Author + information
- Received April 6, 2009
- Revision received May 19, 2009
- Accepted June 11, 2009
- Published online October 13, 2009.
- Jimmy MacHaalany, MD⁎,
- Yeung Yam, BSc⁎,
- Terrence D. Ruddy, MD⁎,‡,
- Arun Abraham, MBBS⁎,
- Li Chen, MSc†,
- Rob S. Beanlands, MD⁎,‡ and
- Benjamin J.W. Chow, MD⁎,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Benjamin J. W. Chow, University of Ottawa Heart Institute, 40 Ruskin Street, Ottawa, Ontario K1Y 4W7, Canada
Objectives We sought to determine the incidence, clinical significance, and potential financial impact of noncardiac incidental findings (IF) identified with cardiac computed tomography (CT).
Background Cardiac CT is gaining acceptance and may lead to the frequent discovery of extracardiac IF.
Methods Consecutive patients undergoing cardiac CT had noncardiac structures evaluated after full field of view (32 to 50 cm) reconstruction. IF were categorized as clinically significant (CS), indeterminate, or clinically insignificant. Patient follow-up was performed by telephone, and verified with hospital records and/or communication with physicians.
Results Of 966 patients (58 ± 16 years of age, 55.4% men, >98% outpatients), 401 (41.5%) patients had noncardiac IF. A total of 12 (1.2%) patients had CS findings, and 68 (7.0%) patients had indeterminate findings. At follow-up (18.4 ± 7.6 months), none of the indeterminate findings became CS. Although 3 patients with indeterminate findings were diagnosed with malignant lesions, they were unrelated to the IF. After adjusting for age, IF were not an independent predictor of noncardiac death. Noncardiac death and cancer death in patients with and without IF were not statistically different. One patient suffered a major complication related to the investigation of an IF. The total direct cost associated with investigating IF was Canadian $57,596 (U.S. $83,035).
Conclusions Although noncardiac IF are common, clinically significant or indeterminate IF are less prevalent. Rates of death were similar in patients with and without IF, and IF was not an independent predictor of noncardiac death. The investigation of IF is not without cost or risk. Larger studies are required to assess the potential mortality benefit of identifying IF.
- multidetector computed tomography
- incidental findings
- noncardiac death
- lung nodules
- noninvasive imaging
Cardiac imaging with multidetector computed tomography (CT) is gaining clinical acceptance. In addition to providing important cardiac information, the same CT dataset may be used to evaluate noncardiac structures and identify incidental findings (IF), which may be responsible for patient symptoms or would potentially alter patient diagnosis or management. Several groups have reported a high prevalence of noncardiac IF in patients undergoing cardiac CT; however, the clinical and economic ramifications of the IF have not been well defined (1–4). Currently, there is concern that noncardiac IF may lead to further unnecessary investigations resulting in inappropriate resource utilization, augmentation of health care costs, and increased patient anxiety (5). Demonstrating patient benefit and cost-effectiveness would support such practice.
Clinically significant (CS) IF (e.g., pulmonary embolism), resulting in a new diagnosis and alterations in treatment, are very important, especially if shown to improve patient outcome. “Indeterminate” findings (e.g., pulmonary nodule) often require additional radiography or diagnostic procedures, and the clinical and economic effects of investigating indeterminate findings are unclear. The objective of this study is to prospectively determine the incidence of CS and indeterminate noncardiac IF in a large population referred for cardiac CT and to determine the clinical and economic impact of detecting IF.
Between February 2006 and February 2007, 966 consecutive patients underwent cardiac CT coronary angiography and were prospectively enrolled into the Cardiac CT registry.
Image acquisition and reconstruction
Retrospective electrocardiogram (ECG)-gated datasets were acquired (General Electric Volume CT, Milwaukee, Wisconsin) with 64 × 0.625 mm slice collimation and a gantry rotation of 350 ms (400 to 800 mA, 120 kVp). Pitch was individualized according to each patient's heart rate, and ECG-gated X-ray tube modulation was utilized in systole. A biphasic timing bolus (15 to 25 cc intravenous contrast; 40 cc normal saline) protocol was used, and final images were acquired with a triphasic intravenous contrast administration protocol (contrast, 40%/60% contrast/saline [50 cc], and normal saline [40 cc]). The volume and rate of contrast administration were individualized according to scan time and patient body habitus (6).
For noncardiac evaluation, the multidetector CT datasets were reconstructed at the 75% phase using mediastinal and lung windows. The full field of view (32 to 50 cm) was reconstructed with a slice thickness of 2.5 mm and increment of 2.5 mm. Scan range varied according to clinical indication, and all available CT images were reviewed by radiologists with expertise in thoracic imaging.
