Cardiac metastases are more prevalent than primary cardiac tumors, and although rare, the incidence is anticipated to increase with the extended survival of oncology patients.
To estimate the current incidence of cardiac metastasis from solid tumors in adult autopsies.
Adult autopsy cases from 1984 through 2019 from patients diagnosed with any type of solid cancer were retrieved. The medical charts and pathologic autopsy data were reviewed in detail.
A total of 1294 adult autopsies performed on patients diagnosed with any type of cancer within the past 35 years were reviewed. We found 124 secondary cardiac tumors. Eighty-five were due to cardiac involvement by solid tumors. Of these, 61 were true cardiac metastases of solid cancers. We focused on these 61 cases. The age range was 32 to 85 years. Forty-four patients were men and 17 were women. The lung was the most common primary site, with 21 cases (34.43%). The most frequent histologic type was carcinoma, with 54 cases (88.52%). The predominant layer of the heart involved was the pericardium, with 35 cases (57.38%). Twenty-one cases (34.43%) had pericardial effusion, with 4 being hemorrhagic. All cases had multiple extracardiac metastases, with 56 cases (91.8%) having distant metastases in 4 or more different organs.
Cardiac metastasis is a rare occurrence, with an incidence of 4.71% (61 of 1294 cases) in our series. Lung cancer accounted for most of the cardiac metastases seen, and carcinomas were the most frequent histologic type. The pericardium was the most frequent location. Cardiac metastases occurred most frequently in cases of massive metastatic dissemination.
According to earlier autopsy studies, the frequency of primary cardiac tumors ranges from 0.001% to 0.3%.1,2 Most primary cardiac tumors (∼75%) are benign, with atrial myxoma being the most common tumor encountered. The remaining (∼25%) are malignant; most of these are either rhabdomyosarcoma or angiosarcoma.
Secondary tumors in the heart or cardiac metastases are more common than primary tumors. Although they are considered rare, when sought for, the incidence seems to be higher than expected. The incidence of cardiac metastases reported in literature is highly variable, ranging from 2.3% to 18.3% (average incidence of 7.1%) among autopsies of cancer patients.3–5 This incidence is anticipated to increase with extended survival rates among oncology patients. Though these metastases may originate from any malignant tumor, previous studies indicate that the most common are melanoma and carcinomas of the lung, breast, and esophagus.3–7 Most secondary cardiac tumors remain clinically silent (>90%) and are often diagnosed postmortem.
This report describes our experience in autopsies performed on adults who had been diagnosed in our hospital with any type of solid tumor during a period of 35 years. Although hematologic cancers can sometimes form a mass, in this article the terms solid tumor or solid cancer will refer to any tumor that can form a solid mass except for hematologic tumors.
METHODS
Cohort Selection
We carried out a retrospective autopsy study from 1984 through 2019. We selected the autopsies of adults who had been diagnosed in our hospital (Hospital Universitario de Navarra, Pamplona, Spain) with any type of cancer. We classified them into the following categories: no cardiac involvement, primary cardiac tumors, leukemia/lymphoma involvement, and secondary cardiac involvement by solid tumors. We reviewed all of the cases of secondary cardiac involvement by solid tumors and separated the cases that spread directly to the heart from the cases of true cardiac metastasis of solid cancers. We focused our study on the cardiac metastases of solid cancers. This study was conducted in accordance with the principles of the Helsinki Declaration.
Clinical and Histopathologic Features
The medical records were reviewed to obtain clinical information such as age, sex, and clinical presentation. Autopsy pathologic data such as primary tumor site, histologic type, extent of cardiac involvement, direct spread or true metastasis, intracavitary tumor thrombus, and the presence or absence of pericardial effusion were reviewed in detail and analyzed with the SPSS software. The gross images found in our multimedia files were selected. Hematoxylin-eosin (H&E), histochemistry, and immunohistochemistry (IHC) slides were selected, and the diagnosis of each case was confirmed by 2 pathologists. In the older cases, we obtained new H&E- and IHC-stained slides according to current protocols. Some cases were scanned for photography.
