Management of intracranial cavernous malformations using conservative vs. surgical and/or radiosurgical treatment: A systematic review and meta‑analysis
- Authors:
- Published online on: March 20, 2024 https://doi.org/10.3892/etm.2024.12503
- Article Number: 215
-
Copyright: © Fotakopoulos et al. This is an open access article distributed under the terms of Creative Commons Attribution License.
Abstract
Introduction
Intracranial cavernous malformations (CMs) are vascular lesions that have an annual bleeding rate of ~0.2 to 3% per individual per year (1). In the literature, there are three management options for addressing CMs: Operative resection, radiosurgery (SRS) and conventional treatment; however, the debate regarding the treatment options for CMs has a long history and remains controversial (2).
The effects of SRS on cavernoma remain hypothetical (3-5). The outcome of CM management can be stated only as a decreased bleeding rate for a large number of patients, which then necessitates dependable data on the natural course. On an individual level, the treatment result is extremely hypothetical, as it is necessary to have knowledge of the natural history of the condition to ensure that a good benefit is achieved (6,7).
Surgery is an alternative treatment option for CMs, with a complete resection to reach a temporary morbidity rate varying from 29 to 67% and a 1.9% combined post-operative re-bleeding and surgery-related mortality rate (1,8). In addition, 58% of incomplete resection cases re-bleed. Hence, concerning surgical management, the proportion of no active treatment cases, the direct morbidity and mortality rates, and the risk of partial removal of CMs with the prospect of re-hemorrhage have to be recalculated (1).
In addition, a number of CMs are considered untreatable due to their placement in eloquent areas. Thus, surgically approachable cavernomas consist of a detailed assortment, whereas a number of untreatable cases can eventually be managed with SRS. Hence, the effectiveness of surgery and SRS can only be estimated based on an accurate designation of exclusion criteria and the exact risks for complications and re-hemorrhage for both treatment options.
The present systematic review and meta-analysis compared with previous reports (9,10), aimed to evaluate the safety of surgical or SRS treatment for the management of CMs and also evaluate their potential outcomes compared with conservative treatment.
Materials and methods
Literature research strategy
The present meta-analysis investigated the relative articles involving intracranial cavernous malformation (CMs) natural history vs. surgical or radiosurgical (SRS) treatment option through electronic databases, counting the Cochrane Library, PubMed (until June, 2023), Embase (until June, 2023), and MEDLINE (until June, 2023). For the study protocol establishing and design, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were applied. ‘Cerebral Cavernous Malformation’, ‘Cerebral Cavernus Malformation natural history vs. surgical or radiosurgical treatment’, ‘Intracranial Cavernus Malformation natural history vs. surgical or radiosurgical treatment option’, and ‘Cerebral Cavernous Malformation natural history vs. surgical treatment’, were used in the MeSH list as keywords.
Selection of studies
In the present study, two authors (VEG and GF) separately pulled out data from the contained articles, following the guidelines for the epidemiology of meta-analysis. The subsequent crucial information was attained: The main authors, year of publication, sample size in the CM natural history vs. surgical or SRS treatment option groups, study type, outcome indicator, etc. The extracted data were contributed to a designed, standardized table according to the Cochrane Handbook. The flow of the study selection process is presented in Fig. 1.
Inclusion and exclusion criteria
If an article fulfilled the subsequent population, intervention, comparison, outcomes and study (PICOS) design criteria, it was eligible for inclusion in the present meta-analysis: i) Population: Limited to patients with intracranial CMs; ii) Intervention: Limited to patients with CMs natural history vs. surgical or/and SRS treatment option; iii) Comparison: Studies comparing the outcomes between the CM natural history (conservative treatment; Cons) vs. the surgical or/and SRS (surg/SRS) treatment option; iv) the comprehensive data of these articles are presented in Table I. To avoid publication bias, the final aim was to collect a homogeneous pool of manuscripts, including articles that compare only two modalities: Patients with intracranial CMs treated with the Cons vs. Surg/SRS treatment option.
All retrospective and prospective studies that assessed these two modalities together were included, whereas editorials, reviews, case reports and articles focusing on the pediatric population, unrelated outcomes, co-morbidities, experimental techniques, or one of the two modalities separately from that article pool were excluded. Additionally, to decrease the risk of bias in the included articles, a quality assessment tool [the Newcastle-Ottawa Scale (NOS)] was used (Table II) (11).
