Assessment of Papillary Muscle Infarction with Dark-Blood Delayed Enhancement Cardiac MRI in Canines and Humans

Abstract

Background

The relationship between papillary muscle infarction (papMI) and the culprit coronary lesion has not been fully investigated. Delayed enhancement cardiac MRI may detect papMI, yet its accuracy is unknown. Flow-independent dark-blood delayed enhancement (FIDDLE) cardiac MRI has been shown to improve the detection of myocardial infarction adjacent to blood pool.

Purpose

To assess the diagnostic performance of delayed enhancement and FIDDLE cardiac MRI for the detection of papMI, and to investigate the prevalence of papMI and its relationship to the location of the culprit coronary lesion.

Materials and Methods

A prospective canine study was used to determine the accuracy of conventional delayed enhancement imaging and FIDDLE imaging for detection of papMI, with pathology-based findings as the reference standard. Participants with first-time myocardial infarction with a clear culprit lesion at coronary angiography were prospectively enrolled at a single hospital from 2015 to 2018 and compared against control participants with low Framingham risk scores. In canines, diagnostic accuracy was calculated for delayed enhancement and FIDDLE imaging.

Results

In canines (n = 27), FIDDLE imaging was more sensitive (100% [23 of 23] vs 57% [13 of 23], P < .001) and accurate (100% [54 of 54] vs 80% [43 of 54], P = .01) than delayed enhancement imaging for detection of papMI. In 43 participants with myocardial infarction (mean age, 56 years ± 16 [SD]; 28 men), the infarct-related artery was the left anterior descending coronary artery (LAD), left circumflex coronary artery (LCX), and right coronary artery in 47% (20 of 43), 26% (11 of 43), and 28% (12 of 43), respectively. The prevalence of anterior papMI was lower than posterior papMI (37% [16 of 43 participants] vs 44% [19 of 43 participants]) despite more LAD culprit lesions. Culprits leading to papMI were restricted to a smaller “at-risk” portion of the coronary tree for anterior papMI (subtended first diagonal branch of the LAD or first marginal branch of the LCX) compared with posterior (subtended posterior descending artery or third obtuse marginal branch of the LCX). Culprits within these at-risk portions were predictive of papMI at a similar rate (anterior, 83% [15 of 18 participants] vs posterior, 86% [18 of 21 participants]).

Conclusion

Flow-independent dark-blood delayed enhancement cardiac MRI, unlike conventional delayed enhancement cardiac MRI, was highly accurate in the detection of papillary muscle infarction (papMI). Anterior papMI was less prevalent than posterior papMI, most likely due to culprit lesions being restricted to a smaller portion of the coronary tree rather than because of redundant, dual vascular supply.

© RSNA, 2022

Online supplemental material is available for this article.

See also the editorial by Kawel-Boehm and Bremerich in this issue.

Summary

As identified by flow-independent dark-blood delayed enhancement cardiac MRI, anterior papillary muscle infarction was less prevalent than posterior because culprit lesions were restricted to a smaller portion of the coronary tree.

Key Results

• ■ This prospective canine study (n = 27) used pathology-based findings as the reference standard and showed that flow-independent dark-blood delayed enhancement cardiac MRI was more sensitive (100% vs 57%, P < .001) and accurate (100% vs 80%, P = .01) than conventional delayed enhancement cardiac MRI for detecting papillary muscle infarction (papMI).

• ■ In participants with first-time myocardial infarction (n = 43), papMI culprit lesions were restricted to a smaller coronary “at-risk” region for anterior (five of 17 segments) versus posterior papMI (nine of 17 segments).

• ■ The at-risk portions predicted anterior and posterior papMI at a similar rate (anterior, 83% vs posterior, 86%).

Introduction

With acute myocardial infarction, necrosis may extend into one or both papillary muscles. A rare, catastrophic complication of papillary muscle infarction (papMI) is papillary muscle rupture (1), but even without rupture, papMI or papillary muscle scarring may cause ventricular arrhythmias (2) and impact ventricular geometry and function (3,4). In fact, the working papillary muscles play an integral role in cardiac mechanics—in diastole by limiting left ventricle distension and in systole by increasing longitudinal shortening, which in turn increases left ventricle wall stretch and generates greater tension and stroke volume (3,4).