IFs and patient follow-up
Based upon the radiology reports, noncardiac findings were considered “incidental” if an abnormality was identified without antecedent clinical suspicion or previously known disease. For example, the finding of an aortic dissection in a “triple rule-out” study was not considered an IF.
“Indeterminate” findings were defined as radiographic abnormalities where additional investigations or procedures were recommended by the radiologist to clarify diagnosis or for surveillance (7–9). Findings were considered CS when a radiographic finding was clearly pathological and/or diagnostic (e.g., pneumonia, malignancy, and so on) resulting in a change in patient diagnosis or therapy.
Patient follow-up was performed by telephone interviews and/or correspondence with primary care physicians. All follow-up investigations, treatment, and outcomes attributed to noncardiac IF were verified by reviewing hospital records or through communication with primary care physicians. All noncardiac deaths and cancer-related deaths were reviewed by a clinical events committee.
The costs reported are the direct and indirect hospitalization and procedural costs attributed to the investigations resulting from the recommendations stemming from noncardiac IF. Canadian (CDN) costs associated with diagnostic procedures or interventions were calculated using local departmental budgets and the provincial physician reimbursement fee schedule. U.S. costs were calculated using the 2008 Medicare and Medicaid Services physician and hospital fee schedules. A geographic practice cost index of 1.0 was used.
Cost estimates terminated once a final diagnosis was made or no further investigations were medically indicated. Therefore, costs do not include those for medical or surgical therapy/intervention after diagnosis (e.g., chemotherapy) nor does it include societal costs.
The study was approved by the Institutional Human Research Ethics Board, and all patients provided informed written consent.
Analysis was performed using SAS version 9.1.3 (SAS Institute, Cary, North Carolina) and SPSS version 16.0 (SPSS Inc., Chicago, Illinois). Continuous variables with normal distributions were expressed as mean ± SD. Variables with skewed distributions were expressed as median (25th, 75th percentile). Categorical variables were expressed as frequency (percentage). Comparison of categorical measures was performed using the chi-square tests and a value of p < 0.05 was considered statistically significant.
The prognostic value of IF for unadjusted and adjusted noncardiac death was assessed for the study population. All unadjusted comparisons of noncardiac death were performed on survival analysis log-rank tests. For the risk-adjusted analysis, Cox proportional hazards models were used to assess the independent prognostic value of IF controlling for age.
A total of 966 consecutive patients (58 ± 16 years of age, 55.4% men, >98% outpatients) who underwent cardiac CT at a tertiary care institution were prospectively enrolled into the cardiac CT registry (Table 1).Patients with IF were older, had greater cardiac risk factors, and higher prevalence of myocardial infarction and coronary revascularization. Follow-up was successfully obtained for 939 (97.2%) patients (mean follow-up 18.4 ± 7.6 months).
Prevalence of IF
Of the 401 patients with noncardiac IF, 12 (1.2%) patients had a CS finding that required intervention, and 68 (7.0%) patients had an indeterminate finding requiring further diagnostic imaging or diagnostic procedures (Fig. 1). Additional procedures were recommended to investigate 69 indeterminate findings (pulmonary nodules [n = 45], liver nodules/cysts [n = 9], pulmonary infiltrates [n = 5], thyroid nodule/enlargement [n = 2], esophageal dilation/thickening [n = 2], bronchial nodule [n = 1], renal calculus [n = 1], pancreatic mass [n = 1], thymus nodule [n = 1], intrahepatic bile duct dilation [n = 1], and a vascular abnormality [n = 1]) (Table 2).
During surveillance of indeterminate findings, a total of 76 CT scans (71 thorax, 4 abdomen/pelvis, and 1 head) were performed. Each of the 68 patients was exposed to an estimated mean effective dose of 9.4 mSv.
At follow-up, none of the indeterminate findings initially discovered at the time of the cardiac CT became CS (Fig. 1). However, during the course of additional investigations, 3 patients were serendipitously diagnosed with malignancies unrelated to the IF being monitored. In 1 patient, the indeterminate lesion resolved spontaneously; however, during surveillance a new lesion was discovered and diagnosed as nonsmall cell lung carcinoma (NSCLC). Additional imaging, recommended for the second patient, demonstrated that the indeterminate lung nodule was stable but 2 new lesions were identified, biopsied, and proven to be NSCLC. After diagnosis, this patient refused further investigations and treatment (Table 3).A third patient, with a history of breast cancer, was followed for an initial indeterminate pulmonary lesion that was subsequently diagnosed as a granuloma. However, during the course of follow-up, new metastatic bony lesions were discovered.
One IF patient suffered a significant complication after transthoracic biopsy resulting in empyema and abdominal abscesses requiring hospitalization, therapeutic drainage, and intravenous antibiotics.
