Size Matters. Prevalence and Clinical Implications of Left Ventricular Apical Aneurysms in Hypertrophic Cardiomyopathy: A Systematic Review and Meta‐Analysis

Abstract

Background

The aim of this study was to systematically review the recently published literature and determine the prevalence of left ventricular apical aneurysm (LVAA) formation in hypertrophic cardiomyopathy and its association with sudden cardiac death, systemic embolization, and heart failure.

Methods

The protocol was registered with the International Prospective Register of Systematic Reviews (registration number: CRD42023453640). MEDLINE and manual searches for articles published up to August 2023 were performed. Longitudinal, observational cohorts of unselected adult patients with hypertrophic cardiomyopathy were considered. Data were pooled using a random‐effects model.

Results

A total of 321 articles fulfilled the search criteria, and 10 retrospective observational studies were selected for the meta‐analysis. The pooled prevalence of LVAA was 3% (95% CI, 2%–5%), and 57% of LVAAs were small (<2 cm). Small LVAAs had a lower prevalence of sudden cardiac death end points (4.71% [95% CI, 1.5%–9%) than bigger (≥2 cm) LVAAs (22% [95% CI, 15%–31%), with an odds ratio of 4.65 (95% CI, 2.14–10.10). The prevalence of systemic emboli was also higher in bigger LVAAs (17% [95% CI, 9%–28%) when compared with small LVAA (9% [95% CI, 4%–16%), with an odds ratio of 1.78 (95% CI, 0.53–5.99). Left ventricular thrombi were also more frequently detected in bigger LVAAs (30% [95% CI, 20%–42%) than small LVAAs (2% [95% CI, 0%–6%), with an odds ratio of 10.92 (95% CI, 3.75–31.84). There are scant data on heart failure deaths.

Conclusions

The available data suggest that patients with bigger LVAAs (>2 cm) have the highest risk of poor outcomes and could be preferentially targeted for primary prevention of sudden cardiac death and systemic embolization.

Nonstandard Abbreviations and Acronyms

HCM

hypertrophic cardiomyopathy

LVAA

left ventricular apical aneurysm

MOOSE

Meta‐Analysis of Observational Studies in Epidemiology

NSVT

nonsustained ventricular tachycardia

SCD

sudden cardiac death

Clinical Perspective.

What Is New?

• Left ventricular apical aneurysms are uncommon, and patients with a big left ventricular apical aneurysm (>2 cm) seem to face the highest risk of sudden death and systemic embolization.

What Are the Clinical Implications?

• Patients with a big left ventricular apical aneurysm (≥2 cm) should preferentially be considered for primary prevention implantable cardioverter‐defibrillator and screening for left ventricular apical aneurysm thrombi to allow prompt anticoagulation and primary prevention of systemic embolization.

Hypertrophic cardiomyopathy (HCM) is an inherited cardiac condition associated with sudden cardiac death (SCD). ¹ , ² Ventricular arrhythmias have been associated with left ventricular apical aneurysm (LVAA) formation, where an akinetic or dyskinetic chamber in the distal left ventricle develops in a subgroup of patients with apical or midventricular hypertrophy. ³ , ⁴ Initially identified during histopathological examinations in 1975, ⁵ an LVAA as small as 2 mm can now be detected by cardiovascular magnetic resonance (CMR) imaging and contrast transthoracic echocardiography. ⁶

In 2020, the American College of Cardiology/American Heart Association recommended that patients with an LVAA of any size should be considered for a primary prevention implantable cardioverter‐defibrillator (ICD). ¹ In contrast, the 2023 European Society of Cardiology (ESC) guidelines do not consider LVAA an independent risk factor for SCD and counsel against using LVAA as the sole determinant of risk for primary prevention ICD implantation. ² The dissonance between international guidelines has the potential to hamper patient confidence and frustrate clinicians. LVAA may also be associated with other adverse cardiovascular outcomes, such as heart failure and systemic embolization, which merit consideration. ⁷ , ⁸

The aim of this study was to systematically review the recently published literature and determine the prevalence of LVAA and its association with SCD, systemic embolization, and heart failure.

METHODS

The data underlying this article will be shared on reasonable request to the corresponding author.