Anterior papMI is less common than posterior papMI (5,6). Despite extensive studies on papillary muscle vasculature dating back to 1885 (2,7–9), the reason for this difference has not been definitively determined. One potential explanation suggests the role of dual vascular supply, which is more common in the anterior papillary muscle (6,9). However, dual vascular supply does not equate with redundant vascular supply. Two vessels may perfuse different portions of a papillary muscle rather than provide overlapping perfusion, which protects against infarction. Additionally, most studies of papMI are postmortem investigations in patients with fatal myocardial infarction in whom the prevalence of papMI may not be the same as in patients surviving myocardial infarction (5,10). The prevalence of anterior and posterior papMI and their relationship to the culprit coronary lesion in patients who have survived acute myocardial infarction is unclear.

Delayed enhancement cardiac MRI may help detect papMI; however, its accuracy is unknown and often there is poor contrast between hyperenhanced papMI and bright blood pool. Flow-independent dark-blood delayed enhancement (FIDDLE) cardiac MRI improves the detection of left ventricular wall myocardial infarction adjacent to blood pool (11–13). However, there are no validation studies comparing FIDDLE imaging to a pathology-based evaluation for the diagnosis of papMI.

In the present study, we sought to assess the diagnostic performance of delayed enhancement and FIDDLE cardiac MRI for the detection of papMI. We employed a canine model so that pathology-based findings could be used as the reference standard. Then, in patients with first-time myocardial infarction, the prevalence of anterior and posterior papMI and their relationship to the location of the culprit coronary lesion was investigated.

Materials and Methods

Canines

Protocol overview.— The care and treatment of canines (20–30 kg) followed the position of the American Heart Association on the use of research animals (14). The protocol was approved by the Duke University Institutional Animal Care and Use Committee. We produced left ventricle papMI in 27 animals by occluding the left anterior descending coronary artery (LAD) or an obtuse marginal branch of the left circumflex coronary artery (LCX) under sterile conditions through a left thoracotomy (15). A range of occlusion times (40–90 minutes, followed by reperfusion) was employed to create a range of papMI sizes. Following MRI, the heart was removed and the presence, location, and extent of papMI was confirmed using triphenyl tetrazolium chloride staining, as previously described (16,17). Briefly, the process involved slicing the heart in short-axis orientation every 1 cm throughout the left ventricle, and slices were matched against cardiac MRI scans by using myocardial landmarks such as right ventricle insertion sites, right ventricle trabeculations, and left ventricle papillary muscle contours.

Cardiac MRI of canines.—Cardiac MRI was performed over a range of time points following myocardial infarction to test whether infarct age might affect the performance of FIDDLE imaging. Fourteen canines were scanned within the first 2 weeks of myocardial infarction and 13 after 2 weeks (median, 8 weeks). Images were acquired using a 3-T MAGNETOM Verio (Siemens Healthineers) scanner during ventilated breath holds. Fifteen minutes after intravenous administration of 0.15 mmol/kg of gadoterate meglumine (Dotarem; Guerbet), delayed enhancement and FIDDLE cardiac MRI were performed immediately one after another, in random order, using matched parameters as detailed in Appendix E1 (online). FIDDLE parameters were optimized as previously described (12,13).

Image analysis.—Gross anatomic, delayed enhancement, and FIDDLE images were interpreted (ImageJ; National Institutes of Health) separately by two readers (D.W. and H.W.K., each with more than 10 years of experience reading cardiac MRI scans) who were blinded to other data. The window and level for cardiac MRI was preset so that image noise was still detectable and infarcted regions were not oversaturated (18). The presence of anterior and/or posterior papMI as demonstrated by gross anatomic, delayed enhancement, and FIDDLE images was determined by visual inspection. The extent of papMI was scored for each papillary muscle according to a five-point scale whereby 0 indicates no myocardial infarction, 1 indicates 1%–25%, 2 indicates 26%–50%, 3 indicates 51%–75%, and 4 indicates 76%–100% infarcted.