There were a total of 7 (1.7%) noncardiac deaths in the 401 patients with noncardiac IF with only 1 (1.0%) death attributable to the IF detected (Fig. 1, Table 4).This was not statistically significant from the 3 (0.5%) noncardiac deaths in the population without IF (Figs. 1 and 2).⇓Of the deaths in the IF group, only 1 death could be directly attributed to the IF. A total of 6 cancers were diagnosed with 4 attributable to the IF (Table 3). Multivariate analysis demonstrated that age was an independent predictor of noncardiac death but IF was not an independent predictor of noncardiac death.
A total of 164 additional diagnostic imaging and procedures were performed for the 80 patients with CS and indeterminate incidental findings (Tables 5 and 6).⇓The total costs associated with investigating CS findings and indeterminate findings were CDN $6,600 (U.S. $8,888) and CDN $16,941 (U.S. $34,550), respectively (Tables 5 and 6). The patient who suffered a significant complication after transthoracic biopsy required additional treatment at an expense of CDN $34,055 (U.S. $39,597). The total cost of working up all patients with CS findings, including the patient who suffered a complication, was CDN $40,655 (U.S. $48,485). The total direct cost of investigating all IF (including the procedural complication) was CDN $57,596 (U.S. $83,035).
With the rapid emergence and acceptance of cardiac CT, the volume of CT studies is anticipated to rise, resulting in a concomitant increase in the identification of noncardiac IF, many of which appear to be benign (10). Such findings may have significant clinical and economic consequences.
Our study confirms the high prevalence of incidental findings (41.5%) in our patient population with 7.0% of indeterminate findings requiring further clinical or radiological follow-up (1,3,4,10,11). None of the indeterminate findings became CS, and only 1 death could be attributed to an IF. There were 2 serendipitous diagnoses of new malignancies (NSCLC) and 1 new metastatic lesion (in a patient with pre-existing breast cancer), none of which were related to the initial IF identified at the time of cardiac CT. Moreover, 1 patient with an IF experienced a significant adverse event with follow-up investigations resulting in significant morbidity.
The results of our study suggest that there may be significant cost and morbidity associated with investigating incidental findings without a clear mortality benefit. Though the total cost for investigations was CDN $57,596 (U.S. $83,035), it underestimates the total economic impact in our cohort because it does not include cost of physician follow-up, cost of treatment after diagnosis (e.g., surgery or chemotherapy), societal costs (e.g., lost wages), and patient quality of life.
Our prospective observational study reports all clinical events in our practice. Admittedly, complications associated with lung biopsy are likely rare occurrences, but such events are typically subject to reporting bias. Several studies have demonstrated that the more common complications (pneumothorax and hemoptysis) of lung biopsy vary between 6% to 31% (12,13) and infections rates are not commonly reported. Assuming that no complications occurred in our cohort, the total cost for investigations would have been much smaller (CDN $18,722 [U.S. $36,808]).
Other than screening for cervical cancer, breast cancer, and colorectal cancer, routine screening of asymptomatic individuals for various types of cancers have not been shown to be beneficial (14,15). Some guidance may be drawn from previous CT-based lung cancer screening studies. While they demonstrated a wide prevalence (22% to 74%) of lung nodules in asymptomatic patients (16–19), the frequency of lung cancer at follow-up remained very small (0.1% to 0.2%). Similarly, Iribarren et al. (10) followed asymptomatic lung nodules in a population of 60- to 69-year-old subjects and determined that 35% of lesions resolved, and 62% were stable or decreased in size. The prevalence of lung cancer in CT coronary angiography studies was small even when Onuma et al. (4) reconstructed a large field of view to assess for incidental extracardiac findings (Table 7).In fact, we found a similarly low malignancy detection rate and low overall cancer death rate.
CT screening studies have demonstrated a very high false-positive rate (>90%) (noncalcified nodules proven to be benign) (19–21). False-positive findings may lead to increased costs, increase patient morbidity, and a reduction in patient quality of life. These, combined with an absence of clear mortality benefit of screening, may partly explain why screening is not recommended nor has it been adopted into routine clinical practice. Our study demonstrated that there was no statistically significant mortality difference between patients with and without incidental findings but may be underpowered to detect a meaningful difference. Multivariate analysis confirmed that age was the only independent predictor of noncardiac death. Further studies are needed to further understand the implications of noncardiac IF and to determine if the identification of IF can reduce patient morbidity or mortality.
Screening of the asymptomatic population does appear to result in a shift in the proportion of stage I NSCLCs (16,18), but has not been shown to be cost-effective or to significantly improve patient survival in population-based studies (16,18). Mahadevia et al. (22) performed a decision and cost-effectiveness analysis from a societal perspective and demonstrated that the societal cost for annual CT screening is very high without substantial reductions in mortality. Similar to our current study, they demonstrated that the diagnosis of a pulmonary nodule potentially leads to subsequent imaging, which may result in the increase of the cost of health care. Our study demonstrates that repeat imaging in 68 patients with indeterminate findings resulted in the detection of 2 new malignancies and 1 metastatic lesion. Of these, 1 patient died, 1 patient underwent successful lobectomy, and 1 patient is receiving chemotherapy.