Protocol and Registration

The systematic review was performed in accordance with the Preferred Reporting Items for Systematic reviews and Meta‐Analyses amendment to the Quality of Reporting of Meta‐analyses statement and Cochrane Collaboration and Meta‐Analysis of Observational Studies in Epidemiology guidelines. The systematic review protocol was registered with the International Prospective Register of Systematic Reviews in March 2023 (registration number: CRD42023453640). All authors read, critically appraised, provided feedback, and approved the final article.

Eligibility Criteria

Study Designs

All longitudinal observational cohort studies reporting on LVAA in patients with HCM were included, irrespective of setting. Abstracts and case reports were excluded.

Participants

HCM was diagnosed in the presence of maximum left ventricular (LV) wall thickness ≥15 mm unexplained by abnormal loading conditions or in accordance with published criteria for diagnosis in relatives of patients with unequivocal HCM. ¹ , ² Only studies examining adult patients (≥16 years of age) were considered.

Type of Exposure

LVAA was defined as an akinetic or dyskinetic chamber in the distal left ventricle. ³ , ⁴ , ⁹

End Points

The annual rates, prevalence and hazard/odd ratios were collected for the following end points:

1. Sudden cardiac death: SCD was defined as witnessed sudden death with or without documented ventricular fibrillation or death within 1 hour of new symptoms or nocturnal deaths with no antecedent history of worsening symptoms. Aborted SCD during follow‐up and appropriate ICD shock therapy were considered equivalent to SCD. ICD shocks were considered appropriate if the treated tachyarrhythmia was ventricular in origin. Antitachycardia pacing when considered as an equivalent to SCD was noted.

2. Heart failure deaths: Any death preceded by worsening heart failure requiring hospitalization or an urgent visit that resulted in intravenous therapy for heart failure within the previous month. Any death that was due to progressive heart failure was also included in the definition of heart failure death.

3. Systemic embolization: The occurrence of a new or worsening embolic ischemic event in any vascular territory during follow‐up period. These can include events such as stroke, transient ischemic attack, peripheral embolism, or visceral embolization.

Information Sources

PubMed was searched, and the reference lists of reviews, letters, and editorials were scrutinized for relevant material.

Search Strategy

The MEDLINE search strategy was: (hypertrophic cardiomyopathy) AND aneurysm. Only articles in English were included for the analysis, and relevant titles in other languages were recorded. The search was undertaken on August 1, 2023.

Study Records

Data Management

The initial literature search results were uploaded to Rayyan, which was used to manage the retrieved abstracts. ¹⁰

Selection Process

The retrieved studies were independently reviewed at the title or abstract level for the inclusion and exclusion criteria. Full articles were obtained for all reports meeting the inclusion criteria or when there was ambiguity. The full articles were then reviewed to see if the inclusion criteria were satisfied. A detailed study of authors, dates, and locations was used to reduce redundancy. The reviewers were not blinded to the journal titles or to the study authors or institutions. Reasons for exclusion were documented.

Data Collection Process

Data were extracted from the full‐length articles and were transcribed to a purpose‐built relational database (Access; Microsoft Corporation). Demographic data and methodological characteristics were collected. LVAAs were classified as small (<2 cm), medium (2–4 cm), or large (>4 cm), and the number of end points in each size category were collected when available. In addition, for SCD, the number of patients with an LVAA in each of the following risk categories, as defined by the 2020 ESC guidelines, ² was recorded:

1. Low risk (ICD not recommended): predicted 5‐year risk of SCD <4%.

2. Intermediate risk (ICD may be considered): predicted 5‐year risk of SCD 4% to <6%.

3. High risk (ICD should be considered): predicted 5‐year risk of SCD ≥6%.

SCD end points per risk group were extracted. All extracted data were verified independently by another author, and disagreements were resolved by consensus. Variations in definitions of SCD end points were recorded.

Outcomes

The primary outcomes were the prevalence of LVAAs in HCM and the occurrence of SCD, heart failure, deaths, and systemic embolization in patients with an LVAA.

Assessing the Methodological Quality of the Studies

The Joanna Briggs Institute critical appraisal tool for case series was used. ¹¹ Two independent reviewers assessed the eligible studies, with a third reviewer adjudicating if no decision could be reached. Each article was scored on 10 questions. If yes was answered for half or more of the questions, the study was classified as low risk of bias. If no was answered to half or more of the questions, the study was classified as at high risk of bias. If unclear was answered to half or more of the questions, the risk of bias could not be assessed.