To identify and confirm the primary tumor site in autopsies with metastases in various organs, we first checked the clinical history of the patients. Then based on the morphology of the tumor on H&E, we performed (if necessary) the following basic IHC panel: lung (TTF1, CK7, P40, CK5/6), gastrointestinal and pancreatobiliary tract (CK20, CK7, CDX2, synaptophysin, chromogranin), genitourinary (GATA3, P63, NKX3.1, PAX8, RCC), skin (P40, P16), endocrine organs (TTF1, thyroglobulin, PAX8, vimentin, melan-A, inhibin), breast (GATA 3, mammaglobin), pleura-peritoneum (calretinin, D2-40, WT1), malignant melanoma (S100, HMB-45, melan-A, SOX-10), and soft-tissue sarcomas (CD31, factor VIII, TLE1, desmin, MyoD1).
RESULTS
Cohort Demographics
A total of 1294 autopsies of adults diagnosed in our hospital with any type of tumor in the last 35 years were reviewed. We found 133 cases of cardiac tumors. Nine cases were primary cardiac tumors, and 124 were secondary cardiac tumors. Of the secondary tumors, 39 cases had lymphoma/leukemia involvement, and thus we excluded them from this study, and 85 cases had cardiac involvement by solid cancers. Of the remaining 85 cases, 24 cases were excluded as they were due to direct spread to the heart and were not true incidences of cardiac metastases. The 61 cases of true cardiac metastasis of solid cancers were the focus of our study. The incidence of cardiac metastases of solid cancer in our study was 4.71% (61 of 1294) of all autopsies (Figure 1). The patients' ages ranged between 32 years and 85 years, with an average age of 60.9 years. Forty-four patients were male and 17 were female.
Histopathologic Features
Cardiac metastasis by organ site of primary malignancy is given in Table 1. Lung was the primary site in 21 of 61 cases (34.43%). Next in frequency was the gastrointestinal/pancreatobiliary tract, with 14 cases (22.95%). Other primaries were genitourinary with 5 cases (8.2%), soft tissue with 4 cases (6.56%), skin with 3 cases (4.92%), endocrine organs with 4 cases (6.56%), breast with 3 cases (4.92%), pleura-peritoneum with 2 cases (3.28%), head and neck with 2 cases (3.28%), and unknown origin with 3 cases (4.92%). Of the 24 autopsies where the tumor spread directly to the heart, 16 originated from the lungs, 4 from pleura, and 4 from bronchus.
The most frequent histologic type was carcinoma, with 54 of 61 cases (88.52%) that included adenocarcinoma (26 cases), neuroendocrine carcinoma (10 cases), squamous cell carcinoma (7 cases), urothelial carcinoma (4 cases), mucoepidermoid carcinoma (1 case), clear cell carcinoma (1 case), adenosquamous carcinoma (1 case), and undifferentiated carcinoma (n = 4). Other histologic types were mesothelioma with 2 cases (3.28%), sarcoma with 4 cases (6.56%), and melanoma of unknown primary with 1 case (1.64%) (Table 2). In 22 cases, more than 1 layer of the heart was involved. The predominant layer of the heart involved by cardiac metastases was the pericardium in 35 cases (25 alone and 10 associated with other layers), the myocardium in 32 cases (19 alone and 13 associated with other layers), the epicardium in 5 cases (3 alone and 2 associated with other layers), and the endocardium in 5 cases. One case showed involvement of both the myocardium and epicardium. Four cases showed involvement of both the myocardium and endocardium, and 1 case showed involvement of both the endocardium and the epicardium. The endocardium alone and the valves were not involved in any case (Figure 2, A through I) (Table 3).
The laterality of cardiac involvement by cardiac metastases was as follows (n = 61): 11 cases (18.03%) affected the left side of the heart, 3 cases (4.92%) affected the right side, 6 cases (9.84%) were bilateral, and 16 cases (26.23%) had no data available. Twenty-five cases (40.98%) had pericardial involvement without our being able to identify laterality (Figure 3, A through L, and Figure 4, A through L). Intracavitary (auricular or ventricular) tumor thrombus was present in only 5 cases (8.2%).
