consisted of all patients with histological confirmation of a primary cardiac tumor and no history (immediate and past medical history) of an embolic event treated surgically at our institution during the same time interval.
A retrospective case–control study was performed, and medical records were reviewed for patient demographics and the cerebrovascular risk factors of hypertension,11 diabetes mellitus,12 and hyperlipidemia13; these factors were classified according to present-day criteria. Due to the retrospective nature of this study and the variable reporting of smoking and alcohol use in patient histories, these factors were not studied directly. Cardiac medical and surgical history, operative procedure, pre-/postoperative data, early and late morbidity, and survival were also analyzed.
Statistical Analysis
Demographic and other patient-related data were obtained from Mayo Clinic (Rochester, Minn) medical records. Follow-up information was obtained from subsequent clinic visits, written correspondence from local physicians, and mailed questionnaires to patients or families. The results were analyzed in multiple phases. Initially, descriptive statistics and univariate analyses were used to compare epidemiological, cerebrovascular, and cardiac tumor variables between the case and control groups. Continuous variables are expressed as a mean±SD. Student t tests and Pearson’s 
2 tests were used to analyze continuous and categorical variables, respectively. During the second phase of analysis, candidate risk factors that included cerebrovascular, epidemiological, and cardiac tumor variables were entered into a logistic regression model, and a backward logistic regression selection method was used to develop the most significant model. Additional variables were subsequently entered into the model by order of univariate probability value to assess for confounding. To attain confidence in the robustness of the model, bootstrap bagging14,15 was used for a subset of 200 samples; variables with a probability value of <0.05 in >50% of the models were kept in the final regression model. Based on the final logistic regression model, a clinical prediction scoring system was developed by using the optimal c-statistic; prediction scores were developed based on the lowest common multiple of the odds ratios (ORs) derived from the logistic regression model.
To analyze survival after surgical resection, a Kaplan-Meier survival curve was constructed to assess long-term outcome and a log-rank test was used to determine significant differences in survival characteristics between the case and control groups. A Cox proportional hazards model was used to identify factors associated with poor outcome in the case group. Because this study spans nearly 50 years, surgical era was trichotomized according to general advances in surgical technique (1955 to 1980, n=53; 1981 to 1995, n=104; 1996 to present, n=166) and included in the Cox proportional hazards model to control for survival advantages that might exist according to surgical era. Statistical significance was considered at P<0.05. SAS version 9.1 and JMP version 7.0 (both SAS Institute Inc, Cary, NC) were used for statistical analysis. Early operative mortality was defined as death occurring within 30 days of surgery or at any time during the index hospitalization. The Mayo Foundation Institutional Review Board approved this study, and all patients or their families gave written informed consent.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Of the 323 patients with primary cardiac tumors treated surgically at our institution, 80 (25%) presented in the context of a recent embolic event and met the criteria to be included as cases; 31 patients presented with stroke, 30 with transient ischemic attack, and 19 with other systemic/pulmonary emboli. The remaining 243 (75%) patients served as the control group. Since 1957, there has been a steady increase in the resection of cardiac tumors in both groups; however, there was no significant difference in the year of resection between the groups (P=0.69). Patients in the case group had a similar age distribution, gender proportion, and cardiovascular risk profile as control subjects (Table 1); however, cases more often presented with a greater cardiovascular functional capacity than did controls. There was no significant difference in the incidence of atrial fibrillation or use of preoperative anticoagulation between the case and control groups.
|
Ischemic Event Diagnosis and Classification
Patients who experienced strokes and systemic/pulmonary embolic events were further classified as to the location of the ischemic event. Because the timespan of this study is large, there were various methods of diagnosis that are beyond the scope of this analysis; however, the events described represent the findings of clinical diagnoses (made at the time of ischemic event identification) as well as results obtained from advanced diagnostic imaging. Of the 31 patients with stroke, 71% (n=22) had partial anterior circulation strokes, 13% (n=4) had strokes in multiple circulation distributions, 10% (n=3) had partial posterior circulation strokes, and 6% (n=2) had partial inferior circulation strokes. Of the 19 patients who experienced a systemic/pulmonary embolic event, 63% (n=12) had an ischemic event in the lower extremities, 11% (n=2) in the upper extremities, 11% (n=2) in the coronary circulation, 11% (n=2) in the retinal artery, and 5% (n=1) in the pulmonary artery.
Cardiac Tumor Diagnosis
The majority of patients in both the case and control groups were diagnosed by echocardiography (85% [n=210] versus 86% [n=68], respectively). The remaining patients were primarily diagnosed by cardiac catheterization (case group 13% [n=10], control group 10% [n=23]); however, use of this modality appeared to be time-dependent because these patients were primarily operated on in an early surgical era. (1976 versus 1993, P=0.004).