Statistical Analysis

The study characteristics are presented descriptively. The pooled observed prevalence for LVAA, LVAA size, SCD, nonsustained ventricular tachycardia (NSVT), systemic embolization, and LVAA thrombi stratified by LVAA size are also presented. Meta‐analysis models are fitted based on the Freeman‐Tukey double arcsine transformation to mitigate for low prevalence. The study statistics and the pooled statistics are given on the (back transformed) proportions scale. The inverse variance method on the transformed scale was used for pooling of studies. The Clopper‐Pearson confidence intervals (also known as the exact binomial interval) for individual study results was used throughout for all models, both fixed and random. Heterogeneity between studies was assessed by performing the Cochran Q test and estimating the between‐study variance of the estimates, τ, using the restricted maximum likelihood method, which is more robust to small samples. ¹² The standard deviation of the rates in the untransformed scale is also presented in the plots. To examine the magnitude of the variation between studies due to heterogeneity rather than chance, and whether it impacts the conclusions of the meta‐analyses, the heterogeneity was quantified by the I ² measure, τ, and standard deviation. The included studies are affected by clinical heterogeneity with differences in participants and reported outcomes. We have presented both random‐effects models (presented in the main article) and fixed‐effects models (Figures S1 and S2). The statistical programming software R with the package meta was used to perform the meta‐analysis and create the plots. ¹³ The exact 2×2 R package was used to calculate the exact study‐specific confidence intervals for the odds ratios.

RESULTS

A total of 321 articles fulfilled the search criteria in PubMed, and 10 retrospective observational studies were selected for the systematic review/meta‐analysis, as shown in Figure 1 and Table 1. ⁷ , ⁸ , ⁹ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ All included studies involved adult patients diagnosed using conventional or familial criteria, and the cohorts were not restricted to a specific HCM phenotype or geographical location. All selected studies were assessed to be at low risk of bias using the Joanna Briggs Institute questionnaire (Table S1). Four studies did not provide the criteria used to diagnose LVAA, ¹³ , ¹⁴ , ¹⁵ , ¹⁶ and the uptake of CMR imaging was highly variable (Table 1). The typical patient with an LVAA is a man in the sixth to seventh decade of life, unaffected by atrial fibrillation. A paucity of data precluded the comparison of baseline characteristics of patients with and without LVAA, because only 1 study ¹⁸ reported this information on the non‐LVAA cohort of patients.

Figure 1. Study selection.

The flowchart shows each step of the systematic search to identify studies. *Redundant studies can be found in Table S2.

Table 1.

Included Studies With the Prevalence and Characteristics of Patients With LVAAs

Study	Imaging	CMR use (%)	Cohort size (n)	LVAA (n)	LVAA prevalence (%)	Small LVAA (n)	Medium LVAA (n)	Large LVAA (n)	Age (y)	Women (%)	AF (%)	NSVT† (%)	LA size (mm)
Ichida 2014 9	CMR+TTE	1.2	247	21	8.5	8	13	0	60±14	42.9	14.3	28.6	43±6
Minami 2014 20	CMR+TTE	10.1	544	24	4.4	●	●	●	52.8±14.4	33.3	25.0	58.3	40.7±5.7
Rowin 2017 7	CMR+TTE	●	1940	93	4.8	53	30	10	56±13	31.2	●	35.5	43±8
Neubauer 2019 15	CMR	100.0	2651	79	3.0	●	●	●	50±11	40.5	●	●	●
Liu 2020 16	CMR+TTE	60.7	1369	13	0.9	●	●	●	●	●	●	●	●
Parcharidou 2020 14	CMR+TTE	36.2	690	16	2.3	●	●	●	●	●	●	●	●
Dong 2021 17	CMR+TTE	3.9	511	3	0.6	●	●	●	●	●	●	●	●
Lee 2022 8	CMR+TTE	●	5300	160	3.0	102	52	6	63±14	28.8	31.3	51.3	42±7
Strachinaru 2022 19	CMR+TTE	●	1332	35	2.6	●	●	●	59±13	62.9	●	60.0	●
Kim 2023 18	CMR	100.0	458	42	9.2	●	●	●	59±●	52.4	16.7	16.7	40 mL/m2 *

Continuous variables are presented as mean±SD or median (25th–75th percentile). ● indicates not available; AF, atrial fibrillation; CMR, cardiovascular magnetic resonance; LA, left atrial; LVAA, left ventricular apical aneurysm; NSVT, nonsustained ventricular tachycardia; and TTE, transthoracic echocardiography.