More recently, Henschke et al. (23) examined the potential of mass screening in asymptomatic patients that are at high risk for lung cancer and healthy enough to withstand surgery. They demonstrated that a 10-year survival rate of 412 patients identified with stage I lung cancer was 88% and compared it with the SEER (National Cancer Institute-Surveillance Epidemiology and End Results Registry), which had a 10-year survival rate of 38%. However, the apparent difference in mortality may be partially explained by differences in the 2 study cohorts. Further studies demonstrating mortality benefit and cost-effectiveness are required to understand the impact of mass screening in asymptomatic patients.
Moreover, a proportion of patients with indeterminate findings required additional diagnostic investigations exposing them to additional ionizing radiation (mean effective dose = 9.4 mSv) without known benefit. There is data from the Japanese atomic bomb survivors (24–26) and recently from a study of radiation workers in the nuclear industry (27,28) showing that ionized radiation doses as low as 5 mSv (average dose ranged between 20 to 40 mSv) may be harmful and may increase the risk of cancer. More clinical data are required to guide the follow-up of patients with incidental findings to maximize diagnosis and minimize patient risk.
Using larger bowtie filters permits the reconstruction of the entire chest at the expense of increasing radiation exposure as much as 40% (29). Such practice is justifiable if incremental patient benefit can be demonstrated. In the absence of such benefit, some centers have elected to use the smallest bowtie filter to reduce patient radiation exposure.
We recognize that a limitation in our study is the relatively short length of follow-up (18.4 ± 7.6 months) resulting in the potential underestimation of IF becoming CS. A small proportion of patients are still undergoing surveillance (because of “new” indeterminate findings on follow-up imaging) thus potentially underestimating the benefit of surveillance recommendations.
We report the costs associated with investigating each patient with IF using both CDN and U.S. cost estimates. We recognize that reimbursement differs regionally and thus our estimates may not reflect reimbursement in all regions. Though some centers incur additional fees from “over-reads,” the current practice in Canada is to share reimbursement. Since there are no additional fees for over-reads, these costs could not be factored into our analysis.
Since our cost analysis terminated once a diagnosis was made or when surveillance was terminated, our cost analysis may underestimate the true economic burden and prevents the accurate calculation of “cost/life-saved.” Further studies are needed to better understand the “cost/life-saved” associated with detection of IF.
The total number of CS incidental findings on cardiac CT is small but important. Though incidental findings are relatively common, recommended investigations have potential economic consequences and are not without risk. In our small cohort, the detection of incidental findings did not appear to predict noncardiac death or cancer death. Larger studies are needed to better understand potential morbidity and mortality benefit of detecting IF.
The authors extend their gratitude to Joanne Bussell, Kathryn Calladine, Roger DesPrez, Micheala Garkisch, Debbie Gauthier, Patricia Grant, Sandina Jamieson, and Richard Tessier for their expertise and dedication to cardiac CT research.
This study was supported, in part, by the Ontario Research Fund: Imaging for Cardiovascular Therapeutics Project #RE02-038 and the Canada Foundation for Innovation #11966. Dr. Ruddy is the Vered Chair of Cardiology. Dr. Abraham is supported by the Heather and Whit Tucker Research Fellowship in Cardiology. Dr. Beanlands has received research grants from GE Healthcare and MDS Nordion. Dr. Chow is supported by CIHR New Investigator Award #MSH-83718 and receives research and fellowship support from GE Healthcare, educational support from TeraRecon Inc., and research support from Pfizer.
- Abbreviations and Acronyms
- clinically significant
- computed tomography
- incidental finding(s)
- nonsmall cell lung cancer
- Received April 6, 2009.
- Revision received May 19, 2009.
- Accepted June 11, 2009.
- American College of Cardiology Foundation
- Horton K.M.,
- Post W.S.,
- Blumenthal R.S.,
- Fishman E.K.
- Hunold P.,
- Schmermund A.,
- Seibel R.M.,
- Grönemeyer D.H.,
- Erbel R.
- Onuma Y.,
- Tanabe K.,
- Nakazawa G.,
- et al.
- ↵(2008) Screening and early detection of cancer. World Health Organization. Ref Type: Internet Communication. Accessed March 10, 2009. http://www.who.int/cancer/detection/en/.
- Stewart B.W.,
- Kleihues P.
- Novello S.,
- Fava C.,
- Borasio P.,
- et al.
- Strauss G.,
- DeCamp M.,
- Dibiccaro E.,
- et al.
- Cardis E.,
- Vrijheid M.,
- Blettner M.,
- et al.