Echocardiography
As shown in Table 2, of those patients diagnosed by echocardiography, univariate analysis indicated that patients who experienced an embolic event had a higher ejection fraction (P=0.02) and were less likely to present with mitral regurgitation (P=0.001) and mitral obstruction (P=0.02) secondary to tumor prolapse. Echocardiographic images of the case group were reviewed to determine if tumors appeared to have abnormal contours or dimensions secondary to embolization of tumor fragments, thereby resulting in a significantly smaller tumor volume at the time of diagnosis. In all cases, echocardiographic images revealed cardiac tumors with naturally appearing contours.
|
Surgical Resection
All patients in the case group were treated surgically within 10 days of tumor diagnosis; the operative parameters are shown in Table 2. Patients who recently experienced an embolic event had significantly shorter cardiopulmonary bypass (P=0.0002) and crossclamp (P=0.0004) times. There was a trend toward increased concomitant mitral valve surgery in the control group (P=0.05); however, there were no differences in the incidence of other concomitant operations. Histological diagnoses were confirmed after surgery; among patients who were operated on with a recent embolic event, papillary fibroelastomas (P=0.04) and aortic valve tumors were (P=0.004) more likely to have embolized than other histological variants or tumor locations.
A backward logistic regression model, including patient demographics, cerebrovascular risk factors, tumor location, pathology, and size characteristics, revealed a significant embolic prediction model (Table 3). Left atrial tumor tumors were 1.95 times and aortic valve tumors were 4.17 times more likely to result in an embolic event than other cardiac tumors. After adjustment for tumor location, the volume of excised tumor (dichotomized according to the overall distribution of the tumor volumes) was inversely related to embolic risk. Thus, tumors between 0.02 cm3 and 13.13 cm3 were 2.2 times more likely to result in an embolic event compared to tumors >13.3 cm3. When adjusted for other factors, patients who presented with mitral regurgitation (OR, 0.12) and those with decreased functional capacity (New York Heart Association [NYHA] III/IV) were less likely (OR, 0.31) to have experienced an embolic event. No additional confounders were identified in the model building process; however, it was clear from the analysis that left atrial location was confounded by tumor burden, NYHA functional class, and the presence of mitral regurgitation. After adjustment of these variables, left atrial location was brought into statistical significance. A bootleg bagging14,15 statistical procedure confirmed the robustness of the logistic regression model with greater than 50% of all models demonstrating statistical significance of the variables shown in Table 3.
|
Clinical Prediction Model
In an effort to enhance the clinical applicability of these results, a prediction model demonstrating the risk of tumor embolism was derived (Figure 1) and showed good discriminatory ability (c=0.75). A plot of prediction scores versus probability of embolism is shown in Figure 1. Scores were generated based on the lowest common multiple of the ORs developed from the logistic regression model shown in Table 3. The most significant predictive factor of embolization is the presence of an aortic valve tumor, which individually translated into a 14% risk of embolization. However, using this model, patients who scored 
7 had a 45% risk of embolization and patients who achieved a score of 9 had a 79% risk of embolism.
|
Postoperative Course
Postoperative complications included stroke, transient ischemic attack, pulmonary embolus, pneumonia, respiratory insufficiency, renal insufficiency (requiring temporary dialysis), and wound infection. Overall, there was no significant difference between the occurrence of total postoperative complications (n=7 [9%] versus n=24 [10%]; P=0.77) or individual complications between the case and control groups. Specifically, no patients in either group experienced stroke or transient ischemic attack in the postoperative period. Operative mortality (30-day mortality) was similar between the case and control groups (n=1 [1%] versus n=7 [3%]; P=0.42).
Follow-Up
Follow-up was obtained in 82% of early survivors in the case group and 80% of the control group (P=0.62) at a mean follow-up time of 6.62±7.4 and 4.8±4.9 years, respectively (P=0.08). All patients in whom follow-up was available responded to a standardized questionnaire, were followed up in our clinic, or were confirmed to have died by family members who provided death certificates. A Kaplan-Meier survival curve is shown in Figure 2. A log-rank test indicated no difference in the survival characteristics between the case and control groups (P=0.78). Median survival after resection was 25.3 years and 16.5 years in the case and control groups, respectively. No patients in the case group experienced neurological or other systemic embolic events at last follow-up. After adjusting for the year of surgery, a Cox proportional hazards model of all-cause mortality in the case group indicated that patients with myxomas had a survival advantage compared to the remainder of the cohort (hazard ratio, 0.15; P=0.029), whereas patients who experienced an embolic event secondary to a malignant tumor had poor survival characteristics (hazard ratio, 74; P=0.001).