^*LA diameter not available; LA volume is presented instead.

^† The number of patients completing a Holter is only available for Ichida et at 2014.

Prevalence and Size of LVAA

The reported prevalence of LVAA was highly variable, ranging from <1% to 9%. The pooled prevalence was 3% (95% CI, 2%–5%), as shown in Figure 2A. LVAA size was reported in 3 studies using different imaging modalities and views. ⁷ , ⁸ , ⁹ Small LVAAs were the most common form and constituted 57% (pooled prevalence 95% CI, 45%–68%) of all LVAAs, as shown in Figure 2B. The fixed‐effects model yielded similar results and is shown in Figures S1A and S1B.

Figure 2. Forest plots.

A, The prevalence of left ventricular aneurysms in patients with HCM. B, The prevalence of small left ventricular aneurysms. C, The prevalence of sudden cardiac death in patients with left ventricular aneurysms. D, The prevalence of nonsustained ventricular tachycardia in patients with left ventricular aneurysms. E, The prevalence of systemic embolization in patients with left ventricular aneurysms. F, The prevalence of left ventricular thrombi in patients with left ventricular aneurysms. HCM indicates hypertrophic cardiomyopathy.

SCD End Points

Only 6 studies provided SCD data, as shown in Table 2. Antitachycardia pacing was considered a surrogate for SCD in 3 studies, ⁷ , ⁸ , ¹⁹ and 2 studies included patients with resuscitated cardiac arrest before baseline evaluation. ⁷ , ⁸ The pooled prevalence of SCD was 16% (95% CI, 6%–29%), as shown is Figure 2C (the follow‐up period of these studies was 3 to 6 years, with the exception of a single study with follow‐up of >10 years, as shown in Table 2). The annual rate of SCD was reported in only 2 studies (1.8% and 4.7% per year). ⁷ , ⁸ In all studies, the prevalence of SCD was lowest in small LVAA (Table 2). The fixed effects models of the prevalence of SCD yielded similar results (Figure S1C). SCD events were more commonly observed in big LVAAs (≥2 cm), with an odds ratio of 4.65 (95% CI, 2.14–10.10) when compared with small LVAAs (<2 cm), as shown in Figure 3A. NSVT was detected in 41% (95% CI, 27%–56%) of patients with an LVAA, as shown in Figure 2D, with the fixed‐effects model showing similar findings, as illustrated in Figure S1D.

Table 2.

SCD, Embolic Events, and Thrombi in Left Ventricular Aneurysms

Study	LVAA (n)	Mean FU (y)	SCD (%)				CVA (%)				Thrombus (%)
			Total	Small	Medium	Large	Total	Small	Medium	Large	Total	Small	Medium	Large
Ichida 2014 9 †	21	4.7	1 (4.8)	0 (0.0)	1 (7.7)	0	4 (19.0)	0 (0.0)	4 (30.8)	0	3 (14.3)	0 (0.0)	3 (23.1)	0
Minami 2014 20	24	11.6	8 (33.3)	●	●	●	2 (8.3)	●	●	●	●	●	●	●
Rowin 2017 7 †	93	4.4	21* (22.6)	6 (11.3)	10 (33.3)	5 (50.0)	5 (5.4)	●	●	●	13 (14.0)	●	●	●
Neubauer 2019 15	79	0	●	●	●	●	●	●	●	●	●	●	●	●
Parcharidou 2020 14	16	8.6	●	●	●	●	●	●	●	●	●	●	●	●
Liu 2020 16	13	3.3	2 (15.4)	●	●	●	●	●	●	●	●	●	●	●
Dong 2021 17	3	4.7	2 (66.7)	●	●	●	●	●	●	●	●	●	●	●
Lee 2022 8 †	160	6.2	14* (8.8)	4 (3.9)	8 (15.4)	2 (33.3)	21 (13.1)	12 (11.8)	9 (16)		23 (14.4)	4 (3.9)	19 (32.7)
Strachinaru 2022 19	35	4	●	●	●	●	2 (5.7)	●	●	●	2 (5.7)	●	●	●
Kim 2023 18	42	6.3	●	●	●	●	4 (9.5)	●	●	●	3 (7.1)	●	●	●

● indicates not available; CMR, cardiovascular magnetic resonance; CVA, cerebrovascular accident; FU, follow‐up; LVAA, left ventricular apical aneurysm; NSVT, nonsustained ventricular tachycardia, and TTE, transthoracic echocardiography.