Participants

Study sample.—Participants presenting with a history of myocardial infarction were recruited prospectively. The diagnosis of myocardial infarction was based on the universal definition (19). Consecutive participants 18 years or older who underwent coronary angiography during admission for myocardial infarction in which the culprit infarct-related artery was clearly identified and who agreed to participate were enrolled (n = 55). Participants with incomplete cardiac MRI scans (n = 2) or multiple infarcts on scans (n = 10) were excluded, resulting in 43 participants included in the final analysis (Fig 1). The control group (n = 10) consisted of participants with no known coronary disease and low probability for developing disease over the next 10 years (lowest Framingham risk score: 1% for women, 2% for men). All participants gave written informed consent, and the study was approved by the Duke University Health System Institutional Review Board.

Figure 1:

Study flow diagram. Participants with a clinical diagnosis of myocardial infarction (MI) who were referred for MRI and had a clear culprit infarct-related artery (IRA) at coronary angiography were enrolled. CMR = cardiac MRI, FIDDLE = flow-independent dark-blood delayed enhancement.

Cardiac MRI of participants.— The cardiac MRI protocol was the same as that in the canine study, and the same sequences were used at similar settings. Participants underwent MRI at 1.5 T (MAGNETOM Avanto; Siemens Healthineers [n = 28]) or 3 T (MAGNETOM Verio [n = 15]). FIDDLE and delayed enhancement cardiac MRI scans were acquired with matched parameters as described in Appendix E1 (online). Images were obtained 10 minutes after intravenous administration of 0.15 mmol/kg of gadoterate meglumine (Dotarem; Guerbet) in multiple short-axis (every 10 mm throughout the left ventricle) and in three long-axis views.

Image analysis.—Cardiac MRI analysis was performed the same as for the canines. FIDDLE and delayed enhancement images were interpreted independently, masked to participant identity and clinical information. Readers analyzed x-ray–based coronary angiograms blinded to all other information. The infarct-related artery and location of the culprit coronary lesion were classified according to American Heart Association (AHA) guidelines (20). The standard AHA 15-segment coronary tree was slightly modified to distinguish between the distal LCX, third obtuse marginal branch of the LCX, and, if present, left posterior descending artery branches (Fig 2); the ramus intermedius branch, if present, was considered equivalent to the first obtuse marginal branch of the LCX (21). Culprit lesions were also classified according to whether they subtended specific target coronary branches (see Fig 2) that are known to perfuse the anterior papillary muscle (first diagonal branch of the LAD or the first obtuse marginal branch of the LCX) or the posterior papillary muscle (posterior descending artery or third obtuse marginal branch of the LCX and/or distal LCX) (9,22). This led to a smaller predefined “at-risk” portion of the coronary tree for the anterior versus the posterior papillary muscle. Specifically, the at-risk portion was 29% of the coronary tree (five of 17 segments) for the anterior papillary muscle compared with 53% (nine of 17 segments) for the posterior papillary muscle.

Figure 2:

Segmental model of the coronary artery tree. The standard American Heart Association 15-segment model was slightly modified to distinguish between the distal left circumflex coronary artery (LCx), third obtuse marginal branch of the LCX (OM3), and, when present, left posterior descending artery (L-PDA) branches, as well as to distinguish between right posterior descending artery and right posterolateral (PL) branches. This resulted in 17 total segments, not including the left posterior descending artery (dashed line), which was rare. (A) Illustration shows the target coronary branches known to perfuse the anterior papillary muscle (first diagonal branch [D1] of the left anterior descending coronary artery [LAD] and first obtuse marginal branch of the LCX [OM1]) highlighted in green. The predefined “at-risk” portion of the coronary tree is shown as green dashed lines for the anterior papillary muscle. (B) Illustration shows the target coronary branches known to perfuse the posterior papillary muscle (third obtuse marginal branch of the LCX and posterior descending artery) highlighted in blue. The predefined at-risk portion of the coronary tree is shown as blue dashed lines for the posterior papillary muscle. The at-risk portion reflects five of 17 segments for the anterior papillary muscle and nine of 17 segments for the posterior papillary muscle. D2 = second diagonal branch of the LAD, IRA = infarct-related artery, LM = left main coronary artery, mid = middle, OM2 = second obtuse marginal branch of the LCX, prox = proximal, RCA = right coronary artery, R-PDA = posterior descending coronary artery from the right.