Of the 61 cases, 21 (34.43%) had pericardial effusion. Of these, 15 cases were associated with pericardial metastases, 4 cases with myocardial metastasis, and 2 cases with epicardial metastasis. The quantity of the pericardial fluid was widely variable. In 12 cases the quantity was less than 1000 mL (8 cases had ≤100 mL), and in 2 cases the quantity was at least 1000 mL. Seven cases did not have data available about the quantity of the pericardial fluid. Of the 21 cases, the pericardial effusion was nonhemorrhagic in 17 cases (80.95%) and hemorrhagic in 4 cases (19.05%). Fluid analysis was not available in any of the cases. All cases of true cardiac metastases of solid cancers had multiple extracardiac metastases. Fifty-six of 61 cases (91.8%) had distant metastases in 4 or more different organs.
DISCUSSION
Cardiac metastases are rather rare,3–7 but they are more common than primary cardiac tumors. The diagnosis of cardiac metastases antemortem is seldom made because more than 90% are clinically silent. The incidence of cardiac metastasis at the time of autopsy varies between 2.3% and 18.3% according to previous studies.3–16 However, these studies were published more than 20 years ago. In our study of 1294 autopsies of adults (older than 18 years) who died with any type of tumor, we found that the incidence of cardiac metastases from solid tumors is 4.71%. In the last decade, there are only case series reports and case reports about cardiac involvement from specific secondary tumors.17–26
Secondary cardiac involvement by solid tumors can occur by hematogenous or lymphatic dissemination.9 In our study, we grouped the forms of secondary cardiac involvement by solid tumors into those that spread directly to the heart and those that were true metastases. In 24 of 85 cases, the cardiac involvement was due to the direct extension of malignant tumors of the lung, pleura, and bronchi. Interestingly, this form of cardiac involvement was observed most frequently in the first half of the study period (between 1984 and 2001), where 20 of 62 autopsies (32.26%) showed direct extension to the heart, whereas it was less frequent in the second half of the study period (between 2002 and 2019), where only 4 of 23 autopsies (17.39%) showed that form of cardiac involvement. It may be that in the last decades the management of cancer patients has been changing owing to new technologies that allow accurate diagnosis at early stages, which allows for early treatment.
In 61 of 85 cases, cardiac involvement was due to true cardiac metastasis of solid cancers. The most frequent location of cardiac metastases was the pericardium, with 35 cases (57.38%), which was associated with other layers in 10 cases. These data are in accordance with previous studies reporting that nearly two-thirds of cardiac metastases are located in the pericardium and one-third in the myocardium and/or epicardium, with only 5% of cases in the endocardium.7 Myocardial or epicardial metastasis is usually due to hematogenous or lymphatic dissemination, whereas the endocardium is usually invaded through systemic circulation and, in fewer cases, through myocardial involvement.9,10 In our study, the endocardium was affected in 5 cases (8.2%). Four of those cases were associated with myocardial involvement and 1 was associated with epicardial involvement. However, the endocardium alone was not affected in any case.
According to the study carried out by Bussani et al9 from 1994 to 2003, the sites of origin of the most frequent primary tumors of cardiac metastases were the lung, the skin, the breast, and carcinomas of unknown origin. These data vary in published articles since some authors include cardiac involvement by lymphomas/leukemia in their studies9,11–15 ; however, we excluded lymphomas/leukemia from our study and focused on cardiac involvement by solid cancer.
In our study, the site of origin of the most frequent primary tumor was the lung with 21 cases (34.43%), followed by gastrointestinal/pancreatobiliary tract with 14 cases (22.95%), genitourinary with 5 cases (8.2%), soft tissue with 4 cases (6.56%), skin with 3 cases (4.92%), endocrine organs with 4 cases (6.56%), breast with 3 cases (4.92%), pleura/peritoneum with 2 cases (3.28%), head and neck with 2 cases (3.28%), and unknown origin with 3 cases (4.93%). As in previous reports, the lung was the most frequent site of primary tumor origin.