^*Patients with resuscitated cardiac arrest before the first evaluation are included.

^† Lee et al 2022 used the maximum transdimensional width obtained in mid‐ to end‐systole in the 4‐, 2‐, and 3‐chamber views on echocardiography and CMR. Rowin et al 2017 reported the maximum transverse dimension measured by CMR or TTE in the 4‐chamber long‐axis view. Ichida et al 2014 do not report how the LVAA size was determined.

Figure 3. SCD and embolic events based on LVAA size.

A, The prevalence of sudden cardiac death stratified by left ventricular aneurysm size. B, The prevalence of systemic embolization stratified by left ventricular aneurysm size. C, The prevalence of left ventricular thrombi stratified by left ventricular aneurysm size. LVAA indicates left ventricular apical aneurysm; OR, odds ratio; and SCD, sudden cardiac death.

LV Thrombi and Systemic Emboli

Six studies provided data on systemic emboli, as shown in Table 2. The pooled prevalence of systemic emboli was 9% (95% CI, 6%–13%), as shown is Figure 2E. Compared with big LVAAs (≥2 cm), patients with a small LVAA had a lower prevalence of systemic emboli (8.8% [95% CI, 3.5%–15.7%] versus 17.4% [95% CI, 8.9%–27.6%]), and the odds ratio of systemic emboli in big LVAAs was 1.78 (95% CI, 0.53– 5.99) when compared with small LVAAs, as shown in Figure 3B. The pooled prevalence of LV thrombi was 12% (95% CI, 9%–16%), as shown in Figure 2F. The pooled prevalence of LV thrombi was also lower in small LVAAs compared with big LVAAs (1.75% [95% CI, 0%–6.24% versus 30.6% [95% CI, 20.0%–42.3%]), with an odds ratio of 10.92 (95% CI, 3.75–31.84) for LV thrombi in big LVAAs compared with small LVAAs, as shown in Figure 3C. Fixed‐effects models for the above analyses are shown in Figures S1E, S2B, S1F, and S2C, and yielded comparable results.

Heart Failure Deaths

Heart failure deaths were reported in 3 studies: Ichida et al 2014: 0/21; Rowin et al 2017: 2/93 (2%), and Lee et al 2022: 1/160 (0.6%). ⁷ , ⁸ , ⁹ There are no data examining heart failure deaths and LVAA size.

DISCUSSION

This meta‐analysis is based on a small number of studies and demonstrates that LVAA are uncommon and often associated with NSVT. The prevalence of SCD, systemic embolization, and LV thrombi during medium‐term follow‐up may be high, and larger LVAAs (≥2 cm) are potentially associated with worse prognosis. The scarcity of available data, combined with frail methodology, hinders robust clinical recommendations.

The totality of the data used in this meta‐analysis indicates that LVAAs are infrequent. However, the reported prevalences are highly variable due to the lack of uniform diagnostic criteria, which are often not stated. Detection is modality dependent, with CMR imaging being more sensitive but less frequently used than noncontrast transthoracic echocardiography. ⁶ , ⁷ Even when CMR imaging is used, interobserver variability hinders diagnosis, especially when LVAAs are small, and there is no standardized methodology to determine the size. ⁶

The prevalence of SCD end points in LVAAs is approximately 7 times higher compared with phenotypically unselected HCM cohorts of a similar follow‐up period (17% versus ~2.5%). ²¹ This is expected, because most patients with HCM with LVAA have other SCD risk markers, ⁷ the exemplar being NSVT. With a prevalence of <20% in contemporary unselected HCM cohorts, ¹⁰ , ¹¹ , ¹⁸ this analysis shows that NSVT is ubiquitous in LVAAs, despite underreporting relating to missing Holter data. End‐stage HCM is also highly prevalent (left ventricular ejection fraction ≤50% was encountered in 29% and 36% of patients with an LVAA with SCD end points ⁷ , ⁸ ). The observed SCD prevalence is, however, exaggerated by SCD end point inflation due to incorporation of antitachycardia pacing as an SCD equivalent, inclusion of patients with a prior resuscitated cardiac arrest, and referral bias (with patients with LVAAs and risk factors of SCD preferentially referred to HCM centers).