Statistical Analysis

Continuous data are presented as means ± SDs or medians and IQRs, as appropriate. Comparisons of continuous data were made using two-sample or paired t tests. Sensitivity, specificity, and accuracy were calculated with 95% CIs for delayed enhancement and FIDDLE imaging. The McNemar test was used to compare the diagnostic performance of these two methods. Interobserver and intraobserver agreements were tested in a subset of participants, as detailed in Appendix E1 (online). The diagnostic accuracy of predefined “at-risk” regions of the coronary tree for predicting anterior and posterior papMI was calculated with 95% CIs. The Fisher exact test was used to compare the proportion of anterior and posterior papillary muscles with infarct sizes less than 50% of the papillary muscle. Statistical tests were two tailed; P < .05 was considered indicative of a statistically significant difference. Statistical analyses were performed using SAS version 9.4 (SAS Institute).

Results

Canines

All canines that successfully completed surgery (n = 27) underwent cardiac MRI. None were excluded on the basis of poor image quality. The pathology-based findings indicated that 15 canines had anterior-only papMI, four had posterior-only papMI, two had both anterior and posterior papMI, and six had no papMI. None had papillary muscle rupture. The extent of papMI ranged from minimal to near complete infarction of the papillary muscle: 1%–25% (n = 12), 26%–75% (n = 7), and 76%–100% (n = 4). Typical delayed enhancement and FIDDLE images are shown in Figure 3. Overall, FIDDLE imaging was more sensitive (100% [23 of 23 papillary muscles] vs 57% [13 of 23 papillary muscles], P < .001) and accurate (100% [54 of 54 papillary muscles] vs 80% [43 of 54 papillary muscles], P = .01) than delayed enhancement MRI for detection of papMI (Table 1). Both techniques had high specificity (100% [31 of 31] vs 97% [30 of 31], P = .3). The sensitivity of delayed enhancement imaging was 42% (five of 12), 71% (five of 7), and 75% (three of four) for papMIs that involved 1%–25%, 26%–75%, and greater than 75% of the papillary muscle, respectively, and 67% (nine of 14) and 50% (six of 13) for papMIs that were less than or equal to 2 weeks old (n = 14) and more than 2 weeks old (n = 13), respectively. FIDDLE cardiac MRI enabled detection of all papMIs less than or equal to 2 weeks old and more than 2 weeks old, and had high interobserver and intraobserver agreement (see Appendix E1 [online]).

Figure 3:

(A–C) Short-axis delayed enhancement cardiac MRI images (left), short-axis flow-independent dark-blood delayed enhancement (FIDDLE) cardiac MRI images (middle), and stained gross anatomic images as the reference standard (right) matched in three canines. (A) Images show papillary muscle infarction (papMI) (arrows) could be detected with both delayed enhancement and FIDDLE cardiac MRI. (B) Images show papMI (arrows) could be detected with FIDDLE imaging only. (C) Images show papMI (arrows) could be detected with FIDDLE imaging only.

Table 1:

Diagnostic Performance of Delayed Enhancement and FIDDLE Cardiac MRI with Pathology-based Findings as Reference Standard in the Canine Model

Participants with Myocardial Infarction and Controls

Comparison between delayed enhancement and FIDDLE cardiac MRI.— Overall, 53 participants were included in the final analysis; 43 with myocardial infarction (mean age, 56 years ± 16 [SD]; 28 men) and 10 controls (mean age, 32 years ± 10; six women). None were excluded due to poor image quality. Table 2 summarizes baseline clinical characteristics. All participants with myocardial infarction had a single infarction according to both delayed enhancement and FIDDLE imaging. None of the 10 control participants demonstrated myocardial infarction (either left ventricle wall or papillary muscle) according to delayed enhancement or FIDDLE imaging.