Another interesting finding in our study was the low percentage of cardiac metastases due to breast cancer throughout the study period. This could reflect the fact that different developed countries have implemented breast cancer screening programs.27,28 In 1990, Navarra became the first region in Spain to initiate a breast cancer screening program.29 Currently, according to statistics, the screening programs in Spain cover more than 90% of the target population.29 In our study, we found only 3 cases of cardiac metastases of breast cancer. We did not observe a significant difference between the first and second half of the study periods.
Previous studies have observed that the most frequent histologic types of cardiac involvement were adenocarcinoma and neuroendocrine carcinoma (mostly of the lung), and squamous cell carcinoma of the esophagus.11 In our study, we have observed that the most frequent histologic type associated with cardiac metastasis was carcinoma, with 54 cases (88.52%). Of these the most frequent subtypes were adenocarcinoma (26 cases), with 8 cases originating in the lung, and neuroendocrine carcinoma (10 cases), with 8 cases originating in the lung. We also observed exceptionally rare cases of primary tumor site such as mucoepidermoid carcinoma of the vulva and adenosquamous carcinoma of the rectum. In all reported series,3–15 sarcomas were infrequent in cardiac metastases and accounted for only 4 cases (6.56%) of secondary cardiac tumors in our study (angiosarcoma, synovial sarcoma, rhabdomyosarcoma, and epithelioid hemangioendothelioma).
Pericardial effusion is present in a wide variety of pathologic conditions. It can be caused by iatrogenesis, malignancy, complications of a myocardial infarction, idiopathic causes, uremia, tuberculosis, aortic dissection, trauma, and infections, among others. Historically, the principal causes of hemorrhagic pericardial fluid were myocardial infarction, malignancy, and tuberculosis.30–32 In 2001, Chiu et al33 performed a study of 125 adult patients comparing the differences in the etiologies of hemorrhagic and nonhemorrhagic pericardial effusion and to determine whether a bloody pericardial effusion is more likely to be associated with malignancy. They studied 29 cases of nonhemorrhagic effusions and 96 cases of hemorrhagic effusions. They found that the 3 leading causes of hemorrhagic effusions were iatrogenesis, malignancy, and complications of a myocardial infarction and that the 2 most common causes for serous (nonhemorrhagic) pericardial effusions were idiopathic causes and malignancy. Results by Chiu et al33 were consistent with the results observed by Atar et al34 and Spodick and Worcester.35 Chiu et al33 also found that malignancy-associated effusions occurred in hemorrhagic and nonhemorrhagic groups with a similar frequency. In our study, we found 21 cases (34.43%) with pericardial effusion. Only 4 of these had hemorrhagic effusion, and the other 17 cases had nonhemorrhagic (serous) effusion. Interestingly, 6 cases that had pericardial effusion did not have pericardial metastases (4 had myocardial metastasis and 2 had epicardial metastases).
Finally, we observed that 56 cases (91.8%) of cardiac metastases occurred in advanced cancers that affected 4 or more different organs. This indicates that cardiac metastases generally occur in disseminated cancer and indicate advanced-stage disease.
Throughout the period of our study, we did not identify changes in the prevalence of the different primary sites or in the histologic appearance.
In conclusion, to our knowledge, this is the largest study in the last decade on cardiac metastases of solid cancers. The incidence of cardiac metastases in our study was 4.71% of all adult autopsies. Lung cancer accounted for most of the cardiac metastases seen, and carcinomas were the most frequent histologic type. Pericardial metastases were the most common type of cardiac metastases, followed by myocardial metastases; cardiac metastases occurred more frequently in cases of massive metastatic dissemination. Furthermore, we have realized over time that the new technologies developed in the diagnosis and treatment of cancer have significantly decreased metastatic cardiac involvement.
The authors express their gratitude to all the staff of the pathology department of our hospital. We also thank Lauren Tucker, AS, for her collaboration with English.
References
Author notes
The authors have no relevant financial interest in the products or companies described in this article.
Part of this work was presented as a poster at the United States and Canadian Academy of Pathology 109th Annual Meeting 2020; March 3, 2020; Los Angeles, California.