The coexistence of established SCD risk markers and SCD end point inflation, coupled with the small number of patients with events, cast a shadow on the independence and effect size of LVAAs on SCD, which merits critical appraisal. Bigger aneurysms (>2 cm), representing a state of more advanced disease, are likely to offer a more fecund substrate for ventricular arrhythmias and are reported in patients requiring ventricular tachycardia (VT) ablation. ²² , ²³ , ²⁴ This meta‐analysis supports this notion, because patients with small LVAAs have the lowest prevalence of SCD end points, suggesting that such patients can be treated conservatively if HCM risk‐SCD is low. Patients with intermediate/high HCM risk‐SCD, irrespective of LVAA size, should be considered for an ICD as per current guidelines. ¹ , ² Consideration for a primary prevention ICD could be given preferentially to patients with medium or large LVAAs (ie, >2 cm) despite low HCM risk‐SCD following a comprehensive discussion of the risks, benefits, and the considerable uncertainty.

LVAAs also pose a threat by harboring a low flow environment that promotes thrombus formation. This meta‐analysis shows a high prevalence of LVAA thrombi, which provides a mechanistic link that partially accounts for the high prevalence of systemic embolization (approximately double of that reported in unselected HCM cohorts of similar follow‐up), with atrial fibrillation and left atrial enlargement also contributing. ²⁵ , ²⁶ , ²⁷ Prevention of thromboembolic events, a core tenet of clinical management, can be achieved by prompt anticoagulation of thrombi either with a direct oral anti‐coagulant or vitamin K antagonist. ²⁸ , ²⁹ Surveillance with contrast transthoracic echocardiography or CMR imaging, in preference to routine noncontrast transthoracic echocardiography as currently recommended, ¹ , ² may help achieve this. Preventative anticoagulation in LVAAs ≥2 cm is another attractive option supported by the meta‐analysis, because patients with small LVAAs had the lowest prevalence of LV thrombi. This is another setting where shared decision‐making becomes the cornerstone of clinical management.

Heart failure deaths are not frequently encountered in patients with LVAAs in the short and medium term but are likely to contribute to cardiovascular mortality in the long term.

Finally, LVAA size has been arbitrarily quantified by a single transverse dimension of a complex 3‐dimensional structure. The risk of adverse events may not have a linear association with the transverse diameter but may be more strongly associated with LVAA volume in a polynomial (cubic) relationship. The relationship of LVAA volume and outcomes could not be determined by the pooled data of this meta‐analysis. The small number of patients with medium LVAAs prevented a meaningful analysis of outcomes in the 3 groups (small, medium, and large).

The major limitation of this meta‐analysis is the limited number of eligible studies, each with a small cohort size and events. The results should be interpreted with a degree of caution due to the limitations of the random‐effects approach in small sample settings. The random‐effects models do not eliminate the problems introduced by heterogeneity; the random and (exact) fixed‐effects analysis yielded similar results. The choice of diagnostic tool can have a significant impact on the diagnosis, and underuse of CMR imaging is likely to underestimate the prevalence of LVAA. Not all eligible studies reported the outcomes of interest, which has the potential to introduce bias. Most, but not all, studies explicitly excluded subjects with a history of myocardial infarction, raising the possibility that some included patients had a LVAA due to coronary artery disease. Comparison of simple demographic characteristics and comorbidities in patients with and without LVAAs was not possible. Similarly, the diverse reporting of outcomes (eg, not all the studies reported annualized rates) limits the meta‐analytic output. Establishing standardized imaging protocols for the detection and measurement of LVAA size will help establish more accurately the association with adverse outcomes.

CONCLUSIONS

This meta‐analysis highlights the limited available data on LVAA cardiovascular outcomes that should be interpreted with caution. Our analysis establishes the early evidential basis suggesting that not all LVAAs carry the same prognostic implications. Patients with bigger LVAAs (>2 cm) seem to face the highest risk of unfavorable outcomes and consequently may be prioritized for therapeutic interventions with shared decision‐making.

Sources of Funding

This work was undertaken at St. Bartholomew's Hospital, which received a proportion of funding from the United Kingdom Department of Health's National Institute for Health Research Biomedical Research Centers funding scheme.

Disclosures

None.

Supporting information

Tables S1–S2

Figures S1–S2

References 30–33