Outcomes' definition
The primary outcome was ‘poor outcome’, defined as at least two successive ratings of the Oxford Handicap Scale (OHS) (12) and OHS 2-6 (suggestive of ‘some restraints to lifestyle, but the patient can look after themselves’, or worse). It was used only for OHS ratings after the initial presentation to time progression to this event at the midpoint between the last OHS score of 0-1 and the first of the successive OHS 2-6 ratings for the conservatively managed group (12).
Secondary outcomes were the frequency of seizures in the surgical or + SRS and cons patients, neurological deficit, re-bleeding and mortality. Information regarding age, sex, localization (lobar, deep, brainstem, cerebellum), size (<2, 2-6 mm) and use of hemosiderin in the pre-surgical MRI is presented in Table I. Re-bleeding was defined as hemorrhage clearly demonstrated on imaging at the time of admission.
Evaluation of the risk of bias
The Cochrane Collaboration tool was used to assess the risk of bias and was used by two authors (GF and VEG) for each study (13). The assessment contained allocation concealment, random sequence generation, the blinding of outcome evaluation, the blinding of participants and assessors, unfinished outcome data, discriminating reports and other biases. The evaluated results were classified into three levels: Low risk, high risk and unclear risk. In the case of a discrepancy, another author with authority gave the final solution.
Statistical analysis and assessment of heterogeneity
All analyses were carried out using Review Manager Software (RevMan), version 5.4. Heterogeneity across trials was identified using I2 statistics; considering I2 >50% as high heterogeneity, a meta-analysis was conducted using a random-effect model according to the Cochrane Handbook for Systematic Reviews of Interventions (version 5.1.0) (14). Otherwise, the fixed-effect model was performed. The continuous outcomes were expressed as a weighted mean difference with 95% confidence intervals (CIs). For discontinuous variables, odds ratios (OR) with 95% CIs were applied for the assessment. A P-value <0.05 was considered to indicate a statistically significant difference.
Results
In total, six articles (1,15-19) met the eligibility criteria. The total number of patients was 399 (157 in the Surg/SRS group and 242 in the Cons treatment group). The study sample was based on six studies (Table I). Of these six studies, two studies were retrospective and four studies were prospective.
Epidemiological and clinical features
The mean age of the patients among the included studies ranged from 34.9 to 56 years. The male-to-female ratio was 1.5 for the Surg/SRS group and 0.9 for the Cons treatment group (95/62 and 119/123) (Table I).
Location
Lobar. Information regarding lobar brain location was available in three articles (1,17,18). No significant difference was found between the Surg/SRS and Cons groups (OR, 1.03; 95% CI, 0.55 to 1.93; P=0.07), with heterogeneity (P=0.92 and I2=62%) (Table III and Fig. S1).
Deep. Information regarding deep brain location was available in three articles (1,17,18). No significant difference was found between the Surg/SRS and Cons groups (OR, 0.62; 95% CI, 0.27 to 1.43; P=0.87), without heterogeneity (P=0.26 and I2=0%) (Table III and Fig. S2).
Brainstem. As regards brainstem location, information was available in three articles (1,17,18). No significant difference was found between groups (OR, 0.24; 95% CI, 0.03 to 1.92; P=0.18) (Table III and Fig. S3).
Cerebellum. As regards cerebellum location, information was available in three articles (1,17,18). No significant difference was found between groups (OR, 1.29; 95% CI, 0.39 to 4.32; P=0.68) (Table III and Fig. S4).
Hemosiderin in the pre-MRI
As regards the use of hemosiderin in the pre-MRI, information was available in six articles (1,15-18), and this demonstrated a statistically significant result (OR, 2.56; 95% CI, 1.20 to 5.47; P<0.05), with very low heterogeneity (P=0.22; I2=32%) (Table III and Fig. 2). The use of hemosiderin in the pre-MRI was found in 106 of 132 (80.3%) patients in the Surg/SRS group and in 99 of 207 (47.8%) patients in the Cons group. When examining the funnel plot of the same parameter, no publication bias was found.