Table 2:

Characteristics of Participants with Myocardial Infarction Compared with Controls

FIDDLE imaging helped detect papMI in more than twice as many participants as delayed enhancement imaging (28 vs 12, P = .012). Figures 4–7 show typical delayed enhancement and FIDDLE images. Even among participants with papillary muscles that were more than 75% infarcted according to FIDDLE imaging (n = 9), 67% (six of none) were missed with delayed enhancement imaging.

Figure 4:

Images in a 63-year-old woman with myocardial infarction. Both short-axis delayed-enhancement (left) and flow-independent dark-blood delayed enhancement (FIDDLE) (right) cardiac MRI scans show anterior papillary muscle infarction (arrows).

Figure 7:

Images in a 39-year-old woman with myocardial infarction. Short-axis delayed enhancement cardiac MRI scan (left) does not show papillary muscle infarction (papMI), while the short-axis flow-independent dark-blood delayed enhancement (FIDDLE) MRI scan (right) shows posterior papMI (arrow).

Figure 5:

Images in a 71-year-old man with myocardial infarction. Short-axis delayed enhancement cardiac MRI scan (left) does not show papillary muscle infarction (papMI), while the short-axis flow-independent dark-blood delayed enhancement (FIDDLE) MRI scan (right) shows anterior and posterior papMI (arrows).

Figure 6:

Images in a 44-year-old woman with myocardial infarction. Short-axis delayed enhancement cardiac MRI scan (left) does not show papillary muscle infarction (papMI), while the short-axis flow-independent dark-blood delayed enhancement (FIDDLE) MRI scan (right) shows posterior papMI (arrow).

FIDDLE imaging, papMI, and the infarct-related artery.— Among participants with myocardial infarction, the infarct-related artery was the LAD, LCX, and right coronary artery in 47% (20 of 43), 26% (11 of 43), and 28% (12 of 43), respectively. Figure 8 shows the prevalence of papMI as detected with FIDDLE imaging according to the infarct-related artery. The majority of participants with LCX (10 of 11, 91%) and right coronary artery (10 of 12, 83%) lesions, but not LAD lesions (eight of 20, 40%), had papMI (P < .001). LAD culprits led to anterior-only papMIs, except in a single participant in whom both papillary muscles were infarcted. Over half (six of 11, 55%) of LCX culprits resulted in infarction of both papillary muscles, and the remaining LCX-related papMIs were evenly divided into anterior only (two of 11, 18%) and posterior only (two of 11, 18%). Right coronary artery culprits led entirely to posterior-only papMIs.

Figure 8:

Bar graph shows the prevalence of detecting papillary muscle infarction (papMI) with flow-independent dark-blood delayed enhancement (FIDDLE) imaging according to the infarct-related artery. LAD = left anterior descending coronary artery, LCx = left circumflex coronary artery, RCA = right coronary artery.

Therefore, despite more frequent LAD than right coronary artery culprits, the prevalence of anterior papMI was lower than that of posterior papMI (37% [16 of 43 participants] vs 44% [19 of 43 participants]). On average, the extent of anterior papMI was smaller than that of posterior papMI (35% ± 26 [SD] vs 63% ± 26, P = .003), and a greater proportion were partial, which was defined as less than 50% infarction (82% [13 of 16 participants] vs 32% [six of 19 participants], P = .008).