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We present the findings of a case–control study identifying the risks of cardiac tumor embolization and outcome after surgical resection, the largest series in the scientific literature to do so. Previously, reports of cardiac tumor embolization have estimated this risk to range from 12% to 45%.16–19 According to this study, the overall rate of tumor embolization is 25%; however, our principle finding is that structural and anatomic tumor characteristics highly influence the risk of embolization. Our second major finding is that long-term survival after cardiac tumor resection in patients who present in the context of an embolic event is excellent. This has implications for optimizing treatment because many patients with cardiac tumors who experience a recent embolic event may not be offered immediate surgical treatment given their assumed high-risk profile. Based on these findings, patients who experience an embolic event and undergo immediate surgery have short- and long-term survival that is comparable to a cohort of patients with cardiac tumors being operated on for other reasons.
We believe our regression model accurately illustrates the natural history of cardiac tumors. Left atrial tumors are most often diagnosed when they cause hemodynamic (ie, mitral valve obstruction/regurgitation) and symptomatic complications, the general presentation characteristics of our control group. Comparatively, small left atrial tumors are unlikely to warrant investigation because patients rarely have concomitant cardiovascular or hemodynamic alterations, and consequently appear to be asymptomatic. Similarly, aortic valve tumors (predominantly papillary fibroelastomas) rarely cause symptoms given their characteristically small dimensions and minimal hemodynamic effects. Thus, left atrial and aortic valve tumors that do not reach anatomic dimensions great enough to alter cardiovascular function may lie dormant for long periods of time, thereby increasing the probability of embolization, whereas larger tumors may undergo diagnosis and resection early in the tumor life cycle. In light of these findings, caution should be exercised when recommending conservative (nonsurgical) treatment for cardiac tumors; larger tumors pose an increased risk of irreversible cardiac and hemodynamic deterioration, whereas smaller, deceptively "innocent" tumors are at high risk of systemic embolization.
Echocardiography has had a tremendous impact on the timely diagnosis and excellent long-term outcome of patients with cardiac tumors. Although cardiac tumors are rare, occurring with a lifetime incidence of 0.0017% to 0.02%,3,20–23 they represent a distinct, curable source of systemic embolism that are easily identified by echocardiography. The American Heart Association guidelines propose that during evaluation of an embolic event, intracardiac masses should be suspected and appropriately pursued with echocardiography based on a suspicious clinical presentation.24 Our results suggest that patients who experience an event due to cardiac tumor embolization are unlikely to be suspected of having cardiac tumors based on clinical suspicion given their minimal symptomatology and normal hemodynamic profile. Thus, it is possible that many patients who experience an embolic event secondary to a cardiac tumor may go undiagnosed, thereby emphasizing the liberal use of echocardiography when a distinct source or cause of ischemia has not been identified.
Short- and long-term outcomes after surgical resection of benign primary cardiac tumors are consistently reported as being excellent with a low incidence of tumor recurrence and embolic events in the follow-up period.17,25–27 Conversely, the prognosis for primary malignant tumors of the heart has historically been dismal with the vast majority of patients dying within 1 year,28–31 a finding confirmed in this study. Thus, patients with malignant cardiac tumors represent the only subgroup of patients who experience an embolic event that may be best served with conservative nonsurgical treatment and aggressive adjuvant chemotherapy.32,33 Malignant cardiac tumors are often easily characterized by echocardiography because they are highly invasive at the time of diagnosis and patients most frequently present with constitutional symptoms.34 In cases in which a malignant tumor is suspected, a biopsy is appropriate to obtain histological verification, thus better directing treatment strategy.
Study Limitations
This is a retrospective case–control observational study; a major limitation of such studies is appropriate representation of both the case and control groups, thereby avoiding bias. We attempted to account for this by analyzing all patients who presented to the Mayo Clinic with a diagnosis of any cardiac mass, and of note, only 6 patients with cardiac tumors were treated without surgery of which none had experienced an embolic event. Based on this comprehensive analysis, we believe that the results reported here are accurate and account for the possibility of treatment bias. Additionally, although left atrial location did not approach statistical significance in the univariate analysis due to significant confounding, we believe our use of bootstrap bagging reinforces the robustness of our logistic regression model.