Free of seizures parameter
Information regarding the free of seizures parameter was available in four articles (1,15,16,18) and demonstrated a statistically significant result between the patients in the Surg/SRS and Cons groups (OR, 3.49; 95% CI, 1.79 to 6.83; and P<0.05), but with heterogeneity (P<0.05 and I2=82%) (Fig. 3A). For testing the sensitivity, the ‘leave out one’ model was used, and one study was removed at a time (Table III). Low heterogeneity (P=0.18 and I2=41%) was achieved only after removing the article by Fernández et al (1); again, a statistically significant difference was found (OR, 5.27; 95% CI, 2.60 to 10.68; P<0.05) (Fig. 3B). When examining the funnel plot of the same parameter, it was found that the study results without the study by Fernández et al (1) displayed better dispersion with a low publication bias (Fig. 4).
Neurological deficit parameter
As regards neurological deficit, information was available in four articles (1,15-18), and this demonstrated a statistically significant result (OR, 0.57; 95% CI, 0.32 to 1.00; P=0.05) with no heterogeneity (P=0.34; I2=11%) (Table III and Fig. 5). A neurological deficit was found in 31 of 148 (20.9%) patients in the Surg/SRS group and in 45 of 136 (33.0%) patients in the Cons group. When examining the funnel plot of the same parameter, no publication bias was found. Thus, the Surg/SRS (experimental) group exhibited superiority over the Cons (control) group.
Re-bleeding
Information regarding the re-bleeding parameter was available in four articles (15-18) and demonstrated a statistically significant result between the patients with Surg/SRS and the Cons groups (OR, 2.84; 95% CI, 1.25 to 6.46; and P<0.05), but with heterogeneity (P<0.05 and I2=67%) (Fig. 6A). For testing the sensitivity, the ‘leave out one’ model was used, and one study was removed at a time (Table III). No heterogeneity (P=0.49 and I2=0%) was achieved only after removing the articles by Tarnaris et al (16) and Moultrie et al (17); again, a statistically significant difference was found (OR, 5.56; 95% CI, 1.02 to 6.67; P=0.05) (Fig. 6B). When examining the funnel plot of the same parameter, it was found that the study results without the studies by Tarnaris et al (16) and Moultrie et al (17) displayed better dispersion with a low publication bias (Fig. 7).
OHS 2-6
As regards OHS 2-6, information was available in five articles (15-19). No significant difference was found between groups (OR, 1.44; 95% CI, 0.83 to 2.49; P=0.19) (Table III and Fig. S5). In addition, after applying the ‘leave out one’ model, no statistically significant result was obtained (Table III).
Mortality
Information regarding mortality was available in three articles (15-19) and demonstrated a statistically significant result between the patients with Surg/SRS and the Cons groups (OR, 4.68; 95% CI, 1.97 to 11.09; and P<0.05) (Table III and Fig. 8A). Mortality was found in 21 of 65 (32.3%) patients in the Surg/SRS group and in 32 of 152 (21.1%) in the Cons group of patients. When examining the funnel plot of the same parameter, it was found that the study results had low heterogeneity (P=0.15 and I2=47%) and a low publication bias (Fig. 8B). Thus, the Cons (control) group exhibited superiority over the Surg/SRS (experimental) group.
Discussion
The present study suggests that surgical or SRS management may be a safe procedure as regards the outcomes of patients with CMs compared with conservative treatment. More precisely, neurological deficit was a statistically significant parameter in these patients, exhibiting the superiority of surgery and/or SRS over conservative management. Of note, the of hemosiderin in the pre-MRI, the free of seizures parameter, and the re-bleeding and neurological deficit parameters yielded statistically significant results, predicting a better outcome in the surgical or SRS group of patients. On the other hand, the mortality rate was lower in the Cons group of patients compared with the surgical or SRS treatment groups.
The treatment of CMs has been an ongoing topic of debate due to the issues concerning their management. The major difficulty with obtaining a clear perspective of their natural history is determining when these injuries should be operated on. In addition, published surgical series may ignore cases that never hemorrhage and are only found on the follow-up for other reasons. SRS is applied for the obliteration of the CMs, preventing any risk of re-bleeding, and is an alternative to surgical treatment (20,21). However, it appears that CMs following SRS re-hemorrhage repeatedly and thus no benefit has been observed (22,23). The present meta-analysis demonstrated that the re-bleeding rate was a statistically significant parameter in patients with CM who underwent surgical or/and SRS management.