Culprit lesion locations within the coronary tree.— The distribution of the culprit lesions on the coronary tree is shown in Figure 9. Although the predefined “at-risk” portion of the coronary tree was far smaller for anterior papMI (subtended first diagonal branch of the LAD and/or first obtuse marginal branch of the LCX) than posterior papMI (subtended posterior descending artery and/or third obtuse marginal branch of the LCX), the sensitivity of the respective at-risk portions for papMI according to FIDDLE imaging were high for both (anterior, 94% [15 of 16 participants];posterior, 90% [18 of 20 participants]) (Tables 3, 4). Table 3 also demonstrates that culprits within their respective coronary at-risk portions were predictive of anterior and posterior papMI at a high and similar rate (positive predictive value, 83% [15 of 18 participants] vs 86% [18 of 21 participants]). Likewise, culprits outside their respective at-risk portions were strongly predictive of the absence of anterior and posterior papMI, respectively (negative predictive value, 96% [24 of 25 participants] vs 91% [20 of 22 participants]).

Figure 9:

(A, B) Coronary artery trees show segments with culprit lesions (red dashed lines) leading to anterior (A) and posterior (B) papillary muscle infarction (papMI) according to flow-independent dark-blood delayed enhancement (FIDDLE) imaging. Note that culprits leading to posterior papMI involved a much larger portion of the coronary tree than those leading to anterior papMI. (C) Example short-axis FIDDLE images in participants with myocardial infarction show culprit lesions in various locations, with papMI (arrows) or without. IRA = infarct-related artery, LAD = left anterior descending coronary artery, LCx = left circumflex coronary artery, mid = middle, prox = proximal, RCA = right coronary artery.

Table 3:

Diagnostic Performance of Predefined “At-Risk” Portions of the Coronary Tree in Predicting PapMI with FIDDLE Cardiac MRI

Table 4:

Description of Cases where the Predefined “at-risk” Portions of the Coronary Tree Resulted in False-Positive and False-Negative Results

Discussion

In this study, we aimed to validate an imaging method for the in vivo diagnosis of papillary muscle infarction (papMI). Using canine pathology-based findings as the reference standard, we found that flow-independent dark-blood delayed enhancement (FIDDLE) cardiac MRI, but not conventional delayed enhancement imaging, was accurate in the diagnosis of papMI (100% vs 80%, P = .01). Specifically, in a canine model, FIDDLE imaging enabled detection of all papMIs, whereas 43% were missed with delayed enhancement imaging. Accordingly, papMI was detected in more participants with first-time myocardial infarction using FIDDLE imaging than delayed enhancement imaging (28 vs 12, P = .012).

In vivo detection of papMI is challenging. A few investigations that used myocardial contrast echocardiography or PET have evaluated papillary muscle perfusion (9,23), but these did not assess papMI. More recently, delayed enhancement cardiac MRI has been used to detect papMI (22,24,25); however, the technique itself was used as the reference standard and no validation was performed. Our study shows that delayed enhancement imaging is insensitive to papMI. This was true regardless of the size of the papMI. This is likely due to the papillary muscles often being completely encircled by left ventricle blood pool, which, from a visual perspective, renders all portions of the papillary muscle as "subendocardial" (ie, there is no epicardial portion) and infarcted areas difficult to distinguish from the bright blood pool. The results of our study indicate that prior delayed enhancement literature regarding the prevalence of papMI and its relationship with the infarct-related artery, mitral regurgitation, and remodeling should be interpreted with the understanding that a substantial proportion of the papMIs in those studies were likely missed.

In our study, the prevalence of anterior papMI was lower than that of posterior papMI (37% vs 44%) despite more frequent LAD than right coronary artery culprits. The common explanation is that the anterior papillary muscle frequently has dual vascular supply and the posterior does not. This means that in a patient with a proximal LAD culprit that subtends the first diagonal branch of the LAD and has dual vascular supply to the anterior papillary muscle from both the first diagonal branch of the LAD and the first obtuse marginal branch of the LCX—the target coronary branches known to perfuse the anterior papillary muscle (8,24)—the first marginal branch of the LCX could "protect" the anterior papillary muscle from infarction. However, our data do not support this explanation. Specifically, a protective effect from dual vascular supply should result in a far lower predictive value for culprits within the “at-risk” portion of the coronary tree for anterior papMI than for posterior papMI. Instead, we found that culprits within their respective coronary at-risk portions predicted anterior and posterior papMI at a similarly high rate (83% vs 86%).