A significant limitation of this study is the timespan it covers given the tremendous advancement in medical imaging technology and surgical techniques. We attempted to compensate for this by analyzing the methods of diagnosis and only including variables that could be accurately identified in all patients, thereby enhancing internal validity. Other limitations include the imprecise ability to capture pertinent aspects of patient histories. It is plausible that patients in the control group may have experienced prior systemic ischemic events that were not well documented in patient histories given more severe cardiovascular symptoms at the time of presentation. Efforts were made to include variables known to predispose thrombotic events; most notably, hyperlipidemia, hypertension, and diabetes mellitus. Although these factors were similar between the case and control groups, our study did not assess for the risk factors of smoking and alcohol use, both of which are known to predispose to the risk of transient ischemic attack and stroke. Despite these limitations, we believe the results presented here represent an accurate profile of patients who experience embolic events secondary to tumor embolization.
Conclusion
Although cardiac tumors are rare, they present a significant risk of embolization. Anatomically, tumors of the aortic valve and left atrium present the greatest risk of embolization. Furthermore, patients with minimal cardiac symptoms, hemodynamic alterations, and smaller tumors are more likely to experience an embolic event in comparison to patients with larger cardiac tumors and greater cardiovascular and hemodynamic effects. In light of these findings, evaluation of all patients with echocardiography is strongly recommended in patients who present with a recent systemic ischemic event and no identifiable source or cause. Surgical resection in the context of a recent embolic event is safe with excellent short- and long-term outcome in all patients except those with malignant cardiac tumors.
|
Acknowledgments |
|---|
Disclosures
None.
|
Footnotes |
|---|
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
Presented at the American Heart Association: International Stroke Conference 2008.
Received May 12, 2008; accepted June 4, 2008.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
1. Larrieu AJ, Jamieson WR, Tyers GF, Burr LH, Munro AI, Miyagishima RT, Gerein AN, Allen P. Primary cardiac tumors: experience with 25 cases. J Thorac Cardiovasc Surg. 1982; 83: 339–348.[Abstract]
2. Vander Salm TJ. Unusual primary tumors of the heart. Semin Thorac Cardiovasc Surg. 2000; 12: 89–100.[Medline] [Order article via Infotrieve]
3. Wold LE, Lie JT. Cardiac myxomas: a clinicopathologic profile. Am J Pathol. 1980; 101: 219–240.[Medline] [Order article via Infotrieve]
4. St John Sutton MG, Mercier LA, Giuliani ER, Lie JT. Atrial myxomas: a review of clinical experience in 40 patients. Mayo Clin Proc. 1980; 55: 371–376.[Medline] [Order article via Infotrieve]
5. Goodwin JF. The spectrum of cardiac tumors. Am J Cardiol. 1968; 21: 307–314.[CrossRef][Medline] [Order article via Infotrieve]
6. Goodwin JF. Diagnosis of left atrial myxoma. Lancet. 1963; 1: 464–468.[Medline] [Order article via Infotrieve]
7. Greenwood WF. Profile of atrial myxoma. Am J Cardiol. 1968; 21: 367–375.[CrossRef][Medline] [Order article via Infotrieve]
8. Gowda RM, Khan IA, Nair CK, Mehta NJ, Vasavada BC, Sacchi TJ. Cardiac papillary fibroelastoma: a comprehensive analysis of 725 cases. Am Heart J. 2003; 146: 404–410.[CrossRef][Medline] [Order article via Infotrieve]
9. Ngaage DL, Mullany CJ, Daly RC, Dearani JA, Edwards WD, Tazelaar HD, McGregory CG, Orszulak TA, Puga FJ, Schaff HV, Sundt TM 3rd, Zehr KJ. Surgical treatment of cardiac papillary fibroelastoma: a single center experience with eighty-eight patients. Ann Thorac Surg. 2005; 80: 1712–1718.
10. The World Health Organization MONICA Project (monitoring trends and determinants in cardiovascular disease): a major international collaboration. WHO MONICA Project Principal Investigators. J Clin Epidemiol. 1988; 41: 105–114.[CrossRef][Medline] [Order article via Infotrieve]
11. The sixth report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure. Arch Intern Med. 1997; 157: 2413–2446.
12. Genuth S, Alberti KG, Bennett P, Buse J, Defronzo R, Kahn R, Kitzmiller J, Knowler WC, Lebovitz H, Lernmark A, Nathan D, Palmer J, Rizza R, Saudek C, Shaw J, Steffes M, Stern M, Tuomilehto J, Zimmet P; Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Follow-up report on the diagnosis of diabetes mellitus. Diabetes Care. 2003; 26: 3160–3167.
13. Summary of the second report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA. 1993; 269: 3015–3023.
14. Breiman L. Bagging predictors. Machine Learning. 1996; 24: 123–140.
15. Blackstone EH. Breaking down barriers: helpful breakthrough statistical methods you need to understand better. J Thorac Cardiovasc Surg. 2001; 122: 430–439.