Based on the literature data, it appears evident that the microsurgical management of CMs is the best option for patients with epilepsy (1). Nevertheless, there are no randomized studies evaluating pharmaceutical and surgical treatment in patients with CMs and epilepsy. In the present meta-analysis, the free of seizures parameter yielded a statistically significant result, predicting a better outcome in the surgical or SRS group of patients.
Researchers have asserted that patients with CMs have an amplified risk of re-hemorrhage following an initial bleed (21). Since a brain MRI is needed to diagnose CMs without other pathological examinations, apart from the use of hemosiderin in the MRI, intracranial hemorrhage in the pre-MRI brain imaging may increase the risk of re-hemorrhage (24). These risks may help to determine whether to treat CMs with neurosurgical excision and/or SRS or to opt for conservative management. In the present meta-analysis, the use of hemosiderin in the pre-MRI was a statistically significant parameter, predicting a better outcome in the surgical or SRS group of patients than in patients in the Cons group.
As regards outcomes due to surgical or/and SRS-related morbidity or CM-related mortality, a poorer outcome in surgically managed patients, such as that detailed in a previous study (17), was also found in the present meta-analysis. This may be explained by the complications associated with CM excisions. In the present meta-analysis, mortality was a statistically significant parameter, predicting a better outcome in the Cons group of patients.
The present meta-analysis has certain limitations, which should be mentioned. Half of the included studies were retrospective, and a limited number of cases were presented, particularly in the surgically treated group. In addition, there is an argument suggesting that surgical outcome depends on the time of the intervention, and some researchers have advocated for surgery at 4 weeks after ictus (15,23). Thus, the present meta-analysis did not include the time-dependent intervention parameter.
In conclusion, the present meta-analysis proposes that surgical or SRS management is a safe procedure as regards the out outcomes of patients with CMs compared with conservative treatment. More precisely, neurological deficit was statistically significant parameter in these patients, exhibiting the superiority of the surgical or/and SRS option over conservative management. Of note, the use of hemosiderin in the pre-MRI, and the free of seizures, re-bleeding and neurological deficit parameters yielded statistically significant results, predicting a better outcome in the surgical or SRS group of patients. On the other hand, the mortality rate was lower in the Cons group of patients compared with surgical or SRS treatment. Future studies are required to examine more precisely the outcomes in the natural history cases (conservative treatment), as the debate regarding the management of CMs remains controversial.
Supplementary Material
(A) Forest plot for lobar (location). The results demonstrated no statistically significant difference between the surgical or/+ SRS and Cons (OR, 1.03; 95% CI, 0.55 to 1.93; P=0.07). (B) Funnel plot, testing the sensitivity of the lobar (location); there was heterogeneity (P=0.92 and I2=62%). In addition, after the applying ‘leave out one’ model, no statistically significant result was obtained (Table III). SRS, radiotherapy; Cons, conservative management group; OR, odds ratio; I2, the percentage of total variation across studies that is due to heterogeneity rather than chance; CI, confidence interval.
(A) Forest plot for deep (location). The results demonstrated no statistically significant difference between the surgical or/+ SRS and Cons (OR, 0.62; 95% CI, 0.27 to 1.43; P=0.87). (B) Funnel plot, testing the sensitivity of deep (location); there was no heterogeneity (P=0.26 and I2=0 %). In addition, after the applying ‘leave out one’ model, no statistically significant result was obtained (Table III). SRS, radiotherapy; Cons, conservative management group; OR, odds ratio; I2, the percentage of total variation across studies that is due to heterogeneity rather than chance; CI, confidence interval.
(A) Forest plot for brainstem (location). The results demonstrated no statistically significant difference between the surgical or/+ SRS and Cons (OR, 0.24; 95% CI, 0.03 to 1.92; P=0.18). (B) Funnel plot, testing the sensitivity of the brainstem (location): there was no heterogeneity. In addition, after the applying ‘leave out one’ model, no statistically significant result was obtained (Table III). SRS, radiotherapy; Cons, that is due to heterogeneity rather than chance; CI, confidence interval.