Our results suggest an alternative explanation for the lower prevalence of anterior papMI. Simply, the “at-risk” portion of the coronary tree for anterior papMI is far smaller than that for posterior papMI. Specifically, the at-risk portion is 29% of the coronary tree (five of 17 segments) for anterior papMI compared with 53% (nine of 17 segments) for posterior. This means that in a sample of patients with acute myocardial infarction, where culprits were uniformly distributed throughout the coronary tree, there would be nearly double the number of culprits within the at-risk portion for posterior versus anterior papMI. Moreover, given the nearly identical positive predictive values of the respective at-risk portions, there would be nearly double the number of posterior versus anterior papMIs. Although recent studies have suggested that the spatial distribution of culprits is not strictly uniform throughout the coronary tree—there is an increased prevalence of culprits in the proximal portions of each of the major coronary arteries (26)—such a phenomenon would reduce, but not eliminate, the preponderance of culprits within the far larger at-risk portion of the coronary tree for posterior papMI.

We emphasize that our data do not contradict the concept that the anterior papillary muscle is more likely to have dual vascular supply. As noted previously, dual vascular supply is not synonymous with redundant vascular supply. Two vessels may perfuse different portions of a papillary muscle rather than provide overlapping perfusion. In the first situation, dual vascular supply limits the size but not the occurrence of papMI, whereas in the second, the prevalence of papMI itself is reduced. Our results are consistent with the first situation. Specifically, when papMI was present, the size of anterior papMI was significantly smaller than that of posterior papMI (mean, 35% vs 63%) and a greater proportion were partial, defined as less than 50% infarction (82% vs 32%).

The study by Voci et al (9) that evaluated papillary muscle perfusion patterns in patients who underwent coronary artery bypass graft surgery is often interpreted as evidence that the anterior (compared with the posterior) papillary muscle is more likely to have redundant vascular supply and, thus, reduced prevalence of papMI. However, the authors never evaluated the redundancy of papillary muscle perfusion. Instead, selective bypass graft injections of echocardiography contrast material were limited to only one of the two potential target vessels (eg, graft leading to the first diagonal branch of the LAD or the first obtuse marginal branch of the LCX, but not both). Opacification of just part of a papillary muscle was defined to represent dual vascular supply, whereas opacification of the entire papillary muscle was assumed to represent single vascular supply. Moreover, the presence or absence of papMI was never determined in the Voci et al study. Instead, the authors reported that dual vascular supply to the anterior papillary muscle was associated with reduced papillary muscle dysfunction, as suggested indirectly by the absence of mitral regurgitation.

Our study has limitations. First, we did not evaluate the impact of papMI on the development of mitral regurgitation as it was beyond the scope of this study. Second, various nonischemic cardiomyopathies could also result in papillary muscle scarring. Therefore, the relationship between papMI and the culprit coronary lesion location may not be valid in these patients or in patients with a mixed disorder with both ischemic and nonischemic components. Third, clinical factors that may potentially influence the presence of papMI, such as reperfusion time and adjunctive medical therapy, were not controlled for in this study. Finally, the number of study and control participants was relatively small. Future investigations with larger sample sizes are warranted to verify the relationship between papMI and the culprit coronary lesion.

In conclusion, flow-independent dark-blood delayed enhancement (FIDDLE) cardiac MRI was highly accurate in the detection of papillary muscle infarction (papMI), while conventional delayed enhancement imaging significantly underestimated the prevalence of both anterior and posterior papMI. Anterior papMI was less prevalent than posterior papMI, most likely due to the culprit lesions leading to papMI being restricted to a smaller portion of the coronary tree rather than because of redundant, dual vascular supply. FIDDLE imaging could be used in future studies to identify the prevalence and prognostic significance of papillary muscle scarring in diverse patient populations, including those with various nonischemic cardiomyopathies.