16. Lee VH, Connolly HM, Brown RD Jr. Central nervous system manifestations of cardiac myxoma. Arch Neurol. 2007; 64: 1115–1120.
17. Knepper LE, Biller J, Adams HP Jr, Bruno A. Neurologic manifestations of atrial myxoma. A 12-year experience and review. Stroke. 1988; 19: 1435–1440.
18. Pinede L, Duhaut P, Loire R. Clinical presentation of left atrial cardiac myxoma. A series of 112 consecutive cases. Medicine (Baltimore). 2001; 80: 159–172.[CrossRef][Medline] [Order article via Infotrieve]
19. Sandok BA, von Estorff I, Giuliani ER. CNS embolism due to atrial myxoma: clinical features and diagnosis. Arch Neurol. 1980; 37: 485–488.
20. Silverman N. Primary cardiac tumors. Ann Surg. 1980; 191: 127–138.[Medline] [Order article via Infotrieve]
21. Chitwood WR Jr. Cardiac neoplasms: current diagnosis, pathology, and therapy. J Card Surg. 1988; 3: 119–154.[Medline] [Order article via Infotrieve]
22. Benjamin H. Primary fibromyxoma of the heart. Arch Pathol. 1939; 27: 950.
23. Heath D. Pathology of cardiac tumors. Am J Cardiol. 1968; 21: 315–327.[CrossRef][Medline] [Order article via Infotrieve]
24. Cheitlin MD, Alpert JS, Armstrong WF, Aurigemma GP, Beller GA, Bierman FZ, Davidson TW, Davis JL, Douglas PS, Gillam LD. ACC/AHA Guidelines for the Clinical Application of Echocardiography. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Clinical Application of Echocardiography). Developed in collaboration with the American Society of Echocardiography. Circulation. 1997; 95: 1686–1744.
25. ElBardissi AW, Dearani JA, Daly RC, Mullany CJ, Orszulak TA, Puga FJ, Schaff HV. Analysis of benign ventricular tumors: long-term outcome after resection. J Thorac Cardiovasc Surg. 2008; 135: 1061–1068.
26. Roeltgen DP, Weimer GR, Patterson LF. Delayed neurologic complications of left atrial myxoma. Neurology. 1981; 31: 8–13.
27. Sandok BA, von Estorff I, Giuliani ER. Subsequent neurological events in patients with atrial myxoma. Ann Neurol. 1980; 8: 305–307.[CrossRef][Medline] [Order article via Infotrieve]
28. Burke AP, Cowan D, Virmani R. Primary sarcomas of the heart. Cancer. 1992; 69: 387–395.[CrossRef][Medline] [Order article via Infotrieve]
29. Poole GV Jr, Breyer RH, Holliday RH, Hudspeth AS, Johnston FR, Cordell AR, Mills SA. Tumors of the heart: surgical considerations. J Cardiovasc Surg (Torino). 1984; 25: 5–11.[Medline] [Order article via Infotrieve]
30. Reece IJ, Cooley DA, Frazier OH, Hallman GL, Powers PL, Montero CG. Cardiac tumors. Clinical spectrum and prognosis of lesions other than classical benign myxoma in 20 patients. J Thorac Cardiovasc Surg. 1984; 88: 439–446.[Abstract]
31. Wiske PS, Gillam LD, Blyden G, Weyman AE. Intracardiac tumor regression documented by two-dimensional echocardiography. Am J Cardiol. 1986; 58: 186–187.[CrossRef][Medline] [Order article via Infotrieve]
32. Leu KM, Ostruszka LJ, Shewach D, Zalupski M, Sondak V, Biermann JS, Lee JS, Couwlier C, Palazzolo K, Baker LH. Laboratory and clinical evidence of synergistic cytotoxicity of sequential treatment with gemcitabine followed by docetaxel in the treatment of sarcoma. J Clin Oncol. 2004; 22: 1706–1712.
33. Fata F, O'Reilly E, Ilson D, Pfister D, Leffel D, Kelsen DP, Schwartz GK, Casper ES. Paclitaxel in the treatment of patients with angiosarcoma of the scalp or face. Cancer. 1999; 86: 2034–2037.[CrossRef][Medline] [Order article via Infotrieve]
34. Bakaeen FG, Reardon MJ, Coselli JS, Miller CC, Howell JF, Lawrie GM, Espada R, Ramchandani MK, Noon GP, Weilbaecher DG, DeBakey ME. Surgical outcome in 85 patients with primary cardiac tumors. Am J Surg. 2003; 186: 641–647; discussion 647.[CrossRef][Medline] [Order article via Infotrieve]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||