(A) Forest plot for cerebellum (location). The results demonstrated no statistically significant difference between the surgical or/+ SRS and Cons (OR, 1.29; 95% CI, 0.39 to 4.32; P=0.68). (B) Funnel plot, testing the sensitivity of the brainstem (location): there was no heterogeneity. In addition, after the applying ‘leave out one’ model, no statistically significant result was obtained (Table III). SRS, radiotherapy; Cons, conservative management group; OR, odds ratio; I2, the percentage of total variation across studies that is due to heterogeneity rather than chance; CI, confidence interval.
(A) Forest plot for OHS 2-6. The results demonstrated no statistically significant difference between the surgical or/+ SRS and Cons (OR, 1.44; 95% CI, 0.83 to 2.49; P=0.19). (B) Funnel plot, testing the sensitivity of the brainstem (location): there was no heterogeneity. In addition, after the applying ‘leave out one’ model, no statistically significant result was obtained (Table III). OHS, Oxford Handicap Scale; SRS, radiotherapy; Cons, conservative management group; OR, odds ratio; I2, the percentage of total variation across studies that is due to heterogeneity rather than chance; CI, confidence interval.
Acknowledgements
Not applicable.
Funding
Funding: No funding was received.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Authors' contributions
GF and VEG conceptualized the study. VEG, DAS, NT, PS, PP, GF and KNF analyzed the data, and wrote and prepared the draft of the manuscript. VEG and GF provided critical revisions. All authors contributed to manuscript revision and have read and approved the final version of the manuscript. GF and VEG confirm the authenticity of all the raw data.
Ethics approval and consent to participate
Not applicable.
Patient consent for publication
Not applicable.
Competing interests
DAS is the Editor-in-Chief for the journal, but had no personal involvement in the reviewing process, or any influence in terms of adjudicating on the final decision, for this article. The other authors declare that they have no competing interests.
References
Fernández S, Miró J, Falip M, Coello A, Plans G, Castañer S and Acebes JJ: Surgical versus conservative treatment in patients with cerebral cavernomas and non refractory epilepsy. Seizure. 21:785–788. 2012.PubMed/NCBI View Article : Google Scholar | |
Steiner L, Karlsson B, Yen CP, Torner JC, Lindquist C and Schlesinger D: Radiosurgery in cavernous malformations: Anatomy of a controversy. J Neurosurg. 113:16–22. 2010.PubMed/NCBI View Article : Google Scholar | |
Fotakopoulos G, Andrade-Barazarte H, Kivelev J, Tjahjadi M, Goehre F and Hernesniemi J: Brainstem cavernous malformations management: Microsurgery vs radiosurgery, a meta-analysis. Front Surg. 8(630134)2022.PubMed/NCBI View Article : Google Scholar | |
Fotakopoulos G, Kivelev J, Andrade-Barazarte H, Tjahjadi M, Goehre F and Hernesniemi J: Outcome in patients with spinal cavernomas presenting with symptoms due to mass effect and/or hemorrhage: Conservative versus surgical management: Meta-analysis of direct comparison of approach-related complications. World Neurosurg. 152:6–18. 2021.PubMed/NCBI View Article : Google Scholar | |
Fotakopoulos G, Brotis AG and Fountas KN: Dilemmas in managing coexisting arteriovenous and cavernous malformations: Case report. Brain Circ. 8:45–49. 2022.PubMed/NCBI View Article : Google Scholar | |
Robinson JR, Awad IA and Little JR: Natural history of the cavernous angioma. J Neurosurg. 75:709–714. 1991.PubMed/NCBI View Article : Google Scholar | |
Sandalcioglu IE, Wiedemayer H, Secer S, Asgari S and Stolke D: Surgical removal of brain stem cavernous malformations: Surgical indications, technical considerations, and results. J Neurol Neurosurg Psychiatry. 72:351–355. 2002.PubMed/NCBI View Article : Google Scholar | |
Gross BA, Batjer HH, Awad IA and Bendok BR: Brainstem cavernous malformations. Neurosurgery. 64:E805–E818. 2009.PubMed/NCBI View Article : Google Scholar | |
Lippitz B: Treatment of cavernoma: An evidence-based dilemma? Acta Neurochir Suppl. 116:99–101. 2013.PubMed/NCBI View Article : Google Scholar | |
Rinkel LA, Al-Shahi Salman R, Rinkel GJ and Greving JP: Radiosurgical, neurosurgical, or no intervention for cerebral cavernous malformations: A decision analysis. Int J Stroke. 14:939–945. 2019.PubMed/NCBI View Article : Google Scholar | |
Wells GA, Shea B, O'Connell D, Peterson J, Welch V, Losos mM and Tugwell P: The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Hospital Research Institute, 2014. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. | |
Bamford JM, Sandercock PA, Warlow CP and Slattery J: Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 20(828)1989.PubMed/NCBI View Article : Google Scholar | |
Higgins JPT, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, Savovic J, Schulz KF, Weeks L, Sterne JA, et al: The cochrane collaboration's tool for assessing risk of bias in randomised trials. BMJ. 343(d5928)2011.PubMed/NCBI View Article : Google Scholar | |
Higgins JPT and Green S (eds): Cochrane handbook for systematic reviews of interventions version 5.1.0. The Cochrane Collaboration, 2011. www.cochrane-handbook.org. Updated March 2011. | |
Mathiesen T, Edner G and Kihlström L: Deep and brainstem cavernomas: A consecutive 8-year series. J Neurosurg. 99:31–37. 2003.PubMed/NCBI View Article : Google Scholar | |
Tarnaris A, Fernandes RP and Kitchen ND: Does conservative management for brain stem cavernomas have better long-term outcome? Br J Neurosurg. 22:748–757. 2008.PubMed/NCBI View Article : Google Scholar | |
Moultrie F, Horne MA, Josephson CB, Hall JM, Counsell CE, Bhattacharya JJ, Papanastassiou V, Sellar RJ, Warlow CP, Murray GD and Al-Shahi Salman R: Scottish Audit of Intracranial Vascular Malformations (SAIVMs) steering committee and collaborators. Outcome after surgical or conservative management of cerebral cavernous malformations. Neurology. 83:582–589. 2014.PubMed/NCBI View Article : Google Scholar | |
Dammann P, Wrede K, Jabbarli R, Neuschulte S, Menzler K, Zhu Y, Özkan N, Müller O, Forsting M, Rosenow F and Sure U: Outcome after conservative management or surgical treatment for new-onset epilepsy in cerebral cavernous malformation. J Neurosurg. 126:1303–1311. 2017.PubMed/NCBI View Article : Google Scholar | |
Kang K, Ju Y, Wang D, Li H, Sun L, Ma K, Zhao X and Lu J: Cerebral venous malformations in a chinese population: Clinical manifestations, radiological characteristics, and long-term prognosis. World Neurosurg. 120:e472–e479. 2018.PubMed/NCBI View Article : Google Scholar | |
Liscák R, Vladyka V, Simonová G, Vymazal J and Novotny J Jr: Gamma knife surgery of brain cavernous hemangiomas. J Neurosurg. 102 (Suppl):S207–S213. 2005.PubMed/NCBI View Article : Google Scholar | |
Kondziolka D, Lunsford LD, Flickinger JC and Kestle JR: Reduction of hemorrhage risk after stereotactic radiosurgery for cavernous malformations. J Neurosurg. 83:825–831. 1995.PubMed/NCBI View Article : Google Scholar | |
Karlsson B, Kihlström L, Lindquist C, Ericson K and Steiner L: Radiosurgery for cavernous malformations. J Neurosurg. 88:293–297. 1998.PubMed/NCBI View Article : Google Scholar | |
Steinberg GK, Chang SD, Gewirtz RJ and Lopez JR: Microsurgical resection of brainstem, thalamic, and basal ganglia angiographically occult vascular malformations. Neurosurgery. 46:260–271. 2000.PubMed/NCBI View Article : Google Scholar | |
Al-Shahi Salman R, Berg MJ, Morrison L and Awad IA: Angioma Alliance Scientific Advisory Board. Hemorrhage from cavernous malformations of the brain: Definition and reporting standards. Angioma alliance scientific advisory board. Stroke. 39:3222–3230. 2008.PubMed/NCBI View Article : Google Scholar |